, PHYLOGENY AND HOST RELATIONSHIPS

OF THE (PHTHIRAPTERA: ISCHITOCERA)

by

CHRISTOPHER HENRY GOUTTS LYAL, B.Sc.

VOLUME 1

October 1983

A thesis submitted for the degree of Doctor of Philosophy of the University of London and for the Diploma of Imperial College,

Department of Pure and Applied Biology, Imperial College, London SN7

and

Department of Entomology, British Museum (Natural History), London SM7 2

ABSTRACT

The external morphology of the Phthiraptera is discussed with particular reference to the Trichodectidae. Structures of the head, thorax and abdomen are examined and homologised, most attention being given to features of potential use in systematic analysis. The homologies of the component parts of the male and female genitalia, hitherto disputed, are established. The characters used by previous workers for systematic placement of the Trichodectidae, , , Rhyncophthirina, Anoplura, Phthiraptera and are examined, and a cladistic analysis of these groups performed. The and Phthiraptera are found to be holophyletic but the Psocoptera are paraphyletic. The Trichodectidae, Amblycera, Rhyncophthirina and Anoplura are all holophyletic, the Rhynco- phthirina is the sister-group of the Anoplura and the Amblycera the sister-group of all other lice. The Ischnocera is not demonstrably holophyletic, and the exact placement of the Trichodectidae is not determined. A cladistic analysis of the Trichodectidae is performed and the 351 and subspecies reclassified on the basis of this. Five subfamilies are used to partition the twenty genera employed, ten of the latter being sub-divided into twenty-seven subgenera. One subfamily, three genera and four subgenera are described as new. Three genera are placed in synonymy, eight genera and subgenera are raised from synonymy, and four genera are reduced to subgenera,. The generic placements of 106 species and subspecies are changed. Keys to subfamilies and to genera and subgenera are produced. Relationships of the mammalian hosts of the Trichodectidae are compared to the phylogeny (cladogram) of the family in to test Fahrenholz* Rule. This Rule, which holds that host and parasite phylogenies match, is found not to be fully applicable to the Trichodectidae, and inferences are drawn for the rest of the Phthiraptera. The ecological and evolutionary aspects of -hpst relationships are examined to indicate possible patterns of co-evolution. 3

ACKNOWLEDGEMENTS

I must first thank the Director and Trustees of the British Museum (Natural History) for permitting me to carry out this research and submit it for examination under the Public Research Institute Scheme. I would also like to thank Br R. G. Davies of Imperial College and Br L. A. Found and Br G. 3. Vfhite of the British Museum (Natural History) for acting as my supervisors, and for their invaluable advice during the course of the stud;/. I am very grateful to all those who have given advice and read parts of the manuscript, particularly Dr T. Clay, Dr P. S. Cranston, Mr U. R. Dolling, Mr D. Hollis, Mr A. M. Hutson, Dr I. F. Hitching, Dr R. P.- Lane, Dr A. Vi. Gentry and Mrs J. M. Palmer of the British Museum (Natural History), Dr H. B. Boudreaux of Louisiana State University and Dr B. Feming of the University of Alberta. During my visit to the United States the following people were very helpful and hospitable: Dr K. C. Emerson, Dr and Mrs K. C. Kim of Pennsylvania State University, Dr and Mrs R. D. Price of the University of Minnesota, Dr 0. Flint of the United States National Museum, Washington, Dr R. Traub and Miss H. Starcke of the University of Maryland School of Medicine, Dr W. A. Drew and Dr D. Peters of Oklahoma State University. I would also like to thank Mrs J. Cranston for typing the thesis and my wife, Mrs J. Lyal, for typing the data matrix. Finally, I must thank my parents and my wife for supporting and encouraging me during the production of this thesis. 4

CONTENTS

Page

VOLUME 1 ABSTRACT ------2 ACKNOWLEDGEMENTS ------3 CONTENTS ------4 I.IST OP FIGURES ------o LIST OP TABLES ------17 INTRODUCTION ------18

1. MATERIALS AND METHODS ------21 1.1. Literature ------22 1.2. Specimens ------26 1.2.1. Introduction ------26 1.2.2. Nature of study material ------26 1.2.3. Collection and preparation of specimens ------27 1.2.3.1. Collecting ------27 1.2.3.2. Mounting ------30 1.2.3.3. Observation ------32 1.3. Systematic analysis and classification ------33 1.3.1 • Introduction ------33 1.3.2. Homology ------33 1.3.3. Methods of systematic analysis ------42 1.3*3*1* Introduction ------42 1.3.3.2. Methods of phyletic analysis ------44 1.3.4. Classification ------54 1.3*4*1* Relationship of classification and systematic analysis- 54 1.3.4*2. Structure of the classification ------55 1.4* Character survey ------63 1.4.1. The taxonomic character ------63 1.4.2. Use of characters for grouping taxa ------63 1.4*2.1. Choice of characters and their relative values - - - 63 1.4*2.2. Determination of polarity for phyletic analysis - - - 63 1.4.2.3* Character weighting in phyletic analysis ----- 66 1.4*2.4* Problems of complex characters — ------68 1.4.2.5* Constitutive and diagnostic characters ------65 1.4*3* Lata recording in this study ------69 1.4.3.1. Procedure ------65 1.4.3.2. Coding ------70 5

Page

2. T'ORPHOLOGY AND CHARACTER ANALYSIS CF TRICEODECTIDAE - - - 75 2.1. Systematic position of Trichodectidae ------75 2.1.1. Introduction ------75 2.1.2. Relationship of Fhthiraptera to other - - - - 78 2.1.2.1. Introduction ------73 2.1.2.2. Apomorphies proposed for relevant groups of P hall one opt era ------78 2.1.2.3. Significance of ahove characters ------89 2.1.2.4. Relationship of Phthiraptera to Liposcelidae - - - - co 2.1.3. Relationship of Trichodectidae to other Fhthiraptera - - 94 2.1.3.1. Introduction ------04 2.1.3.2. Apomorphies proposed for families and supra-familial groups within the Phthiraptera ------05 2.1.3.3. Significance of ahove characters ------103 2.1.3.4. Ranking and classification ------108 2.2. General morphology ------110 2.2.1. Terminology ------no 2.2.2. The head ------m 2.2.2.1. Introduction ------ni 2.2.2.2. Structure of the head capsule ------112 2.2.2.3. The tentorium ------120 2.2.2.4. The antennae ------121 2.2.2.5. The ocelli and compound eyes ------127 2.2.2.6. The mouthparts ------129 2.2.3. The thorax ------137 2.2.3.1. Introduction ------137 2.2.3.2. The neck region ------137 2.2.3.3. Segments of the thorax ------137 2.2.3.4. Description of the trichodectid thorax ------140 2.2.3.5. Goxal articulations ------142 2.2.3.6. The legs ------142 2.2.4. The ahdomen ------148 2.2.4.1. Segmentation ------148 2.2.4.2. Female genital and postgenital segments ----- 154 2.2.4.3. The male ahdomen ------166 2.2.4.4. The male genitalia ------173 2.2.4.5- The tracheal system and spiracles ------

£age

2.3. Character survey of Trichodectidae ------qo6

2.3.1. Introduction ------qc6 2.3.2. List of characters ------jcy 2.4. Character analysis ------229 2.4.1. Introduction ------229 2.4.2. Identification of apomorphic states for phyletic analysis 229 2.4.3. Cladistic analysis ------248

VOLUME 2

3. CLASSIFICATION AND TAXONOMY OF TRICHODECTIDAE ----- 264 3.1. Taxonomic history of Trichodectidae ------265 3.2. Proposed classification ------282 3.3. Descriptions of genera and subgenera ------291 3.3.1. Introduction ------291 3.3.2. Bovicolinae ------292 3.3.2.1. Ewing ------292 3.3.2.2. Genus n. '2 ------303 3.3.2.3. Nerneckiella Eichler ------305 3.3.2.4. Genus n. 3 ------307 3.3.2.5. Fjttberg ------309 3.3.3. Eutrichophilinae ------323 3.3.3.1. Eutrichophilus Mjbberg ------323 3.3.4. Dasyonyginae ------327 3.3.4.1. Cebidicola Bedford ------327 3.3.4.2. Procavicola Bedford ------331 3.3.4.3. Frocaviphilus Bedford ------338 3.3.4.4. Dasyonyx Bedford ------346 3.3.4.5. Eurytrichodectes Stobbe ------352 3.3.5. Trichodectinae ------354 3.3.5.1. Frotelicola Bedford ------354 3.3.5.2. Lutridia Kdler ------358 3.3.5.3. Genus n. 4------362 3.3.5.4. Nerneckodect es Conci ------364 3.3.5.5. Nitzsch ------366 3.3.5.6. Bring ------383 3.3.5.7. Lorisicola Bedford ------357 3.3.6. Subfamily n.------409 3.3.6.1. Neotrichodect es Ewing ------409 3.3-6.2. Geomydcecus Ewing ------421 Page

3.4. Keys to Trichodectidae ------431 3.4.1 • Introduction ------431 3.4.2. Key to subfamilies of Trichodectidae ------431 3.4*3* Key to genera and subgenera of Trichodectidae - - - - 432

4. HOST-PARASITE RELATIONSHIPS ------445 4.1. Introduction ------446 4.2. Species concept ------448 4*3. Host specificity ------450 4.3.1 • Introduction ------450 4-3.2. Predictions of Fahrenholz* Rule ------450 4-3.2.1. Introduction ------450 4.3-2.2. Phylogenetic component of specificity ------451 4.3.2.3. Ecological component of specificity ------466 4.3.2.4. Summary ------472 4.3.3. Factors influencing host specificity ------474 4.3.3.1. Introduction ------474 4.3.3.2. Independence of dispersal ------476 4.3.3.3. Host suitability ------485 4.3.3.4. Availability of suitable hosts ------489 4.3.3.5. Summary ------493 4.4. Comparison of trichodectid and host phylogenies - - - - 494 4.4.1. Introduction ------494 4.4.2. Bovicolinae ------498 4.4.2.1. Damalinia ------498 4.4.2.2. Bovicola ------510 4.4.2.3. Nerneckiella ------812 4.4.2.4. Bovicolinae ------515 4.4.3. Eutrichophilinae ------516 4.4.4. Dasyonyginae ------______516 4.4.5. Trichodectinae ------518 4.4.5.1. Trichodectes ------518 4.4.5.2. Trichodectini ------521 4.4.5.3. Felicola ------522 4.4.5.4. Lorisicola ------525 A.4.5.5« Trichodectinae ------529 4.4.6. Subfamily n. ------53O 4.4.6.1. Oeomydoecus ------530 4.4.6.2. ITeotrichodectes ------530 8

£age

4.4.7 • Trichodectidae ------534

5. SUMMARY ------530

6. BIBLIOGRAPHY ---_----__-_-_-_ 548

7. APPENDICES ------577 7.1. Appendix A - Data matrix ------578 7.2. Appendix B - Glossary ------597 9

LIST OF FIGURES

All legends somewhat abbreviated.

1. Edge—punched index card. 2. Layout of index cards. 3. Magner tree for hypothetical taxa A - D. 4. Cladograms for hypothetical taxa A - H, a - c and A - D. 5. Systematic position of Phthiraptera. 6. The phylogenetic position of the Liposcelidae. 7. Cladistic relationships of the major groups of Fhthiraptera. 8. Alternative systematic arrangements of the major groups of lice. 9. Structures of the trichodectid head. 10. Regions of the trichodectid head. 11. Ventral view of the three types of ischnoceran head (after Symmons, 1952). 12. Variation in form of preantennal margin of head in Trichodectidae. 13. Variation in tentorial form in the Fsocodea (after Symmons, 1952). 14. Antennae of male Ischnocera. 15* Mandibles of Trichodectidae. 16. Diagrammatic representation of generalised trichodectid thorax. 17* Apex of tibia and tarsus of two Trichodectidae. 18. Lateral view of metatarsal claws of three Trichodectidae. 19. Structures of the terminal segments of the male trichodectid abdomen. 20. Distortion of gonapophyses by mounting process. 21. Spermathecae of Phthiraptera (after Blagoveshtchenski, 1956). 22. Diagrammatic three-dimensional representation of female genital chamber and oviduct. 23. Copulation of Eutrichophilus coraiceps. 24. Diagrammatic representations of dorsum of male abdomen, showing different tergite forms. 25. Diagrammatic three-dimensional representation of male genitalia. 26. Male genitalia of Pedicuius humanus, illustrating terms used in this study. 27. Male genitalia of Microthoracius cameli. 28. Male genitalia of Enderleinellus longiceps. 29. Male genitalia of Stenopsocus stigmaticus. 30. Sagittal section of psocodean male genitalia. 31. Male genitalia of showing endophallus spiculation. 32. Male genitalia of Trichodectes canis. 10

33. Postulated, evolution of basi-parameral sclerite. 34• Abdominal setal arrangement of Trichodectidae. 35• Sitophore sclerite of Damalinia neotheileri. 36. Sitophore sclerite of Bovicola hemitragi. 37. Gladogram of Trichodectidae: initial furcations. 38. Cladogram of Subfamily n. 35. Cladogram of Trichodectini. 40. Cladogram of Felicola. 41• Cladogram of Lorisicola. 42. Cladogram of Dasyonyginae• 43. Cladogram of part of Bovicolinae. 44- -Cladogram of D. (Damalinia) and D. (Cervicola). 45• Cladogram of Damalinia (Tricholipeurus). 46. Werneck^s (1948) schematic representation of the relationships of Felicola, Parafelicola and Heofelicola. 47• Reticulate pattern of phenetic relationships of hypothetical taxa

A, Bt C, D, E and F. 48. Classification of Trichodectoidea.according to K41er (1938). 45. Classification of Trichodectiformia according to Eichler (1941). 50. Classification of Trichodectoidea according to K61er (1944). 51. Classification of Trichodecrtiformia according to Eichler (1963). 52. Classification of Trichodectoidea according to K£ler (1969). 53. Cladogram of the genera of Trichodectidae, with subfamily assignments. 54- Terga I and II of male Bovicola (n.l) hemitragi. 55- Bovicola (B.) .jellisoni female: terminalia. 56. Bovicola (L.) breviceps female: terminalia. 57. Bovicola (H.) crassipes female: gonapophysis. 58. Bovicola (B.) caprae female: gonapophysis. 59- Bovicola (B.) caprae male: abdomen. 60. Bovicola (3.) caprae male: terminalia. 61. Bovicola (B.) bovis male: terminalia. 62. Bovicola (B.) concavifrons male: termirialia. 63. Bovicola (H.) crassipes male: terminalia. 64. Bovicola (n.l) hemitragi male: subgenital plate. 65. Bovicola (B.) bovis male: genitalia. 66. Bovicola (3.) caprae male: genitalia. 67. Bovicola (3.) concavifrons male: genitalia. 68. Bovicola (B.) concavifrons male: detail of paramere and mesomere. 69. Bovicola (H.) crassipes male: genitalia. 11

70. Bovicola (n.1) hemitragi male: genitalia. 71. Bovicola (L.) breviceps male: genitalia. 72. n.2 sedecimdec embrii female: terminalia. 73. n.2 seaecimdecembrii male: subgenital plate. 74. n.2 sedecimdecembrii male: genitalia. 75. h'erneckiella ecpii male: terminalia. 76. T.Terneckiella ecrui male: genitalia. 77. n.3 traguli female: head, dorsal. 78. n.3 traguli male: genitalia. 79. Damalinia (D.) crenelata male: abdomen. 80. Damalinia (D.) theileri male: terminalia. 81. Damalinia (D.) theileri female: terminalia. 82. Damalinia (D.) baxi female: head, dorsal. 83. Damalinia (D.) appendiculata male: terminalia. 84. Damalinia (D.) orientalis male: genitalia. 85. Damalinia (D.) neotheileri male: genitalia. 86. Damalinia (D.) crenelata male: genitalia. 87. Damalinia (T.) elongata female: • terminalia. 88. Damalinia (C.) meyeri female: gonapophysis. 89. Damalinia (C.) hendrickxi female: gonapophysis. 90. Damalinia (C.) martinaglia male: scape. 91. Damalinia (C.) natalensis male: subgenital plate and pseudostyli. 92. Damalinia (0.) martinaglia male: subgenital plate and pseudostyli. 93. Damalinia (T.) indica male: terminalia. 94. Damalinia (T.) aepycerus male: terminalia. 95. Damalinia (T-) victoriae male: genitalia. 96. Damalinia (T.) indica male: genitalia. 97. Damalinia (T.) aepycerus male: genitalia. 98. Damalinia (C.) reduncae male: genitalia. 99. Damalinia (C.) hopkinsi male: genitalia. 100. Damalinia (C.) meyeri male: genitalia. 101. Eutrichophilus minor female: head, dorsal. 102. Eutrichophilus maximus female: terminalia. 103. Sutrichophilus setosus male: abdomen. 104. Eutrichophilus setosus male: genitalia. 105. Eutrichophilus moo.jeni male: genitalia. 106. Eutrichoohilus guyanensis male: genitalia. 107. Eutrichophilus guyanensis male: genitalia, detail. 108. Cebidicola armatus male: abdomen. 109. Cebidicola armatus female: terminalia. 110. Cebidicola semiarmatus female: head, dorsal. 11. Cehidicola armatus female: head, dorsal. 12. Oebidicola extrarius male: genitalia. 13. Gehidicola armatus male: genitalia. 14. Procavicola (P.) natalensis female: terminalia. 15. Procavicola (C.) lindfieldi female: terminalia. 16. Procavicola (C.) dissimilis female: terminalia. 17. Procavicola (P.) eichleri male: abdomen. 18. Procavicola (P.) vicinus male: terminalia. 19- Procavioola (P.) natalensis female: head, dorsal. 20. Procavicola (C.) dissimilis male: abdomen. 21. Procavicola (_C.) dissimilis male: genitalia. 22. Procavicola (C.) dissimilis male: genitalia, detail. 23. Procavicola (P.) pretoriensis male: genitalia. 24. Procaviphilus (H. n. baculatus female: terminalia. 25- Procaviphilus (M. n. baculatus female: gonapophvsis. 26. Procaviohilus f. granuloides female: head, dorsal 27. Procaviphilus (M. scutifer female: terminalia. 28. Procaviphilus (K. angolensis male: terminalia. 29- Procaviphilus (M. scutifer male: terminalia. 30. Procaviphilus (K. angolensis male: genitalia. 31. Procaviphilus (£• dubius male: genitalia. 32. Procaviphilus (M. .jordani male: genitalia.

33. Procaviphilus (H. jordani male: genitalia, detail. 34. Procaviphilus (M. serraticus male: genitalia. 35. Procaviphilus (£• f. granuloides male: genitalia. 36. Procaviphilus (E. neumanni male: genitalia. 37. Dasyonyx (£• dendrohyracis female: head, dorsal 38. Dasyonyx (£• v. ugandensis female: terminalia. 39. Dasyonyx (2- v. validus male: abdomen. 40. Dasyonyx (2- guineensis male: terminalia. 41. Dasyonyx (S- nairobiensis male: terminalia. 42. Dasyonyx (2- hopkinsi male: terminalia. 43. Dasyonyx (2- ovalis male: terminalia. '44. Dasyonyx (£• ruficeps male: genitalia. 45. Dasyonyx (D. minor male: genitalia. !46. Dasyonyx (2- guineensis male: genitalia. 47. Dasyonyx (2- validus male: genitalia. 48. Dasyonyx (B. ovalis male: genitalia. 49. Eurytrichodectes paradoxus female: head, dorsal. 13

150. Eurytrichodectes paradoxus male: abdomen. 151. Eurytrichodectes paradoxus male: genitalia. 152. Protelicola hyaenae male: genitalia. 153. Undescribed Protelicola species: male genitalia, 154. Lutridia exilis male: abdomen. 155- Lutridia matschiei female: head, dorsal. 156. Lutridia matschiei female: terminalia. 157. Lutridia matschiei male: genitalia. 158. Lutridia exilis male: genitalia. 159- n.4 lutrae female: terminalia. 160. n.4 lutrae male: genitalia. 161. llerneckodectes ferrisi male: genitalia. 162. Trichodectes (T. galictidis female: terminalia. 163. Trichodectes (T. canis female terminalia. 164. Trichodectes (T. emersoni female: terminalia. 165. Trichodectes (T. Piflguis female: terminalia. 166. Trichodectes (S. emeryi female: terminalia. 167. Trichodectes (n.5) zorillae female: terminalia. 168. Trichodectes (S. erminiae female: terminalia. 169. Trichodectes (S. octomaculatus female: terminalia, 170. Trichodectes (S. potus female: terminalia. 171. Trichodectes (T. canis male: abdomen.

172. Trichodectes (T. 2- euarctidos male: terminalia. 173. Trichodectes (n. ) zorillae male: abdomen. 174. Trichodectes (S. erminiae male: terminalia. 175. Trichodectes (T. emersoni male: abdomen. 176. Trichodectes (T. galictidis male: terminalia. 177. Trichodectes (n.5 ) ovalis male: abdomen. 178. Trichodectes (S. emeryi male: terminalia. 179. Trichodectes (T. canis male: genitalia. 180. Trichodectes (T. galictidis male: genitalia. 181. Trichodectes (T. emersoni male: genitalia. 182. Trichodectes (n.5) ovalis male: genitalia. 183. Trichodectes (n.5) ugandensis male: genitalia. 184. Trichodectes (n.5) zorillae male: genitalia. 185. Trichodectes (S.) octomaculatus male: genitalia, 186. Trichodectes (S.) erminiae male: genitalia. 187. Trichodectes (S.) eneryi male: genitalia. 188. Felicola (F.) zeyIonicus female: terminalia. 189. Felicola (S.) vulpis female: terminalia. 14

190. Felicola (s.) decipiens female: gonapophvsis. 191. Felicola (SO cynictis male: abdomen. 192. Felicola (SO setosus male: abdomen. 193. Felicola (SO minimus male: abdomen. 194. Felicola (EO congoensis male: abdomen. 195- Felicola (SO calogaleus male: abdomen. 196. Felicola (SO hopkinsi male: terminalia. 197. Felicola (SO helogale male: terminalia. 198. Felicola (so pygidialis male: abdominal terga 199- Felicola (so fahrenholzi male: abdomen. 200. Felicola (SO decipiens male: abdomen. 201. Felicola (sO bedfordi male: abdomen. 202. Felicola (SO acutirostris male: abdomen. 203. Felicola (SO inaecrualis male: genitalia. 204. Felicola (p.) calogaleus male: genitalia. 205. Felicola (SO congoensis male: genitalia. 206. Felicola (SO helogale male: genitalia. 207. Felicola (SO setosus male: genitalia. 208. Felicola (SO robertsi male: genitalia. 209. Felicola (SO hopkinsi male: genitalia. 210. Felicola (SO subrostratus male: genitalia. 211. Felicola (SO liberiae male: genitalia. 212. Felicola (so decipiens male: genitalia. 213. Felicola (so acutirostris male: genitalia. 214. Felicola (so vulpis male: genitalia. 215. Felicola (§0 bedfordi male:- genitalia. 216. Felicola (SO minimus male: genitalia. 217. Felicola (sO fahrenholzi male: genitalia. 218. Lorisicola (L.) m.jobergi female: abdomen. 219. Lorisicola (L.) spenceri female: terminalia. 220. Lorisicola (P.) bengalensis female: subgenital 221. Lorisicola (L.) felis female: gonapoohysis. 222. Lorisicola (P.) africanus female: head, dorsal 223. Lorisicola (L.) m.iobergi male: terminalia. 224. Lorisicola (L.) similis male: terminalia. 225. Lcrisicola (L.) spenceri male: terminalia. 226. Lorisicola (F.) acuticeps male: abdomen. 227. Lorisicola (P.) laticeps male: terminalia. 228. Lorisicola (P.) benyalensis male: abdomen. 15

229• Lorisicola- (P.) .juccii male: abdomen. 230. Lorisicola (L.) m.jobergi male: genitalia. 231. Lorisicola (L.) malaysianus male: genitalia. 232. Lorisicola (P.) laticeps male: genitalia. 233. Lorisicola (L.) spenceri male: genitalia. 234. Lorisicola (P.) bengalensis male: genitalia. 235. lorisicola (P.) .juccii male: genitalia. 236. Lorisicola (P.) acuticeps male: genitalia. 237. Neotrichodectes (T.) barbarae male: temple margin. 238. ITeotrichodectes (n.7) semistriates male: flagellum. 239. ITeotrichodectes (N.) mephitidis female: abdominal pleura II and III.

240. ITeotrichodectes (T.) barbarae male: anterior abdominal segments. 241. ITeotrichodectes (IT.) mephitidis male: terminalia. 242. Neotrichodectes (N.) mephitidis female: terminalia. 243. ITeotrichodectes (n.7) chilensis female: terminalia. 244. TTeotrichodectes (T.) barbarae female: terminalia. 245. ITeotrichodectes (L.) gastrodes female: terminalia. 246. ITeotrichodectes (n.6) pallidus female: terminalia. 247. ITeotrichodectes (n.6) pallidus male: genitalia. 248. ITeotrichodectes (N.) mephitidis male: genitalia. 249. ITeotrichodectes (L.) gastrodes male: genitalia. 250. ITeotrichodectes (T.) barbarae male: genitalia. 251. ITeotrichodectes (n.7) chilensis male: genitalia. 252. (T.) potteri female: antenna. 253. Geomydoecus (T.) asymrr.etricus male: anterior of abdomen. 254. Geomydoecus (G.) californicus female: terminalia. 255. Geomydoecus (T.) minor female: abdomen. 256. Geomydoecus (G.) actuosi male: genitalia. 257. Geomydoecus (G.) thomomyus male: genitalia. 258. Geomydoecus (T.) minor male: genitalia. 259. Geomydoecus (T.) wardi male: genitalia. 260. Phylogenies of hosts and associated parasites, showing speciation of the parasite independent of host speciation. 261. Phylogenies of hosts and associated parasites, showing host • speciation without parasite speciation.

262. Phylogenies of hosts and associated parasites, showing secondary

absence of parasites. 263. Distribution of eight species of Geomydoecus. 264. Diagrammatic representation of the factors influencing host specificity of parasites of birds and mamrrals. 16

265• Relationship between percentage of monoxenous species in families of parasites and the independence score1. 266. Cladogram of living Artiodactyla. 267. Phylogeny and classification of the Bovidae. 268. Fhylogeny and classification of the 3ovidae. 269. Cladogram and host associations of D. (Damalinia). 270. Cladograms of lice and their hosts, (a) Cladogram of Alcelaphinae; (b) cladogram of the Damalinia theileri - ornata clade. 271. Cladogram and host associations of Damalinia (Cervicola). 272. Two alternative possible phylogenetic associations between lice in Damalinia (Cervicola)and their hosts. 273. Cladogram and host associations of Damalinia (Tricholipeurus). 274. Cladogram and host associations of Bovicolinae other than Damalinia. 275. Cladograms of living (and one recently extinct) species of Squus. 276. Phylogeny of living Carnivora. 277. Cladogram and host associations of the Trichodectini. 278. Cladogram and host associations of Trichodectes (Stachiella). 279« Cladogram and host associations of Felicola. 280. Cladogram and host associations of Lorisicola. 281. Cladogram and host associations of Subfamily n. 282. Cladogram of the Eutheria. 283. Primary host associations of the major clades of Trichodectidae. 17

LIST OF TABLES

All legends somewhat abbreviated.

I. Data matrix for hypothetical taxa A, B, C and D. II. Distribution of apomorphies within Psocodea. III. Distribution of apomorphies within Phthiraptera. IV. Distribution of abdominal pleural projections in Trichodectidae. V. Distribution of ovipositor elements within the Psocodea. VI. Distribution of number of pairs of abdominal spiracles in the Trichodectidae• VII. Generic concepts in the Bovicolinae. VIII. Generic concepts in the Trichodectini (plus Neotrichodectes and Trigonodectes)• IX. Generic concepts in the Pelicolini (plus Protelicola). X. Numbers of trichodectid species and subspecies parasitising individual host taxa. XI. Percentage of species and subspecies of Trichodectidae parasitising various host groups in different classes of number of hosts attacked. XII. Sister-species of Trichodectidae parasitising the same primary host species. XIII. Possible cases in the Trichodectidae of louse speciation without concomitant host speciation. XIV. Secondary infestations in the Trichodectidae. XV. Percentage of species of families of insects parasitic on mammals and birds in different classes of number of hosts attacked. XVI. Degree to which parasite location is determined by host location in nine families of parasites of birds and mammals. 18

INTRODUCTION

There sore four major groups of lice, the Amblycera, Ischnocera, Rhyncophthirina and Anoplura, the first three of which are sometimes united into the group ''. The 'Mallophaga' ('chewing or biting lice') and Anoplura ('sucking lice') are accorded ordinal status by some workers, but others consider all lice to belong to a single order, the Phthiraptera. Trichodectidae are a family of ischnoceran chewing lice parasitic on mammals. There are 351 described species and subspecies grouped into between 13 and 39 genera, various workers having widely differing views on generic limits. The classification of the Trichodectidae at the generic level is perhaps more confused than that of any other lice (Hopkins, 1949> 1960; Emerson & Price, 1981) and no workable keys to genera are available for the family. The confusion and disagreement surrounding the classification of the family persists despite a-sound basis of taxonomic knowledge at the species level, derived largely from the works of Wemeck (1948, 1950), although this author did not attempt to produce any keys. One of the problems to be addressed in this study is therefore the classification of the Trichodectidae, and the production of a key to genera and subgenera. Lice are obligate ectoparasites of birds and mammals, feeding on feathers, scurf, dermal secretions or blood. Their entire life, from egg to adult, is spent on the host, and they cannot live for long away from the environment produced by the skin and fur or feathers of the host (the 'dermecos'). There is no cyclical change of host, and.the majority of louse species are known only from single host species. The high level of host specificity and the clear dependence of the lice on the dermecos might be expected to lead to a high level of co-evolution of lice and their hosts. Such co-evolution would result in relationships between host species being reflected in parallel relationships between their lice. This assumption has been formalised in 'Pahrenholz* Rule', which states that the phylogenies of host and parasite are topologically identical because of co-evolution and co- speciation. This Rule (and its unwritten precursors) has been accepted 19

as axiomatic by many taxoxiomists and systematists of lice for at least the last fifty years, and classifications proposed for lice have been strongly influenced by the classifications of the hosts. The corres- ponding view, that host relationships will be indicated by the relation- ships of their parasites, is also held, and proposals for the reclassi- fication of hosts have .been made on the basis of louse relationships (e.g. Webb, 1949; Timmermann, 1957). In some cases this argument has been reduced to complete circularity, the postulated relationship of the hosts being used to claim louse relationship, which is in turn used to support the initial hypothesis of host relationship (e.g. Traub, 1980). To test Pahrenholz1 Rule adequately, and thereby to test the basis of much louse classification, a louse phylogeny derived without the use of host date must be compared with such phylogenetic information as is available for the hosts. To this end, a phylogeny of the Trichodectidae, which parasitise Carnivora, Rodentia, Hyracoidea, Artiodactyla, Perissodactyla, and Edentata, is prepared in this study.

No phylogeny can be compiled without accurate morphological information upon which to base hypotheses of homology. The • published morphological work on Phthiraptera is extremely scanty and scattered and, in many cases, contradictory. It was therefore necessary to under- take a morphological study of Phthiraptera with particular reference to Trichodectidae (although this is necessarily limited.In scope). In order to determine .the applicability to Trichodectidae of observations made on the morphology of lice and other insects, the phylogenetic position and relationships of this family must be known. Because of the considerable differences in opinion expressed in the literature regarding the phylogeny and higher classification of lice, these matters, are also examined in some detail. The objectives of the study may therefore be summarised as follows, in their necessary order: (1) To determine the phylogenetic position of the Trichodectidae with respect to other lice, and the position of the lice with respect to other insects (section 2.1.). 20

To examine the morphology of the Phthiraptera with special reference to the Trichodectidae, and.thereby establish a basis for character survey and analysis (section 2.2.)* To produce a cladogram of the Trichodectidae, from which phylogenetic deductions could be made (sections 2.3« and 2.4.). To produce a revised classification of the Trichodectidae based on the cladistic relationships established, and to describe and key all genera and subgenera (section 3»)« To compare the phylogenetic relationships of the species of Tricho- dectidae (derived from the cladogram) with such phylogenetic relationships as are known for the hosts, and thereby to evaluate Pahrenholz* Rule and other predictions of co-evolution (sections 4.2. and 4.4.). As a corollary to objective (5), to examine the phenomenon of host specificity with particular reference to lice (section 4.3*) • SECTION 1

MATERIALS AND METHODS 22

1.1. LITERATURE The major source of taxonomic and descriptive information on the Trichodectidae is a series of three papers by Werneck (1941, 1948, 1950), in which all species of Trichodectidae described at the time are reviewed. The first of these papers is restricted to the Trichodectidae parasitic on hyraxes, the latter two (Os Malofagos de Mamiferos parts I and II) deal with all the rest. Since 1950 only two genera have been revised: Geomydoecus by Price 2c Emerson (1971, 1972) with a subsequent series of papers by Price and his co-workers, and V/erneckiella by Moreby (1978). Other papers have been restricted to descriptions of small numbers of new species, these sometimes being keyed or placed into species groups. A list of all species and their synonyms is provided by Hopkins 2c Clay (1952), with supplements in 1953 and 1955; an updated list, including all generic placements and synonyms of species, is being prepared by the present author. Lists (arranged by host) of the hosts of all Tricho- dectidae have been published by Hopkins (1949) and Emerson & Price (1981). Some of these works contain a considerable amount of information' on other aspects of the family and order, mostly morphological in nature. Additional information on taxonomy, morphology, embryology, distribution and ecology is contained in general review works and numerous papers scattered through the scientific literature. A literature survey was carried out using primarily the bibliographic resources of the British Museum (Natural History). References to scientific papers of relevance were entered on edge-punched index cards. Each card bore the author(s), date, title and source of the paper, the library serial number of the latter and an indication of the contents entered as one or more of 68 categories punched around the edge (Fig. 1). This prevented the inevitable duplication of references necessary in a standard card index (68 of which would have been required) and permitted rapid retrieval of relevant references. To facilitate the taxonomic part of the study two card indices were compiled to the species and genera of Trichodectidae. One contained all specific and generic names available, placed in alphabetical order. Each card bore the name, author, date of description and junior synonyms and homonyms, or, if the species or genus was itself a junior or homonym the identity of the senior synonym or H"* C C IC c: r G|C cloTo C C r C r t » > c r c NEARCTIC • C: NE0TR0PIC n o f rv PALAEARCTSCo 0 fc- O -t AUSTRALIAN _ f) I-Cl ^z. » A O H« S » Ct3 g o n r 70 O f. J o \ c- P O o c ' ? r-, „ VI vi r> G vl 0) A— O d ? n - O O 5 O Cd Ui P P- CL = o H- 5 O d S G p IO « _dj O o S G p P- 5 O d F- CO P CD C P^ (tr» H- ijs-. 1 o d o 2 o P _oi d O cP+ A $ t-O-^L 2 o i

CO s o C O" p «•* o jQi xi r*- i-G

C

^ f G G C 0 G G c; c g q o q o q o q c o q c> c o o c c oilY , — 24

homonym, the reference to the synonomy and the full reference to the original description of the species or genus. For species, all genera under which the species had been placed were also listed (Fig. 2a). The other index contained all currently used specific and generic names (i.e. all species and genera other than junior synonyms and homonyms) arranged with species in alphabetical order behind the relevant (tabbed) generic cards, which were themselves in alphabetical order. Each card bore the name, author(s), date of publication of and reference to the original description; in addition the generic cards bore the name of the species of the genus and the species cards the identity of the host or hosts from which the type material of the species had been collected (Fig. 2b). An index was also maintained to the hosts, each card bearing the name of the host mammal and the lice found upon it: this was compiled from the collections of the British Museum (Natural History) and information in the literature, primarily Hopkins (1949) and Emerson & Price (1981). 25

I

ocellata Piaget, 1880. Trichodectes descr. as subsp. of parumpilosus Bovicola Piaget Damalinia Good species (Harrison, 1916) (Uerneckiella) J. syn.: VJ. equi asini Eichler Werneckiella

a

heaneyi Timm & Price, 1980. (Geomydoecus) J. Med. Entomol., 17(2):136. Type host: Geomys bursarius texensis

Fig. 2. Layout of index cards, a) example of card from alphabetical index of taxa; b) example of card from systematic index of taxa. 26

1.2. SPECIMENS

1.2.1. Introduction The number of specimens examined is given for each species in the check-lists below (section 3). Fourteen thousand, two hundred ana eighty-six specimens representing 327 of the 351 valid species and subspecies of Trichodectidae were examined during the study. Most of the material studied is in the collection of the British Museum (Natural History), but the collections of the National Museum of Natural History (V/ashington), the University of Minnesota, and Oklahoma State University were also examined.

1.2.2. Nature of Study Material

Virtually all taxonomic work on the Phthiraptera is based on the study of cleared specimens mounted in canada balsam on glass microscope slides, and utilises characters made available by this preparative technique. Most of the specimens available for study are prepared in this manner, more than three-quarters of the described species of Trichodectidae being represented in museum collections only by slide- mounted specimens. Specimens mounted in this way may be surveyed for characters of the dorsal and ventral surfaces, genitalia, and internal sclerotisations. However, it is implicit in any preparative technique that as some features are made more accessible for study others are made less accessible or obliterated. The characters made inaccessible by mounting lice on slides are: internal membranous structures, lateral features, membranous surface projections and depressions, setal orientation and details of surface sculpture. Despite this disadvantage, the characters used in a major systematic study such as this must be almost exclusively those available in slide-mounted specimens. Any other course would require a major collecting exercise followed by a taxonomic • revision of the family at the species level, incorporating novel characters made available by different preservative techniques. The limitations imposed by the nature of the material are not as great as might be imagined, however; use of high-power phase-contrast compound micro- scopes made possible the study of many characters that previous workers had apparently not detected. In addition, freshly-collected material 27

and unmounted specimens from museum collections were used for detailed morphological studies elucidating features seen in slide-mounted specimens, during which novel characters were detected which could then be identified on conventionally-mounted specimens.

1.2.3. Collection and Preparation of Specimens

During the course of the study a number of specimens were collected and prepared for examination. An outline is given of the methods employed so that sampling difficulties in col-lection and the nature of characters lost to specimens during preparation might be appreciated.

1.2.3.1. Collecting

Lice were collected from both living and dead mammalian hosts, the latter being either freshly dead, frozen, long-dead and dried (i.e. museum skins) or preserved in alcohol or formalin. One source of error in louse-host associations is the post mortem contamination of hosts by lice from other hosts. To avoid this, bodies and skins were, when possible, kept separate from one another and when collected, individually placed in plastic bags for subsequent treatment. All equipment used in the collection of lice from the hosts was thoroughly cleaned between treatment of different hosts. Several collecting techniques are available and these are described below; in each case the circumstances in which the method is best employed are given, a. Searching (Suitability: living and dead hosts, all preservation types; especially live hosts and museum specimens) The fur of the is carefully searched and the lice picked out. The method is by far the least efficient of the available techniques and the yield generally low, except in cases of very heavy infestation; Hopkins (1949) notes that searching three skins of Madoqua so. (dik-dik) produced 33 specimens of Damalinia victoriae; subsequent dissolving of the hair (see below) yielded a further 908 specimens. Despite this drawback, the method is widely applicable, in that it is usable on dead hosts however prepared, and is the most suitable for use on live hosts 28

(see comments on insecticide use below). Live hosts (deer, cattle, horses, goats, dogs and cats were examined in this study) are tied or held whilst being searched. Dead hosts and skins are searched over a white plastic sheet, so that any lice falling from the hair can be seen and collected. The sheet is washed between host specimens. The suitability of the technique for museum specimens lies in the lack of damage to the skin; much less than beating (see below) and of course dissolving. b. Use of temperature gradient (Suitability: live hosts; very freshly-dead hosts) Lice are usually found within a few millimetres of the skin of the host, and maintain that position, at least partially, through sensitivity to a temperature gradient between the skin and the exterior of the pelage (Murray, 1957a). Should the exterior.of the pelage be warmed, the lice will venture out from the skin. Bertram et al. (1952) used an electrically-heated pillow placed over a pad of surgical lint laid onto the coat of cattle to collect , which was found to congregate on the lint. This technique was not deliberately employed in this study, but in one case a young roe deer was held for a considerable time and numerous lice transferred to the aims of the person holding the animal to be subsequently collected. c. Brushing (Suitability; freshly-dead hosts and fresh skins; preserved skins, if dry) The method may cause some damage to more delicate skins, and was consequently not generally employed for museum skin collections. Parasite specimens are also sometimes damaged, especially if they died by desiccation and are dry and brittle. If the lice are still alive (i.e. on a freshly-dead host or fresh skin), they are first killed by placing the body or skin in an airtight box with a small amount of chloroform or other suitable insecticide. The chloroform is soaked into a small pad of cotton wool, which is not allowed to come into direct contact with the host. The body or skin is then held over a sheet of white plastic and brushed by hand against the lie of the fur. Lice are then picked from the plastic sheet. 29

d. Dri-die (Suitability: freshly-dead hosts and fresh skins) A silica aerogel insecticide, "Dri-die" (Fairchild Chemical Corporation) is frequently used for the collection of bird ectoparasites (Keirans, 1967; 7/at son and Amerson, 1967) • This substance absorbs the waxy layer of the cuticle, causing rapid dehydration and death. Its irritating properties also cause lice to release their hold on the feathers and fall from the host. Although of proven use on birds the substance may not be suitable for use on mammals (especially valuable livestock and family pets) because of the unknown consequences arising from ingestion of the chemical. The substance was used in this study as a substitute for chloroform on the corpses of large mammals, particularly badgers and deer. e. Dissolving (Suitability: any skins not required to be kept ) The method is described by Hopkins (1949). The skin of the host is cut into pieces 1 or 2 inches square. These are soaked in 5% cold NaOH for about 15 minutes or until the hair can -be scraped off the hide with a blunt knife (if the skin is macerated too much it takes on a jelly-like consistency that increases the difficulty of removing the hair). The partially-dissolved hair is then.boiled in 5% NaOH until completely dissolved. This mixture is filtered and the lice recovered. Ordinary filter-paper in a BUchner funnel and flask connected to a vacuum pump was used in this study; the lice are removed from the filter-paper by washing and picking off by hand under a dissection microscope. Hopkins (1949) and Cook (1954) recommend using a fine- mesh stainless-steel or bronze screen which may be folded to a conical shape or used flat in connection with a Btlchner funnel. Though the metal screen is probably superior to the filter-paper, the latter was employed in this study because the equipment was readily available and the technique was not used frequently. Cook (1954) suggests a refine- ment of the technique by which the hide is first digested in trypsin to avoid damage and loss of lice when the hair is removed. The advantage of the dissolving method is that all lice on the skin are recovered. 30

The lice collected, however, are macerated by the hydroxide and suitable only for slide-mounting. As the main purpose in collecting fresh specimens in this study was to provide an alternative to slide-mounted material, the method was not much employed. Some specimens were collected alive for microscopic study whilst others were killed or already dead on collection. Those that were killed or were found freshly-dead were collected into 70?!- ethyl alcohol, whilst those that had died by desiccation (such as those found on dried museum skins) were collected dry. Ethyl alcohol fixes the muscles of Phthiraptera making manipulation and maceration after prolonged immersion increasingly difficult. This problem may be ameliorated to some extent by the addition of a small amount of glycerine to the alcohol before use. It was found that desiccated specimens collected from dried skins (some over 100 years old) could be mounted very easily and made very good preparations, perhaps better than those preserved in alcohol. The only benefit of alcohol preservation seems to lie in protection from mechanical damage.

1.2.3.2. Mounting

Specimens were mounted on glass microscope slides for detailed examination. Temporary mounts of fresh or macerated specimens (see below) were made in lactophenol; the procedure for preparation of permanent mounts is set out below. Numerous mountants are available for small insects, each with slightly different properties. The mountant generally used for Phthiraptera is Canada balsam, and several authors have stressed its suitability, reliability and durability over others (Palma, 1978; Ledger, 1980). This mountant was used, accordingly, for specimens prepared during the study. Mounting specimens may be considered in several stages: a. Maceration The specimens are placed in 10% KOII for about 24 hours at room temperature. About 30 minutes after the specimens are placed in the hydroxide the abdomens are pierced to allow entry to the KOH and later expulsion of the macerated abdominal contents. This cannot be carried out later in the procedure as the integument becomes elastic and difficult to pierce. 31

b • Evacuation When the specimens are sufficiently macerated they are placed in distilled water and the macerated contents carefully squeezed out. Great care must be taken to avoid brushing against the setae as these are easily broken off. This process is best carried out with two mounted micropins, with each point bent to a hook shape. c. Staining If the specimens are colourless and require staining, they are washed in 1.0% KOH, transferred to diluted safranin 0 stain and left for 20-30 minutes. Overstaining can be corrected during stage (e) by increasing the period of immersion in the alcohol. d. Neutralisation. If the specimens are not to be stained they are transferred from distilled water (stage b) to 10% acetic acid and left for 30 minutes. e. Dehydration (Prom stages (c) or (d)). The specimens are passed through 40%, 80%, 90% and 95% ethyl alcohol, remaining in each for about 15 minutes. f. Clearing The specimens are transferred to terpineol or clove oil and left for about 30 minutes, though longer immersion does not harm them. If the male genitalia or any other part of the insect is to be removed for separate mounting, that is done at this stage as the specimens are flexible and there is little danger of the dissected part getting lost. g. Mounting A small amount of canada balsam is placed on a coverslip and spread with a glass rod to cover the whole area. A single specimen (or, a male and a female) is placed briefly on . bristol-board to dry it, then placed into the balsam and pressed with a fine needle so that it sinks. It is arranged so that in the final mount the long axis of the louse is pointing across the slide. A small spot of balsam is placed centrally on a microscope slide, which is then touched to the cover-slip. The coverslip and specimens adhere to the slide and the specimens remain in position. When a long series of specimens are mounted, some are mounted with the dorsal surface upwards and some with the ventral. If only one specimen of a sex is available, the convention is to mount 32

males with the dorsal surface upwards and females with the ventral surface upwards. h. Drying The slides are labelled (2-ply bristol-board is used for the labelling; this also acts as spacing when the slides are stored vertically) and placed in a drying oven at 37 degrees Centigrade for three weeks. Slides should be placed in the oven as soon as possible after being made, to facilitate drying.

1.2.3.3. Observation

Live specimens and dissections were studied using a Vickers 'Steros II1, dissection microscope. Slide-mounted specimens were studied using a Wild Mil compound microscope with phase contrast and a Leitz 'Laborlux 12* compound microscope with phase contrast. Illustrations were prepared using a camera-lucida attachment to the Laborlux, and drawn onto tracing paper. Figures of all or part of the abdomen, and of the head, are drawn so that the right-hand side is a ventral view and the left-hand side a dorsal view, unless otherwise stated. A scale line of 50ym is given for each figure. 33

1.3. SYSTEMATIC ANALYSIS AND CLASSIFICATION

1.3.1. Introduction

Before discussing the character survey carried out in this study it is necessary to consider the concept of homology and to describe the analytical techniques applied to the characters, as the demands of the analysis inevitably dictate to some extent the type of data collected and the manner of their collection. In this study the term 'holophyletic* will be used to denote groups that include an ancestral species (known or hypothetical) and all of its descendants. The choice of this term rather than the more familiar •monophyletic1 is made because of the confusion surrounding the latter term, which is used by many workers of the 'evolutionary school1 to refer to both holophyletic and paraphyletic groups. This and other terns used in the systematic discussions are defined in the glossary (Appendix B).

1.3.2. Homology

The teim homology is used widely in systematics and morphology to refer to the property of 'sameness1 of features in different organisms. A widely-accepted definition of homology is that of Simpson (1961): "resemblance due to inheritance from a common ancestor", but this includes two rather different concepts: phylogentic relationship and phenetic resemblance (referred to below as phyletic and phenetic homology respectively). Phyletic homology has been defined as "the relationship which pertains between features (or states of features), present in two or more organisms, that may be traced back to the same feature (or state) in the immediate common ancestor of those organisms" (Bock, 1969; Mayr, 1969; Scudder, 1971). Phenetic homology may be defined as "the relationship which pertains between features (or states of features), in two or more organisms, that exhibit compositional and structural correspondence" (emended from Sneath Cz Sokal, 1973). In this study the concept of homology is employed in the morphological discussions and the systematic analysis. Because the systematic analysis is based on conclusions reached in the morphological study, the same 34

concept must be used throughout. The systematic analysis is intended to determine the phylogeny of the group, so it would seem that phyletic homology should be used. Sneath & Sokal (1973) maintain that homologies, if used for phylogeny construction, must be defined in terms independent of phylogenetics, i.e. phenetics, and would therefore recommend phenetic homology. If a postulation of homology (of a feature found in two or more taxa) is not also a postulation of holophyly of the taxa concerned hov/ever, then a dendrogram constructed using homologies.cannot show holophyletic groups and a phylogeny cannot be constructed. _ Basing phyletic analyses on phenetic homologies can therefore only be justified if there is an exact equivalence between phyletic and phenetic homologies. The purpose of the following discussion is to determine whether such an equivalence exists and, if not, how phyletic homologies can justifiably be postulated. Both 'types* of homology must be ba.sed on phenetic observations, as pointed out by Jardine (1969) and Sneath & Sokal (1973). The following categories of observation have been suggested as important in the determination of homology (sensu lato) of morphological features: i) Comparison of structural components and anatomical relationships; ii) Examination of ontogeny; iii) Comparison of genetic basis.; iv) Examination of palaeontological history. These categories are discussed in detail below, and the conclusions possible within phyletic and phenetic methodological frameworks are indicated.

i) Comparison of structural components and anatomical relationships. It is self-evident that comparisons can only be made in terms applicable to the features being compared. Sneath & Sokal (1973) propose that the terms themselves be considered as homologues. Thus, dimensional concepts such as 'length* and 'weight' are referred to as 'definitional homologies'. It is considered here that the application of the term 'homology' to general concepts that are defined independently is unnecessary. The term 'definitional homology' is also applied by Sneath & Sokal (1973) to compositional factors. Thus, the presence of a substance such as 35

haemocyanin in two organisms would be considered homologous by definition. Such a correspondence would not be considered as a phyletic homology without further investigation, and Florkin (19&2) proposes the term !isology' for chemical correspondence where little is known of evolutionary origins. Features may be compared with respect to their biological or chemical constituents, their structure, their spatial relationships with other features, and their function. If two features are compared phenetically on the basis of any or all of these factors, the results of that comparison will be in the form of an index of similarity. This index will indicate a position on a continuum of (phenetic) relationship in which total identity is at one extreme and total dissimilarity at the ether. Sneath & Sokal (1973) propose that the general region of high similarity be termed homology. The term homology1, however, if unqualified, is meaningless in this context, and an indication must be given of the level.of similarity deemed significant (or acceptable for comparative purposes). In the following discussion the term 'degree of isomorphism1 is used to refer to phenetic homology derived in this manner. It is clear thai there is no theoretical limit to the 'amount' of evolutionary change possible for a feature. It cannot, therefore, be proposed that features must display a given degree of isomorphism before they can be considered as phyletic homologies. Phyletic homologues may have a high degree of isomorphism, but it is not a necessary concomitance. 3y this criterion of phenetic homology phyletic homologues are not equivalent to phenetic homologues. To find the phenetic homologue of a feature of one organism in a second, the feature is compared with all potentially homologous features derived from the second organism and homologised with the one with which it displays the highest degree of isomorphism (Sneath & Sokal, 1973). Alternatively, all potentially-homologous features derived from both organisms may be compared simultaneously and the features paired in such a way that the summed similarity values are maximal (Jardine, 19&7, 1969; Key, 1967). Key (19^7) proposes an operational definition of phenetic homology that follows the latter principle; "Feature a^ of organism A is said to be homologous with feature b-, of organism 3 if 36

comparison of a^ and b^ with each other, rather than with any third feature, is a necessary condition for minimising the overall difference between A and 3". Feature b^ is not, therefore, necessarily the feature of B that in itself exhibits the highest degree of isomorphism with a^. Phenetic homology, if determined according to Key's definition, will probably equate almost exactly with phyletic homology. In phylogenetic studies decisions are usually reached on the basis of minimum assumptions, as in the phenetic homology of Key and Jardine. Key's definition can be recast to provide an operational definition of phyletic homology, as follows: "Feature a^ of organism A is said to be phyletically homologous with feature b^ of organism 3 if comparison of a^ and b^ with each other, rather than with any third feature, is a necessary condition for minimising the number of character state changes necessarily postulated in the derivation of A and B from a postulated most recent common ancestor". This definition does not demand that A and B are considered in isolation and, as shown below, evidence from other organisms (taxa) is permissable in the determination of phyletic homology. If two taxa are very similar to one another (i.e. very little divergence has taken, place since their postulated common ancestry), features may be assumed to be phyletic homologues on the basis of a high degree of isomorphism, because it is likely that overall parsimony will be maintained in this way. This assumption can be made in certain cases even if the taxa are fairly dissimilar to one another; for example, the head of a cockroach is probably a phyletic homologue of the head of a flea, despite the low level of similarity of the two taxa. Thus, if the degree of isomorphism is sufficiently great at the level of the feature to be homologised, the homology can be postulated even if the degree of isomorphism is very low at a more detailed level. Where the degree of isomorphism is not great at the level under consideration (i.e. there has been great divergence since the postulated common ancestry of the organisms), other taxa believed to be of the same phyletic line may be used to identify homologies. Examination of many taxa of the same phyletic line may reveal the presence of transformation series 37

that enable phyletic homologies at each end of the series to be recognised as such, despite their low degree of isomorphism. These other taxa would be irrelevant in any derivation of purely phenetic homology. Recognition and study of taxa. related phylogenetically to those under consideration may be of additional value in the determination of phyletic homology. Matsuda (1976) points out that a proposal of phyletic homology, like other scientific theories, is a matter of probability. The probability of a given feature in two or more taxa being phyletically homologous is proportional to its frequency of occurrence in other taxa of the same phyletic line. 'When a feature occurs in all or nearly all members of a given phyletic line the feature is likely to have been inherited from a common ancestor (i.e. homology is very probable). Conversely, the probability of two features being homologous becomes lower with the absence of the feature in some members of the line (Matsuda, 1976). The element of circularity possibly inherent in any phylogenetic statement based on- homologies is particularly apparent in Matsuda's argument. This Circularity1 is not necessarily a weakness of the system, however. The initial hypothesis of homology is necessarily a judgement based on phenetic grounds, upon which a further hypothesis of phylogeny is made. This hypothesis is then compared to others concerning the same organisms but based on different supposed homologies and the probability of each homology being Correct' is evaluated with reference to all the others. The initial hypothesis of homology may then, if necessary for reasons of parsimony, be changed. It is possible that no method of determination of homologies is entirely inference-free, and in every case the determination relies upon initial hypotheses that may subsequently be tested (Hull, 1968; Sneath 2c Sokal 1973).

ii) Examination of ontogeny. Ontogenetic processes may be homologised in their own right as discussed above, or may be used to indicate homologies in a later developmental! stage. It is in the latter sense that they will be discussed here, but only in relation to phyletic homology, because for phenetic homology the addition of a temporal 38

dimension to a feature is of significance only in that it permits comparison in four dimensions rather than three. Thus ontogeny may he considered in addition to the structure and relationships of the final feature to determine phenetic homology (maximum isomorphism) of two features. Identity of ontogeny is sometimes considered necessarily concomitant with phyletic homology (Abercrombie, Hickman & Johnson, 1951; Cans, 1969). Matsuda (1976), however, proposes a new term, 1euhomologyt, for phyletic homology in cases where the ontogeny (and primordia) of the structures concerned-are (phenetically, and presumably phyletically) homologous. Before considering whether phyletic homologies must be euhomologies, it should be deteimined if there are any conflicts between phyletic homologies as recognised by the criteria set out in (i) above and euhomologies. Matsuda (1976) lists and discusses numerous cases in the Insecta where this appears to be the case, for example: "different developmental processes forming the posterior abdominal segments including the cercus 7 different developmental processes producing the stylus...., different rudiments forming the ovipositor valvulae different developmental processes forming the gonangulum ...... 15 different developmental processes forming the common oviduct and the lateral •oviducts and the speimatheca". Each of these features is, by the criteria discussed in (i) above, phyletically homologous (in • the imago) throughout the Insecta, but is patently not euhomologous. It is apparent that determination of euhomology would be extremely difficult in any large taxonomic group such as the Insecta where the ontogeny of most members is not known. It is possible to regard phyletic homology as determined by the criteria set out in (i) as a preliminary hypothesis to be tested by work on ontogeny, but in most cases this testing would never be done, and the concept as a basic tool of morphology and systematics is therefore somewhat devalued. The definition of euhomology is not derivable from the definition of phyletic homology in terms of descent from a common ancestor, however, and there is no theoretical requirement for phyletically homologous features to be euhomologous. Such a theoretical requirement has been proposed in the form of the biogenetic law of Haeckel, v/hich states 39

that the ontogeny of an organism is a sequential recapitulation of the adult forms of the organism* s ancestors (de Beer, 1951, 1958). This 'law' is not now considered to hold, though a recapitulation of the ancestral state of a feature may be observed in ontogeny (de Beer, 1951, 1958; Matsuda, 1976). Natural selection must act not only on the imago, but also on the pre-imaginal forms, which bear the precursors of imaginal features (phyletic homologies). It is possible, therefore, for natural selection to act in such a way that such precursors are modified in individual lines,, giving rise to different ontogenies for phyletically homologous end-features. Such difference in ontogeny may be variation in form, time of development (heterochrony), position of primordial cells, or the involvement of different cells as primordia. The requirement that, whatever the path of development, the same cells be involved as primordia (Scudder, 1964) implicitly proposes a determinate embryo (and ultimately, egg), though it is well-established that determination occurs to very different extents in different embryos (Bodemer, 1968). Such a requirement is, as with euhomology, not implicit in the definition of phyletic homology. Euhomology is not, therefore, necessarily concomitant with phyletic homology, so that whilst similarities in ontogeny may provide support for a proposal of homology, lack of^ correspondence is not necessarily evidence against such a hypothesis (Simpson, 1961; Bock, 1969; Scudder, 1971; Matsuda, 1976). Ontogeny may be used to detect phyletic homology even in the absence of euhomology. As Matsuda (1976) points out, during development structures are usually less fused, reduced, differentiated and shifted in position than in the adult (von Baer's rule). Therefore the developmental sequence of structures tends to show intermediate conditions between the least modified (at the beginning of development) and the most modified (in the adult). This array of states helps to identify the adult structures and hence homologies.

iii) Comparison of genetic basis. The genotypes of different organisms may be homologised in their own right by comparison as discussed in section (i) above, although in practice this comparison is generally of inferences drawn from observations of the karyotype and phenotype under experimental conditions. In addition, phenetic homology of a 40

phenotypic feature may include homology of the relevant portion of the genotype. The definition of phyletic homology does not demand (phyletic) homology of the genetic basis of phyletic homologues, and there is some evidence that such genetic homology is not necessarily concomitant with phylet'i.c homology of phenotypic features. Sturtevant (1929) notes considerable disturbance in the chaetotaxy of the F^ hybrids of Drosophila melanogaster and D. simulans, despite the identical chaetotaxy of the two parental species. Rendel (1959? 1967) investigates the effect of selection on the number of scutellar bristles in D. melanogaster. He reports the existence of a 'plateau1 at the wild-type number, over which a constant selection pressure has a depressed effect on the phenotype. Rendel identifies the genetic basis of bristle number as a single major gene modified by a number of minor genes; suppression of the minor genes has a much greater effect if the bristle number differs from the wild- type than if it does not. Mayr (1963, 1970) and Matsuda (1976) use this observation as evidence that genetic 'substitution' can occur without phenotypic change, although it seems from Rendel's paper that the bristle number is always under the prime control of the same major gene more or less repressed by the activity of minor genes. The precise genotype (i.e. of the minor genes) controlling the development of a phenotypic feature may differ independently of the phenotype, however, so two organisms (and hence two species) may have a different arrangement of minor genes controlling the development of the same phenotypic feature. It is possible that this is the explanation of the results obtained by Sturtevant (1929) mentioned above. Precise genotypic homology is therefore not necessary for phyletic homology of a phenotypic feature. In practice, i the determination of genetic homology for every phenotypic feature used in systematic analysis is not realistic.

iv) Examination of nalaeontological history. For determination of phenetic homology (as discussed in (i) above) it is immaterial whether the organisms studied are modern or fossilised. Determination of phenetic homology of features of two or more modern forms is not affected by information derived from fossils of the putative ancestor of those forms, however, unless phyletic homology is invoked to explain the relevance of those fossils. 41

As noted in (i) and (ii) above, phyletic homology of features (in modern forms) may be obscured by great divergence in the phenotype of one of the forms concerned. In such cases fossils may provide intermediate stages in transformation series enabling homology in the modern forms to be recognised. In addition, study of the relevant fossils may provide evidence of homoplasy in modern forms (Hennig, 1966; Patterson, 1981). The fossil record is limited in its applicability to the determination of phyletic homology by its lack of completeness, the modifications to specimens during fo.ssilisation, and by the problems introduced by the presence of stem-groups. It is.not possible to determine the polarity of transformation series unless the fossil record is believed to be complete (as in the Foraminifera, for example) (Patterson, 1981). Not only is the fossil record likely to be incomplete in most cases (it is non-existent for the Phthiraptera), but fossil specimens are never as complete, or can provide as much information as, modern specimens (Hennig, 1981; Patterson, 1981; see Seeger in Hennig, 1981 for a dissenting view). The uncertainties attached to fossils that have undergone numerous and ill-understood taphonomic processes (Olson, 1971) lowers the value of the fossils in homology determination. The difficulties in distinguishing between paraphyletic stem-groups and holophyletic 1 crown-groups' can lead to incorrect hypotheses of homology (and homoplasy) (Herrnig, 1981). For example, the Permopsocida have been identified as the stem-group of the Psocoptera (Smithers, 1972), of the Psocodea (Kim & Ludwig, 1982) and as an assemblage comprising members of the stem-groups of Acerearia and Psocoptera and other acercarian orders (Hennig, 1981) . These different concepts have led to different interpretations of features common to Psocoptera and other Acercaria (see discussion of acercarian systematics in section 2.1. below).

Conclusion. It has been shown that there is no theoretical requirement for phyletically homologous features to be phenetic homologues or vice-versa, though it is probable that many phyletic homologues will be phenetically homologous by the criteria of Key (1967) and Jardine (1967, 1969). Evidence of phyletic homology is, in practice, generally taken from phenetic study of the organisms concerned and, in the case of morphological features, postulated on the basis of the most parsimonious 42

overall topographic homology (Jardine, 1969). Other evidence of phyletic homology may also be admitted from ontogeny and genotype, though neither of these provides a conclusive foundation; phenetic homology, however, while it may apply to ontogeny and genotype, cannot utilise evidence derived from them in support of homology of imaginal features as in phyletic homology. Phyletic homology may also be indicated through study of transformation series in fossils or modern organisms; such transformation series cannot be used to justify phenetic homology v/ithout incorporating a phyletic element in the analysis.

1.3-3- Methods of Systematic Analysis

1.3.3-1- Introduction

Systematic analysis is carried out in order to determine the 'relationships1 between different taxa, the relationships discovered being employed to permit study of some aspect of the biology or evolution of the taxa and/or to provide a framework for their classification. The term 'relationship' is ambiguous, however, as taxa may be deemed to be related according to a number of different criteria. The two forms of relationship that are considered here are 'phenetic' and 'phyletic'. Phenetic relationship is determined by the distribution of all character states so that, given three taxa, the two having most character states in common are considered more closely related to each other than either is to the third. Ho evolutionary connotations need be inferred from the results of such an analysis, although Sneath & Sokal (1973: 47) believe that "phenetic groups are usually monophyletic". The 'monophyly' of Sneath and Sckal includes both holophyly and paraphyly, and, as it is clear that a large number of autapomorphies of any one member of a holophyletic group will probably remove it from a phenetic group, phenetic groups may frequently be paraphyletic. Prom this it can be seen that phenetic analysis cannot be used to construct phylogenies, which rely on the recognition of holophyletic groups. Phyletic relationship is determined by the distribution of derived (apomorphic) character states so that, given three t axa, the two having most 43

apomorphies in common are considered more closely related to each other than either is to the third. The results of a phyletic analysis are expressed in a cladogram which, according to some authors, reproduces the branching pattern of the phylogenet ic tree of the group (Wiley, 1979s-» 1981). This interpretation is sometimes challenged (Platnick, 1977), but in this study it will be considered that there is a one-to-one relationship between dichotomously branching cladograms and phylogenetic trees (following Wiley, 1979a, 1981), though a polychotomy in-the cladogram is not taken to imply the same in the corresponding phylogenetic tree. It is always possible to perform a phenetic analysis on a group, but it may not be possible to carry out a phyletic analysis. If the polarity of character state transformation series cannot be identified, a necessary requisite for phyletic analysis, phenetic analysis is the only available option. In many groups where homoplasy is believed to be very common, phenetic analysis may be preferred to phyletic analysis. For taxa where phenetic and phyletic analyses are both possible, the type of analysis employed must depend on the type of research to be performed on the results of the analysis. There has been much discussion in the literature concerning the relative merits of phenetic and phyletic analyses and the classifications based on them (e.g. Mayr, 1969; Sneath & Sokal, 1973; Wiley, 1981), each author claiming that the method espoused is more predictive than others, or more suited to form the basis for other biological or evolutionary research. It is not proposed that this debate be reviewed or continued here. The requirements of this study will be outlined and reasons given for the choice of one technique rather than the other. As noted above, one of the major aims of this study is to evaluate the degree of correspondence between host and parasite phylogenies. The construction of a phylogeny can, as stated above, be performed (with more or less confidence) following a phyletic analysis, but cannot be done on the basis of a phenetic analysis. For this reason a phyletic analysis was carried out on the data collected in this study, and the character survey was undertaken with the aim of detecting suitable 3g>omorphies (see below). 44

It will be shown in Section 4 below that the phylogenies of louse and host do not entirely match, and that a single host taxon is sometimes parasitised by a polyphyletic assemblage of lice. It is possible that of the same host induces convergence of features of the lice, and it would be of interest to determine (a) whether such convergence does occur, (b) which features are particularly susceptible to modification according to the host, and (c) to what extent such convergence has influenced past classification of the Trichodectidae. To answer these questions a phenetic analysis is required. It was intended at the start of this study that such an analysis be performed, and the-data matrix was constructed so that characters would be available for phenetic treatment (see below). The prime aim of the study demanded a phyletic analysis as discussed above, however, and characters were selected for inclusion in the data matrix partially on the basis of their (intuitively assessed) lack of adaptive response to the environment provided by the host. After a consideration of the characters surveyed it was concluded that a phenetic analysis of them would no.t provide the data necessary to test adequately any correlation of louse morphology with host type. A further character survey would have been time-consuming, and the problem tackled is in any case outside the initial aims of the study. A further difficulty would anise in the interpretation of the results of the phenetic analysis with relation to the hosts. A valid comparison could only be made with a phenetic analysis of the hosts, particularly utilising characters of the dermecos (the environment provided by the host) and, if possible, not utilising characters not associated with the dermecos (such as the skull and other skeletal characters, all of major importance in most mammal systematics) . Such analyses are not available. A phenetic analysis of the lice was not, therefore, carried out.

1.3.3.2. Methods of phyletic analysis

The first step in any analysis is the construction of a data matrix for storage of the characters used. The data matrix constructed in this study is presented in Appendix A. As. noted above, phyletic analysis depends on the determination of apomorphic character states. This process 45

has been discussed in great detail by Hennig (1966) and, more recently, by Kluge & Farris (1969) and Hecht & Edwards (1977). The identification of apomorphies and their coding in this study are discussed in section 1.4.3.. In all techniques of phyletic analysis, to ensure that holophyletic groups of taxa are produced, the taxa used must themselves be holophyletic. In the case of a group such as the Trichodectidae, where no rigorous phylogenetic studies have been performed in the past, the analysis must generally be done at the species level as higher taxa may not be holo- phyletic. In this study no higher taxa were accepted as holophyletic without autapomorphies justifying that acceptance. There are a number of different methods of determining phyletic relationships. The three most generally-known types of analysis are compatibility, Wagner tree and 'phylogenetic' (sensu Hennig, 1966; in this study it will be referred to as 'cladistic'). The distribution of characters (or character states of a transformation series) over a number of taxa is, in the absence of homoplasy, consistent with the 'true* phylogenetic tree of those taxa. In practice homoplasy does exist, and not all characters will be consistent with the 'true' phylogenetic tree in the distribution of their states. The more independent characters that share the same distribution pattern of their states, however, (i.e. are compatible in the distribution of their states) the more probable it is that the tree supported is the 'true-1 phylogenetic tree. Compatibility analysis is performed by identifying such 'cliques' of compatible characters. If there is no homoplasy in a group there will be only one clique, but in almost all cases the existence of homoplasy leads to the formation of more than one clique for a group of taxa, and therefore more than one possible phylogeny. A choice is made between alternative cliques on the basis of size, the largest clique generally being deemed most likely to be correct. If most or all of the characters of the largest clique are associated with a particular feature, however, this clique may be rejected in favour of a smaller clique comprising characters involving differing adaptations (see also discussion of complex characters below). The choice of any clique automatically excludes any 46

characters displaying homoplasy, and thus in a group with a large amount of homoplasy in the characters, a compatibility analysis may explain only a small proportion of the data set (Wiley, 1981). Meacham (1980) points out, however, that following the initial analysis and choice of a clique (and hence a phylogenetic tree) smaller parts of the tree, unresolved by the first analysis, may be analysed using the entire data set, including characters eliminated from the first analysis because of homoplasy. In this way most if not all of the data can be explained. A confusing factor in discussions of compatibility is the advocacy by many of its proponents of 'convex1 taxonomic groups (Estabrook et al., 1977; Estabrook, 1979; Meacham, 1980). A set of taxa is said to be 'convex' if there exists a path on the tree between every pair of members in that set that passes through members of the set (ancestral taxa being accepted and placed at nodes on the tree) (Meacham, 1980). Estabrook et al. (1977) demonstrate that whilst holophyletic groups are convex, convex groups may be holophyleti or paraphyletic. The characters of a clique are also convex, in that they do not include homoplasy, and this has led to suspicion of the ,method by some phylogeneticists (Wiley, 1981). As explained above, however, repeated analysis of the tree and its component parts eventually allows inclusion of all homoplasy, and the technique does not of itself lead to the recognition of paraphyletic groups. A computer program for compatibility analysis was not readily available until after the manual dadistic analysis had been completed, and the one then available in any case can not be run on a data set as large as the one. generated in this study. It was not considered desirable to devote a large amount of time at this point in the study to this technique and accordingly a compatibility analysis was not performed. A Wagner tree analysis is designed to construct the most parsimonious tree consistent with the data (i.e. a rooted (approximation to a) Steiner tree linking the taxa and the hypothetical ancestors to those taxa generated during the analysis). A similar analysis can be performed to produce an unrooted network, but this will not be discussed here. The Wagner tree analysis produces a cladogram directly comparable with 47

one derived using the more 'traditional' technique of Kennig (1966), though the Wagner tree starts from the bottom (the 'root') whilst the hennigian cladogram starts from the top (the 'branches'). Whilst it is possible to perform a Wagner tree analysis by hand, with a large data matrix it is more efficient to use a computer, a number of different programs being available. The process of analysis is outlined below in stages, modified from the algorithm proposed by Kluge & Harris (1969). 1) The first step in constructing the rooted tree is the provision of the root. The root may be specified as either the sister-group of the group under analysis or a hypothetical ancestor (ANC) with all characters set to their most plesiomorphic observed state (' groundplan' of Wagner, 1961). 2) The amount of difference of each taxon from the ancestor or sister- group is determined. This difference (sometimes known as the Manhattan distance or advancement index) is calculated by

D(A, A1IC) = £|X(A,i) - X(ANC,i)|

where A is a taxon to be analysed ANC is the hypothetical ancestor or sister-group X(A,i) is the state of character i in taxon A SX(A,i) is the array of characters displayed for taxon A D is the Manhattan distance The taxon showing the least amount of difference from the hypothetical ancestor or sister-group is connected to the ancestor, the connection being knov/n as the 'interval* . The interval is deemed to have a length equal to the difference (in Manhattan metric) between the terminal taxon and the most recent ancestor, and the overall 'length* of the final tree is determined by summation of the lengths of all the intervals. •3) Of the remaining taxa, the taxon exhibiting the least Manhattan distance from the hypothetical ancestor is determined by examination of the data matrix as described above. 4) The place where the taxon selected in step 3 is to be joined to the tree is then determined. Because taxa are all terminal in a Wagner 48

tree, taxa are always added (by intervals) to intervals and not to other taxa, even though the length of an interval may be zero if no apomorphies distinguish one taxon from the hypothetical ancestor of another. The interval on the tree that shows the least difference from the taxon to be placed is therefore located. To do this the following formula is used:

D(A,INT(B)) = (D(A,B) + D(ATAIIC(3)) - D(3,ABC(B»)

where A is the unplaced taxon B is a taxon already placed on the tree, joined to its most immediate common ancestor, AITC(B) by the interval IHT(B) D is the Manhattan distance This calculation is performed for 'A1 and all intervals on the tree. 5) The taxon selected in step 3 is attached to the interval found in step 4« This is done by introducing a hypothetical common ancestor of the terminal taxa so that its* characters are the median of the first taxon, the original ancestor of that taxon, and the new taxon. 6) If any taxa are left unplaced, steps 3» 4 and 5 are repeated. The Wagner tree constructed for three taxa A, B, C and D (data matrix in Table I) is presented in Fig. 3. Wiley (1981) provides two worked examples of Wagner tree analysis. Steps 1, 3 and 4 in the list above each rely on determination of least difference. In each case there may be more than one taxon with the same minimum Manhattan distance, and a choice must be made between them. The necessity for such a choice indicates that the analysis may produce any of several different trees depending on the (frequently arbitrary) criteria employed for the choice (Jensen, 1981). The most desirable tree is the most parsimonious, i.e. the one requiring the smallest number of character transformations. This is the tree with the shortest overall length. It is possible for there to be more than one tree with the same minimum length, between which a choice must be made, and it may in any case be felt that choosing the most parsimonious tree if the next most parsimonious is only slightly longer (say, one character out of one hundred) is unwarranted and that both should be considered on an equal basis. With a large data set, and a large number of possible 'choices1 in the construction of the tree, a computer program may not provide the most 49

TAX ON CHARACTER STATES A 0 0 1 1 0 B 0 0 0 1 0 C 0 1 1 1 1 D 0 0 1 1 1

Table I. Data matrix for hypothetical taxa A, B, C and D. f Apomorphies are scored as •1*f plesiomorphies as *0 .

ANC (00000)

Fig. 3. Cladogram of hypothetical taxa A, B, C and D, produced by Vlagner tree analysis. X, Y, Z and ANC are hypothetical taxa produced during the analysis. The relevant character states are listed in parentheses after each taxon. Note that the intervals between Y and B, X and A, and Z and D all have zero length. . 50

parsimonious tree nor, if it does, will it necessarily indicate all the equally parsimonious and slightly less parsimonious trees available (Jensen, 1981). Some programs will produce different trees if the order of input of the taxa is changed, so to ensure that all the most parsimonious trees are produced the taxa must be fed in in every possible order, clearly an impossible operation with a large number of taxa (.Jensen, 1981). It was intended initially to carry out a Wagner tree analysis in this study, and the data matrix was constructed so that this would be possible (see section 1.4.3. below). To ensure that all the most parsimonious trees had been produced, however, 167 f. analyses would have had to be undertaken. In addition, the computer program available would not permit analysis of all the taxa at the same time, so only sections of the cladogram could be resolved, and these without reference to other sections, so that more analyses would have been necessary to ensure that all equally parsimonious trees had been produced. The final choice of a tree, upon which to base the classification would, in any case have been somewhat subjective. It was felt that to obtain results which could be considered with confidence (or given a mathematical level of confidence) much more labour and time would have had to be invested than was justified by the study, and the technique was not used. The final type of analytical technique to be considered (and the one employed in this study) is cladistic analysis, by which is meant the 'classical1 phylogenetic analysis espoused by Hennig (1966). In this technique a cladogram is produced by identification of 'sister-groups' of taxa on the basis of shared apomorphies. Cladistic analysis differs from compatibility and Wagner tree analyses in that the taxa are grouped first rather than last (i.e. the analysis proceeds from the branches down rather than from the root up). The process of cladistic analysis has been described many times (e.g. Hennig, 1965, 1966; Brundin, 1966; Wiley, 1981). The description below is based mainly on the procedure followed in this study. The primary data matrix may be analysed directly or, if it is too large to be manipulated conveniently by hand (as in this study) it may be subdivided. It was possible in this study, by inspection of the primary data matrix and general observation of the taxa, to form plausibly 51

holophyletic groups of moderate size on the basis of presumed autapomorphies. These first approximations of holophyletic groups were inclusive rather than exclusive, and taxa with a low probability of membership were included. Secondary data matrices were constructed for each of these possible holo- phyletic groups (some taxa, because of the inclusive nature of the groups, appearing on more than one matrix). These matrices were of more manageable size, as fewer taxa were included. Only characters with apomorphic states in the taxa included needed to be entered, which also reduced the necessary size of the matrices. During the construction of the matrices taxa that were initially outside the possible holophyletic groups were added if they were found to share apomorphic states with taxa initially included. From the secondary matrices sister-group pairs of taxa were deteimined by their possession of synapomorphies not shared with other taxa. The process is as follows: A taxon is compared with all other taxa (or groups of taxa believed to be holophyletic) in the matrix for the possession of synapomorphies not shared with any other taxa. If the taxon exhibits unique synapomorphies with more than one other taxon (thus indicating more than one possible sister-group relationship due to homoplasy) the sister- group relationship indicated by the greatest number of-apomorphies is chosen (c.f. compatibility analysis). In the same way as described above for compatibility analysis, large groups of apomorphies all"linked to the same adaptation might be disregarded in favour of a smaller group comprising apomorphies related to different adaptations; in this study however, large groups of apomorphies were seldom available. As discussed in section 1.4.2. below, some apomorphies are given greater weight than others, and decisions made on alternative hypotheses of holophyly were influenced in this study by the weight of the characters. Because the five-category weighting system of Hecht & Edwards (1976) was not always applicable (see section 1.4.3*) the more simple 'gain1 and 'loss' criteria were employed. Taxa were only grouped on loss characters when gain characters were not available, and after gain characters had been employed to resolve the area of the tree where the groups were located as much as possible. It follows that, whereas the distribution of gain characters should be maximally parsimonious over the cladogram, this is unlikely to 52

be the case for loss characters. Where a choice between alternative sister-group relationships could not be made readily the taxa were re- examined in the hope of identifying 'new1 apomorphies. If a sister-group relationship could still not be determined unequivocally, and provided the taxa concerned themselves formed a holophyletic group (indicated by synapomorphies not found in other taxa) they were placed on the cladogram as a polychotomy. Having determined a sister-group pair or other holophyletic group the sister-group to this is then sought on the data matrix, and identified by the same criteria. The composition of each holophyletic group is'retained in the data matrix, so that any further apomorphies identified during the course of the analysis can be entered for each taxon, and any consequent change in parsimony can be recognised. Should such a change be noted, new hypotheses of sister-group relationships can be formulated, and the characters supporting the old hypotheses re-examined. Initial hypotheses of the polarity of transformation series can also be tested as different sister-groups are proposed, and in some cases during this study the first polarity hypothesis was rejected during the analysis. This process whereby all hypotheses of sister-group relationships and of character polarity are continually tested during the process of analysis was termed 'reciprocal illumination' by Hennig (1966). Prom the secondary data matrices, after all possible sister-group relationships of the included taxa have been determined, a final data matrix can be developed and the sister-group relationships between the holophyletic groups already determined can be found. Prom this the complete cladogram can be developed. The cladogram produced is a dendrogram displaying all sister-group relationships and the apomorphies supporting them, as well as such homoplasies as are present, though for discussion purposes the character states are omitted and only the branching-pattern of the taxa shown. The advantage of a manual cladistic analysis lies largely in its 'availability'. The analysis can be carried out with constant reference to specimens, so that if resolution of any part of the tree is difficult, further observations can be made to check anomalous character states, locate new characters, and test estimates of polarity. These new 53

observations can then be incorporated directly into the analysis. This immediacy of response to problems is an appealing feature of any manual analysis, but perhaps more so of cladistic than Wagner tree. As indicated above, a full Wagner tree analysis, if performed manually, would have been very time-consuming for a data matrix of the size prepared in this study. A computerised Wagner tree analysis would also have been time- consuming if all the most parsimonious trees were obtained and compared, and the necessity for the matrix to be split and the resultant trees combined would have made the analysis too cumbersome, even though paralleling to some extent the method employed in this study for the cladistic analysis. 'Whilst the Wagner tree is maximally parsimonious, the tree produced by the cladistic analysis is not (at least for loss characters). This lack of parsimony is an inevitable result of character weighting (see section 1.4.2. below) as, if the tree were maximally parsimonious for all apomorphies (gain and loss equally) the two categories of apomorphy would •have been treated as if they were equally reliable as phylogenetic indicators, which they are not. Recognition of this difference would have been possible by running the Wagner tree analysis on gain characters first, then adding in the loss characters, as was done with the'manual cladistic analysis. Not all gain characters were assigned equal weight, though in this study it was not found possible to give confidently a weight to each character. The difference in weight would have been employed in the Wagner tree analysis in the final decision between equally parsimonious trees, and this process was exactly duplicated in this study on the equally parsimonious trees produced manually. It is notable that the cladogram produced in this study is not maximally parsimonious for gain characters. The reasons for this are discussed in detail in section 2.4.2. below. This option would not have been open in a Wagner tree -analysis, but the final maximally parsimonious tree would have had to be treated manually to produce the same tree as has been presented in this study. In view of the greater convenience of the manual cladistic analysis, its greater speed, and the necessity of detailed manual treat- ment of the Wagner tree following its production, the manual cladistic analysis was the technique practised in this study. 54

1.3.4. Classification

1.3.4.1. Relationship of classification and systematic analysis

Biological classifications are designed not only to enable the taxa classified to be located, but also to store information about those taxa. In many cases the nature of this information is not clear, and the taxa are grouped together on the basis of some ill-defined combination of phenetic and phyletic relationship. It is possible to indicate much more precisely the type of information stored in the classification (i.e. the type of relationship used in its construction), and thus permit the employment of the classification as an efficient data-retrieval system. Phenetic classifications will not be discussed here, as it has been decided above that a phenetic analysis, upon v/hich such a classification must be based, will not be carried out in this study. Phyletic classifications all reflect some aspect of evolution, but there has been prolonged debate in the literature as to what this aspect should be, and consequently v/hat information the classification should convey. Two types of phyletic classification may be distinguished: phylogenetic or cladistic (Hennig, 1966;G. Nelson, 1972, 1974a, 1974"b; Harris, 1980; Wiley, 1979b, 1981) and evolutionary or phylistic (Mayr, 1969, 1974; Johnson, 1970; Bock,.1974, 1977; Ashlock & Brothers, 1979). Broadly speaking, two aspects of evolution may be discerned: splitting or branching of lineages (cladogenesis) and evolutionary change through time (anagenesis). Clado genesis is detected and ordered by analysis of apomorphies, as described above. Anagenesis is detected in the character- state changes on the tree between taxa and quantified by their enumeration; the greater the 'amount* of anagenesis between two taxa (degree of evolutionary divergence) the greater the inferred dissimilarity in their genotype, irrespective of their cladistic relationship (Mayr, 19&9). Patterson (1981) presents evidence indicating that the latter assumption is not valid. Phylogenetic classifications are constructed to reflect only clado- genesis and therefore contain only holophyletic groups; only cladistic information is conveyed. Evolutionary classifications utilise both cladogenesis and anagenesis, so that the classification "reflects the 55

evolution, not just the phylogeny" of a group (Bock, 1974). Whilst polyphyletic groups are avoided in an evolutionary classification, the inclusion of anagenesis as a criterion leads to the recognition of para- phyletic groups (both holophyly andparaphyly being included in the phylist concept of monophyly). This is believed to be a strength of the system by its proponents, who claim that it results in a classification that is maximally predictive and provides a "foundation for all comparative studies in biology" (Bock, 1974: 379). Cracraft (1974) points out that though both cladistic and anagenetic information are put into an evolutionary classification, no method exists whereby either may be extracted with any precision. The user of such a classification must always be in doubt whether a given taxon (above the species level) is accepted because of the overall phenetic (inferred genetic) similarities of its members, or because it is a holophyletic group. As holophyletic and paraphyletic groups may have quite different properties, this ambiguity must lower the efficiency of the classification as a data- retrieval system, and hence its desirability. This problem is avoided in a phylogenetic classification, which should be constructed so that the original phylogenetic (cladistic) relationships postulated for the organisms classified may be reconstructed from it by any user, whether or not the user is familiar with the group classified. Should morphological relationships be required (whether or not they are deemed to indicate 'genetic relationships') these can be provided by separate tabulation, and this in far greater detail than could ever be stored in a Linnaean classification. Studies that depend on phylogenetic evidence, such as the present study of the Trichodectidae, are facilitated by the recognition in the classification of holophyletic groups alone.

1.3.4.2. Structure of the classification

It has been decided that only holophyletic groups will be recognised in the classification, the rationale given for this decision being that the phylogeny of the group classified will be fully recoverable from the classification. By the Rules of Zoological Nomenclature the classification must itself be in the form of a Linnaean hierarchy. The way in which 56

;bhis hierarchy is formed according to the phylogeny, and the way that the phylogeny may be recovered from the classification, must be discussed. The parallel between the Linnaean hierarchy and the hierarchy of a phylogenetic tree (or cladogram) has led to attempts to represent the latter in the former directly (Hennig, 1966; Brundin, 1966; Boudreaux, 1979), with the condition that sister-groups be given the same absolute rank (Hennig, 1966; Brundin, 1966). Reliance on this aspect of the Linnaean Hierarchy ('subordination') leads to the employment of a large number of names to represent adequately the branching pattern of the phylogeny. Taking the phylogeny in Pig. 4a, a subordinated classification that expresses all the relationships would be: Order AH Suborder H Pamily H Suborder AG Infraorder PG Pamily G Pamily P . Infraorder AE Superfamily E Pamily E Superfamily AD Pamily CD Subfamily C Subfamily D Pamily AB Subfamily A Subfamily B

Some classifications of this type (e.g. Boudreaux' 1979 classification of the hexapods), contain a large number of 'redundant' categories, higher taxa (e.g. Subclass, Infraclass, Cohort, Superorder) that contain only the terminal taxon (Order, in the case of Boudreaux, 1979). Naming of all these categories, which is sometimes unavoidable with the subordinating convention, burdens the literature unduly, and makes the classification cumbersome. 57

Fig. 4. . Cladograms for hypothetical taxa A - H (Fig. 4a), a - c (Fig. 4b) and A - D (Fig. 4c). For explanation see text. 58

The assignment of rank must also be considered. Hennig (1966), following the assumption that holophyletic groups being ordered in the classification according to relative rank are arranged also according to their relative age of origin, recommends that the absolute rank of a group be determined by the absolute age of that group. Thus taxa arising in the Precambrian might be deemed to be of 'Phylum stage' , those arising in the Cambrian, Ordovician, Silurian and Devonian of 'Class stage' (i.e. either Class, Subclass, Infraclass or Kicroclass), those arising in the Carboniferous and Permian of 'Family stage' and so on. At least partially due to the difficulty of assigning absolute age to many taxa, this convention has found little support, and the rank of taxa in a subordinated phylogenetic classification has been generally determined on an _ad hoc basis. If all the furcations of a phylogeny from ordinal to species level were to be expressed on a subordinated classification, the number of categories required would be far higher than are actually allowed for in the Linnaean hierarchy. Hennig (1981) attempts to combat this problem by adoption of the numerical prefix system as an adjunct to the Linnaean hierarchy (although adoption of this convention in fact makes the Linnaean hierarchy unnecessary). The phylogeny in Fig. 4a, if represented as a classification using the numerical prefix system,-would be: —1. AH 1.1. H 1.2. AG 1.2.1. • FG 1.2.1.1. G 1.2.1.2. F —1.2.2. AE 1.2.2.1. E 1.2.2.2. AD 1.2.2.2.1. CD* —1.2.2.2.1.1. C 1.2.2.2.1.2. D 1.2.2.2.2. AB 1.2.2.2.2.1. A 1.2.2.2.2.2. B 59

It is clear that this system can accommodate as many ranks as reauired, but it is very cumbersome, and a list of numbers is not as memorable as the names employed in the Linnaean hierarchy. The system has not found even limited acceptance. L^vtrup (1977) suggests a contracted binary system of numeration that reduces the number of prefixes required. This system, however, is still cumbersome and difficult to operate (Wiley, 1980). To minimise the number of categories required in a phylogenetic classification furcations below subgenus level are not usually expressed, and genera and subgenera may comprise any number of species (concomitant with their being holophyletic groups). The phylogeny of genera is usually expressed independently of the classification in diagrammatic form. The criteria utilised in genus construction in this study are discussed below on pp. 282 - 283. A further difficulty with subordinated phylogenetic classifications is vulnerability to change with the discovery of new taxa (especially in groups with a fossil history). Recognition of previously unknown furcations, perhaps at high taxonomic levels, may cause radical alteration to a classification (G.Nelson, 1974a). Failure to accommodate such new discoveries by a modification of the classification, however, may allow the introduction of paraphyletic groups and destroy the rationale of the system. It is evident that a phylogenetic classification utilising only the subordination aspect of the Linnaean hierarchy is not fully satisfactory, as it may be clumsy, incorporating a number of redundant taxa and numerous hierarchical levels, it may not fully express the phylogeny in a large group where insufficient hierarchical levels exist, and it may lack stability. There is a second aspect of a formal classification which it shares with any other list of names, and that is the sequence in which those names are written. 3y applying stated conventions, a classification may be used to store phylogenetic information in the sequence in which taxa of equal rank are listed (G.Nelson, 1972, 1974a;. Cracraft, 1974; Wiley, 1979b, 19S1)j. This process has been termed phyletic sequencing (Cracraft, 1974). The convention employed is that within a classification holophyletic taxa of equal rank are listed ('sequenced') so that each taxon is the sister- 60

group of all those taxa of the same rank (and within the same taxon of immediately higher rank) listed below it in the classification (modified from Cracraft, 1974). Thus a classification based on Fig. 4b might be: Order ac Family a Family b (or c) Family c (or b) Naturally the order of the taxa in the terminal furcation is interchange- able, as each is the sister-group of the other. Most phylogenies are more complex than this, and must be represented in a classification by a combination of sequencing and subordination. The phylogeny in Fig. 4a might accordingly be represented in the classification: Order AG Suborder K Family H Suborder FG Family G (or F) Family F (or G) Suborder E Family E Suborder CD (or AB) Family C (or D) Family D (or C) Suborder A3 (or CD) Family A (or 3) Family 3 (or A) Again, it can be seen that some parts of the sequence can be 'rotated1 about symmetrical furcations. This classification is less cumbersome than the first (subordinated) classification derived from the phylogeny in that it contains only three categories rather than six. It does, however, still contain redundant categories for H and 3. These can be eliminated and the classification left with no redundant names, whilst retaining the storage of the phylogeny of the group: 61

Order AH Family H Family FG Subfamily G Subfamily F Family E Family CD Subfamily C Subfamily D Family AB Subfamily A Subfamily B Yet another alternative would be to have the suborders H, FG and AE, the latter containing the families S, CD and AB. This also stores the " phylogeny. It can be seen, therefore, that the convention provides a logical framework for decisions on ranking holophyletic groups, and is sufficiently flexible to allow retention of names at established rank-level should this be required, and to permit elimination of redundant names. The precise form of the classification will depend on the viewpoint of the classifier (as with all classifications) but the phylogeny of the group will always be recoverable from the formal classification. A further advantage is the stability of the classificiation. Should a new taxon be discovered it may be included in the classification at the appropriate point without altering the status of other groups. For example, taking the phylogeny depicted in Fig. 4a, should a group I be discovered between groups FG and E, in the subordinated classification it would be necessary to downgrade all taxa in AE and include a new infraorder AI, or to raise the order to superorder, with order AG to include suborders FG and AI. In a sequenced and subordinated classification, however (say, the second given for Fig. 4a), it would be necessary only to include a family I between FG and S, and no ranks need be altered. Holophyletic groups with a trichotomous or polychotomous inter- relationship must also be accommodated in a classification. Wiley (1979b) recommends that these be placed in the classification with equal rank and 62

be noted as 1 sedis mutabilis1 to denote that their order is unknown or interchangeable. Thus for the phylogeny shown in Pig. 4a, the classificati would be: Pamily AD Genus A Genus BD Subgenus B sedis mutabilis Subgenus C sedis mutabilis Subgenus D sedis mutabilis or: Pamily AD Genus A Genus BD (all subgenera sedis mutabilis) Subgenus B Subgenus C Subgenus D It must be remembered that this convention is for a formal classification, not for regular use in discussion. Wiley (1979b) terms a classification presented in this way an "annotated Linnaean hierarchy". He proposes a number of other conventions for use in an annotated Linnaean hierarchy which are not of relevance to this study and will not be discussed here.' All hierarchical groups may be sequenced, but it is probably more efficient to employ informal groups below the subgenus level, especially with larger genera. A sequenced and subordinated classification, in its formal presentation will enable reconstruction of the phylogeny of the group classified (which is, according to Cracraft, 1974, all that any classification can do). New taxa may be placed precisely in it without the necessity of re-drawing the phylogeny, which may be time-consuming and irrelevant in a small paper describing, say, a new genus. It is preferable to a subordinated classification without sequencing due to its greater stability, its need for fewer hierarchical levels, and its requirement of fewer redundant categories. 63

1.4. CHARACTER SURVEY

1.4.1. The Taxonomic Character

Taxonomic characters are aspects of organisms that are utilised in the process of grouping or distinguishing those organisms. Such characters must be comparatively invariant at taxonomic (hierarchical) levels below the one under consideration. For taxa being distinguished characters must vary at the level of the taxa, but for taxa being grouped such variation is unnecessary; characters used for grouping must, however, be variable at taxonomic levels higher than the taxa grouped.

1.4.2. Use of Characters for Grouping Taxa

1.4.2.1. Choice of characters and their relative values

Characters selected (by the criteria of their variability mentioned above) for possible use in analysis to group taxa may not all be equally suitable. Characters may be unsuitable because they do not provide information of the type required for a particular analytical technique, or because the information provided is too 'unreliable' for conclusions to be drawn. In addition, an analysis may provide several alternative groupings of taxa, each supported by different characters, between which a choice must be made. All characters selected are therefore given a 'weight', a measure of relative reliability or value.for any given analysis, though such a weighting may not be consciously applied, and may be equal for all characters. The criteria used for deciding the weight will depend on the type of analysis to be performed.

1.4.2.2. Determination of polarity for phyletic analysis

For phyletic analysis, the different states of a character (distributed among two or more taxa) must be arranged into a morphological transformation series (e.g. present - absent; large - medium - small; one pair of spiracles - two pairs - six pairs) and a 'polarity' assigned of plesiomorphic (ancestral) to apomorphic (derived). Plesiomorphic character states and characters for which the polarity of the transformation series is undetermined provide no evidence for grouping taxa, and are 64

automatically given zero weight and excluded from analysis. In a transformation series there can only be one plesiomorphic condition (in the sense of 'most ancestral'), but there may be more than one apomorphic state (Hecht & Edwards, 1977). These apomorphies may be independently derived from the plesiomorphic st ate, or may be increasingly derived conditions in an evolutionary developmental sequence. The change' in a transformation series may thus be uni- or multi-directional (Marx & Rabb, 1972). Kluge & Farris (1969), following Wagner (1961), propose three criteria whereby the plesiomorphic character state of a transformation series (and thus the polarity) may be identified. "In order of reliability these criteria are: (1) The primitive state of a character for a particular group is likely to be present in many of the representatives of closely related groups. (2) A primitive state is more likely to be widespread within a group than is-any one advanced state. (3) The primitive state is likely to be associated with states of other characters known from other evidence to be primitive." (Kluge & Farris, 1969: 5). Kluge & Farris note that for the application of the first criterion the "closely related groups" may be identified as such by estimates of overall similarity that make no assumptions about plesiomorphic states. The assumption is made, however, that constituent taxa of these "closely related groups" are not more properly included in the group under consideration; that is, that the. group is holophyletic. Even when some assumptions of polarity are involved in the choice of outgroups, these are in the form of hypotheses that can and must be callenged by all other characters discovered and analysed. This process of 'reciprocal illumination' has been discussed in section 1.3«3»2. above. Hecht & Edwards (1977) note that the closer the relationship between two groups, the greater the likelihood that they may evolve the same character state in parallel from the same (plesiomorphic) state. Failure to recognise such honoplasies will lead to confusion in the estimation of the polarities of other transformation series that are not similarly homoplastic. 65

As a corollary to criterion (1), it is probable that character states found only in a holophyletic group are apomorphic within the group. This suggestion is supported by Hecht & Edwards (1977), who suggest that in a transformation series the apomorphic states may be recognised by their restricted distribution among taxa. This criterion is particularly useful where all of the character states observed are derived from an unknown plesiomorphic state, though there is clearly a danger of assuming the unrecognised plesiomorphic state to be apomorphic if it is present but retained by only a few taxa. V/ithin this study the criterion of limited distribution has been employed frequently at the smaller species-group level, particularly with reference to the male genitalia. This criterion is apparently related to the second of Kluge & Farris (1969), listed above, and the concept that 'common equals primitive' employed by some workers (Estabrook, 1977). If character states are identified as plesiomorphic on a simple taxon count, however, states associated with recent, relatively successful, phyletic lines would be considered plesiomorphic whilst plesio- morphic states present only in relatively ancient but unsuccessful taxa would erroneously be considered apomorphic. Kluge & Farris (1969) point out that by 'widespread' they do not mean a simple taxon count, but refer to a singl-e state found in taxa that otherwise have little in common. Thus the process they recommend is an analysis of degree of concordance between different transformation series. The greater the degree of concordance the greater the probability that the hypotheses of polarity are correct (Hecht & Edwards, 1977). The analysis of concordance also applies to the third criterion of Kluge & Farris. It must be remembered, however, that correlation of states between different characters demands transformation series of at least three states, as correspondence between series of only two states cannot be predictive.

A measure of likelihood of homoplasy for a particular character ma:/ be gained by an appreciation of its "'complexity1 , in that detailed and complex apomorphies are presumed to be less likely to have arisen convergently than £re simple ones, and secondary resurrection of such states following loss is of low probability (Hecht & Edwards, 1977) . As with any criterion, however, the possibility of it being a false guideline must not be neglected. 66

1.4.2.3. Character weighting in phyletic analysis

As noted above, only apomorphic character states are utilised in phyletic analysis, and these may be weighted according to their 'reliability*. An apomorphic character state is considered 'reliable' and may therefore be heavily weighted, if there is a complete correspondence between the set of organisms that have the state and the set of members of a holo- phyletic group (Wilson, 1965; Inger, 1967; Kluge & Farris, 1969) . Thus a 'reliable' apomorphy defines a holophyletic group, and is not secondarily lost or homoplastically present in other groups. The greater the degree of homoplasy of a character state, the less reliable the apomorphy. Clearly, any weighting of apomorphies should be related to their reliability, but this cannot be directly assessed before the analysis has been performed. Apomorphies must, therefore, be weighted on the basis of other properties, presumed to be related to the reliability or 'uniqueness'. Inger (1967) suggests four criteria by which 'unique' (apomorphic) character states may be recognised: "1) There is no obvious selective difference between the states of a character. 2) The state occurs in many taxa of the group being studied. 3) The character has low variability within taxa. 4) The unique state has an unusual developmental pattern." The first and last criteria are rejected by Kluge 2c Farris (1969) and Hecht 2c Edwards (1977) on the grounds that there is rarely enough evidence upon which to base objective decisions on relative selective difference and 'usualness' of developmental pattern. The second criterion is dependent in part on prior analysis, and does not allow for parallelisms, which may be frequent (Hecht 2c Edwards, 1977) • The third criterion identifies a correlation between the variability of a character and its evolutionary rate of change, so that conservatism in observed variation is indicative of conservatism on an evolutionary timescale. Thus, the more variable a character (and the more states that exist), the less reliable the character is as an indicator of a holophyletic group, the lower the weighting of that character (Kluge 2; Farris, 1969). Hecht 2c Edwards (1977) point out that this suggestion is based on the rather 67

dubious assumption that variation and selective pressure are uniform throughout the temporal and spatial distribution of a character or character state. Hecht & Edwards (1976) suggest five categories of weighted apomorphies, and give some 'rules of thumb1 for developing this weighting. The categories, slightly modified from Hecht & Edwards (1976, 1977) are, arranged in order of increasing information content: I. Characters and character states of the lowest value are those involving loss of a structure, with no developmental information to indicate the pathway by which the loss occurred. This character weighting group has zero information (as to holophyly) because there is no way of determining whether the state has been derived by a single change or by two or more independent processes. II. The second category includes simplificiation or reduction of complex characters, indicated by comparative or developmental anatomy. Loss characters, such as eyelessness in cave salamanders or fish, in which the developmental mechanism leading to the loss is known, should be included in this category. Independent reduction of the same character by two closely related taxa may show a different develop- mental process in some minor detail. III. The third category includes those character states that are the result of common growth processes for the taxa being compared. These similarities are due to growth and developmental processes dependent on size, age or hormonal and other physiological relation- ships, such as allometry or neoteny. IV. The fourth category includes all those character states which a.re part of a highly integrated functional complex. The complexity of characters in this weighting state makes them important indicators of polarity and useful to distinguish parallelism. V. The fifth and most informative type of weighting group is that which is innovative and unique for a transformation series.and, therefore, most useful as a shared and derived character state to distinguish a new lineage. The more complex the innovative character state, the more reliable an indicator of lineage it is. At the higher hierarchical levels, this weighting group usually indicates new 68

functional or adaptive trends. If the characters, or character states are complex enough, they can preclude parallelism. Eecht & Edwards (1977) note that these categories overlap, and suggest that their number might be varied, depending on the requirements of the analysis. In the present study the weighting scheme of Eecht & Edwards (1976) is used as far as possible, although there are some difficulties in its application, some necessary information (e.g. developmental) frequently being unavailable. To supplement the weighting system of Hecht & Edwards (1976) a more simple division into apomorphic character states involving the acquisition of novel structures ('gain characters') and apomorphic character states involving the loss of structures ('loss characters') is employed. This is in effect a division of the system of Hecht & Edwards into categories I plus II versus III - V. The weighting of the character states used in this study is given in section 2.3.3., both of the above systems being used.

1.4.2.4. Problems of complex characters

The most difficult of the^ categories of Hecht & Edwards (1976) to interpret, and that which includes the most important source of error in weighting, is the fourth, dealing with complex characters. A ••complex character may be defined as a "set of closely-integrated characters that change together in order to maintain biological efficiency or permit the organism-to remain viable" (Hecht & Edwards, 1977). The separation of such functional complexes into structurally distinct (or distant) components may result in undue weighting. For example, Hecht & Edwards (1976) suggest, in their study of proteid salamanders, that nine apomorphies considered separate by earlier workers are all aspects of a single apomorphy: the development of paedogenesis. Hot all complex characters are readily discemable as such, and it is apparent that the less well- understood the functional interrelationships of features, the greater the likely weighting of the character that represents the functional totality of those features. Though high weighting of complex characters is recommended by Hecht & Edwards (1976, 1977), this 'hidden' weighting 69

may well be greater than that properly applied to the complex character were it to be recognised. In addition, weighting must be recognisable in order for the reliability of the analysis to be assessed. In many cases, particularly in phenetic analyses, weighting is purportedly equal for all characters ('unweighted characters'), and numerous techniques have been developed to ensure that equal weighting is applied (Sneath 8c Sokal, 1973). Sneath & Sokal (1973) recommend the use of 'unit characters' , which they define as "taxonomic characters of two or more states, which within the study at hand cannot be subdivided logically, except for the subdivision brought about by the method of coding". The application of 'equal weighting' to unit characters would appear to give very heavy weighting to complex characters, unless great care is taken to exclude all but one aspect of them. Cain & Harrison (1958) recognise the problems associated with complex characters in their definition of a taxonomic character: "anything that can be considered as a variable independent of any other thing considered at the same time".

1.4.2.5. Constitutive and diagnostic characters

Characters used for grouping are of two types, ' constitutive' and ' diagnostic' , constitutive characters being autapomorphies of the group and diagnostic characters distinguishing symplesiomorphies (Hennig, 1981). For example, Amblycera form a distinct group within the Phthiraptera, united by their possession of maxillary palps, horizontally-articulated mandibles, and a pedunculate first flagellomere. Of these characters only the last is constitutive - autapomorphic for the Amblycera - the first two being plesiomorphic within the Phthiraptera but lost or modified in the Ischnocera, Rhyncophthirina and Anoplura.

1.4.3. Data Recording in this Study

1.4.3.1. Procedure

Many of the features of trichodectid species that have been utilised in this study are simple two-state characters, which are easily recorded and readily-discernible from a simple inspection of the insect. Examples 70

of characters in this class are the abdominal spiracles (present or absent on each segment from III to VIII), the apophysis of abdominal sternum II (present or absent) and the ventral teeth on the tarsal claw (present or absent; broad or narrow). Such characters were recorded direct onto a data matrix, many later to be incorporated into the main matrix reproduced below. Some characters showed too much or too little variation and were not included in the main matrix. Other features, however, (antennae, male genitalia, male terminalia, female vulva and gonapophyses), are of greater morphological complexity and the characters derivable from each are not obvious on simple appraisal. Discrimination of these characters, and survey of all species for each character, would have been very time-consuming, so the process was carried out in two stages. For each feature, species of similar structure were associated into T equivalence groups'. This process relied on simple perception of similarity, so that groups are purely phenetic, and do not carry an implication of holophyly of included taxa. A small number of characters was employed to give some structure to the groups, but the variation within and between the groups was somewhat arbitrary, depending on the complexity of the feature, the known development and function of the feature, and various subjective factors. The number of species in an equivalence group varied from one to over one hundred, but was generally around four or five. In the second stage, further characters were proposed, and all species surveyed for states of these characters-. The association of species into equivalence groups allowed this survey to be more rapid than would otherwise have been possible. All characters, both those that had been employed in the construction of the groups, and those that had been formulated subsequently, were recorded on a data matrix and evaluated, as were the more simple characters above, for inclusion in the main matrix.

1.4.3.2. Coding

The form of the code given to character states for subsequent data retrieval depends on the type of analysis to be performed, as each technique imposes unique constraints on this form. Data were recorded 71

in this study so that cladistic and Wagner tree (phyletic) and average linkage (phenetic) analyses could each he performed if required. Both Wagner tree and phenetic analyses require a machine-readable code for the data, whilst the cladistic analysis, performed by hand, employs a sometimes less clumsy notation. The character state survey was carried out prior to any form of phyletic analysis (identification of transformation series and determination of their polarities), and it was in any case (correctly) anticipated that not all characters would prove suitable for phyletic analysis. For these reasons the data matrix (see Appendix A) is coded for phenetic analysis, with alternative Wagner tree codings where required; coding for cladistic analysis is obtained by cross-reference from the data matrix to the list of characters (page 197)• Coding for Phenetic Analysis. Many of the characters employed in this study are 'two-state1; that is, only two alternative states of the character are available (e.g. present or absent, long or short, symmetric or asymmetric). Such characters are coded as follows: CHARACTER STATS CODE .abdominal tergite absent 1 present 0 It is immaterial for the purposes of phenetic analysis which state is coded.by '0' and which by '1' ; to conserve space on the data matrix and in the computer, however, the coding of two-state characters was arranged to be identical for phenetic and Wagner tree analysis, as described below. Some characters - 'multistate characters' - occur, as the name suggests, in more than two states. In some cases the order in which these states are arranged is arbitrary, as they do not form part of a logical sequence; such characters are known a.s 'qualitative' or 'nominal* multi- state characters, and may be coded as follows: CHARACTER STATE CODE colour brown 0 0 0 0 red 10 0 1 or black 0 10 2 white 0 0 1 3 As is apparent from the binary matrix, nominal multistate characters may be treated as a set of two-state characters (brown or not brown, red or 72

not red etc.), though this increases the space taken by the matrix. The reduction of the binary code to a numerical code saves matrix and computer space, but the analysis must be run in such a way that 0, 1, 2 and 3 are treated as different, unrelated symbols and not as a secuence. Multistate characters may also be ordered, the states forming a logical sequence the arrangement of which is not arbitrary; such characters are /known as 'quantitative' or t ordinal' multistate characters. Such characters may be coded as follows: CHARACTER STATE CODE number of setae 0 0 0 0 0 2 10 0 1 or 4 110 2 6 1113 The space-saving numerical code may be read as a true sequence. Suitable instructions given to a computer will ensure that ordinal and nominal multistate characters, which when coded numerically are superficially indistinguishable, are each treated appropriately during the running of a program. Observations could not in this study be made of all characters of all taxa, either because of the unavailability of undamaged specimens, or because of the restriction of the available sample to one sex only. In such cases the characters concerned are deemed to be 'non-comparable' for the relevant taxa, and the number '9' is inserted in the matrix. A second category of non-comparable characters (also coded '9' in the matrix) is that comprising characters of which all states are precluded by other characters. For example, the male mesomeral arch may have the median extension 'present' or 'absent'; in the latter case the character 'apex of extension shape' ('acuminate' or 'bifurcate') is not applicable arid the taxa non-comparable for that character. Coding for Wagner Tree Analysis. In Wagner tree analysis characters are employed only if one or more states are recognised as apomorphic. This must be discernable from the code applied to the states, so the practice is for plesiomorphic states to be coded zero (0) and apomorphic states to be coded one (1) (Wiley, 1981). To conserve space in the data matrix, 73

two-state characters were entered only once, the coding applied being the same for the phenetic analysis as for the Wagner tree analysis. Multistate characters had to be entered as a binary code, however, so could not be reduced to numerals as could the phenetic; the form of the binary code for phyletic use differed in some cases from that derived for phenetic analysis, so parallel coding had to be used for multistate characters suitable for Wagner tree as well as phenetic analysis. The use of the binary code for phyletic characters enables transformation series of apomorphic states to be distinguished from alternate apomorphic states: CHARACTER STATE CODE apex of basal apodeme straight 0 0 0 concave 10 0 deeply concave 110 acuminate 0 0 1 1 Apex of basal apodeme straight - concave - deeply concave' and ' apex of basal apodeme straight - acuminate' are two different transformation series. The first category of.non-comparable characters as described above is employed, but the second is slightly modified to include only those characters for which the apomorphic state cannot be discerned because of other apomorphies. Coding for Cladistic .Analysis. As with Wagner tree analysis, apomorphies must be distinguished from plesiomorphies and transformation series of apomorphic states from alternative apomorphies. Because the cladistic analysis was carried out by hand a machine-readable code was not required. Plesiomorphic states are identified by the code zero (0) and apomorphic states by higher numbers. Whilst states in a transformation series are represented by a numerical sequence (0, 1, 2, 3), alternate apomorphies are indicated by the 'prime' symbol (0, 1, 1'): CHARACTER STATE CODE apex of basal apodeme straight 0 concave 1 deeply concave 2 acuminate 1' As noted above, this code is obtained by cross-reference from the data- 74

matrix (machine-readable code) to the list of characters. An attempt was made to weight the cladistic characters according to the criteria proposed by Hecht & Edwards (1976, 1977) (see above), and the weighting category (I - V) is shown for each character state in the list of characters. It was found, however, that the criteria are very difficult to apply to most characters, especially in the absence of developmental information, and that categories IV and V are not always readily distinguishe These weighting categories were accordingly used only as a general guide during analysis. A broader assessment of weighting was utilised in the distinction between apomorphic 'gain' and 'loss1 character states, which are indicated by • and • respectively on the cladogram and ' g» and 'iT in the list of characters. The outlines • and o indicate the plesiomorphic form of the character, and these are used on the cladogram to indicate reversals. SECTION 2

MORPHOLOGY AND CHARACTER ANALYSIS

OP TRICKGDECTILAE 76

2.1. SYSTEMATIC POSITION OF TRICHODECTIDAE

2.1.1. Introduction

Systematic (phylogenetic) relationships within a holophyletic group are determined "by study of the distribution of apomorphic (derived) character states (see discussion of cladistic methodology above). Apomorphies may be identified by their restriction to the group under study, whilst plesiomorphies (primitive character states) occur both within the group and in other, related, groups (Kluge and Farris, 1969). Assignment of apomorphic status to a character state must, therefore, be preceeded by a survey of the character in groups related to the one under study. Consequently, studies on phylogenetic relationships within a group depend in part upon assumptions concerning the phylogenetic relationships of the group to others.* Morphological investigation, especially of a group such as the Trichodectidae for which few published studies are available, may be greatly aided by morphological, anatomical and embryological studies on related groupsw- Thus neither the morphology of Trichodectidae nor the phylogeny (and hence systematic arrangement) of trichodectid genera can be investigated until the systematic position of the family itself is determined. Relationships of the Trichodectidae are investigated here using cladistic methodology (see above). The analysis is presented in two parts: relationships of the Phthiraptera- to the rest of the Insecta, and relationships of the Trichodectidae to other Phthiraptera.

* The identification of two groups as phylogenetically related (and holophyletic) implies an a priori assignment of polarity to transformation series, and might be held as an indication of circularity of reasoning. This problem is discussed more fully above, but two matters arising in this study are worth noting. Firstly, the holophyly of groups including and deemed to be related to the Phthiraptera is tested and, in the case- of the Psocoptera, challenged. Secondly, the holophyly of the Acercaria, initially accepted on the basis of the characters known, is subsequently challenged on the basis of characters discovered during the studyv (see p. 187). The continual scrutiny of groups 'related to' or including the group under study, and the consequent reassessment of polarity of transformation series, is referred to by Hennig (1966) as 'reciprocal illumination*.

POLYNEOPTERA PHALLONEOPTERA r l r HOLOMETABOLA i r ACERCARIA

PSOCODEA CONDYLOGHATHA 1 r HEMIPTERA r i POLYNEOPTERDUS ZORAPTERA PSOCOPTERA PHTHIRAPTERA THYSANOPTERA STERNORRHYNCHA AUCHENORRHYNCHA COLEORRHYNCHA HETEROPTERA HOLOMETABOLOUS

ORDERS ORDERS 25-29*, liO-lfi, 63 44-52, 56-^0

14 (reversed) 13, 14 (reversed) 15-17, 20. 21

1, 3-1 of 18, 19, 22 I

14

Fig, 5. Systematic position of the Phthiraptera, For explanation of numbered apomorphies see text. * These apomorphies were probably present in the ancestor of the Psocoptera plus Phthiraptera, but lost in the latter group. The Psocoptera are considered to be paraphyletic with respect to the Phthiraptera (see text). 78

2.1.2. Relationship of Phthiraptera to Other Insects

2.1.2.1. Introduction

Gladistic treatments of the Insecta have been produced recently by several authors, notably Hennig (1981), Kristensen (1975, 1981) and Boudreaux (1979)• The Phthiraptera are uniformly placed as sister-group to all or part of the Psocoptera, and these two Orders (superorder Psocodea) as sister-group to the Thysanoptera plus Hemiptera (superorder Condylognatha). The Zoraptera are sometimes placed as sister-group to this assemblage to form the Paraneoptera (Hennig, 1981; Kristensen, 1975), though this assignment may be very tentative (Kristensen, 1981), or are placed in the polyneopterous group Cursoria (= Blattopteroidea) (Heming, 1977; Boudreaux, 1979). • In this study the Zoraptera are tentatively placed as part of the Paraneoptera, as in Kristensen (1981). A cladogram for the infrasection of the Insecta including the Phthiraptera, the Phalloneoptera, is reproduced in Fig. 5, with the group names used in this study. Numerous apomorphies, not all of which are valid, have been proposed by various authors as evidence for the holophyly of groups within the Phalloneoptera. Apomorphies proposed for the Acercaria, Psocodea, Psocoptera and Phthiraptera are listed below", with references and comments where appropriate. The significance and validity of these apomorphies, and the identity of the sister-group of the Phthiraptera, are discussed.

2.1.2.2. Apomorphies proposed for relevant groups of Phalloneoptera

Acercaria B8rner. 1904. (= Hemiptero.idea Handlirsch, 1908; Acercarida Boudreaux, 1979). 1. Loss of abdominal cerci in all stages (BBrner, 1904; KSnigsmann, 1960; Kristensen, 1981). Smithers (1972) notes that cerci are present in some fossil Permopsocida and concludes that the loss of cerci is an autapomorphy of the Recent Psoccptera, though he does not indicate to which group he then considers the Psocoptera 79

•bo belong. Hennig (1981) suggests that the Pemopsocida are not all representatives of the stem-group of the Psocoptera, but are a paraphyletic group containing representatives of jthe stem-groups of the Acercaria, Psocoptera and possibly other acercarian orders. Reduction in number of tarsomeres to three (Hennig, 1981; Boukreaux, 1979)• Hennig (1981) proposes this as an apomorphy of the Paraneoptera Whilst Boudreaux (1979) considers the reduction in number of tarsomeres to be convergent in Zoraptera and Acercaria. Both authors note that the Permopsocida have four or five tarsomeres again indicating the"possible inclusion within this group of members of the acercarian (or paraneopteran) stem-group. Reduction or loss of labial palps (Boudreaux, 1979)• The labial palps may be reduced to two segments or one segment, or lost. Smithers (1972) considers this reduction |to be an autamorphy of the Psocoptera; see comment following i character 1. Enlargement of clypeus to accommodate origins of large cibarial dilator muscles (Boudreaux, 1979). Smithers (1972) considers this enlargement to be an autapo- morphy of the.Psocoptera, which he maintains exhibit it to a greater extent than other Acercaria. It is probable, however, that the functional significance of the cibarial muscles is the same in Psocoptera and Phthiraptera (see discussion of hypopharynx morphdbgy below), indicating that the character is synapomorphic in these two groups. Though the function of the enlarged cibarial muscles is slightly different in the Condylognatha, enlargement of the clypeus (or of the dorsum of the head, where the clypeus is not readily distinguished) does occur, and the most parsimonious hypothesis is that the enlargement evolved only once, and is apomorphic for the Acercaria. Modification of lacinia (Hennig, 1981; Boudreaux, 1979; Kristensen, 1981). The lacinia, when recognisable at all, is elongate, with a 80

slender styliform or toothed tip, and is detached from the maxilla. 6. Retraction of lacinial base into head capsule (Hennig, 1981). This character is presumably related to character 5« 7. Reduction of abdominal sternum I (Kdnigsmann, 1960; Hennig, 1981; Boudreaux, 1979; Kristensen, 1981). Hennig (1981) and Kristensen (1981) note sternum I as absent rather than reduced; the membranous nature of the abdomen in many Acercaria makes it difficult to ascertain its precise state. 8. Fusion of gonangulum to tergum IX (Scudder, 1961; Kristensen, 1975, 1981). 9. Concentration of nervous system in thorax (KOnigsmann, 1969; Hennig, 1981; Boudreaux, 1979). Hennig (1981) proposes this character as an apomorphy of the Paraneoptera, but admits that the degree of concentration is far less in the Zoraptera than in the Acercaria; as an apomorphy of the Acercaria he suggests "Abdominal ganglionic cord reduced to a single ganglionic mass.". See also character 47. 10. Reduction in number of malphigian tubules to four (K8nigsmann, 1960; Hennig, 1981; Boudreaux, 1979; Kristensen, 1981). In some members of the Acercaria the number of malphigian tubules is further reduced. 11. "Generally anteromotoral in modern forms" (Boudreaux, 1979). This proposed apomorphy refers to the dominance of the fore wings over the hind wings, and the concomitantly greater development of the mesothorax than the metathorax. Though this apomorphy is true of the Psocoptera and the Hemiptera, it does not apply to the Thysanoptera, in which the hind- wings are only slightly smaller than the fore-wings. It is possible, therefore, that 1anteromotoria* is convergent in the Psocoptera and the Hemiptera (Boudreaux, oers. comm.). 12. Spermatozoa basically biflagellate or with doubled axial filament complexes in one flagellum (Boudreaux, 1979; Kristensen, 1981). 81

Psocodea Herrnig, 1953. (= Psocoidea Weber, 1933; Psocopteroidea Jeannel, 1945; Psocodida Boudreaux, 1979). 13. Development of structural mechanism for rupturing the antennae (Seeger, 1975; Kristensen, 1981). This structure has been demonstrated by Seeger (1975) to occur in both Psocoptera and Phthiraptera, but it almost certainly does not have the same function in the two groups. 14. Ovarioles polytrophic (KBnigsmann, 1960; Kristensen, 1975, 1981; Seeger in Hennig, 1981). Heming (1977) suggests that, as polytrophic or acrotrophic ovarioles (the latter being derived from the former) are found in the Psocodea, Hemiptera and throughout the Holometabola, the plesiomorphic panoistic ovariole was lost in the stem- group of the Acercaria + Holometabola (he excludes Zoraptera from the Phalloneoptera) and secondarily regained only in the Thysanoptera. If Zoraptera are placed as sister-group to the Acercaria, reversal must be postulated to have occurred twice (once in Zoraptera, once in Thysanoptera),- still a more parsimonious proposal than convergent gain of the polytrophic condition in Psocodea, Hemiptera and Holometabola. 15. Reduction of ovipositor (Boudreaux, 1979). 16. Loss of cardo (Matsuda, 1965; Boudreaux, 1979). Kim and Ludwig (1978b) consider that the cardo is fused to the stipes rather than lost. 17. Loss of loral arm (Matsuda, 1965). The homology of the loral arm in Acercaria with that of Orthopteroidea is doubtful (Matsuda, 1965), but the structure in Condylognatha known by this term is absent in the Psocodea. 18. Modification of sitophore to produce cup-shaped (•cibarial*) sclerite (Ktfnigsmann, 1960; Boudreaux, 1979). See comments following characters 19 and 21. 19. Development of epipharyngeal process fitting into sitophoral cavity (Kristensen, 1975). 82

The sitophore sclerite and the epipharyngeal crest together form the cibarial pump, and are present not only in Psocodea, but also in Thysanoptera (Matsuda, 1965; Heming, 1978). The development of the cibarial pump is probably plesio- morphic within the Acercaria (c.f. character 4) and secondarily modified in the Hemiptera. See comment following character 21. 20. Expansion of lingual sclerites (Matsuda, 1965). Boudreaux (1979) refers to these as "ovoid sclerites of the hypopharynx", apparently in the belief that they are entirely novel structures. See comment following character 21. 21. Development of filamentous ligaments connecting lingual and sitophore sclerites (Matsuda, 1965)• The so-called 'ligaments' are tubular ducts (Rudolph, 1982b). Characters 20 and 21 with 18 and 19 are features of a single functional system in the Psocodea (see discussion of hypopharyngeal morphology below), and perhaps should not be considered as separate apomorphies. These characters are presented as apomorphic for the Psocodea, secondarily obscured in the highly-modified Rhyncophthirina and Anoplura (though seen in the anopluran embryo by Seeger, 1979). Hennig (1981), however, points out that the mouthparts of the Condylognatha are also highly modified, and it is possible (though not demonstrable) that characters 20 and 21, like characters 18 and 19, (discussed in the same terms by Hennig (1981), are apomorphic not for the Psocodea but for the Acercaria, and are secondarily lost in the Condylognatha. 22. Lacinia protrusable and retractable, serving as 'pick' (Boudreaux, 1979). Boudreaux (1979) includes in his original statement of this character the detachment of the lacinia from the maxilla, despite having already cited this as apomorphic for the Acercaria. Heming (1978) notes that the lacinia of Thysanoptera is also protrusable and retractable, whilst 83

Hennig (1981) suggests that some degree of lacinial trans- formation is characteristic of all Acercaria. Smithers (1972) holds a conflicting view, maintaining that this character is an autapomorphy of the Psocoptera and is "convergent in some Mallophaga". The view of Hennig (1981) is followed here, and the character taken as a possible apomorphy of the Acercaria. 23. Aedeagus not formed by fusion of mesomeres (Boudreaux, 1979). By the definition accepted in this study (that of Snodgrass, 1935) the aedeagus, if present, is always formed by the fusion of the mesomeres. The aedeagus is apparently absent in the Psocodea, though the mesomeres do fuse apically in many species. Absence of the aedeagus may be plesiomorphic or an apomorphic reversal (secondary loss) for the Psocodea; the matter is discussed in detail in the morphology section (P. 173). 24. Parameres and raesomeres both clasping (Boudreaux, 1979)• Boudreaux may be suggesting this apomorphy following his erroneous conclusion that the mesomeres do not fuse; certainly there is no clasping function of the mesomeres when they do fuse. In many cases it is apparent that the parameres do not clasp the female either. There are very few observations on mating mechanics in the Psocodea, and very little evidence to support Boudreaux' contention. Significance of male genitalic characters to the holouhyly of the Psocodea. It is tempting to use the structure of the male genitalia as support for the holophyly of the Psocodea, though the assignment of apomorphic status to any particular character state is difficult. In phenetic terms, male genitalia of Psocoptera and Phthiraptera are so similar that (in the author's view) a specimen could not be placed confidently in one order rather than the other by study of the male genitalia alone and without knowledge of the species or genus concerned. Smithers (1972), however, considers the male genitalia of Psocoptera and Phthiraptera totally different and supporting the holophyly of the former group with respect to the latter. The 84

morphology of the psocodean male genitalia is discussed in detail below. Psocoptera Shipley, 1904. (= Gorrodentia Burmeister, 1839). 25« Reduction in thickness of chorion (Seeger, 1979; Seeger in Heirnig, 1981). The very thin chorion is offset by a thick serosal cuticle (suborder ) or added suprachorionic layers such as coverings of silk, encrustations from anal secretions and other debris (suborders and Psocomorp-ha) (Seeger, 1979). 26. Loss of micropyles (Seeger, 1979; Seeger in Hermig, 1981). The absence of micropyles is doubtless linked to the thinness of the chorion. The contention of KBnigsmarm (1960) and Haub (1980) that the absence of micropyles is plesiomorphic in the Acercaria, and their presence in the Phthiraptera an apomorphy for that group, does not seem tenable in view of their presence in the Condylognatha. 27. Loss of aeropyles (Seeger, 1979; Seeger-in Hennig, 1981). The absence of aeropyles is doubtless linked to the thinness of the chorion. 28. Adoption of unusual (dorsal) position in egg by embryo, with no rotation of egg and embryo (Seeger, 1979; Seeger in Hennig, 1981). Goss (1953) maintains that the embryology of Liposcelis divergens (Psocoptera: Liposcelidae) reveals affinities with the 'Mallophaga1 (Phthiraptera: Ischnocera plus Amblycera plus Rhyncophthirina), but does not refer to the positioning of the 'mallophagan' embryo in the egg. 29. Adoption of unusual manner of folding of embryonic appendages (Seeger, 1979; Seeger in Hennig, 1981). Seeger (in Hennig, 1981) states that the folding of the appendages is "very different from the ground-plan of the Psocodea and all other Paraneoptera". 30. Suppression or reduction of prophragma and mesophragr.a (Boudreaux, 1979)• 85

The phragmata are reduced in the Phthiraptera, but the mesophragma, at least, is not reduced in the Psocoptera. 31. Insertion of dorsal longitudinal mesothoracic muscles anteriorly on strongly-arched mesoscutum (Boudreaux, 1979). 32. Reduction of prothorax (Smithers, 1972; Boudreaux, 1979:226). Boudreaux (1979:280) lists "Metathorax small" - presumably an error for prothorax. The character was limited by Smithers (1972) to winged forms only, and does not apply to the apterous members of the order. 33* Spermatozoa biflagellate and uniflagellate in each sperm bundle (Boudreaux, 1979). The combination of biflagellate and uniflagellate spermatozoa (plesiomorphic within the Acercaria and Insecta respectively) may be considered apomorphic. However, the extent of the condition and the state in the Phthiraptera is not known, as published infoimation is very limited. 34. Media of forewing three-branched (Boudreaux, 1979). This feature is not found in some Permopsocida (Hennig, 1981), presumably for the reasons outlined following character 1. See comment following character 37. 35. Cu and Mp united at base (Boudreaux, 1979). See comment following character 37. 36. Rs forked (Boudreaux, 1979). Hennig (1981) refers to this as a plesiomorphic character for the Psocodea. See comment following character 37. 37. Wings coupled at rest by blunt projection of stigma of forewing (Boudreaux, 1979; Seeger, 1979). This character is rather more variable than implied by Boudreaux (1979) (New, 1974) but it is unlikely that it arose more than once, and thus may be taken as apomorphic within the Acercaria (see below). The wing-coupling mechanism utilised during flight is separate, operating in the same way as does than in Heteroptera and Coleorrhyncha. Though the four characters 34 - 37 are suggested as 86

apomorphies for the Psocoptera it is clear, despite the assertions of Smithers (1972) and Boudreaux (1979), that they do not provide evidence precluding paraphyly of the group with respect to the secondarily apterous Phthiraptera. 38. Antennal flagellum slender, setiform (Boudreaux, 1979). The antennal flagellum of many Condylognatha might also be described as slender and setiform, rendering the exact nature of this supposed autapomorphy rather obscure. Some Psocoptera have the highest number of flagellomeres within the Acercaria (ranging from 10 to 50), contrasting with Phthiraptera (1 to 3), Thysanoptera (4 to 9) and Hemiptera (2 to 23). However, very high numbers are found in some Holometabola and in some of the Orthopteroid orders, and a high number of antennal flagellomeres may be plesiomorphic . for the Acercaria. The form of the psocopteran antenna does not preclude reduction to the phthirapteran condition and does not, therefore, provide evidence against paraphyly of the Psocoptera. 39. Pom of the female genitalia (Smithers, 1972). Smithers (1972) is not precise about the nature of this supposed apomorphy, but there seems no reason to suppose that the very reduced female genitalia of the Phthiraptera (see discussion of morphology, p. 155) could not have been derived from the psocopteran type.

Phthiraptera Haeckel. 1896. (sensu Weber, 1939). 40. Haploid reduction in primary spermatogonia (White, 1957). Apparently a unique autapomorphy for the Phthiraptera and not detected in the Psocoptera (Meinander _et al., 1974), though studies of spermatogenesis are very limited. 41. Development of hydropyle in egg. 42. Development of operculum in egg (KSnigsmann, 1960; Haub, 1980). 43. Egg-cement vaginal, not anal (KSnigsmann, I96O; Haub, 1980). This apomorphy is proposed following the observation that in Phthiraptera the egg-cement is produced in glands apparently 87

absent in Psocoptera (Florence, 1921; Weber, 1936; Mukerji & Sen-Sarma, 1955), but is not linked, to any observations of the actual source of the cement in Psocoptera (or in those Phthiraptera which also lack the glands). Without these latter observations the apomorphy must be regarded as unfounded. See discussion of female genital morphology below. 44. Dorsoventral compression of head (Boudreaux, 1979). Boudreaux (1979) combines prognathy with this character. The Ischnocera are hypognathous and the Amblycera only just prognathous, however, and it is equally parsimonious to suggest that prognathy arose within the group twice, once in the Amblycera and once in the Rhyncophthirina - Anoplura (see discussion of head morphology below). 45. Movement of supraoesophageal ganglion posteriad (KQnigsmann, 1960; Boudreaux, 1979). This feature is probably linked to the alteration of head shape (character 44). 46. Loss of dorsal tentorial arms (Symmons, 1952; Boudreaux, 1979). This feature is probably linked to the alteration of head shape. 47. Virtual loss of discrete abdominal ganglia (KOnigsmann, ' 1960). This incorporation of abdominal and third thoracic ganglia into a composite ventral ganglion is an extreme of a trans- formation series, amalgamation of the abdominal ganglia being a feature of the Acercaria (character 9). 48. Reduction of lacinial stylets (KSnigsmann, 1960; Boudreaux, 1979). The stylets are lost in the Rhyncophthirina and Anoplura. 49. Development of lacinial gland (Symmons, 1952; Boudreaux, 1979). Though used as an apomorphy for the Phthiraptera, the character is not present in the highly-modified Rhyncophthirina and Anoplura, and the presence or absence of the gland in the stem-groups of these suborders cannot be directly determined. 50. Great reduction of maxillae (Matsuda, 1965; Boudreaux, 1979). 51. Reduction of labial palpi (Symmons, 1952; Boudreaux, 1979). 88

In the Rhyncophthirina and. Anoplura the labial palpi are completely suppressed. 52. Reduction of antennal flagellum to three flagellomeres (plus pedicel) (Kdnigsmann, 1960; Boudreaux, 1979). In some cases the flagellum is reduced to fewer than three flagellomeres (see morphological discussion below). 53. Loss of wings (Boudreaux, 1979). See comment following character 55. 54. Loss of ocelli (KOnigsmann, 1960; Boudreaux, 1979). Throughout the Acercaria the presence of-wings and ocelli are linked, and therefore this character cannot be considered as independent from character 53* See comment following character 55. 55- Reduction of compound eyes to two ommatidia (Boudreaux, 1979). Further reduction, to one ommatidium or complete absence, takes place within the group (see character 96). The' three characters above may be regarded as the result of heterochronous development, perhaps linked to the reduction in the number of nymphal stadia (see character 56) (Matsuda, 1976). Although these characters are presented as autapo- morphies for the Phthiraptera, they are all found in at least some members of the Liposcelidae (Psocoptera). 56. Reduction to three nymphal stadia (Kdnigsmann, 1960; Boudreaux, 1979; Haub, 1980). Psocoptera have four to six nymphal stadia and, rarely, three (New, unpublished), not five or six as suggested by Haub (1980). 57. Loss of abdominal spiracles I and II (KBnigsmann, 1969; Boudreaux, 1979). Similar loss is found in some members of the Liposcelidae (Psocoptera). 58. Loss of metathoracic spiracle (KGnigsmann, I96O; Boudreaux, 1979). 59. Loss of trochantin (Boudreaux, 1979) 60. Reduction of ovipositor. Boudreaux (1979) suggests that the ovipositor is completely lost, but gonapophysis VIII has been retained in Trichodectidae, Rhyncophthirina, Anoplura and some Amblycera and Ischnocera 8q

(see morphological treatment below, and discussion following character 101). 61. Reduction of testicular follicle number to three (Boudreaux, 1979). Further reduction to two follicles occurs in the Phthiraptera, the character polarity being supported by histological evidence (Schmutz, 1955) (see character 97). The status of the reduction to three as an apomorphy of the Phthiraptera is more doubtful. Reduction has certainly occurred in the Psocodea, as more follicles are present in other Acercaria and the Orthopteroid orders (Matsuda, 1976). However, within the Psocoptera the number is either one or three (rarely four), the Liposcelidae for example having three. The pattern of variation indicates a considerable degree of homoplasy, and there is disagreement over the plesiomorphic state for the order (Wong and Thornton, 1968; Matsuda, 1976). In view of •this, character 61 cannot be utilised as an apomorphy for the Phthiraptera. 62. Adoption of ectoparasitic habit on vertebrate hosts (Boudreaux, 1979). Kim and Ludwig (1978b, 1982)-argue that the ectoparasitic habit was attained separately in Anoplura and "Mallophaga",' though they present little evidence supporting this view. However, many of the above characters are, or may be, adaptations to ihe parasitic mode of life and thus show the same pattern of distribution as ectoparasitism, which is therefore not utilised as an autapomorphy for Phthiraptera.

2.1.2.3. Significance of above characters

Though a number ofthe above characters are not suitable as apomorphies indicating relationships within the Phalloneoptera, the distribution of acceptable apomorphies still supports the holophyly of the Acercaria, Psocodea and Phthiraptera (but see discussion on p. 187 ). The holophyly of the Psocoptera is less certain, and is discussed separately below (2.1.2.4.). Haub (1973) suggests that the Rhyncophthirina forms the sister-group to all other Psocodea, and 90

that v/ithin the latter group the Trichodectidae forms the sister-group to all the rest. This contention is based solely on a postulated evolutionary sequence of morphological changes in the sitophore sclerite and is not supported by other characters.

2.1.2.4. Relationship of Phthiraptera to Liposcelidae

Introduction. The holophyly of the Psocoptera is supported by five characters: 25 (with 26 and 27) and 28 (with 29). These would be sufficient were it not for a number of synapomorphies of Liposcelidae (Psocoptera: Troctomorpha) and Phthiraptera, which indicate the paraphyly of the Psocoptera with respect to the Phthiraptera. Smithers (1972) suggests that the Liposcelidae are "highly derived", but most of the derived features cited are those in which the Liposcelidae approach the Phthiraptera. Apomorphies proposed by Smithers (1972) for the establishment of the phylogenetic position of the Liposcelidae (Fig. 6), with the exception of those pertaining to the wings, are listed and discussed below. The distribution of these characters, together with others listed above and pertinent to the holophyly of the Psocoptera, is presented in Table II.

Apomorphies proposed for Liposcelidae and associated groups

63. Loss of chorionic sculpturing of the egg. Smithers (1972) proposes this as an apomorphy of Troctomorpha and . Seeger (1979)» however, points out that the sculpturing on trogiomorph eggs is derived from serosal cuticle, not the chorion. Loss of chorionic sculpture is therefore an apomorphy of all three suborders, though doubtless linked to the thinness of the chorion (see discussion following character 25). 64. Development of T-shaped sclerite in female subgenital plate. This sclerite is absent in some members of the Liposcelidae (Smithers, 1972). 65. Development of secondary annulations on antennal flagellomeres distal to flagellomere IV. 91

LIPOSCELIDAE

SPHAER OP S OCIDAE

PACHYTR OOTID AE

AMPHIENOMETAE

PSOCCMCRPHA

TROGICMGRPHA

Fig. 6. The phylogenetic position of the Liposcelidae, according to Smithers (1972). 92

PSOCOPTERA

P3 W

Table II. Distribution of apomorphies within Psocodea. Gain states are indicated by 'g', loss states by flf, and states not present in all members of a taxon by '(g)' and *(l)' respectively. For explanations of characters see text. 93

The absence of flagellomeres distal to flagellomere III in Phthiraptera renders this character irrelevant to any discussion of paraphyly or holophyly of Psocoptera. It is notable, however, that the flagellomeres of some Amblycera (Phthiraptera) do possess secondary annulations. 66. Loss of pilosity on external lobe of gonapophyses. The absence of gonapophyses in most Phthiraptera, and the considerable degree of modification to these structures when they are present, renders this character irrelevant to any discussion of paraphyly or holophyly of Psocoptera. 67. Anterior closure of male basal apodeme. The ontogeny of this structure (see morphological discussion below) is such that the anterior end is always 'closed', whilst the degree of sclerotisation and of anterior divergence of the lateral margins is very variable in Phthiraptera. The character is not suitable for use in phylogenetic studies of higher taxa. 68. Loss of paraproct spine. 69. Loss or reduction of Pearman's organ. 70. Loss of trichobothrial field. 71. Shortening of legs.

Significance of above characters. From Table II it can be seen that the holophyly of the Psocoptera is supported by the distribution of seven characters, of which 25, 26, 27 and 63 are aspects of the same apomorphy and 28 and 29 are probably linked in a similar way. The holophyly of the Liposcelidae plus Phthiraptera (and thus the paraphyly of the Psocoptera) is supported by the distribution of eleven characters, of which 53 and 54 are aspects of the same apomorphy and 95 is of very dubious validity. If the Psocoptera are holophyletic there has been convergence of nine loss characters and one (dubious) gain character; if they are paraphyletic there has been secondary loss of four gain characters (one of which is anyway lost in some Liposcelidae) and, linked with one of these, reversal of three loss characters. The majority of apomorphies therefore supports holophyly of Liposcelidae plus Phthiraptera, but all these apomorphies are loss characters which, 94

as explained above (section 1.3.2.), are given low (or no) phylogenetic 'weight1 . Waage (1979) suggests that the Phthiraptera arose from a nidicolous ancestor much like the modern Liposcelidae in habit, and the holophyly of the Psocodea implies that this ancestor was itself derived from a winged insect similar to the Psocoptera (with the possible exception of apomorphies 25 - 29, 63 and 64). The selective pressures acting on this ancestor would, by reason of its habitat, have been very similar to those acting on modern (and ancestral) Liposcelidae, and may be presumed to have favoured the same set of adaptations. These adaptations include not only some of the 'loss' characters mentioned above, but also behavioural and physiological adaptations that must be accounted 'gains' . If allowance is made for these 'gain' synapomorphies of the ancestors of Liposcelidae and Phthiraptera, the hypothesis that these ancestors were the same becomes more parsimonious than the hypothesis that they were different. Because of this the Liposcelidae and the Phthiraptera are here accepted as sister-groups and the Psocoptera is postulated to be paraphyletic with respect to the Phthiraptera. It is presumed that the adoption of the ectoparasitic habit by the ancestor of the Phthiraptera imposed a new set of environmental pressures on the egg, which reverted to the plesiomorphic thick-chorion form but developed an operculum, maintaining easy egress for the nymph. Embryonic orientation within the egg also reverted to the plesiomorphic state.

2.1.3. Relationship of Trichodectidae to Other Phthiraptera

2.1.3.1. Introduction

Having established the holophyly of the Phthiraptera and identified the sister-group as the Liposcelidae, the systematic relationships of the included sub-groups may be studied. The Phthiraptera comprises four sub-groups: the Amblycera, Ischnocera, Rhyncophthirina and Anoplura. The first three groups are frequently combined as the Mallophaga, and commonly known as 'biting lice', 'chewing lice' or 'bird lice'; Anoplura are commonly referred to as 'sucking lice' or 'mammal lice'. The common epithets based on feeding mechanisms (biting, chewing and sucking) are largely accurate 95

though, as is shown below, represent a grade classification. The restriction of Amblycera, Ischnocera and Rhyncophthirina to birds is incorrect, however, as all of the Rhyncophthirina, two of the five families of Ischnocera, and four of the seven families of Amblycera feed exclusively on mammals. Before establishing the sister-group relationships between these four groups, however, the holophyly of each must be demonstrated. This poses some difficulties in the Phthiraptera, as many of the presumed apomorphies within the Order are loss characters or features intimately associated with the parasitic mode of life, both carrying a high probability of homoplasy. Amblycera, Rhyncophthirina and Anoplura each have at least some gain characters that indicate their holophyly, but this is not the case for Ischnocera, as all of the identified apomorphies of this group might also have been characters of the ancestors of Rhyncophthirina and Anoplura. Nor, within the Ischnocera, is assignment of genera to holophyletic families simple. Eichler (1963) recognises 21 families in three superfamilies (though without giving any characters whereby these might be distinguished), whilst Hopkins and Clay (1952) accept only three families, similar in composition to Eichler1 s sup'er- families. In the following discussion five families of Ischnocera are recognised: Goniodidae (Goniodes and related genera), Heptapso- gasteridae (sensu Hopkins and Clay, 1952), Trichophilopteridae (Trichophilopterus only), Trichodectidae (sensu Hopkins and Clay, 1952; Trichodectiformia of Eichler, 1963) and. (all the rest). Of these families only two are demonstrably holophyletic: Trichophilopteridae (by definition) and Trichodectidae, the holophyly of which is supported by four autapomorphies (fusion of the male flagellomeres, development of modified hook-like setae on the male flagellum, loss of one tarsal claw, and the development of lateral abdominal flecks in the male). Of the other three families, Heptapsogasteriaae and Goniodidae are phenetically-recognisable groups and may be holophyletic, whilst Philopteridae is a miscellaneous collection of genera and almost certainly paraphyletic or polyphyletic. 56

2.1.3.2. Apomorphies proposed for families and supra-familial groups within the Phthiraptera

Numerous apomorphies, not all of which are valid, have been proposed by various authors as evidence for the holophyly of groups within the Phthiraptera. These apomorphies are listed below, with references and comments where appropriate. The distribution of the apomorphies accepted as valid for analytical purposes is summarised in Table III . 72. Development of pedunculate first flagellar 'segment* (Clay, 1970; Boudreaux, 1979). Haub (1980) refers to a "peduncle-like part between the scape and the- flagellum", presumably the pedicel, as being plesio- morphic within the Psocodea, its loss being a synapomorphy of Ischnocera, Rhyncophthirina and Anoplura. However, the pedicel is not pedunculate in any of the Psocodea and the first flagellar 'segment' (first flagellomere) is pedunculate only in the Amblycera, an autapomorphy for this group. See comment following character 102. 73. Development of antennal fossa concealing antennae (Boudreaux, 1979). 74. Prolongation of head anteriorly into rostrum with mandibles terminal and rotated through 180 degrees; reduction of posterior - tentorial arms and loss of tentorial bridge; suppression of tibiotarsal muscles; extension of pretarsal apodeme into femeur without tibial muscle bundle (Boudreaux, 1979)• These and many other synapomorphies strongly indicate the holophyly of the two species of Rhyncophthirina. 75. Development of piercing stylets from hypopharynx and labium; great development of connective tissue septum (obturaculum - see character 89); fusion of pronotum to mesonotum; reduction of mesothoracic and metathoracic terga, with dorsal extension of thoracic pleura (Kim and Ludwig, 1978b; Boudreaux, 1979). These and many other autapomorphies strongly indicate the holophyly of the Anoplura. Boudreaux (1979) includes "thoracic spiracles on dorsal part of the pleurum" in his 97

list of apomorphies, but this condition is found throughout the Phthiraptera (Matsuda, 1970) and thus is not an autapomorphy of the Anoplura. The dorsal position of the spiracle on the anopluran thorax is a result of the dorsal extension of the pleura. A further character listed by Boudreaux (1979), "tibia modified for grasping hair of host" is also found in the Ischnocera, particularly in the Trichodectidae. 76. Separation of small rhombic sclerite from anterior pronotal margin. This sclerite is present in all Ischnocera except some Philopteridae, and occurs in no other group. In some species of Anatoecus (Philopteridae) the sclerite is present but not fully detached from the pronotal margin. The distribution of the character indicates the paraphyly of the Philopteridae with respect .to the other Ischnocera. 77. Absence of mandibulo-hypopharyngeal muscle (Syramons, 1952; Haub, 1973). This loss is proposed as an apomorphy for all Ischnocera except Trichodectidae, in which the muscle is reduced, but relatively few ischnoceran genera have so far been examined. The loss of the muscle in Rhyncophthirina is possibly homo- plastic and connected with the extensive modifications to the mouthparts in this group. 78. Rotation of mandible to operate about a vertical axis (Boudreaux,; 1979). The autapomorphic modifications of the mouthparts of the Rhyncophthirina and Anoplura do not preclude this form of mandibular articulation being ancestral; the character cannot, therefore, be used as an autapomorphy for the Ischnocera (as is proposed by Boudreaux, 1979). 79. Mouthparts of modified chewing type (Kim and Ludwig, 1978b). This apomorphy is proposed for a group comprising the Ischnocera and Rhyncophthirina despite the quite different modifications of the mouthparts in these two groups. The unknown attributes of the stem-group of the Anoplura are not considered by these authors. 98

80. Head 'fixed' in relation to thorax (Kim and Ludwig, 1978b). 81. Loss of anterior tentorial pits (Symmons, 1952). 82. Loss of articulation between presternum and forecoxae (Kim and Ludwig, 1978b). This loss has also occurred in some Amblycera (Matsuda, 1970). 83. Suppression of lateral cervical sclerites (Boudreaux, 1979). Boudreaux (1979) proposes this apomorphy for Rhyncophthirina plus Anoplura. Mayer (1954) considers that this sclerite is absent in Bovicola caprae (Trichodectidae), but examination has shown that it is present in this and all other species of Trichodectidae. 84. Absence of lacinial gland (Kim and Ludwig, 1978b). The gland is not present in Rhyncophthirina and Anoplura. This absence may be primary or secondary in either group and, if secondary in both, may represent a homoplastic loss. See discussion following characters 49, 78 and 85. 85. Loss of lacinia (Boudreaux, 1979). The loss of the lacinia in Rhyncophthirina and Anoplura is associated in each case with considerable modifications to the mouthparts. These modifications being separate autapo- morphies, there must be a high probability that the loss of the lacinia is homoplastic in these two groups. 86. Reduction in number of tarsal claws to one (Kim and Ludwig, 1978b; Boudreaux, 1979). Kim and Ludwig (1978b) consider this autapomorphic for the Anoplura, though admitting some homoplasy in the Amblycera and Ischnocera. Boudreaux (1979) correctly points out that there is only a single tarsal claw in the Rhyncophthirina also, and proposes the reduction as a synapomorphy of the Rhyncophthirina and the Anoplura. The reduction to a single tarsal claw occurs in and Abrocomophagidae (Amblycera), Trichodectidae, Rhyncophthirina and Anoplura, all groups ectoparasitic on mammals. A similar reduction has taken place in those Hippoboscidae (Diptera) ectoparasitic on mammals, indicating the high probability of homoplasy in this feature. 99

87. Development of large cibarial and pharyngeal pumps (Boudreaux, 1979). The cibarial pump is present in all other Acercaria, and is not an autapomorphy of Anoplura and Rhyncophthirina as suggested by Boudreaux (1979). The pharynx does not act as a separate pump, but contributes to the action of the cibarial pump-(Mukerji and Sen-Sarma, 1955; Kim and Ludwig, 1978b). 88. Development of saucer-shaped antennal sensilla (Clay, 1970; Kim and Ludwig, 1978b; Boudreaux, 1979). 89« Development of connective tissue septum (obturaculum) nearly occluding occipital foramen (Symmons, 1952; Kim and Ludwig, 1978b; Boudreaux, 1979). The development of this feature may be allied to the reduction of the tentorial bridge, and is carried to a far greater degree in the Anoplura than in the Rhyncophthirina and Ischnocera. A transformation series thus exists, from the plesiomorphic 'undeveloped' in Amblycera and other Insecta, through 'developed' in Ischnocera and Rhyncophthirina, to 'greatly developed' in Anoplura. The use.of the 'developed' in contrast to the * greatly developed' state as a synapomorphy of Ischnocera and Rhyncophthirina (Kim and Ludwig, 1978b) is not justified. 90. Development of spiracular glands (Kdnigsmann, 1960; Kim and Ludwig, 1978b; Haub, 1980). Spiracular glands are associated with the abdominal spiracles in Ischnocera, Rhyncophthirina and Anoplura. Kim and Ludwig (1982) attempt to dismiss this apomorphy by stating that "information on the Psocoptera and Amblycera is meager.". This view is not maintained here, and the apomorphy is accepted. 91. Development of tensory ligament muscle (Symmons, 1952). 92. Development of occipital apodeme extending into thorax (Boudreaux, 1979). The presence of the occipital apodeme is not always easy to confirm as its size and degree of sclerotisation is very variable. The structure cannot be detected in Trichodectidae 100

and Goniodidae, but in these groups the occipital ring is very heavily sclerotised and may project into the thorax, possibly subsuming the occipital apodeme. It is not present in some Philopteridae, and the absence of the structure in most Anoplura is interpreted by Kim and Ludwig (1978a) as an apomorphic reduction. 93. Transfer of attachment site of antennal muscles to dorsum of head, at least in part. This may be related directly to the reduction of the tentorium. Rhyncophthirina and Anoplura lack the anterior tentorial arms and thus the antennal muscles must arise on the dorsum of the head. In all Ischnocera the anterior tentorial arms are present, though reduced, but the Trichodectidae do not have any of the antennal muscles arising from the dorsum of the head whilst the Philopteridae and the Goniodidae do (Symmons, 1952). The condition in the Heptapsogasteridae and Trichophilopteridae is not known. 94. Presence of symbionts (KiJnigsmann, 1960; Haub, 1980). Use of this character follows the work of Ries (1930, 1931), who records symbiotic bacteria or Rickettsia-like organisms in Rhyncophthirina, Anoplura, seven species of Philopteridae, four out of five species'of Goniodidae and one species of Amblycera (Somenacanthus stramineus). Four species of Trichodectidae and species of eight genera of Amblycera were examined with negative results. Heptapsogasteridae and Trichophilopteridae were not examined. The location of the symbionts and the form of the mycetocytes or mycetomes differ considerably in Amblycera, Ischnocera, Rhyncophthirina and Anoplura (Buchner, 1965). Buchner (1965) draws attention to the connection between the diet of the insect host and the presence of symbiotic bacteria, which are present chiefly in blood- and keratin-feeding insects. Kim and Ludwig (1978b) suggest a number of separate acquisitions of endosymbionts in the parasitic Psocodea, connected with changes in diet. The absence of symbionts in one of the Goniodidae studied suggests either independent acquisition in the other 101

Goniodidae or secondary loss. In view of the high probability of homoplasy and the paucity of information regarding the full distribution of endosymbionts within the Phthiraptera, this character is not given great weight in this study. 95. Fusion of mesonotum and metanotum (Kim and Ludwig, 1978b). This apomorphy is tentatively proposed here for the Ischnocera, Rhyncophthirina and Anoplura, though it is exhibited by some Amblycera (Clay, 1970) and Liposcelidae (Smithers, 1972). Thoracic fusion is completed in the Anoplura, with the pronotum fused to the mesonotum (character 75). Kim and Ludwig (1978b) propose "fusion of mesonotum and metanotum (but not pronotum)" as an autapomorphy of Ischnocera and Rhyncophthirina, but this decision does not seem justified. 96. Reduction of eye to a single ommatidium (or complete loss) (Boudreaux, 1979). See comment following character 55. 97. Reduction in number of testicular follicles to two pairs (KOnigsmann, 1960; Kim and Ludwig, 1978b; Boudreaux, 1979; Haub, 1980). See comment following character 61. Kim and Ludwig (1982) recognise the unsuitability of this character in phylogenetic inference within the Phthiraptera. 98. Loss of maxillary palpi (Boudreaux, 1979). The loss of the maxillary palpi in Ischnocera, Rhyncophthirina and Anoplura may be homoplastic. See comment following character 85. 99. Reduction from two to one pair of labial glands (Kim and Ludwig, 1978b). See comment following character 100. 100. Presence of cephalic labial gland (Kim and Ludwig, 1982). This apomorphy is contrasted with the thoracic placement of both pairs of labial glands in Psocoptera and Anoplura, and is proposed as synapomorphic for Amblycera, Ischnocera and Rhyncophthirina, as was character 99. Both characters 99 and 100 are based on the misconceptions that: a) The lingual 102

sclerites are labial salivary glands (Snodgrass, 1899; Cope, 1940a, 1940b, 1941), and b) The thoracic glands are connected to the alimentary canal, not the salivarium (Snodgrass, 1899; Imras, 1925; Richards and Davies, 1977). Paired thoracic labial glands, with a joint duct opening into the salivarium, are present in the Psocoptera (Weber, 1938), Anoplura (Snodgrass, 1944; Haug, 1952), Amblycera (Haug, 1952; Buckup, 1959) and Ischnocera (Risler, 1951; Haug, 1952). In Rhyncophthirina the dorsal labial glands have migrated to the head (Mukerji and Sen-Sarma, 1955), however, and character 100 is autapomorphic for this group. Character 99 is rejected. The identification and positioning of the labial glands is discussed in the General Morphology on p. 133. 101. Presence of gonapophyses (Symmons, 1952). This character is proposed as an apomorphy of the Trichodectidae,' Rhyncophthirina and Anoplura. Gonapophyses, however, are found in a number of Amblycera and Ischnocera, and are remnants of the greatly-reduced ovipositor. The retention of the gonapophyses cannot be used as a synapomorphy. See discussion following character-60. 102". Antennae homonomous (KOnigsmann, 1960; Haub, 1980). This is proposed as a synapomorphy of Ischnocera, Rhyncophthirina and Anoplura, contrasting with the supposed plesiomorphic heteronomous condition of the antennae of Psocoptera and Amblycera. It has been shown above (see discussion following character 72) that Haub (1980) misinter- prets the morphology of the Smblyceran antenna, and his postulated similarity of Psocoptera and Amblycera is not a reality. Moreover, the scape and pedicel are of greater diameter than the flagellomeres not only in the Psocoptera but also in the Rhyncophthirina, most Anoplura and most Ischnocera, though not in Amblycera. The degree to which the Ischnocera, Rhyncophthirina and Anoplura exhibit the heteronomous condition is not as great as the Psocoptera, 103

but the antennae in these groups can not be considered homonomous. 103* Closure'of posterior border of sitophore sclerite (Haub, 1973). The form of the sitophore sclerite in the ancestor or ancestors of Rhyncophthirina and Anoplura, in both of which it is absent, cannot be known. Haub (1973) presents a phylogeny for the Psocodea based solely on his interpretation of the characters of the hypopharynx. He suggests that the Anoplura forms the sister-group of the mammal-infesting Amblycera, whilst all the Amblycera plus the Anoplura form the sister-group of the Ischnocera less Trichodectidae. The closure of the posterior border of the sitophore sclerite is proposed as an apomorphy of all the Psocodea except Trichodectidae and Rhyncophthirina. Haub (1980) accepts the more traditional views of KBnigsmann (1960).

2.1.3.3. Significance of above characters

The holophyly of the three groups Amblycera, Rhyncophthirina and Anoplura is indicated by characters 72 and 73, 74 and 100, and 75 respectively. Within the fourth group, Ischnocera, the holophyly of the Trichodectidea is indicated by the four characters listed on p. 95 though the Ischnocera itself, with or without the Trichodectidae, is not demonstrably holophyletic. Inspection of Table III reveals a good general agreement of the apomorphy distributions between the groups. The hypothesis that the Rhyncoohthirina is the sister-group to all or part of the Ischnocera (Hopkins, 1949; SchUfer, 19&4; Kim and Ludwig, 1978b, 1982) is supported by character 77 (the loss of the mandibulo-hypopharyngeal muscle) only. Four characters (80-83), in supporting the holophyly of a group comprising Rhyncophthirina and Anoplura, refute this conclusion, which is consequently rejected on the grounds of parsimony. Character 77 is considered as homoplastic in Rhyncophthirina and some Ischnocera (_s. lat.), and Rhyncophthirina and Anoplura are accepted as sister-groups. The holophyly of a group comprising Amblycera and Ischnocera might be inferred from the distribution of 104

ISCHNOCERA

wp pM

Table III. Distribution of apomorphies within Phthiraptera. Gain states are indicated by fg', loss states by 'l', and states not present in all members of a taxon by '(g)' and *(l)f respectively. For explanation of characters see text. 105

the 1 acini al gland, which is absent in Rhyncophthirina and Anoplura but present in the former groups (characters 49 and 84). Ten characters (88-98), in supporting the holophyly of a group comprising all of the Ischnocera, the Rhyncophthirina and the Anoplura, refute this conclusion, which is consequently rejected on the grounds of parsimony. The lacinial gland is inferred to have evolved in the common ancestor of the Phthiraptera, and to have been lost either once, in the common ancestor of the Rhyncophthirina and Anoplura, or twice, once in each group. The holophyly of a group comprising the Ischnocera, Rhyncophthirina and Anoplura, which forms the sister-group to the Amblycera, (Kdnigsmann, 1960; Kim and Ludwig, 1978b; Boudreaux, 1979; Haub, 1980) is accepted. Because the Ischnocera (including or excluding the Trichodectidae) are not demonstrably holophyletic it follows that the Ischnocera (less Trichodectidae) may be paraphyletic with respect to the Trichodectidae and/or the Rhyncophthirina plus Anoplura. The Ischnocera cannot, therefore, be cited as the sister- group of either of the other groups. Three alternative dichotomously- branching cladograms can be drawn indicating possible relationships of the Rhyncophthirina plus Anoplura, the Trichodectidae and the Ischnocera (less Trichodectidae) (Pig. 7a). Apomorphies common to all three groups or characteristic of only one of the groups are omitted from the figure. Of the three cladograms C, indicating a sister-group relationship between Trichodectidae and the other groups, is the most parsimonious, requiring fewest homoplastic gains or reversals. However, characters 92, 93 and 94 are far from satis- factory, as can be seen from the discussions above, and it is felt that character 76 should be more heavily weighted than these three characters. The problem is resolved temporarily by depicting a trichotomy for the three groups (Pig. 7"b ) • Thus the identity of the sister-group of the Trichodectidae is not known, although it is least likely to be the Rhyncophthirina plus Anoplura. This situation is plainly unsatisfactory, but must be accepted at this time. It is hoped that further revisionary work on the genera and families of the Ischnocera will elucidate the systematics of this group, and allow identification of the sister-group of the Trichodectidae. 106

RA

76 i 192 f93 194

B

AMBLYCERA TRICHODECTIDAE ISCHNOCERA RHYNCOPHTHIRINA ANOPLURA

174 175 1100

180 182 183 184 V- 192 193 194

188 189 172 190 173 191 -095 >96 >97 i 198

b

Fig. 7« Cladistic relationships of major groups of Phthiraptera: (a) Three alternative hypotheses of relationships of Trichodectidae (,Tt), Ischnocera less Trichodectidae ('I') and Rhyncophthirina plus Anoplura ('RA'); (h) Cladogram of major groups of Phthiraptera. 'Gain1 character states are indicated by 'loss' states by ' and states not present in all members of a taxon by * n '. For explanation of characters see text. CLAY, 1970

AMBLYCERA SUBORDER AMBLYCERA ISCHNOCERA SUBORDER ISCHNOCERA • ORDER PHTHIRAPTERA RHYNCOPHTHIRINA SUBORDER RHYNCOPHTHIRINA ANOPLURA SUBORDER ANOPLURA

BOUDREAUX, 1979

AMBLYCERA ORDER MALLOPHAGA ISCHNOCERA 1 SUBORDER ISCHNOCERA RHYNCOPHTHIRINA INFRAORDER RHYNCOPHTHIRINA 1 [ 0RDER ANOPLURA Y SUBORDER LIPOGNATHA J ANOPLURA INFRAORDER SIPHUNCULATA J

KIM & LUDW IG, 1978h

AMBLYCERA SUBORDER AMBLYCERA ISCHNOCERA SUBORDER ISCHNOCERA • ORDER MALLOPHAGA RHYNCOPHTHIRINA SUBORDER RHYNCOPHTHIRINA . ANOPLURA ORDER ANOPLURA

FIG. 8. Alternative systematic arrangements of the major groups of lice. 108

In discussion, the family Trichodectidae is referred to the Ischnocera in accordance with current usage, and morphological conclusions are reached with reference more to the Ischnocera than to the more highly- derived Rhyncophthirina-Anoplura line.

2.1.3.4. Ranking and classification

Following production of a cladistic scheme of relationships for the groups of lice (Fig. 7b ), a system of ranking and nomenclature for these groups must be adopted. Several such schemes have been proposed, and these are summarised in Fig. 8 . Clay (1970) and Haub (1980) consider Amblycera, Ischnocera, Rhyncophthirina and Anoplura to be of equal taxonomic status as suborders of the order Phthiraptera. Boudreaux (1979) considers Phthiraptera as a subcohort (Phthiriapterida), comprising two orders: Mallophaga, which he restricts to the Amblycera, and Anoplura. Two suborders of Anoplura are proposed: Ischnocera and Lipognatha, the latter comprising two infraorders, Rhyncophthirina and Siphunculata (Anoplura sensu Clay, 1970). Kim and Ludwig (1978b) imply subordinal status for Amblycera,_ Ischnocera and Rhyncophthirina, combining these into the order Mallophaga, and elevate Anoplura to ordinal status. The same proposal is made by Kim and Ludwig (1982), though based on the premise (rejected here) that the 'Mallophaga1 are a holophyletic group; this will not be discussed further. These three schemes (Clay, 1970^ Kim and Ludwig, 1978b; and Boudreaux, 1979) reflect different approaches to systematic decision- making. Boudreaux (1979), employing a strict cladistic methodology, accepts only holophyletic groups as true systematic entities and applies names in rank order to each group so produced. Kim and Ludwig (1978b) represent the phylist approach, accepting paraphyletic groups as systematic entities and using the 'amount' of morphological difference between taxa as a criterion in assessment of rank. Clay (1970) follows cladistic practice in not accepting paraphyletic groups, and specifically rejects 'Mallophaga' as proposed by Kim and Ludwig (1978b). She does not feel obliged, however, to produce a series of names for each presumed holophyletic group, as does Boudreaux (1979). 109

The paraphyletic group 'Mallophaga' sensu Kim and Ludwig (1978b) which, in common with other paraphyletic groups, cannot be defined by apomorphic character states without reference to excluded groups, is not accepted here (see section 1.4-3• above for a discussion on holophyletic and-paraphyletic groups in systematics). The plethora of names employed by Boudreaux (1979) is also rejected, because the alteration from the traditional usage of Mallophaga and Anoplura will, far from clarifying systematic discussion, serve only to increase confusion. It is possible, however, for a classification to reflect cladistic relationships without including names for all holophyletic groups. The convention that pennits this, known as 'phyletic sequencing', is described in section 1.4.3. above. By using phyletic sequencing the Amblycera, Ischnocera, Rhyncophthirina and Anoplura can each be regarded as suborders within the order Phthiraptera, with sister-group relationships as set out in Figure 7b . This is the classification that will be accepted here. 110

2.2. GENERAL MORPHOLOGY

2.2.1. Terminology

The terminology used in morphological and taxonomic studies of the Phthiraptera is very confused and in some cases inadequate. Problems of three general types have been encountered during the present study: lack of terms, limited applicability of available terms, and multiplicity of terms for single structures. These are considered separately: i. Lack of terms. Because of the limited and rather patchy nature of published morphological work some structures have not been described and named. New terms have been coined for such structures where required, combining, where possible, clarity of meaning and compatibility with existing terminology. ii. Limited applicability of available terms. Species taken as models in morphological studies of higher taxa have sometimes been aberrant in one or more of the features described, and unsuccessful generalisations have been made. K6ler (1938), for example, discusses the male genitalia of the Trichodectidae using as a model Trichodectes canis, an atypical species representing a very advanced condition within the family. Terminology derived in studies with a limited observational base is frequently not widely applicable and where necessary has been discarded or modified during this study. iii. Multiplicity of terms for single structures. Numerous terms have been applied to some structures within the Phthiraptera for a number of reasons: (a) lack of knowledge of previously-published work, (b) renaming of structures without scientific cause, (c) conflicting morphological interpretations and (d) development of different terminologies to perform different functions. Parallel terminologies for identical morphological concepts (a andb) may be resolved on the basis of common usage or applicability. Differing morphological interpretations require value judgements before a choice can be made; such judgements are of course implicit in any morphological study, though must be constrained by the extent of the study. Clay (1951, 1970) attempts 111

to resolve the problem without further morphological investigation by developing a separate terminology for taxonomic use. This is intended to be 'morphologically neutral' and independent of differing morphological interpretations, thus providing a degree of termino- logical stability. In addition it provides names for features of interest to the taxonomist but of little significance to the morphologist. However, morphology and taxonomy are not totally separate disciplines, and terms proposed initially as descriptive conveniences not indicative of homology have been interpreted by some workers as having homological connotations. The very provision of a terminology for general application does, despite disclaimers, imply a homological relationship for structures given the same name. Within this study terms are applied where possible as a result of morphological interpretation, and a glossary is included (Appendix B). Structures given terms of convenience for descriptive use, where no homology is intended to be implied, are indicated as non- homologous in the glossary

2.2.2. The Head

2.2.2.1. Introduction

The phthirapteran head is characterised by a series of reductions from the condition presumed for the common ancestor of all Psocodea. The ancestral head was probably hypognathous with slightly reduced mouthparts, long antennae, large compound eyes, three ocelli and a well-developed tentorium. The typical phthirapteran head is dorso- ventrally compressed and more or less prognathous with reduced or greatly-modified mouthparts, short antennae, small or non-existent compound eyes, no ocelli and a reduced tentorium. Major modifications of the head and mouthparts are present in the Rhyncophthirina and the Anoplura. The head of the Rhyncophthirina is globular with no clear sulci, and extends into a long slender rostrum anteriorly, at the end of which are the mouthparts. The anopluran head is conical with a single transverse sulcus, and the mouthparts consist of three protrusible stylets formed from the fused maxillae, a papilla of the salivary duct, and the labium (Ferris, 1951; Kim and Ludwig, 1978b). 112

Because the complex modifications of the Rhyncophthirina and Anoplura have no bearing on the condition of the head of Trichodectidae these two groups are not treated in detail here. The morphology of the head of Anoplura has been discussed by Ferris (1951), Young (1955), and Matsuda (1965), and that of Rhyncophthirina by Mukerji and Sen-Sarma (1955).

2.2.2.2. Structure of the head capsule

The phthirapteran head is frequently described as dorso-ventrally compressed and prognathous (Symmons, 1952; Matsuda, 1965; Kim and Ludwig, 1978b). Richards and Davies (1977) define prognathous heads as those in which the long axis is horizontal or slightly inclined vertically and the mouthparts nearly or completely anterior in position, whilst hypognathous heads are defined as having the long axis vertical and the mouthparts ventral. The orientation of the long axis of the head and the position of the mouth are not necessarily as interdependent as implied by the definitions, however, and are discussed separately below. The degree of dorso-ventral compression of the phthirapteran head is difficult to assess because most specimens are mounted on their dorsal or ventral surface on glass microscope slides, obscuring the vertical component of the insects' dimensions. K^ler (1960a) figures the lateral aspect of the head and thorax of Nesiotinus demersus Kellogg (Ischnocera): the head is globular,• and measure- ments made from the figure indicate a height to length ratio of 1.0, compared to about 0.8 for a typical psocopteran (an average of measurements of figures from several sources). Kim and Ludwig (1978a) refer to the anopluran head as conical, and many species of this suborder have heads with a roughly circular cross-section. However, examination of numerous specimens of lice preserved in alcohol [ B.M.(N.H.) collection] indicates that dorso-ventral compression of the head, whilst not uniformly as great as is generally assumed, is marked within the Amblycera and Ischnocera. The change in orientation of the long axis of the head from vertical to horizontal is unlikely to have been achieved by dorso- 113

ventral contraction alone, but this must have been accompanied by elongation and broadening of the head. Symmons (1952) shows that in the Amblycera and Ischnocera the gular region of the neck membrane and the postgenal area of the cranium (including the hypo stomal part of the subgenal sulcus) are elongate in comparison with the Psocoptera. These two developments cause elongation of the posterior and median ventral surface of the head, displacing the mouth anteriad. A second development, however, is the extension of the clypeus ventrally, displacing the labium onto the ventral surface of the head and the mouth posteriad (Symmons, 1952). In the Amblycera the labium, though ventral, is not far from the front of the head, and the sub-genal area is also close to the margin. The mouthparts are consequently well anterior of the mid-point of the head, and the insect might, in those teiros, be called prognathous. In Ischnocera the clypeolabral suture is expanded into the 1 pulvinus' (see below) and the mouthparts are displaced posteriad (Symmons, 1952). The degree of displacement depends on the length of the pulvinus; if short, the mouthparts remain anterior (or nearly so); if long, the mouthparts are displaced to about the midpoint of the head. In the latter case the mouthparts must be considered ventral and the head itself hypognathous, although the long axis is horizontal. The insect head is not obviously segmented, but is marked by a number of grooves. These grooves are variously termed sutures (e.g. Clay, 1951; Richards and Davies, 1977) or sulci (e.g. Symmons, 1952; Snodgrass, 1960). Snodgrass (I96O) makes the point that, historically, application of the term 1 suture' follows usage in the study in vertebrate skulls, where similar features represent lines of fusion between distinct centres of ossification. The insect head has no such centres of sclerotisation but is in essence a simple sclerotised box (Snodgrass, I96O). The grooves (and their associated internal carinae, if present), are thus not formed when plates meet, but are functional adaptations to resist or accommodate various strains on the head capsule. Snodgrass (1960) therefore considers the term 'suture' to be inappropriate, and recommends use of the term 'sulcus' (defined in the Oxford English Dictionary as a groove or furrow) for use in referring to most of the grooves of the Fig. 9. Structures of the trichodectid head (female). Terminology modified slightly from Symmons (1952). Mandible, maxilla, hypopharynx and labium omitted. osculum

Fig. 10. Regions of the trichodectid head (female). Terminology modified slightly from Symmons (1952)• Mandible maxilla, hypopharynx and labium omitted. 116

insect head. Snodgrass (1960) makes two exceptions to this, the postgenal suture (where the genae come together to eliminate the gular region, a development not present in Phthiraptera), and the 'Y'-shaped ecdysal cleavage line ('epicranial suture' auctt.) where the head capsule splits during ecdysis. The junction of the clypeus and the labrum should also be referred to as a suture. Chapman (1982) accepts the terminology of. Snodgrass (1960), though refers to the postoccipital sulcus as a suture. The terminology of Snodgrass (1960) is followed in this study. The relative positions of the sulci are fairly constant throughout the insects, and have been given names for descriptive purposes, as have the areas of the head capsule defined by the sulci. Symmons (1952) identifies the sulci and defined areas on the heads of Amblycera and Ischnocera. She notes that whilst some of the sulci are distinct and marked by clear internal carinae, others are united or difficult to discern, and the clypeo-frontal (epistomal) sulcus is not apparent. The morphology of the trichodectid head following Symmons (1952) is depicted in Pigs. 9 and 10* The internal carinae marking sulci are strongly-developed in many Trichodectidae, and the differences in degree of sclerotisation and of position on the head are useful taxonomically. The postoccipital sulcus is marked by a greatly developed internal carina ('occipital ring') in many lice, this development being particularly apparent in the Trichodectidae. Rhyncophthirina, many Ischnocera and some Anoplura possess two occipital apophyses projecting posteriorly from the occipital ring into the thorax (Mukerji and Sen-Sarma, 1955; Ferris, 1951; personal observation), but these are not apparent in the Trichodectidae, though in .this family the ring iself is developed posteriorly and may project into the thorax (personal observation). Symmons (1952) distinguishes three types of head in the Ischnocera, naming each after a family characterised by the type, but failing to indicate the full distribution of the types. These types, and their distribution through the ischnoceran families accepted in this study are; 1) the 'goniodid' (circumfasciate) head, with an unbroken anterior clypeal margin (Pig. 11a) (Goniodidae, Heptapsogasteridae, some Philopteridae and some Trichodectidae); 2) the 'trichodectid' a

KEY

clypeal marginal carina memcrane labrum 11clypeofronta l sulcus clypeus pulvinus Pig. 11. Ventral view of the three types of ischnoceran head (after Symmons, 1952). (a) the 'goniodid' head (b) the 'trichodectid1 head (c) the 'philopterid' head. 118

head, in which the clypeal margin is medially divided by the pulvinus (Fig.11b) (Trichodectidae and Trichophilopteridae); 3) the 1 philopterid* head, in which the median part of the clypeus is detached, forming an anterior median plate (Fig.11c) (Philopteridae). A detailed treatment of the variation of the head in bird-infesting Ischnocera is presented by Clay (1951). As noted above, a few Trichodectidae have heads of the circum- fasciate type, but most have 'trichodectid heads1. Symmons (1952) derives the trichodectid head directly from the circumfasciate type, which she believes to be plesiomorphic for the Ischnocera. It is notable, however, that the median antero-dorsal sclerotisation of the head found in many Trichodectidae is very similar in form to the philopterid anteclypeus. Although Symmons describes the trichodectid head as having the clypeal margin interrupted by the pulvinus, in many cases the pulvinus fails to reach the margin, which is instead interrupted by a small median dorsal or ventral sulcus ('dorsal and ventral preantennal sulci' , following Clay, 1951). The anterior margin of the head may be indented medially, the indentation generally being termed the osculum. The presence of the osculum is normally associated with the dorsal preantennal sulcus and either the ventral preantennal sulcus or, more usually, the anterior extension of the pulvinus (Fig. 9)* When the insect is at rest the pulvinus and osculum have been observed to be applied to the hair of the host (see discussion of mandibular morphology below). The width of the osculum, therefore, is probably influenced by the diameter of the hair in the region of the host's body inhabited by the louse. Between the antennae and the osculum the margin of the head may, in plan, be shallowly or broadly convex, sinuate or straight (Fig. 12); sclerotisation along the margin (forming the ' clypeal marginal carina' ) may be minimal or heavy, but is usually pronounced in the median dorsal region (Fig. 12 ). It is possible that much of the variation is influenced by the density and texture of the hair of the host. The anterior margin of the head in Trichodectidae is produced into a variably sclerotised corns anterior to the antennal socket (Fig. 9 ), but the trabecula (Clay, 1946) is absent. A sclerotised 119

V

13

Fig. 12. Variation in form of preantennal margin of head in Trichodectidae, Dorsal view of head of (a) Bovicola caprae (b) Damalinia lineata (c) Felicola subrostratus (d) Trichodectes zorillae $.

Fig. 13. Variation in tentorial form in the Fsocodea (after Symmons, 1952). (a) Psocoptera, (b) , (c) , , (d) , Gyropidae, (e) Trimenoponidae, Boopiidae, (f) Ischnocera, (g) Rhyncophthirina, (h) .Anoplura. 120

conical projection from the dorsal nodus of the clypeo-frontal sulcus may be present projecting parallel to the margin of the antennal socket. The latter projection frequently exhibits sexual dimorphism in its presence, connected with the sexual dimorphism of the antennae. The postero-lateral margins of the head are more or less rounded and may be produced at the postero-lateral angles in some Trichodectidae. The eyes may be positioned adjacent to the antennae or more posteriorly; in some cases (e.g. the trichodectid Cebidicola) the eyes occupy lateral projections. The cephalic chaetotaxy of the bird-infesting Ischnocera is described by Clay (1951). The five pairs of setae she lists as being "always present and approximately in the same position" can be identified readily in the Trichodectidae. Setae corresponding to the 1 post-temporal1 position of Clay (1951) are also present in male Trichodectidae, but a reduced number (two) are present in females; in some cases a line of setae is present posteriorly across the frons. The number and position of these setae is useful taxonomically, but lacks general systematic application.

2.2.2.3. The tentorium

The psocopteran tentorium is similar to the generalised insectan form, with well-developed anterior and posterior arms joined across the head by a stout bridge. Dorsal tentorial arms are present in Psocoptera, though reduced in size. The most complete phthirapteran tentorium, found in the Menoponidae, is similar to the psocopteran tentorium, though in Phthiraptera the dorsal arms are absent. It is a reasonable inference that the tentorium of modern Psocoptera is similar in form to that of the common ancestor of Psocoptera and Phthiraptera. The tentorium of Phthiraptera is discussed in detail by Symmons (1952) and the variation summarised here in Fig. 13 • The tentorial form is variable in the Amblycera, but comparatively uniform in the Ischnocera. Anterior arms are present and complete in Amblycera, though in Boopiidae they are desclerotised posterior to the site of attachment of the clypeal ligament. In Ischnocera 121

the anterior arms are reduced to short rods disconnected from the rest of the tentorium hut attached to the dorsum of the head by the clypeal ligament. The tentorial bridge is sclerotised in some Amblycera, but reduced to a fine ligament in others and in Ischnocera. The fibrous connective tissue surrounding the head ganglia in Psocoptera and Amblycera (a feature of most insects) is enlarged in the Ischnocera and Rhyncophthirina, possibly as a protective and supportive device in place of the reduced tentorium. In Anoplura this tissue is modified still further to form a structure known as the' obturaculum, which supports the muscles of the trophic stylets (Stojanovich, 1945; Symmons, 1952). The primitive insertion site of the antennal muscles on the dorsum of the head is replaced in most insects by insertion on the dorsal tentorial arms (Symmons, 1952; Matsuda, 1965)• In psocoptera the muscles insert on the dorsal and anterior arms, and in some cases have migrated entirely onto the anterior arms. In Amblycera and Ischnocera the dorsal arms are lacking and the antennal muscles insert on the anterior aims. This is the sole insertion site in the Amblycera and Trichodectidae, but in other Ischnocera the muscles have migrated in part to the plesiomorphic position, the dorsum of the head. In Rhyncophthirina and Anoplura both dorsal and anterior tentorial arms are absent and the anteimal muscles all insert on ' the dorsum of the head (Symmons, 1952).

2.2.2.4. The antennae

Two types of antennae - 1 segmented1 and ' annular1 - are found in the Arthropoda (Imms, 1939). The segmented type, found in the 'lower' (including Diplura and Collembola) comprises a variable number of segments, each of which (except the teiminal) contains intrinsic muscles (Imms, 1939; Matsuda, 1965). The annular type, found in Thysanura-, comprises a basal segment (scape) and an annular flagellum (Imms, 1939), the flagellum being formed by subdivision of a single segment (Matsuda, 1965). The only intrinsic muscles of the annular antennae are found in the scape, insert on the basal segment of the flagellum (the pedicel) 122

(Imms, 1939), and. move the flagellum as a unit; the scape itself is moved by muscles inserting on its basal margin and arising on the tentorium or the dorsum of the head (see discussion above). Primitively there are four muscles inserting at three points on the basal margin of the scape, but in Psocodea one of these muscles may be absent (Matsuda, 1965). The three insertions are clearly depicted for Stenopsocus stigmaticus (Psocoptera) by Badonnel (1934) but in published figures of the musculature of the antennae of Phthiraptera (Stflwe, 1943; Risler, 1951; Ferris, 1951; Mukerji and Sen-Sarma, 1955; Buckup, 1959; Haub, 1967, 1971) only two points of insertion are shown. Comparison of these figures of psocopteran and phthirapteran antennae reveals that the retractor muscle (muscle 51 of Matsuda, 1965) is absent in the Phthiraptera, the muscle inter- preted by StSwe (1943) and Buckup (1959) as the retractor muscle being one of the abductor muscles (muscle 54 of Matsuda, 1965). The muscles of the antero-dorsal basal margin of the scape function as abductors, those on the postero-ventral basal margin as adductors (Matsuda, 1965). The intrinsic muscles of the scape typically insert at two points on the base of the pedicel in insects (Matsuda, 1965) and such insertions are figured for Phthiraptera by Risler (1951), Buckup (1959) and Haub (1967). Badonnel (1934) and StOwe (1943) suggest a third muscle insertion, however, for Psocoptera and Phthiraptera respectively. The positions of muscle insertions on the scape and pedicel are such that the antenna can be moved through 360 degrees, but there is no muscular mechanism for flexing the flagellum.

The pedicel is considered by some authors to be a complete segment analogous to the scape (Matsuda, 1965), but the absence of intrinsic muscles indicates that this is not the case (Imms, 1939)• The pedicel is, however, characterised by the presence of a mechano- receptor ('Johnston1 s Organ') in many Thysanura-Pterygota (Mhtsuda, 1965). Ib. this study the pedicel is considered as distinct from the flagellum, and the term 'flagellum' is applied below only to the annulations (flagellomeres) distal to the pedicel. The Psocoptera have a scape, pedicel and from eight to fifty flagellomeres, whilst 123

the number of flagellomeres is reduced to three or less in the Phthiraptera (Richards and Davies, 1977)• The flagellomeres are here numbered from the base of the flagellum, so that the antenna comprises scape, pedicel, first, second and third flagellomeres. The scape and the head capsule are articulated at two points on the antero-ventral and postero-dorsal basal margin of the scape, whilst the scape and pedicel are articulated at two points at right angles to these. The joints between the flagellar segments are not articulated, but operate in the Psocodea by layers of hard exocuticle and soft endocuticle sliding over one another during movement (Seeger, 1975). The flagellar joints of Psocodea incorporate a collar-like fold identified by Seeger (1975) as an antennal-rupturing mechanism, purported to function in the Trogiomorpha and Troctomorpha (Psocoptera) as a protective device. The structure is much reduced from the psocopteran condition in the Phthiraptera (Seeger, 1975), and its function in the latter group is not clear. Whilst the prime function of the antenna is sensory, secondary modification in the male has taken place in some Ischnocera (including almost all Trichodectidae) and some Anoplura, the antennae being used to clasp the female round the abdomen during copulation (Keler, 1938; Sikora and Eichler, 1941) (Pig. 23 ). This development has led to an increase in length and degree of sclerotisation .of the pedicel and flagellar segments and an increase in strength of the antennal muscles. The intrinsic muscles of the scape are enlarged and the head concomitantly broader in the male than in the female. In most if not all cases the male antennae retain a sensory function and are found to bear the same sensilla as are present in the female. There must, therefore, be free movement of the flagellum and to facilitate this the joint between the scape and pedicel is broad and membranous (other than at the two articulatory points). The movement of the pedicel and flagellum is thus as free as in other insects, and controlled by extrinsic and intrinsic scape musculature. The flagellum is modified to clasp the female, but is not under direct muscular control. The degree of free movement of the flagellum relative to the pedicel must therefore be limited, so that control may be effected by the intrinsic muscles of the scape. 124

Observations made in this study reveal that the apex of the pedicel is angled relative to the long axis of the segment, the longest margin of the pedicel being the antero-ventral (Fig.14a). There is very little membrane between the pedicel and flagellum on the antero-ventral margin, but more on the postero-dorsal, so some flexibility between the pedicel and flagellum is possible, though limited in the anterior direction by the prolongation of the pedicel. During copulation the male clasps the female around the top of the abdomen from underneath (Werneck, 1936; Sikora and Eichler, 1941), the pedicel and flagellum being curved to match the curvature of the abdomen. The antennae are raised above the head of the male and the intrinsic adductor muscles contracted. The pedicel and the flagellum are thus brought down over the abdomen of the female, the flagellum being constrained by the form of its junction with the pedicel. The form of the pedieel-flagellum joint is such that, should the haemolymph pressure be reduced, the joint membrane would collapse and further contract the flagellum against the female. It is not known, however, that lice have any control over haemolymph pressure in their antennae, though Lepidoptera larvae are known to control their antennal movement partially by this means (Matsuda, 1965). The mechanical strength and degree of possible control of a system involving joints as described above is likely to be inversely proportional to the number of joints in succession, with a single joint being most efficient. The degree of curvature attainable using three segments is sufficient to grasp the female. For these reasons only the first flagellomere is required to take on a clasping function; the apical two flagellomeres may be retained in a sensory capacity or lost. In Harrisoniella spp. the scape, pedicel and first flagellomeres are similar in size to those of the female. In Goniodes pavonis the apical flagellomeres are also very similar in both sexes, but in the male arise from the ant ero - dorsal margin of the first flagellomere (Fig. 14b); similar migration of the apical flagellomeres is present in Polyplax (Anoplura). In Trichodectidae the last two flagellomeres have contracted and fused to the first flagellomere in all males (though a small semicircular 125

' r7>> •// //

Pig. 14* Antennae of male Ischnocera. (a) Damalinia indica, right antenna, dorsal view; arrow points to the prolongation of the antero— ventral margin of the pedicel; (b) Goniodes pavonis, right antenna, dorsal view. 126

sclerotisation probably representing the apical flagellomere is present on the male of Eurytrichodectes paradoxus) and in females of Neotrichodectes, Geomydoecus, Trichodectinae, Bovicolinae and most Eutrichophilinae. The sensilla of the two apical flagello- meres are retained on the remaining flagellomere. The firmness with which the female is held may be increased by projections in the form of spikes or denticles on the antennae (Fig.14a ,14"b ), particularly apparent on the .flagellum. Trichodectidae are characterised by the possession of two modified setae apically on the male flagellum, which take the form of sharp, stout teeth (Fig. 14a). Similar modifications to setae are present in some Anoplura. The amblyceran antenna differs from that of other Phthirapt'era in that the first flagellomere is pedunculate and the second and third closely-appressed or fused to it, producing a globular flagellum (Clay, 1970). The significance of this morphological feature is not known. The sensilla of the antennae have been studied by Clay (1970). In addition to the undifferentiated setae scattered generally over the surface of the antennae there is a patch of stout unpointed setae apically on the terminal flagellomere, particularly apparent in the Ischnocera. The last two flagellomeres also bear other sensilla. Menoponidae, Boopiidae and Ricinidae have a sensillum coeloconicum 'on each of the last two flagellomeres, or both on the last when these two flagellomeres fuse. Laemobothriidae have three sensilla coeloconica on the terminal (fused third and fourth) flagellomere. Trimenoponidae and Gyropidae have two to four modified sensilla coeloconica on the terminal (fused third and fourth) flagellomere. Ischnocera have one sensillum coeloconicum and two saucer-shaped pore organs (sensilla placodea of Kim and Ludwig, 1978b) on the terminal flagellomere, and one sensillum coeloconicum and one sensillum placodeum on the second flagellomere; Trichodectidae with all three flagellomeres fused have two sensilla coeloconica and three sensilla placodea on the terminal flagellomere, sometimes closely associated (Clay, 1970; Kim and Ludwig, 1978b) 127

In Lorisicola malaysianus and mjobergi (Trichodectidae) the sensilla are in pits with tongue-like projections around them. (Rhyncophthirina) has a sensillum basiconicum and two sensilla placodea on the terminal flagellomere and a sensillum basiconicum and a sensillum placodeum on the second flagellomere (Clay, 1970). Many Anoplura have a similar arrangement to the Ischnocera, though others have two sensilla basiconica and two sensilla placodea on the terminal flagellomere (Miller, 1970a, b, 1971a, b, c).

2.2.2.5. The ocelli and compound eyes

Psocoptera possess the plesiomorphic condition of the Para- neoptera in having a pair of compound eyes and three ocelli in winged adults, whilst nymphs and apterous adults have a pair of compound eyes only. Phthiraptera have a single pair of eyes laterally on the head, each eye having one (Ischnocera, Rhyncoph- thirina and Anoplura) or two (Amblycera) lenses; some Anoplura completely lack eyes. Wigglesworth (1941) identifies the eye of Pedi cuius hum anus (Anoplura) as an ocellus, and certainly the structure of the eye in this species is very similar to the structure of the psocodean ocellus ( e.g. Mesopsocus immunis as figured by Jentsch, 1940). Y/undrig (1936) notes a similar correspondence between the structure of the eye in Ischnocera and Anoplura and the structure of the ocellus, leading her to teim the feature a 'pseudoocellus1 but she suggests that the eye in Amblycera (and hence, by implication, other Phthiraptera), is a reduction of the compound eye of the Psocoptera. Wundrig (1936) does not depict a crystalline cone (a feature found in the oramatidia of compound eyes and not characteristic of ocelli) in any of her figures: Webb (1948), however, figuring the eye of suis (Anoplura), also depicted by Wundrig, finds a large crystalline cone present, though he still terms the eye an ocellus. The ocellar nerves of Psocoptera enter the protocerebrum medially and dorsally, whilst the optic nerve enters posteriorly and laterally (3adonnel, 1934). In Phthiraptera the nerve from 128

the eyes is the most posterior of the nerve entries to the proto- cerebrum and is lateral, not dorsal (Stojanovich, 1945; Risler, 1951; Ramcke, 1964). The entry of the nerves thus seems to support the contention that the eyes are reduced compound eyes, not ocelli. In the apterous Liposcelinae (Psocoptera) ocelli are absent and the compound eyes are reduced in size to 6 - 8 ommatidia; in Embiopsocinae (Liposcelidae) the ommatidia number only two per eye in apterous forms. The eye of Haematopinus suis as figured by Webb (1948) is very similar to the ommatidium of Liposcelis * divinatorlus* (Psocoptera) as figured by Jentsch (1940), though having a larger number of sensory cells. There seems no reason to suppose that the eyes of Phthiraptera are ocelli and not reduced compound eyes. Compound eyes are primarily concerned with detecting movement, whilst ocelli are more likely to perceive only light intensity (Chapman, 1982). The adoption of a nidicolous habit by the ancestor of the Phthiraptera encouraged the loss of wings, probably by heterochronous development, and the concomitant loss of ocelli. Detection of movement by a nest-dwelling insect is of limited importance, however, and this importance is further reduced . in the adoption of the ectoparasitic habit. At the same time, it might be assumed that detection of changes in light intensity might be of more significance to the insect, either to keep it within the confines of the nest or to prevent it moving too far from the skin of the host into the less hospitable outer regions of the dermecos. It is known that some lice are negatively phototactic (Wigglesworth, 1941) or will orient themselves using light if no temperature gradient is present (Murray, 1957b). The change in 'use1 of the eye would apply selective pressure towards reduction in the number of ommatidia and convergence tov/ards an 0cellar structure, leading to the condition found in many modern lice. 129

2.2.2.6. The mouthparts

The lab rum and the pulviims As noted above, the ventrad extension of the clypeus in Amblycera and Ischnocera has displaced the labrum onto the ventral surface of the head. This movement has been accompanied by a similar displacement of the internal structures, such as the hypopharynx and associated muscles (Matsuda, 1965). In Amblycera the clypeo-labral suture is unmodified and the labrum may be very close to the front of the head, but in Ischnocera this suture is expanded to form a thick bilobed pad of more or less unsclerotised tissue (the pulvinus), and the labrum is more posterior in position (Symmons, 1952). Keler (1938) suggests that the pulvinus may be a clasping structure, acting on a hair or feather during feeding. Should this be the case, it is likely that operation is by variation of haemolymph pressure, since Symmons (1952) shows that the pulvinus has no associated muscles. The pulvinus is long and parallel-sided in most Trichodectidae, but in some it is short and semi-circular; it is possible that the function of this organ is not uniform throughout the family or the suborder. An organ perhaps analogous in function is the "pallette", membranous extrudable outgrowths of the labrum found in the Ricinidae (Amblycera). This organ is discussed in detail by Nelson (1972). "In Philopterus and possibly all other Ischnocera the labrum may be expanded hydrostatically to a position where it covers the pulvinus; retraction is achieved by the labral muscles (Symmons, 1952).

The mandibles. Mandibles are always present in Amblycera and Ischnocera, though highly modified in the amblyceran Trochilocetes (Clay, 1949)- In Amblycera, as in Psocoptera, the mandibles are horizontal in articulation, whilst in Ischnocera the articulation is vertical, though intermediate conditions do exist. Matsuda (1965) suggests that the change in plane of articulation is related to the dorso-ventral compression of the head. The molar region of the mandible, present in Psocoptera, is absent in Phthiraptera, a development associated by Matsuda (1965) with carnivory. The 130

mandibles are variously toothed apically, and may have a blunt projection on the inner basal margin (absent in Trichodectidae). Trichodectidae have three apical teeth on the right mandible and two on the left; on the right the centre tooth is generally longest, whilst on the left the posterior tooth is normally the better developed (Pigs. 15a, 15b, 15c , 15<3). Mandibular asymmetry is general throughout the Amblycera and Ischnocera, and is considered by Snodgrass (1935) to be a feature of insects that masticate their food. The lack of asymmetry in the mandibles of some species of Ricinus (Amblycera) is associated by Nelson (1972) with b'lood-feeding in those species. Trichodectes canis, the only trichodectid known to take blood meals (Bouvier, 1945), has dimorphic mandibles, as do all other Trichodectidae. In addition to collection and preparation of food the mandibles of Trichodectidae are employed in anchoring the insect. When at rest, Trichodectes canis and T^. melis have been observed to -enclose a hair in the pulvinus with the mandibles and whilst in this position may completely release the grip of the tarsal claws and straighten the legs laterally from the body. The insect is then held on only by the mandibles. In addition to the species above, specimens of Pelicola sp. collected from dried museum skins have been found in this position, and specimens of most genera have been found preserved in alcohol clinging to single hairs by their mandibles alone. The selective pressures on the mandibles of Trichodectidae are different, therefore, from those operating on lice that do not have a mandibular anchoring mechanism, and possibly preclude the assumption of monomorphy in blood-feeding species. The 'interior1 face of the right or of both mandibles may be ridged, so that when the mandibles are folded closed, the ridges on one mandible are not covered by the other, and all or most of the exposed mandibular area is ridged. The ridges are therefore not positioned in such a way that they can act against one another, and it is suggested that they are developed to prevent the mandibles slipping on a hair v/hen clasping it (Pigs. 15b ,15c , 15d). When the mandibles are closed, the right is always interior to the left, which may have a concavity to receive it (Pigs.15a, 15c ,15d); the left mandible is not then kept by the right from contact with the Fig. 15. Mandibles of Trichodectidae, dorsal view, (a) Damalinia conectens; (b) Damalinia indica; (c) Eurytriohodectes

paradoxust showing mandibles interlocking; (d) Trichodectes canis, with detail of postero-dorsal margin of left mandible. 132

hair, and the greatest possible mandibular area is utilised. In some species the right mandible has a basal notch on the anterior margin, which receives the tip of the left mandible (Figs. 15a, 15&); it is possible that this feature acts in concert with the restraint of the right mandible by the anterior margin of the hollow in the left to 'lock* the mandibles closed about a hair. Although, as said above, none of the Trichodectidae have monomorphic mandibles, the mandibles of Damalinia (T,.) conectens (Fig. 15a) are very slender and pointed, and the centre tooth of the right mandible of Dasyonyx spp. and Burytrichodectes spp. (Fig. 15c) is long and pointed, suggesting in each case a piercing function and thus possible haemophagy. In contrast, the mandibles of Damalinia (T^.) indicus (Fig.15"b) are blunt and broad, suggesting an adaptation to grinding and chewing.

The maxillae. The maxillae are present but greatly reduced in Amblycera, Ischnocera and Rhyncophthirina, the reduction being most pronounced in the latter group. In the Anoplura the maxillae are highly modified into stylet guides (Young, 1953). In Amblycera and Ischnocera, as in all other Acercaria, the lacinia is detached from the rest of the maxilla and slender in form. It resembles the lacinia of the Psocoptera in being long and stylet-like and lies in a deep sac in the lobate galea (itself set in a mandibular cavity). (Matsuda, 1965)* -The lacinia is absent in Rhyncophthirina and unrecognisable (presumably absent) in Anoplura; it exists in its least reduced form in the Amblycera, though it may be absent even in this group, as Stdwe (1943) did not locate it in Trimenopon, but erroneously identified the galea as the lacinia. In Amblycera and Ischnocera a lacinial gland is attached to the anterior end of the lacinia after it has entered the head. This is not known from other Psocodea, though a precursor may be present in the Psocoptera (Symmons, 1952). No secretions have been found within the lacinial gland and its function is not known. There is no duct or outlet from the gland through the lacinia in the Ischnocera, though in the Amblycera a minute opening in the tip of the lacinia may provide a route to the exterior (Symmons, 1952). Matsuda (1965) suggests that the lacinia itself functions as a valve or pump for the 133

lacinial gland, though he does not give further details. Snodgrass (1905) notes that the maxilla is generally fused to the labium in Amblycera and Ischnocera, a condition that has led to some confusion in identification of the constituent features. The cardo is lost or fused to the stipes in Psocodea (Matsuda, 1965; Kim and Ludwig, 1978b), and both may be completely lost, as in Bovicola (Risler, 1951; Matsuda, 1965); the galea is always present (except in Rhyncophthirina and Anoplura) (Matsuda, 1965). Maxillary palps are present and four-segmented in Psocoptera and Amblycera but are absent in Ischnocera.

The labium. The labium is relatively generalised in Amblycera and Ischnocera, consisting of a prementum and a postmentum, the former bearing the labial palps, paired paraglossae and fused median glossae (Matsuda, 1965). The highly modified Rhyncophthirina and Anoplura will not be discussed. The labial palps, reduced in all Acercaria, comprise a single segment in most Amblycera and Ischnocera, and are absent in some Amblycera (e.g. Ricinidae). As in Psocoptera the prementum forms the ventral wall of the salivary canal, and is hollowed to accommodate the ventral faces of the lingual sclerites. Two pairs of labial glands (dorsal or acidophil and ventral or basophil) are present in the Psocoptera (Weber, 1938; Finlayson, 1949), the Ischnocera (Risler, 1951; Haug, 1952), the Amblycera (Haug, 1952; Buckup, 1959), the Rhyncophthirina (Mukerji and Sen-Sarma, 1955) and the Anoplura (Snodgrass, 1944; Ferris, 1951; Haug, 1952). Snodgrass (1899) locates and describes the labial glands of Docophoroides brevis (Ischnocera) (as Burymetopus taurus), but erroneously figures the efferent salivary duct as opening into the alimentary canal, not the salivarium. The lingual sclerites have been believed to be glands (Snodgrass, 1896, 1899) and sometimes identified as the labial glands (Cope, 1940a, 1940b, 1941); this misconception is probably the source of Richards and Davies1 (1957, 1977) reference to "a pair of labial salivary glands, lying beneath the suboesophageal ganglion" in the Ischnocera and Amblycera, though they also note that the lingual sclerites are not glands. Kim and Ludwig (1978b) maintain that the Amblycera, Ischnocera and Rhyncophthirina have only one pair of 134

labial glands, attributing this information to KSnigsmann (1960). Kim and Ludwig (1982) note correctly that two pairs of labial glands are found in the thorax of Anoplura and Psocoptera (though they cite Cope, 1940a as an authority for the position of the glands in Psocoptera, whilst this author in fact considered the lingual sclerites to be the labial glands and failed to find any glands in the thorax). Kim and Ludwig (1982) go on to assert that a single pair of labial glands is found beneath the suboesophageal ganglion in 'Mallophaga* , and a single pair of salivary glands in the thorax of Ischnocera and Rhyncophthirina. As authorities for the first statement they cite Cope, 1940b (see above), 'Weber, 1938 and Risler, 1951, though the latter two authors make no such statement. The presence of a single pair of thoracic salivary glands in Ischnocera is recorded, according to Kim and Ludwig (1982), by Cope (1940b) and Risler (1951), though Cope (1940b) makes no mention of thoracic glands and Risler (1951) demonstrates two pairs of thoracic labial glands. Richards and Davies (1977) is cited as a general supporting reference, and the phrase "beneath the subo esophageal ganglion" employed by Kim and Ludwig (1982) suggests that the information came from the latter work. The positioning of the salivary glands in the Rhyncophthirina is noted by Kim and Ludwig (1982) as demonstrated by Mukerji and Sen-Sarma (1955). In this group one of the glands (probably the dorsal gland, from the sections illustrated in Mukerji and Sen-Sarma, 1955) has migrated from the thorax to the head. Kim and Ludwig (1978b) and Kim and Ludwig (1982) are both incorrect in their description of the labial glands in the Phthiraptera, and their ' use of these structures in phylogenetic reconstruction unjustified. The Psocoptera, Amblycera, Ischnocera and Anoplura have two pairs of labial salivary glands in the thorax and none in the head, whilst the Rhyncophthirina have one pair in the thorax and one in the head. Weber (1938) shows that in silk-spinning Psocoptera the ventral glands produce saliva whilst the dorsal glands are responsible for the production of silk. The dorsal glands, however, are always present whether or not silk is produced. A further possible function of the dorsal glands is discussed below. 135

The hypo pharynx. Immediately anterior and dorsal to the labium lies the hypopharynx. This is uniquely modified in the Psocodea (additional major modifications in the Rhyncophthirina and Anoplura will not be discussed here), in that the loral aim is absent, the lingual sclerites are swollen and connected by a filamentous structure to a cup-shaped sclerite on the ventral surface of the sitophore, and a production of the dorsal face of the sitophore (the 'epipharyngeal crest') is developed, fitting the depression in the middle of the sitophore sclerite (KBnigsmann, 1960; Matsuda, 1965; Kristensen, 1975) • The anterior margin of the hypopharynx may be smooth or covered in hairy projections (Cummings, 1913). In Gliricola oorcelli (Amblycera) the anterior is produced into two divergent serrate projections (Ewing, 1924; misdetermined as lingual sclerites by Risler and Geisinger, 1965). Mjdberg (1910) and Ewing (1924) identify these serrations as cutting devices which the insect uses to wound the host around the hair follicles to obtain sebaceous exudations as food. Bouvier (1945) fails to detect any evidence of blood meals having been taken by this species. The sitophore sclerite is very variable in form in the Psocodea, particularly in the Amblycera and the Ischnocera (Cummings, 1913; Haub, 1967, 1972, 1973, 1977). The posterior border is heavily sclerotised in the Amblycera but not the Ischnocera; uniquely, the posterior border is not closed in the Trichodectidae (Haub, 1973)• There is some difficulty in seeing all the variation, as the orientation of the sclerite is such that the posterior border is obscured in slide- mounted specimens, and differential inclination of the specimens may change the apparent dimensions (Haub, 1977). The functional significance of the variation is unclear, but Haub (1973) attaches phylogenetic value to it, and it may be utilised in taxonomic descriptive studies at the species and genus levels. The lingual sclerites lie posterior to the sitophore in Psocoptera but anterior in the Phthiraptera, the apodemes of the lingual sclerites being elongated in the latter group; this change in relative positions is correlated with the antero-ventral rotation of the clypeolabral area in the Phthiraptera (Matsuda, 1965). The lingual sclerites 136

are interpreted by Cope (1940a, b, 1941) as salivary reservoirs, but the absence of a lumen (Weber, 1938) precludes this possibility. Rudolph (1982b) shows that in the Psocoptera the hypopharynx is intimately concerned with the uptake of water vapour from the atmosphere. During this process, the lingual sclerites are rotated through 90° to form a flat plane between the labrum and labium, probably by contraction of the muscles running from the apodemes of the lingual sclerites to the distal margin of the prementum ('dorsal dilator of salivarium' in Phthiraptera). Contraction to rest is by hypopharyngeal retractor muscles inserted onto the apodemes of the lingual sclerites and originating on the tentorial bridge. Stretched between the lingual sclerites is the membranous portion of the hypopharynx. During uptake of water the surface of -the membrane and sclerites is covered with a thin irridescent film of presumably hygroscopic liquid, which Rudolph (1982b) suggests is secreted in the dorsal labial glands. Also during water uptake, the clypeo-cibarial muscle regularly and rapidly contracts, causing an in and out movement of the epipharyngeal crest against the sitophore sclerite. Weber (1936) suggests that the epipharyngeal crest and sitophore sclerite function as a 'mortar and pestle', breaking down food particles. Risler (1951) maintains that the mechanism is not a grinding one in Ischnocera, and Keler (1966) points out that no food particles have ever been found in the pit of the sitophore sclerite, though the cibarium is often full up to the sclerite. Buckup (1959) and Keler (1966) suggest that the crest and sclerite act together as a pump, the up-stroke of the crest producing a negative pressure and sucking saliva into the cibarium. Risler (1951) and Keler (1966) suggest that saliva is sucked through the filaments running from the lingual sclerites, though Risler (1951) invokes capillary action to provide the pressure, whilst Buckup (1959) maintains that the 'saliva passes over the anterior part of the hypopharynx. Weber (1938) contends that the filaments are without a lumen, and therefore cannot be ducts, but Rudolph (1982b) demonstrates the hollow nature of the filaments by means of serial sections. Rudolph (1982b) goes on to suggest that in Psocoptera water absorbed on the lingual sclerites is sucked through 137

the filaments by the cibarial pump, and from the sitophore sclerite passes into the alimentary canal. Because of the great similarity of the hypopharyngeal structures in Psocoptera and Phthiraptera, and since water-vapour absorbtion of similar efficiency and rate is known to occur in both groups (Williams, 1971; Rudolph, 1982a), it is likely that the function of the hypopharyngeal structures is the same in Psocoptera and Phthiraptera.

2.2.3. The Thorax

2.2.3.1. Introduction The thoracic structure of the Phthiraptera has been discussed by Ferris (1951), Mayer (1954) and Matsuda (1970), the latter author attempting to summarise all previous work. Mayer (1954) describes a few species of Amblycera and Ischnocera in detail, but does not provide general conclusions on the basis of her observations, though . Ferris (1951) discusses the anopluran thorax in general terms. Matsuda (1970) treats the 'Mallophaga1 and Anoplura much as do Mayer (1954) and Ferris (1951) respectively.

2.2.3.2. The neck region The lateral cervical sclerite is absent in Rhyncophthirina and Anoplura, but present in all other Psocodea (Boudreaux, 1979). Mayer (1954) fails to find this sclerite in Bovicola caprae (Tricho- dectidae), but examination has confirmed its presence in this and all other species of the family. In Harrisoniella (Ischnocera) a dorso-lateral cervical sclerite is also present (Cope, 1940b; Matsuda, 1970), the homologue of which may, in Tetrophthalmus (Amblycera), be fused to the lateral cervical sclerite (Matsuda, 1970). The origin and distribution of the dorso-lateral cervical sclerite are not known.

2.2.3.3. Se^mients of the thorax

The prothorax is reduced in winged Psocodea, but not in apterous forms (Smithers, 1972). The dorsal divisions into prescutal, scutal, scutellar and postnotal areas are largely lost 138

in the Phthiraptera, though a small postnotum, lying posterior to the main pronotal sclerite, is present in many Amblycera of the families Menoponidae, Boopiidae, Gyropidae and Trimenoponidae (Mayer, 1954). A small rhombic sclerite, not homologised with any of the psocodean prothoracic sclerites, is developed from the anterior margin of the pronotal sclerite in most Ischnocera (Fig.. 16a), the two sclerites being fused in some Philopteridae (e.g. Anatoecus). Keler (1938) named this plate the 'mucro', a term which is not adopted here due to its established usage as the terminal claw-like structure of the collembolan furcula. • The pronoturn is free in Ischnocera, Rhyncophthirina and most Amblycera, but fused to the mesonotum in some Amblycera of the families Trimeno- ponidae and Gyropidae. In all but one genus of Anoplura the prothorax and mesothorax are completely fused, intersegmental boundaries being lost (Matsuda, 1970). The exception is Mirophthirus. in which the pronotum and "mesonotum are incompletely fused, a feature which leads Chin (1980) to place the genus,in a new suborder, the Protanoplura. In other characters Mirophthirus appears to-be closely allied to the anopluran family Polyplacidae (Kim and Ludwig, 1982), indicating that the thoracic subdivision is secondary, and may not even correspond to the true (plesiomorphic) pronotal-mesonotal division. Keler (1938) considers that in Trichodectidae the prothorax and mesothorax are united to form a * skelothorax1 but, though the prosternum and mesosternum may be fused, the mesonotum is fused not to the pronotum but to the metanotum (Mayer, 1954; Matsuda, 1970). The pronotum and propleura are fused in some Phthiraptera, including Trichodectidae and at least some Amblycera. This fusion, together with the presence of a longitudinal and a transverse notal ridge in the Amblycera, leads Cope (1941) to propose that in the amblyceran genus Tetrophthalmus the pronotal area is limited to the anterior of the transverse ridge, the area posterior to this being pleural in origin. The- postnotum is interpreted as being the mesonotum and the mesonotum as dorsal extensions of the mesopleura. A similar hypothesis, if applied to the Ischnocera, would suggest that the rhombic sclerite is the true pronotum whilst the putative pronotum, which frequently is sclerotised only laterally (Fig.. 16a), is pleural in origin. The meso- and meta-notum 139

may be absent, or fused to the pleural elements of those segments. Ferris (1951) proposes similar developments for the thorax of Anoplura; in this group the mesothoracic spiracle has been displaced onto the dorsal surface of the thorax, presumably being carried by the invasive pleurum. Matsuda (1970) rejects Ferris' (1951) proposal, suggesting that the structures identified as pleural ridges by Ferris are homologous with the notal ridges of Amblycera, indicating the hotal rather than pleural origins of the supporting plates. Matsuda (1970) does not discuss the origins of the notal ridges, nor does he comment directly on the proposal of Cope (1941) that they mark the pleural boundaries, though he states that they cannot be used as landmarks in determining the division of a segment. For the purposes of this study the traditional concept espoused by Matsuda (1970) will be accepted, though further study of a wide range of Phthiraptera is necessary before a confident statement as to the true composition of the phthirapteran thorax can be made. The mesothorax lacks any dorsal division into prescutal, scutal, scutellar and postnotal areas in the Phthiraptera. The mesonotum and metanotum are fused in Anoplura, Rhyncophthirina, Ischnocera, Laemobothriidae and Ricinidae (Amblycera), some Trimenoponidae and Gyropidae (Amblycera) and some Liposcelidae (Psocoptera) (Clay, 1970; Smithers, 1972)• In Anoplura the boundary between mesothorax and metathorax is absent, the segments being completely fused. The mesothorax and metathorax are sometimes referred to as the ' pterothorax' in Phthiraptera, though wings are not present in the Order. The mesonotum and mesopleura are fused in some Phthiraptera. Only one pair of thoracic spiracles is present in the Phthiraptera, this lying on the proepimeron (Cope, 1941; Matsuda, 1970). The metathorax lacks any dorsal division into prescutal, scutal, scutellar and postnotal areas in the Phthiraptera. The fusion of the mesothorax and metathorax has been discussed above. The reduction or suppression of the anterior and mesal phragmata in Psocoptera has been proposed as an autapomorphy for that Order, with the implication that phragmata are present in the Phthiraptera (Boudreaux, 1979). True phragmata are transverse infoldings of the 140

intersegmental sclerites (postnota), and lie between the pro-, meso- and metathorax and the first abdominal segment; no phragma is ever borne by the prothorax (Snodgrass, 1935; Richards and Davies, 1977). The main function of the phragmata is to provide increased areas of attachment for the dorsal longitudinal muscles, and thus phragmata are best developed in winged insect s, especially those that fly actively (Richards and Davies, 1977). In winged Psocoptera the dorsal longitudinal muscles of the mesothorax attach anteriorly onto the mesoscutum, but posteriorly onto the mesophragma; the longitudinal dorsal muscles of the metathorax extend between the mesophragma and the postphragma (Badormel, 1934). The prophragma is reduced, but not the mesophragma. In Phthiraptera the dorsal longitudinal muscles are relatively small and the phragmata reduced or entirely suppressed (Mayer, 1954). Ferris (1951) refers to the thoracic pleural ridges as phragmata, and in this sense Matsuda (1970) homo- logises them with the notal ridges in Amblycera. The 1 phragmata* referred to in some discussions of the phthirapteran thorax are not homologous with the phragmata of other insects, but are notal ridges, the homologies of which are uncertain (see above). True phragmata are reduced or absent in the Phthiraptera, but are present in the Pscoptera. The sternal modifications of the Phthiraptera are diverse and poorly studied. To avoid confusion, only the sterna of Trichodectidae are discussed in this study.

2.2.3.4. Description of the trichodectid thorax Figures 16a and 16b depict a generalised trichodectid thorax, based upon Damalinia (Tricholipeurus) indie a The rhombic sclerite anterior to the-pronotum is always present; the pronotum itself generally bears two lateral sclerites, sometimes narrowly joined medially. The mesonotum and metanotum are fused, with prominent pleural ridges extending onto them; there may, as with the pronotum, be a medial zone of desclerotisation. The lateral cervical sclerite is always present, bearing two anterior setae. The pronotum is fused to the propleuron, which in turn is fused to the pro sternum. 141

a

lateral

b

Fig. 16. Diagrammatic representation of generalised trichodectid thorax, (a) dorsal; (b) ventral. 142

The presternum may extend, anteriorly between the fore-coxae and. be unsclerotised medially. The mesosternite, fused to the prosternite, may also be medially divided. The mesepisternum is difficult to delimit, but rarely extends unbroken between the metastemite and the metanotum. The metasternite, if present, is only rarely fused to the mesosternite, and is never sclerotised medially. Posterior to the metacoxa may be a semicircular sclerite ('postcoxale' of Matsuda, 1970). In Procaviphilus (Meganarionoides) this is very heavily sclerotised and fused to abdominal pleurum II; the two postcoxales may also fuse medially. One, two or three short setae are always present on the anterior margin of the thorax posterior to the temple margin of the head. The lateral margins of the thorax also bear setae, which may be more or less abundant. A row of setae is generally present across the posterior dorsal margin of the prothorax and pterothorax, sometimes interrupted medially ('median gap'). Setae are sometimes present on the dorsal disc of the prothorax and pterothorax ('anterior setae'), and may be present medially on the mesosteraum.

2.2.3.5. Coxal articulations The articulations of the legs with the corresponding thoracic segments in the Phthiraptera may be pleurocoxal, sternocoxal or, more generally, both (Matsuda, 1970). Cope (1940b) suggests that the Psocodea are unique in the "retention of the ventral coxal articulation on the prothorax"; this articulation may be ventro- lateral or even lost, however, (Richards and Davies, 1977) and does not constitute an autapomorphy for the superorder. The trochantin is not discernable as a separate sclerite, and may be either fused to the pleurum (Matsuda, 1970) or lost (Boudreaux, 1979).

2.2.3.6. The Legs

Introduction. The legs of Phthiraptera are variously modified for locomotion amongst the hair or feathers of the host. In mammal-infesting species, these modifications are for grasping the hair of the host, though Haematomyzus (Rhyncophthirina), parasitic 143

on the relatively hairless elephant and the wart-hog, has long slender legs apparently lacking any modifications of this type (Ferris, 1931). The setal patterns of all of the leg segments may be of taxonomic value (Keler, 1938; Clay, 1969, 1970), but are difficult to evaluate for cladistic purposes.

The coxa. The coxae are attached ventrally or ventrolaterally to the thorax (Fig. 16b), Clay (1970) suggests that the form of the first coxa and its articulation with the propisternum are apomorphic for the Boopiidae and the bird-infesting Amblycera, setting these apart from the mammal-infesting Amblycera of the New World.

The trochanter. Ventrally on each trochanter there are one or two setae and several well-marked campaniform sensilla, the number of

the latter varying from three to five in the Amblycera (Clay, 1970) # In Trichodectidae the first trochanter generally bears two setae and two campaniform sensilla, arranged linearly with the setae at either end. The second and third trochanters in this family generally bear a set'a and three campaniform sensilla arranged in a rough square ventrally, with a second seta positioned posteriorly.

The femur. Apart from the modifications in the Gyropidae (Amblycera), which are discussed separately below, the femur is not greatly modified in the Phthiraptera. In many Menoponidae (Amblycera), the third femora are equipped with setal brushes or combs similar to those found on the abdominal sterna of the same species (Clay, 1969); the function of these structures is unknown. The femora of Trichodectidae are short and broad, especially the first pair, and are not greatly modified.

The tibia. In Ischnocera and Anoplura the. tibia has a projection on the ventro-apical angle, opposite the tarsal insertion (Fig. 17 )• Rhyncophthirina and Amblycera lack such a projection, though Latumcephalum (Amblycera) approaches the condition found in the Trichodectidae (Keler, 1971). This projection (the "thumb" of Keler, 1938 and Ferris, 1951), opposes the tarsal claw and, in 144

18

Fig. 17- Apex of tibia and tarsus of two Trichodectidae. (a) Bovicola longicornis; (b) Trichodectes canis. Fig. 18. Lateral view of metatarsal claws of three Trichodectidae. (a) Dasyonyx (D.) smallwoodae; (b) Dasyonyx (IT.) diacanthus; (c) Eurytrichodectes paradoxus. 145

Trichodectidae and Anoplura, permits the louse to grasp a hair of the host. In the Anoplura particularly, the thumb may be very large and the tarsus and claw massively developed, especially on the third pair of legs. In bird-infesting Ischnocera the thumb frequently bears three stout, blunt, somewhat hyaline setae. These may be greatly modified, as in Struthiolioeurus, or increased in number, lor example to six in Austrogoniodes and over twenty in Docoohoroides (Ledger, 1980). In Trichodectidae the thumb is tipped with a single large seta ("thumb-spur" of Keler, 1938) always apically rounded and frequently partially hyaline (Fig. 17"b) • On the second and third pair of legs there may be a secondary subapical seta of a similar type. Keler (1938) suggests that the claw can "close like the blade of a pocket-knife" between these setae, though, were the claw to be closing on a hair of the host, as presumably would normally be the case, it would be unlikely to reach the setae. The function of the setae in Trichodectidae is probably to detect the presence and position of the hair during locomotion of the louse. The form and number of these setae are employed by Keler (1938) as characters of taxonomic value.

The tarsus. The insect tarsus corresponds to the penultimate segment of a generalised limb, and in some hexapods (e.g. Protura, Diplura, most holometabolous larvae) it retains the plesio- morphic form of a single segment (Snodgrass, 1935; Richards and Davies, 1977; Chapman, 1982). In most adult insects, however, the tarsus is subdivided into two to five 1tarsomeres' (Snodgrass, 1935). As the tarsomeres contain no intrinsic muscles, being moved by muscles at the apex of the tibia, they are not true segments. Primitively the Psocoptera have three tarsomeres per leg, which are reduced to two tarsomeres in a number of clades (Smithers, 1972; Hennig, 1981). The Phthiraptera generally have two tarsomeres, though these may be partially fused to each other (as in most Anoplura and the trichodectid Lutridia) or completely fused to form a single tarsal segment. In the Anoplura the tarsi may be virtually immovable with respect to the tibia (Ferris, 1951). The shape of the tarsus, whether long and slender or short and conical, may be of taxonomic 146

use. Busvine (1978) suggests that the lengths of claw, tarsus and tibia are related to the size and structure of the hair to be grasped by the louse, and demonstrates that human head lice (Pedicuius capitis) and body lice (P. humanus) can be separated by measurements of these structures. The probability of the shape and size of the tarsus being directly related to structural features of the host hair must be considered when employing these characters for systematic analysis. The ventral surface of each tarsomere frequently bears a plantula (euplantula of Keler, 1952, 1969; Clay, 1969). Clay (1969, 1970) finds that in the Amblycera the ventral surface of the plantula may be unstructured or, beneath a covering membrane, composed of a honey- combed area within which may be a framework of vertical or vertical and horizontal strands. The plantulae of tarsomeres I and II may have quite distinct structures and are frequently of different sizes (Clay, 1969). The plantulae of both tarsomeres are present in Trichodectidae; generally they are small and nipple-like, but in some of the Bovicolinae the plantula of the first and sometimes second tarsomere is long, sickle-shaped and pointed (Pig. 17a) (Keler, 1938).

The pretarsus. The pretarsus plesiomorphically bears two claws, but this number is reduced to one in Gyropidae and Abrocomophagidae (Amblycera), Trichodectidae, Rhyncophthirina and most Anoplura, a rudimentary second claw being present in a few of the latter group. All of these lice are parasites of mammals, but three families of mammal-infesting lice: Trichophilopteridae (Ischnocera) and Boopiidae and Trimenoponidae (Amblycera) have two claws, as do all bird-infesting forms. The reduction to a single claw is believed here to be an adaptation to ectoparasitism on mammals; a parallel reduction is present on those Hippoboscidae (Diptera) ectoparasitic on mammals. Mayer (1954) maintains that claws are absent in the Gyropid subfamily Gliricolinae (Amblycera), but Keler (1955) demonstrates that the leg terminates in an enlarged second plantula, dorsal to which is a small seta-like claw. The claw in Psocodea may be ventrally toothed, as in many Psocoptera, Haematomyzus (Rhyncophthirina) and Dasyonyx 147

(Trichodectidae) (Fig.l8a,b), ridged, as in Eurytrichodectes (Tricho- dectidae) (Fig. 18c) or smooth. At the base of the claw in many Trichodectidae is a small hyaline projection, which may be pointed or blunt (Fig. 17a, 18a , 18b ). Mayer (1954), in her study of Bovicola caprae, terms this structure a "pulvillus11 but, as a true pulvillus is generally paired (Richards and Davies, 1977; Chapman, 1982), this term is probably inappropriate. It is more probable that the projection is an empodium or arolium, though it is possible that at least in some cases it is nothing more than a simple basal tooth. Kim and Ludwig (1978b, 1982) maintain that the pulvillus and empodium do not occur in Phthiraptera, but Clay (1969) demonstrates the presence of an empodium in Menoponidae (Amblycera), and Clay (1970) figures structures in Boopia (Amblycera) that almost certainly are pulvilli, though she follows Keler (no reference given) in teiroing them plantulae of the second tarsomere.

Function. Most mammal-infesting lice apparently clasp the hair of the host between the tarsal claw and the broadened apex of the tibia, but members of the subfamilies Gyropinae and Protogyropinae (Amblycera: Gyropidae) use the second tarsomere, tibia and femur of the second and third pairs of legs (Ewing, 1924: Keler, 1955). The inner surface of each of these segments is strongly ridged to provide a frictional surface. During locomotion the leg is clasped round a hair, the tarsus locking into a tentaculum at the base of the femur (Ewing, 1924). In the subfamily Gliricolinae (Amblycera: Gyropidae) the tarsus is much reduced, as described above, but the femur and the tibia are still strongly striate. In this case small hairs are clasped between the tibia and femur of the same leg, but large hairs are held between opposing legs of the same thoracic segment (Ewing, 1924). In species of Phthiraptera where the antennae are not modified in the male to clasp the female during copulation (see above), this function may be performed by the legs. In Pedicuius (Anoplura) the front pair of legs is used for this purpose (Sikora and Eichler, 1941), whilst in Gyropidae (Amblycera) the hind two pairs are used (Ewing, 1924), but no special modifications have been observed. 148

It is possible that the legs have a sensory function in some lice. Stenram (1956) suggests that the median and hind pairs of legs in Columbicola columbae (Ischnocera) are used to detect the lay of the feather barbules, so that the louse can maintain its precise orientation with respect to the feather. This hypothesis has not been tested, as any attempt to inhibit the sensilla of the legs is sufficient to prevent the insect from moving at all, for mechanical reasons. In most Ischnocera and Anoplura the front legs are appreciably smaller than the mid and hind pairs, and in some Anoplura the claw type is very different, the front legs not being greatly adapted for clasping hairs whilst the other two pairs are. It is possible "that the front legs have a sensory function in these cases, but the hypothesis has not been tested and the significance of the dimorphism is not known.

2.2.4. The Abdomen 2.2.4.1. Segmentation Introduction. In many Phthiraptera one or more of the abdominal segments are partly or wholly suppressed. Failure to recognise this has led to considerable confusion regarding the correct numeration of segments and the position of abdominal structures. Bedford (1932a) places the six pairs of abdominal spiracles "either on segments 3-8 or 2 - 7". The opening of the male genitalia has been variously placed as posterior to sterna eight, nine or ten (Piotrowski, 1961; Keler, 1938, 1971) and Mukerji and Sen-Sarma (1955) use the presumed opening of the female genitalia behind sternum seven as one character justifying ordinal status for Rhyncophthirina. If allowance is made for segmental loss the structures mentioned above are found to be constant in their position, and can be used to identify segments. The first pair of abdominal spiracles, if present, is alv/ays on abdominal pleurum III (the single exception, the anopluran Neolinognathus, has its only pair of abdominal spiracles on the pleura of segment VIII). The male genitalia always open behind sternum IX and the female genitalia behind sternum VIII. 149

In this study the true segmental number is referred to by roman numerals to distinguish it from the apparent number. The terms ' stemite', 'pleurite' and *tergite' are used for the sclerites of the sternum, pleurum and tergum respectively of each segment.

Segment I. In all Acercaria sternum I is reduced or suppressed (KHnigsmaim, 1960); the identification of the area between the hind coxae as the first sternum (Keler, 1938) in some Phthiraptera is almost certainly incorrect. Tergum and pleurum I are fully expressed in all Psocoptera, some Amblycera, Ischnocera and Anoplura, and possibly the Rhyncophthirina. Full expression in the Amblycera is found in the Laemobothriidae, Gyropidae and almost all Menoponidae; in Trimenoponidae the pleura are absent (presumed fused to the metathoracic pleura) and the tergum is reduced though free; in Boopiidae, Ricinidae, Protogyropinae and Gliricolinae the pleura are absent (presumed fused to the metathoracic pleura) and the tergum is fused to the metanoturn (Clay, 1970). In the Rhyncophthirina the first tergum ("mid-dorsal tergite" of Mukerji and Sen-Saima, 1955) is free and large. Pleurum I of the Rhyncophthirina is difficult to identify. The '"pleurum one' of Mukerji and Sen-Sarma (1955) is pleurum II, and pleurum I is apparently absent. However, pleurum II has a pair of long stout setae postero-dorsall^ one of which is broad and slightly hyaline, and it is notable that a similar pair of setae is present on either side of tergum I without serial homologues on the other terga. Although Clay (1963) figures these setae as on tergite I, they are on separate lateral sclerites, which are here identified as the first pleura, displaced to the dorsal surface of the abdomen. In Ischnocera tergum I is present only in Trichodectidae and male Trichophilopteridae, though in both families it is reduced in size. Pleurum I is always absent. Wilson (1936) notes that the first apparent tergum of the first instar nymph of Cuclotogaster heterogranhus (Philopteridae) is wider than the succeeding terga and bears two transverse setal rows instead of only one. This leads Wilson to suggest that the first apparent tergum is a fusion product of terga I and II. Wilson (1939) describes a similar situation for 150

Lipeurus caponis (Philopteridae) and Clay (1958) identifies the first tergum of Degeeriella (Philopteridae) as terga I and II fused. An anterior reduced setal row is present on the first apparent tergum of many Ischnocera in the Philopteridae, Goniodidae and female Trichophilopteridae, indicating that terga I and II fuse in most Ischnocera. The fate of pleurum I in the Ischnocera is not apparent. The Anoplura vary in the dqgree of reduction of the first segment, though the loss of abdominal sclerites and the proliferation of setae make it difficult to determine the degree of reduction in many cases. In Polyplacidae and Hoplopleuridae tergal and pleural sclerites are present for segment I, whilst in the presence of the tergal element of segment I is inferred from a doubled row of setae on the first apparent tergum. It is suggested that terga I and II fuse in some Anoplura as they do in some Ischnocera.

Segment II. Sternal, pleural and tergal elements are present in all Phthiraptera, though fusion of the tergum to tergum I, as discussed above, takes place in some Ischnocera and Anoplura and fusion of all elements to segment III takes place in some Anoplura. In the trichodectid genus Procavicola sternum II is greatly thickened and articulated with pleurum II, forming a 'sternal apophysis* (Pigs. 117 , 120 ). The function of this structure is unknown, but in its morphology it is convergent on a structure formed by the enlargement of the postcoxale of the hind leg (see above) which is developed in some other hyrax lice.

Segments III - VIII. Sternal, pleural and tergal elements are present in all Phthiraptera. Some fusion of these segments may take place, especially in the Anoplura. Primitively, a pair of spiracles is present on each of these segments, though some or all may be lost (see below). Segments VII and VTII may be modified in connection with the genitalia; these modifications are discussed below. 151

Segment IX. This segment is greatly modified in both sexes, as the female genitalia open to its anterior and the male genitalia to its posterior. In most Psocodea it is fused to segment X, though this may be obscured ventrally in the male by the genital opening (Wilson, 1936; Gttnther, 1974). 'Wilson (1936) notes that in Cuclotogaster heterographus (Ischnocera) the segmental fusion takes place at the final ecdysis.

Postgenital segments. The two terminal segments are treated together because of their intimate association, particularly in the Trichodectidae. There is some disagreement concerning the identity of the final abdominal segment. Snodgrass (1935) and Richards and Davies (1977) consider that the final segment in most insects is XI, with a posterior telson normally lost. Matsuda (1976) disagrees, however, suggesting that XI is generally lost or retained as cerci, whilst XII is retained as the epiproct and paraprocts. In this study the more traditional view of Richards and Davies (1977) is followed. In both sexes of Trichodectidae segments X and XI are fused. In females XI lies caudally, posterior to X (Fig. 164), but in males the modifications of segment IX have led to the displacement of both X and XI onto the dorsal surface of the abdomen. In most male Trichodectidae the genital opening is posterior or postero-dorsal (see discussion below). The postgenital segments are reduced to form a small anal cone arising from the dorsal (anterior) wall of the genital chamber (Keler, 1957) (Fig. 19).

Pleural projections. In some Phthiraptera the posterior angles of the abdominal pleura are produced dorsally and ventrally. The function of this modification is unknown. In many Anoplura, such as Polyulax and Hoplopleura, pointed or planate projections are present on most pleura, whilst in Pedicuius schaeffi pleurum VT only is conically produced dorsally and ventrally. In many Trichodectidae conical projections may be present either dorsoposteriorly or ventro- posteriorly on any of the first three pleura (II, III and IV). The distribution of the pleural projections in the Trichodectidae is Fig. 19. Structures of the terminal segments of the male trichodectid abdomen. VJl ro 153

TAXA SEX PLEUHUM PLEURJM PLEUFUM

II III IV

D V S D V S D V S

Neotrlchodectes mephltldls 9 +

Cebidlcola + +

Procavocola - Eurytrichodectes clade. Lorlsicola 9 2 + + +

rajobergl, Fellcola zeylonlcus -viverriculae clade

Fellcola bedfordl. F. congoensls. F. cooleyi. F. cynlctls . 8 2 + +

F. decipiens. F. helogale. F. minimus. F. liberiae -

subrostratus clade. Lorlsicola aspldorhynchus. L. caffrai

L. fells. L. hereynianus, L. malayslanus. L. mungos.

L. spencerl. L. sumatrensis. L. werneckl,

L. bengalensis - juccll clade

Lorlsicola acutlceps. L. afrlcanus. L. lenlcomis, + ?

L. neoafrlcanus

Fellcola occldentalls. F. quadratlceps. F. vulpls + +

Fellcola calogaleus. F. setosus

Trlchodectes zorlllae 8 + + + + + + + + + o + ?

Oeomydoecus (G.) most species

Geomydoecus (G.) thomomyus - dakotensls clade. 8 + + + + • + • + + + + + + Geomydoecus (T.) 2

Table IV. Distribution of abdominal pleural projections in Trichodectidae. For pleura II, III and IV an indication is given whether a dorsal (fDf), or ventral (•V*) projection is present, and whether those projections are sclerotised ('S1). In each case the presence state is indicated by f+f. Very light sclerotisations are indicated by f?f. 154

summarised in Table IV. The projections of Trichodectidae, like those of the Anoplura, are directed dorsoposteriorly and ventro- posteriorly, but in slide-mounted specimens the vertical component of the projections is lost, and they lie posteromedially (Pigs 150, 218 , 255 )•

2.2.4.2. Pemale genital and postgenital segments Introduction. The female genital chamber in Phthiraptera opens to the posterior of sternum VIII, the sclerite of which forms the subgenital plate, and may be fused to the sclerite of sternum VXE. Laterally on sternum VIII may be a pair of projecting lobes termed gonapophyses. The postgenital segments are fused and may bear further gonapophyses.

Subgenital plate. There has been considerable confusion regarding the composition of the subgenital plate, much of it arising from incorrect numeration of the abdominal segment's. Thus, though Strindberg (1916) correctly identifies the subgenital plate in Gliricola and Gyropus (Amblycera) as comprising sternites VII and VIII or VIII alone, Heberdey (1931) maintains that in these genera the subgenital plate is formed from sternite VTI, with sternum VIII fozming all or part of the dorsal wall of the genital chamber. The latter interpretation involves the assumption that the seventh apparent sternum is sternum VII, when it is in fact sternum VIII. Keler (1938) suggests that in Trichodectidae the subgenital plate and the ventral wall of the genital chamber are formed from the ninth sternum reflexed upon itself, the gonapophyses being the transformed ninth pleura. This interpretation is based upon the misinterpretation of a line of flexion across sternum VIII as an intersegmental suture, and is corrected by Keler (1957). Ferris (1951) and Piotrowski (1961) report that in the Anoplura the subgenital plate is developed from the sclerite of sternum VIII (sternite VIII) and Keler (1971) states that in Boopidae (Amblycera) the plate is developed from either sternite VIII alone or sternites VII and VIII in conjunction. Observations made in this study 155

indicate that in the Trichodectidae the situation is identical to that reported by Keler (1971) in the Boopiidae.

Ventral vulval margin and subgenital lobe. The posterior margin of sternum VIII forms the ventral margin of the vulva. This margin may be unmodified, emarginate or produced. In many Trichodectidae it is expanded posteriad and slightly laterad (Pig. 109 ), an apomorphic modification listed below as "vulval margin expanded". In some species of Trichodectidae the centre of the ventral vulval margin is greatly expanded into a flat lobe termed the 1 subgenital lobe* (Pigs 164 , 166 , 188 ). The subgenital lobe may be smooth or spinous, and the variations in its form are of taxonomic and systematic value. The subgenital lobe appears to have evolved at least three times in the Trichodectidae, but the precise function of the structure is not known.

Gonapophyses. It has been stated that the Phthiraptera lack an ovipositor (Scudder, 1971; Matsuda, 1976; Boudreaux, 1979). However, although there is no fully-developed ovipositor, structures derived from the ovipositor of Psocoptera are frequently present (Table V.) and in some cases are clearly involved in oviposition. These structures have been collectively termed copulatory valves or gonapophyses, though no clear idea of their homologies has been •developed in the literature. Keler (1938) shows that in Trichodectidae the posterior margin of the so-called copulatory valves of segment VIII are continuous with the vulval margin, and his erroneous identification of the subgenital plate as sternum IX leads him to identify the copulatory valves as the modified pleura of segment nine. However, as the subgenital plate is sternum VIII, and the pleura of segment VIII are present and distinct from the copulatory valves, this identification is not tenable. Piotrowski (1961) shows that in Pedicuius humanus (Anoplura) the copulatory valves of segment VIII arise as outgrowths within the hypodermal invagination on sternum VIII at the end of the first instar, moving outwards to lie on the subgenital plate in the adult. Piotrowski interprets these structures as 'uropods1 - paired segmental appendages serially homologous with AMBLYCERA ISCHNOCERA A A 0 0 P 0 9 H M O p SEGMENT STRUCTURE M 0 m P fe IH 0 w H P EH p 8 pLi CD H 53 53 b P O P 8 in e « < EH fe fcH P M pi A EH fe g fe g O Js; M 8 g Si fe p te O P M fe H p o o P O fe M 8 H M H m M S O 8 w P P o P O p g « %

VIII GONAPOPHYSIS VIII

IX GONAPOPHYSIS IX P - GQNOPLAC P

•GENITAL LOBE' P x (structure of unknown origin,

developed, from one or "both of

the above)

Xl/XII PARAPROCT x X

Table V. Distribution of ovipositor elements within the Psocodea. Structures are scored positively if present in a"t least one member of a taxon. The presence of the structure, however modified, is indicated by 'p', the presence only of a group of specialised setae on the site by 'a', absence by 'x1, and non-comparability (unidentifiable) by 157

the thoracic legs. Clay (1970) notes that structures similar to and probably homologous with the copulatory valves of Trichodectidae are found in some genera of Goniodidae and Philopteridae (Ischnocera), Rhyncophthirina and Anoplura. Such valves are also found within the Amblycera (Table V ). The most parsimonious explanation for the distribution of these valves of the subgenital plate throughout the Phthiraptera is that they.are all homologous with the structures in this position in Psocoptera, the gonapophyses of segment VIII. The teim 'copulatory valve' is therefore discarded in favour of 'gonapophysis VIII' for use throughout the Phthiraptera. The 'gonapophyses' of segment IX present a more difficult problem, and it is not clear whether these structures are homologous throughout the Order. Keler (1971) identifies the 'gonapophysis' of Boopiidae (Amblycera) with the 'outer gonapophysis' (gonoplac) of Psocoptera, though in a footnote suggests that this is incorrect, a more correct homology being with the paraproct or a fusion of this with the gonoplac. In Psocoptera the paraprocts usually lie posterior to lateral extensions of tergite IX + X (Keler, 1971; Matsuda, 1976; and see Gilnther, 1974, Fig. 12). Keler (1971) in his footnote compares this situation to that found in the Boopiidae, where lateral extensions of tergite IX + X approach the 'gonapophyses' in a very similar manner. (Keler, 1971, refers to tergite IX + X as IX throughout in the belief that tergum X is absent; this failure to recognise the fusion of tergites IX and X in no way affects the validity of his arguments.) Keler (1971) also shows that the anal margin is dorsally continuous with the 'gonapophyses', as is true of the psocodean paraprocts. The correspondence of position and associated sclerites between the para- procts of Psocoptera and the 'gonapophyses' of Boopiidae strongly indicates the homology of these two structures. Sclerites interpreted as paraprocts and epiprocts have been observed in other groups of Phthiraptera (Table V ). Keler (1971) states in the footnote mentioned above: "... probable that the "gonapophyses" of Boopiidae correspond either to paraprocts or to a fusion of them with the outer gonapophyses [of the Psocoptera] The fact that the posterior arms of tergite IX [ + X] in Boopiidae reach in some species to the sides of the postgenital sternite and others to the lateral margins of the 158

"gonapophyses" speaks rather for the bivalent nature of the "gonapo- physes" in Boopiidae." Keler (1971) does not state in which species he considered the extensions of tergite IX + X as reaching the post- genital sternite, and examination of Keler1s figures and of specimens of many species of Boopiidae has failed to reveal any case in which this is so. This discrepancy may be due to differing interpretations of the posterior dorsal sclerites in Boopiidae. Tergum VIII has in some species an anterior and a posterior tergite, the latter being isolated from the pleura by the former. The intersegmental suture between terga VIII and IX is clear, and there is no doubt that the posterior tergite is really that of segment VTII. On tergum IX + X there is a median and two lateral'sclerites. The median sclerite, here interpreted as tergite IX + X, has in some species two postero- lateral extensions towards the paraprocts. The lateral plates articulate with pleura VIII, but do not represent the gonapophyses of segment VIII. K§ler (1971) interpreted these plates as tergopleural in origin. In some cases the lateral plates approach the postgenital stemite, but this is not homologous with the posterolateral extension of the median tergite; it is possible, however, that this is the arrangement referred to by Keler (1971) as the reaching of the posterior arms of tergite IX[,+ X] to the postgenital stemite. If this is the case, the association does not provide any support for the inclusion of any element of the gonoplac with the paraproct in the formation of the boopiid ' gonapophysis'. The postgenital sternite in Boopiidae is laterally setose and slightly produced; it may lie in close apposition to the paraprocts, though it is separated from them by clear sutures. Keler (1971) identified the lateral projections as the inner gonapophyses (gonapo- physis IX). The terminal sterna of female Psocoptera have four pairs of sclerites that are relevant to the possible composition of 'phthirapteran terminalia: the paraprocts, the gonopla'cs, gonapophyses XI and gonapophyses VIII. Of these, gonapophyses VIII are not present in the Boopiidae, and the paraprocts are, and have been identified. The Boopiidae have then a single pair of unhomologised sclerites, fused to the postgenital stemite (itself unhomologised). Keler1 s (1971) assumption that the homology of the processes is with the- 159

gonapophyses of sternum IX is linked to his initial homology of the gonoplacs with the so-called "gonapophyses" (paraprocts). This being rejected, there is no reason for homologising the lateral extensions of the postgenital sternite with one of the paired structures of segment IX rather than the other. Both gonapophyses IX and the gonoplacs arise from the same structure (Scudder, 1971), and both are present in the Liposcelidae (Smithers, 1972), the presumed sister- group of the Phthiraptera. There seem to be no criteria upon which homology with one or the other can be made, though the posterior lateral extensions on sternum IX in the Boopiidae may be homologous with other extensions similarly placed in other Phthiraptera. All these structures can only be referred to as of unknown homology, and it is proposed that the neutral term "genital lobes" (Kim, 1965) be applied throughout the Phthiraptera (Table V ). In view of the above discussion, it is evident that attempts to divide the Phthiraptera on the basis of gonapophysis type (Clay, 1970) are neither possible nor justified. The structure of the gonapophysis (gonapophysis VIII) in Trichodectidae is variable. In all cases a basal internal apophysis is produced, presumably as a muscle attachment (Keler, 1938). The gonapophysis lies longitudinally, parallel to the abdomen; it may curve on to the dorsal surface apically. It may be long, slender and apically acute (Pig. 244), long and apically 'squared off' (Pig. 167 ), broad and membranous (Pig. 254), produced into a rounded, toothed or rectangular medial lobe with an apical ' spur' (Pigs. 188, 243a» 245), lobulate but with the spur reduced (Pig. 246) or absent (Pig. 221), sometimes with serrations laterally (Pig. 221). Setae may be present on the gonapophyses and the ventral vulval margin in various configurations, occasionally based on sclerotised tubercles. In some cases a sclerotised band links the gonapophyses and the vulval margin, but in most cases this is not present and the connection is membranous. The curvature of the gonapophyses about both longitudinal and lateral axes produces a complex three-dimensional structure difficult to interpret on slide-mounted specimens. Distortion due to mounting can be seen by comparison of Figures 20a and 20b ; both are of the 160

Fig. 20. Distortion of gonapophyses by the mounting process, (a) Specimen of Lorisicola bengalensis in collection of the British Museum (Natural History); (b) Specimen of L. bengalensis figured by her neck (1948) • Fig. 21. Spermathecae of Phthiraptera (after Blagoveshtchenski, 1956). (a) Laemobothrion cerci; (b) Ricinus fringillae. 161

gonaphysis of the same species, in each case distorted by the mounting process. A similar apparent difference might result if two species possessed gonapophyses of the same shape but differently oriented. A further problem arises with species in which the gonapophysis is curved dorsally and has a membranous base; pressure from the cover- slip may distort the specimen and give a false impression of the orientation. The function of the gonapophyses in Trichodectidae is at least partially as an ovipositor. Murray (1957a) describes oviposition behaviour in Bovicola ovis (as Damalinia ovis): "First, the gonopods were raised away from the abdomen and by sweeping movements a fibre was 'caught1 and held next to the abdomen". Extrusion of the cement and egg followed. Murray (1957b) shows that if a hair cannot be held firmly next to the abdomen by the gonapophyses, the diameter of the hair being too small or too large, no egg is laid. In Bovicola ovis the gonapophyses have not been observed to have any function except the one described above, but in the Anoplura stenopsis and Haematopinus eurysternus the gonapophyses not only 'grasp' the hair but also act as moulds for the cement. Murray (1957a) states "Variations in the shape of the cement attachment appeared to be related to the use of the gonopods. If the gonopods were not used as a mould for the cement, relatively untidy attachment resulted as with D. ovis. When the gonopods were used as a mould, the attachment was tidy as with Li. stenopsis and H. eurysternus. The shape and manner of use of the gonopods as moulds for the cement undoubtedly leads to attachments which may be characteristic for a species." Ferris (1951) depicts a number of eggs of Anoplura together with the cement attachments; it is clear that there is considerable variation in the shapes produced in the cement. It is not known whether any Trichodectidae utilise the gonapophyses to mould the egg cement. The observation that Bovicola ovis can move the gonapophyses independently of the abdomen may apply to most Trichodectidae, though in no case can a clear articulation between the gonapophyses and the vulval margin be seen. In many species of the Trichodectes-Lutridia clade the continuous sclerotised margin of the gonapophysis and the vulval margin may limit such independent movement. The requirement 162

of Bovicola ovis for a hair of the correct diameter for oviposition and the ability to detect unsuitable hairs implies selection on the gonapophyses directly related to the structure of the hair of the host.

Postgenital sclerites. Posterior to the vulva in Phthiraptera are several sternal and pleural sclerites, the homologies of which are not clear. There are several sclerites which must be considered when making homologies: the gonangulum, sternite IX + X, pleurite IX + X, and the paraproct. The paraprocts have been discussed in some detail above, but are not always readily identifiable. In the Trichodectidae tergite IX + X is generally present, and frequently has extensions that approach or are fused to lateral sclerites of uncertain homology. In some species the extensions are directed posterolaterally, and the lateral sclerites are positioned posteriorly on projections of the abdomen (Figs- 162, 164); though distant from the anus, the lateral sclerites in this case appear to be readily homologisable with the paraprocts. In other species, however, the sclerites are more laterally positioned on the projection (Fig. 246) and appear more as pleura than- as paraprocts. Homology of these sclerites being uncertain, they are referred to in this study as 'postgenital pleurites' (p-g.p.). In Boopiidae two posterior sclerites are present: the genital sternite and the postgenital stemite (terminology of Keler, 1971). The genital sternite extends into the genital chamber and may surround the entry to the spermathecal duct; it meets the genital lobes (gonapophyses IX or the gonoplacs) posteriorly and pleurotergite IX + X posterolaterally. These associations may be compared to those of the gonangulum, which is a triangular plate found in many Thysanura - Pterygota, articulating anteriorly with gonapophysis VIII (a structure absent in the Boopiidae but which arises laterally to the genital chamber in some other Phthiraptera), posteroventrally with the second gonocoxa (from which arises both gonapophysis XI and the gonoplac) and, in the Acercaria, fused posterodorsally with tergum IX (Scudder, 1961, 1971). The gonangulum is present in Psocoptera, though generally fused to other sclerites and difficult to delimit (Scudder, 1971: Matsuda, 1976). The correspondence between the 163

associations of the genital stemite in Boopiidae and the gonangulum as described by Scudder (1971) strongly indicates the homology of these two sclerites. The homology of the genital sternite with sternite IX, proposed by Keler (1971), is rejected. The remaining sclerite of the boopiid postgenital sternum, the postgenital sternite, is homologised with sternite IX + X, though it must be noted that this sclerite is apparently absent in the Psocoptera. In the Trichodectidae there is only one sternal postgenital sclerite, the 'post-vulval sclerite1 (Moreby, 1978), which lies posterior to the vulval margin and may be fused to the postgenital pleurite. The sclerite may be single, medially divided, narrow and strip-like, broad and triangular, or absent (Pigs 167, 188,2459 246 ,254) . it is not considered possible to homologise this sclerite with either gonangulum or sternite IX + X, and the term 'post-vulval sclerite1 is retained.

Female genital chamber, oviduct and associated structures. The female genital chamber is an oval, dorso-ventrally compressed, lightly- sclerotised invagination of the integument lying between sterna VTII and IX. The "common oviduct enters the chamber anteriorly and a spermatheca may be present posteriorly. Accessory glands of the ninth segment are absent, but accessory glands associated with the common oviduct may be present. A spermatophore, or free spermatozoa, is deposited in the genital chamber by the eversible or protrusible male genitalia. The dorsal wall of the chamber may be heavily sculptured and bear sclerotised spicules, ridges or spines; this sculpturing may extend on to the postgenital sterna. The" ventral wall may be similarly sculptured, though usually to a lesser extent; where the subgenital lobe is present the sculpturing of the genital chamber may extend on to its dorsal surface. The sculpturing of the walls of the chamber probably provide a highly frictional surface against which the spicular surface of the endophallus acts to provide a firm union during copulation. The nature of the sculpturing and the distributions of spicules or spines may be taxonomically useful at the specific or subspecific level. The common oviduct opens into the anterior end of the genital chamber and curves sharply posteriad to lie dorsally to the chamber. Dorsally to the vulva Fig. 22. Diagrammatic three-dimensional representation of female genital chamber and oviduct (internal). ON 165

the common oviduct curves sharply anteriad and divides into the two paired oviducts (Pig. 22). The genital chamber, though assuming a more or less circular cross-section during passage of an egg and perhaps during copulation, is at rest a dorso-ventrally flattened, fairly rigid structure, the minimum width of which is governed by the diameter of the egg. The oviduct, however, is an elastic, folded membranous tube, compressed and folded to reduce volume and expanding only to allow passage of the egg. The common oviduct must, at the junction with the genital chamber, be the same diameter as the chamber. At its division into the paired oviducts, however, it is when at rest narrow and greatly folded. Between these two points it narrows more or less abruptly, and folds may be apparent in its walls. Emerson and Price (1971) interpret these folds in Geomydoecus (Trichodectidae) as striations of the dorsal wall of the genital chamber, and term this apparently membranous structure the 'genital sac'. The true dorsal wall is interpreted as the ventral wall and the true ventral wall is apparently not observed. The form of the folds of the 'genital sac' (common oviduct), as well as its length and the width of the junction with the genital chamber, have been utilised by Price and his co-workers as specific and subspecific characters in their revisionary work on Geomydoecus (e.g. Emerson and Price, 1971; Price, 1974; Price and Hellenthal, 1976; Timm and Price, 1980). The apparent constancy of the dimensions of the oviduct is related to the size of the egg and perhaps to the restrictions in variability of size of the genital chamber imposed by selection. The apparent constancy of the folds is more puzzling though it is possible that tubes of identical length and diameter have an optimum folding pattern if compressed in the same way. It is also possible that the wall of the common oviduct is very lightly sclerotised, at least near to the junction with the genital chamber. The folds of the common oviduct have not been utilised as a taxonomic character elsewhere in the Phthiraptera. A spermatheca is present in many Phthiraptera (Heberdey, 1931; Blagoveshchenski, 1956), and may be in the form of a single globular sac (Pig. 21b) or a tube with two lateral diverticula (Fig.2la). Blagoveshchenski (1956) examined several species of Trichodectidae but failed to find evidence of a spermatheca in this family. In 166

the present study examination of slide-mounted specimens of most species has failed to reveal a spermatheca in any genus apart from problematically in Dasyonyx, where a lightly-sclerotised sac, differing in form between species, is developed from the wall of the common oviduct. The relatively anterior position of this sac in comparison to the spermatheca of other Phthiraptera suggests that it may not be homologous with the spermatheca. No histological or anatomical studies have been carried out on the sac. Accessory glands of ninth segmental origin are absent in the Psocodea (Heberdey, 1931; Matsuda, 1976); the so-called accessory gland described by Strindberg (1916) is shown by Heberdey (1931) to be the spermatheca. Accessory glands arising from the common oviduct are found in a number of Phthiraptera (Ferris, 1951; Blagoveshchenski, 1956; Piotrowski, 1961) but have not been observed in Trichodectidae (Blagoveshchenski, 1956). The accessory glands of the oviduct have been proposed as the secretors of the cement used for attaching the egg to the hair or feather (Florence, 1921; Ries, 1932; Ferris, 1951; KBnigsmann, I96O). Yfeber (1936) fails to find accessory glands in the Psocoptera, but figures cement in proximity to the anus in an ovipositing psocid. Kdnigsmann (1960) cites Weber (1936) in proposing secretion of egg-cement as "vaginal" not "anal" as an apomorphy of the Phthiraptera. The origin of cement in Trichodectidae and other Phthiraptera in which no accessory glands are present is not known, and it is possible that the true source of cement in all Psocodea has yet to be discovered.

2.2.4.3. The male abdomen Abdominal morphology and copulation. The opening of the male genital chamber ('genital opening') is always posterior to sternum IX, and primitively is ventrally positioned and distant from the anus, which is terminal. During copulation the male and female are usually oriented the same way, with male ventral to the female. The tip of the male's abdomen is curved dorsally and anteriorly so that the genital openings of male and female meet (Fig. 23) • The male genitalia consequently enters the female genital chamber 'upside-down* , with the ventral side of the former coming into contact with the 167

Pig. 23. Copulation of Eutrichophilus cordiceps (from photograph in VI er neck, 1536). 168

dorsal wall of the latter. Should sclerotised tergites be absent (as in the Trichodectidae Neotrichodectes and Geomydoecus) or greatly reduced (as in the Trichodectidae Trichodectes galictidis and Felicola) flexion of the male abdomen during copulation is evenly distributed along the membranous dorsal surface. If sclerotised tergites are present and fully-developed, however, flexion must be about the sclerite-membrane-sclerite joints of the dorsal surface, these thus functioning as 'hinges1. At each such hinge, there will be considerable deformation of the internal structures during flexion, whilst between the hinges there will be no such deformation. Increasing the number of 'hinges' on the dorsal surface permits smoother curvature of the abdomen and reduces the degree of internal deformation at each hinge. Many male Trichodectidae and some Anoplura (e.g. Polyplacidae, Hoplopleuridae) have tergal sclerites divided into an anterior and posterior plate on at least some segments (Figs 117, 120, 150), possibly for this reason. The degree of internal deformation may be further reduced by modification of the shape of the plates. In order to determine the effect of different shapes of sclerite a transparent flexible model was constructed and rigid plates of various shapes were attached to one surface. The form of plate which, when placed in series, caused maximum internal deformation and required maximum membrane area: sclerite area ratio for flexion was a simple rectangle, with anterior and posterior margins at right angles to the long axis of the abdomen (Fig. - 24a). These factors were diminished by introducing curvature in the anterior and posterior margins of the plates, either with both margins parallel (Fig. 24b) or opposite, producing a series of alternating biconvex and biconcave plates (Fig. 24c). Both these patterns are approached in the Trichodectidae (Figs 79, 117, 120). The flexibility required of the ventral surface of the abdomen is developed, should the sterna be sclerotised, by increasing the lengths of the sclerites and allowing them to overlap when the abdomen is at rest (Figs 117, 120). Despite the increased flexibility of the abdomen achievable by modifications of the tergal and sternal plates, the position taken by the male during copulation is compatible with a ventral genital opening only if the abdomen is long and slender. A short, broad 169

a

Fig. 24. Diagrammatic representations of dorsum of male abdomen, showing different tergite forms, (a) rectangular tergites; (b) lunular tergites; (c) alternating binconvex and biconcave tergites. 170

abdomen will not deform sufficiently to turn back on itself and bring the male genital opening into contact with the vulva. Such a short, broad abdominal form has been developed independently several times in the Phthiraptera, however, including at least once in the Tricho- dectidae. In most cases the limitation imposed on the degree of curvature of the abdomen has been met by a posteriad or even dorsad migration of the genital opening, thus reducing the degree of curvature necessary. This displacement has been effected by an increase in length of sterna VII, VIII and IX and a concomitant shortening of the corresponding terga. The resultant configuration of terminal sclerites is very variable; tergite IX + X may be complete or medially divided, the two halves in the latter case lying anterior or laterally to the genital opening (Ledger, 1980). Tergite VIII, similarly, may be complete or medially divided by segment IX (Pigs 117, 120, 224). The variation of the male terminalia of bird-infesting Ischnocera has been discussed by Clay (1953) and Ledger (1980). The re-positioning of the genital opening has increased its proximity to the anus, and led to the displacement of the anus on to the dorsal surface of the abdomen in many species. In Trichodectidae the anus is contained within the genital chamber and the reduced segments X and XI project from the genital opening (Pig. 19 ) • Tbe curvature required of the abdomen has been further limited in some Trichodectidae by two distinct adaptations. In some species of Felicola segment IX is developed into a long, slender posterior process and the genital opening is dorsal and apical (Fig. 200). The basal apodeme and parameres are elongate and slender (Fig. 212). It is postulated that most of the curvature required is developed at the base of segment IX and the junction of the parameres and the basal apodeme. Constriction of the endophallus at the latter fold is prevented by the presence of a reduced, circular mesomeral arch, lacking an extension, which, being fused to the endophallus, prevents this structure from being pinched shut. In some trichcd&btid species segment IX projects dorsally from segment VIII, the genital opening being dorsal (Fig. 128). This positioning limits the degree of curvature required of the abdomen by increasing the proximity of the genital opening to the vulva on minimum curvature. 171

Subgenital plate. A subgenital plate is generally present in male Phthiraptera. This may comprise any combination of sternites IX, VIII and VII, fused or unfused. In Trichodectidae the plate is fundamentally uniform in construction, though great superficial differences may exist between species. The plate is here considered to comprise eight discrete elements: sternites VII, VIII, IX (and, occasionally, VI); a pair of rods (referred to as 'subgenital plate rods' or 's.g.p.r.' in this study) which lie laterally to the sternites and sometimes fuse them together (Figs. 19, 174, 224); the two pleurites of segment IX, and the post-genital sclerite, which is of uncertain homology (Fig. 19 )• Any of these elements may be present, reduced or absent, or fused to adjacent sclerites. The sternites and the post-genital sclerite may be whole or medially divided. In cases of extreme reduction only the lateral rods may be left (Fig. 176) or all sclerites absent (Fig. 200). in the most complete form (Fig. 79 ), all "the sclerites are fused, forming a squared plate; usually there are membranous areas surrounding the sternal setae ('perisetal gaps'), but these may be absent (Fig. 129 ). The subgenital plate rods on sternum VIII are always connected to the ventral wall of the genital chamber in Trichodectidae (Fig. 25), a feature also observed in some Anoplura (e.g. Enderleinellus, ). The function of this attachment is not known, although in species with ventrally-positioned parameres and a median posterior extension to the mesomeral arch the difference in lengths between the dorsal wall of the genital chamber and the membrane bet ween the basal apodeme and the subgenital plate might cause divergence of the apices of the parameres and mesomeral arch extension during extrusion of the genitalia to allow evertion of the endophallus (see below for discussion of terms).

Pseudostyli. In some Trichodectidae the posterolateral angles of the subgenital plate are greatly extended into setose lobes termed ' styli' by Eichler (1963) (Figs. 60 , 79 ). Abdominal sternal sclerites are probably derived from a fusion of the sternal plate and the coxal elements of the paired abdominal appendages (coxopodites) (Matsuda, 1976; Richards and Davies, 1977). In the subgenital plate 172

of the Acercaria, the paired nature of the gonocoxopodite component may be expressed as a concavity in the posterior margin of the plate and concomitant projection of the posterolateral angles (Matsuda, 1976). True abdominal styli are derived from the paired abdominal appendages and are serially homologous with either the shaft of the thoracic legs or the coxal styli (spurs of Matsuda, 1976) of the thoracic legs as present in the Machilidae (Thysanura) (Matsuda, 1976; Richards and Davies, 1977). Styli are not, therefore, homologous with the coxopodites, but arise from them, being separated by a clear sulcus. Though true styli are present in some Psocoptera (Matsuda, 1976), the posterior extensions of the subgenital plate of Trichodectidae are not demarcated by a sulcus and thus are not homologous with true styli. The so-called styli of the Trichodectidae are an indication of the gonocoxopodite component of the subgenital plate, and are here termed 'pseudostyli'• Taxonomic use may be made of the form of the pseudo- styli, which is very variable in the Trichodectidae, but the difficulty of assigning polarity to transformation series limits the applicability of this character within phylogenetic studies. The distribution of the pseudostyli through the Trichodectidae, however, is utilised in the construction of the phylogeny of the family (see below).

1 Abdominal flecks' . Moreby (1978) notes the presence of so-called 'abdominal lateral flecks' in males of the genus Werneckiella (Trichodectidae). These he describes as "cuticular patches on either side of the intersegmental line in the lateral-tergal regions of the abdomen, most developed in mid to posterior segments". These structures are found in all male Trichodectidae, but not other Phthiraptera. They consist of small pits positioned on the antero- dorsal angles of pleura III-VII, occasionally on II and VIII, and anterolaterally on terga III-V in many species (Fig. 22°). In Werneckiella the cuticular patches described by Moreby (1978) are positioned anteriorly and posteriorly to the pit; in Trichodectes canis there are no sclerites bordering the pit but a small sclerite is present at the base of the pit; in Neotrichodectes there are no sclerites associated with the pit at all. The function of these structures is unknown. 173

Specialised setae and projections. In some species of Phthiraptera, terga II and/or III of the males are ornamented with large setae or projections. Such developments are numerous in the Trichodectidae. Many species of Felicola have a pair of long setae medially on male tergum II (Pigs. 192, 200 , 202 ), whilst Geomydoecus (Thomomydoecus) spp•, Trichodectes ovalis, _T. ugandensis and T. galictidis undescribed subsp. have paired 'combs' of long setae on male terga II and III (Pig. 177) and Bovicola multispinosa and B. hemitragi have paired circular 'brushes' of setae on male tergum II (Pig. 54)* Damalinia omata has sclerotised blunt projections on male terga II and III. It is suggested that these specialised setae and projections in some way assist the male to hold the female during copulation. In no case, however, have the setae of -terga II or III been observed to be damaged, as they might be expected to be should they operate against the female abdomen, and there is no observational evidence to support (or refute) the hypothesis. Males of Neotrichodectes species have a pair of small medial setae on terga II-VII, sometimes separated by a seta of normal length (Pig. 241). This feature, not found in females or males of any other group, is of unknown function.

2.2.4.4. The male' genitalia Male genitalia of Psocodea. Considerable disagreement exists over the homologies of the component structures of the phthirapteran genitalia (for examples, see entries in glossary for 'parameres', 'mesomeres' and 'mesomeral arch'). This difficulty is not limited to the Phthiraptera, and homologies within and between other Orders are difficult to determine (see Tuxen, 1970). Study of the ontogeny of the genitalia throughout the Insecta does, however, provide evidence of homology in some of the imaginal features (Snodgrass, 1957; Scudder, 1971; Matsuda, 1976). Before examining the Phthiraptera and Trichodectidae, the ontogeny and structure of the genitalia in the rest of the Thysanura - Pterygota will be briefly reviewed. The rudiments of the male genitalia arise from the conjunctival membrane posterior to sternum IX, or within a pouch formed on the 174

median posterior part of sternum IX (Matsuda, 1976). The form of the rudiments is nearly always as simple paired lobes ('primary phallic lobes') (Snodgrass, 1935, 1957; Scudder, 1971; Matsuda, 1976). The occasional development of the primary phallic lobes on sternum X leads some workers to assign them to this segment (Snodgrass, 1957), but they are more generally accepted as outgrowths of segment IX, sometimes displaced during development to segment X (Scudder, 1971; Matsuda, 1976). The single exception to this is the Acrididae (Orthoptera), in which the primary phallic lobes are apparently derived from the embryonic appendages of segment X (Matsuda, 1976). The evolutionary origin of the primary phallic lobes, whether from abdominal paired appendages or as novel sternal structures, is a matter of some contention (Scudder, 1971; Matsuda, 1976) and will not be discussed here. The genital opening of the male has not remained in the same position throughout the evolution of the Insecta, but has progressed posteriad through the addition of posterior ducts of ectodermal origin. In Thysanura and Ephemeroptera the paired vasa deferentia open through the penis or penes as in more primitive forms. In the Neoptera a novel structure, the ductus ejaculatorius, is formed posteriad to the vasa deferentia, which join to meet it; the posterior opening of the ductus ejaculatorius is known as the gonopore, and lies between the primary phallic lobes (or their derivatives). In the Dermaptera the ductus. ej aculatorius is paired, each gonopore opening at the tip of a primary phallic lobe. In the Polyneoptera the ductus ejaculatorius may be eversible (Brinck, 1956; Boudreaux, 1979), and in the Plecoptera the openings of the vasa deferentia are the effective gonopores (Brinck, 1956) . In the Phalloneoptera a novel structure, the ' endophallus', is formed distal to the gonopore. The endophallus may be fully eversible, so that the gonopore is presented at its tip, or it may be less eversible, with partially sclerotised walls. Matsuda (1976) suggests that the endophallus arose early in insect development, being represented by the paired preputial sacs of the Dermaptera (one of which is lost in higher forms within the Order). The placement of this Order within the Polyneoptera (Heming, 1977; Boudreaux, 1979; Hennig, 1981), hov/ever, indicates that this particular form of the male genitalia is 175

autapomorphic for the Dermaptera, and the eversible sacs distal to the vasa deferentia are not homologous with the endophallus of the Phalloneoptera. Matsuda (1976) also applies the term 'endophallus1 to the eversible membranous structure of the Plecoptera from which open the paired vasa deferentia; this feature is clearly homologous with the genital chamber - the invagination of the conjunctival membrane containing the phallic organs - of other Thysanura - Pterygota (Brinck, 1956). The primary phallic lobes develop in very different ways to give the wide range of genitalic structures apparent in "adult insects. Matsuda (1976) lists fifteen different generalised 'end-products' of the developmental process undergone by the primary phallic lobes. In Odonata, Embioptera, Isoptera, and some species of Plecoptera the adult genitalia of the ninth segment are reduced or vestigial. In the Zoraptera the genitalia are complex but have not been investigated properly; their structure as observed in this study does not provide any evidence to support or refute hypotheses proposed below. In the Ephemeroptera the vasa deferentia invaginate both primary phallic lobes, which form paired penes, whilst in the Thysanura the primary phallic lobes fuse and the twin openings of the vasa deferentia coalesce to become a single apical opening. In the Dermaptera the primary phallic lobes are invaginated by paired ejaculatory ducts, and form paired penes in a manner analogous to the Ephemeroptera. In Neoptera the primary phallic lobes divide longitudinally to form four phallomeres (two parameres and two mesomeres), though delay in separation (heterochrony), secondary fusion, or further subdivision of some or all of the phallomeres, and partial or complete suppression of the parameres. all occur within the group (Matsuda, 1976). In Dictyoptera, Notoptera, some Plecoptera and some Orthoptera the phallomeres, from two to nine in number, surround the gonopore, whilst in the other Orders (other than Diptera, Psocodea and Thysanoptera, which are discussed more fully below), the mesomeres fuse to form a median aedeagus surrounding the genital opening (' phallotreme') through which everts the endophallus. Matsuda (1976) states that such an ' aedeagus-type penis' arose independently many times in the Insecta, but if the Orders in which such a structure occurs are plotted on a phylogeny of the Insecta, it can be seen that the ' aedeagus-type penis' 176

evolved once in the Thysanura (though in this case both parameres and mesomeres are involved, and the structure is invaginated by the vasa deferentia not the endophallus), and once or perhaps twice in the Phalloneoptera. The Diptera lack an aedeagus, or derive one independently of the primary phallic lobes, and in Hymenoptera and Mecoptera further structures are derived from the mesomeres in addition to the aedeagus (Matsuda, 1976). In the Phalloneoptera the parameres and mesomeres may not be entirely separate from each other basally, * and in such cases the common base of the structures is termed the 'phallobase* (Snodgrass, 1935). The phallobase, if present, surrounds the endophallus or ductus ejaculatorius, and may be provided with a basal apodeme lying ventrally to these structures. The aedeagus may be partially retracted into the phallobase, and the phallobase produced into a tubular sheath, the 'phallotheca' . Reduction of the aedeagus may lead to functional replacement of the aedeagus by the phallotheca, in which case its eversible lining (or the proximal part of it) is known as the endotheca (Snodgrass, 1935)• Few ontogenetic studies of the phthirapteran male genitalia have been made, so developmental information on which to base homologies is limited. Comparison of the available accounts indicates that the ontogeny is not uniform within the Phthiraptera, but the differences noted do not affect the homologies developed below. In all cases the foraation of the genital pouch is noted (Nussbaum, 1882; Strindberg, 1916; Piotrowski, 1961), but Nussbaum (1882) finds that the initial invagination in Lioeurus and Goniodes (ischnocera) is paired, the rudiments later fusing to form a single sac, whilst Strindberg (1916) and Piotrowski (1961) find only a single invagination in Gliricola and Gyropus (Amblycera: Gyropidae) and Pedicuius (Anoplura) respectively. Nussbaum (1882) and Strindberg (1916) describe the formation of the seminal vesicles from the original epidermal infolding, but Piotrowski (1961) finds that in Pedicuius the seminal vesicles are developed from mesodermal ampullae. Ontogenetically, therefore, the seminal vesicles of Pedicuius are not homologous with those of the Ischnocera and Amblycera examined, but by the congruence of the imaginal structures phyletic homology is indicated. 177

Of the available accounts of the development of the male genitalia in Phthiraptera, the most complete and detailed is that of Pedicuius humanus (Anoplura) by Piotrowski (1961), and examination of the external genitalia of Trichodectidae and other Psocodea with reference to this work enables homologies of the major parts of the genitalia to be established. The ontogeny of the male genitalia' of Pedicuius humanus as described by Piotrowski (1961) is summarised below; the imaginal genitalia are depicted in Pig. 26, labelled with the terms accepted in this study. In the first instar nymph the genital pouch forms as an invagination of the epidermis between sterna IX and X. At the first ecdysis a pair of ventrally-directed lobes forms from the dorsal wall of the genital pouch; these, termed 'uropods' by Piotrowski, are the primary phallic lobes. During the second instar the distal ends of the lobes fuse to form a 'V'-shaped structure, and the wall of the genital pouch between the bases of the lobes extends dorsad and anteriad to form the ductus ejaculatorius. At the second ecdysis a prolongation of the duct extends posteriad into the pouch to form the 'penis' , and "there is invagination of [the] vault [of the' dorsal wall of the pouch} into its lumen". This observation requires clarification. The 'penis' is a two-walled structure, the inner wall being the ductus ej aculatorius, the outer wall an ingrowth of the dorsal Wall of the genital pouch. During the second ecdysis the dorsal wall of the genital pouch around the base of the 'penis' , especially posterior to it, grows rapidly and extensively, so that the opening of the 'penis' in the dorsal wall is displaced'anteriad. The 'penis', however, stays in the same position relative to the primary phallic lobes. Growth of the outer wall of the pouch, therefore, must be matched by equal growth inside the invagination of the wall that forms (leads to) the 'penis'. At the end of the second ecdysis the genital pouch has a long tubular invagination from the anterodorsal wall that terminates with the 'penis' between the primary phallic lobes, and contains, the ductus ejaculatorius. During the third instar a paired infolding appears in the antero-lateral walls of the pouch, extending to cut off a smaller antero-ventral part from the antero-dorsal portion. By the end of the third instar the antero-ventral part, completely cut

179

Fig. 26. Male genitalia of Pedicuius humanus (after Ferris, 1951)f illustrating terms used in this study. 180

off, has become greatly compressed dorso-ventrally, and is producing cuticle (especially laterally). This leads to virtual loss of identity as a two-layered structure as it becomes a rather trough- shaped 'basal plate' (Fig. 25). The antero-dorsal part of the genital pouch forms a membranous sac ('vesica penis' of Poitrowski) and connects with the ductus ejaculatorius. During the third ecdysis the bases of the primary phallic lobes must move laterally and - ventrally until they abut the posterior ends of the lateral arms of the 'basal plate', for in the imago the 'V'-shaped structure lies posterior to the 'basal plate' and articulates with its lateral arms. The 'vesica penis' arises between the branches of the 'V' and, when not everted, lies on the 'basal plate' . The eversible 'vesica penis' distal to the ductus ejaculatorius must be considered homologous with the endophallus of other Phallo- neoptera, and will be referred to as such. A 'basal plate' is found throughout the Psocodea. In all cases it lies ventrally to the genital duct and does not extend dorsally to it. The plate is usually faintly sclerotised medially and strongly sclerotised laterally, frequently appearing as two disconnected rods (Fig. 27 ); the anterior of the plate may be sclerotised (Fig. 181) or not (Figs 180 , 205 )» degree of sclerotisation is not considered to be of phylogenetic importance. Posteriorly the 'basal plate' articulates v/ith or is fused to at least some of the other sclerites of the genitalia (products of the primary phallic lobes); further sclerites may be present on the walls of the endophallus, and these are more or less distant from the plate. The positional and structural correspondence of the 'basal plate' throughout the Psocodea strongly indicates the homology of this feature throughout the superorder. In the Psocoptera the term ' phallobase' is applied to the 'basal plate' (Matsuda, 1976). As detailed above, the 'true' phallobase is derived from the primary phallic lobes and extends dorsally to the genital duct. Such a structure is found in the Condylognatha (Matsuda, 1976), but neither of the criteria are fulfilled by the 'basal plate' in the Psocodea. The homology of the 'basal plate' may be determined by its position relative to the products of the primary phallic lobes and to the genital duct. It is suggested that the 'basal plate' is phyletically 181

homologous with the-basal apodeme of other Phalloneoptera, and this teimwvill be used in this study. The primary phallic lobes give rise in Pedicuius only to the 1V1- shaped structure, which is interpreted by Piotrowski (1961) as the apically-fused parameres. Ferris (1951), however, distinguishes the parameres as a pair of processes anterolaterally positioned on the ,V'- shaped structure, the remainder of which he terms the 'pseudopenis'. In Microthoracius cameli (Anoplura) the relationship of the 1 parameres1 and 1 pseudopenis' is similar to that of Pedi cuius, but the structures are separated (Fig. 27), whilst in Enderleinellus longiceps (Fig. 28) the pseudopenis is separate from the 'parameres1 and arising at their base, adjacent to the basal apodeme. This configuration, of two lateral 'parameres' and a median 'V' or 'Y' -shaped 'pseudopenis' lying between them, is found in many Anoplura, Trichodectidae, Amblycera and Psocodea. In all cases the 'pseudopenis', if present, extends dorsally over the endophallus opening, and the ' parameres' lie laterally or ventrally. In some cases the 'parameres' may be fused together (Figs 181, 233), or the two arms of the 'pseudopenis' unfused apically, the latter structure then appearing as two' separate sclerites ('internal parameres' of Psocoptera) (Fig. 29 ). This indicates that four structures are present surrounding the opening of the endophallus, which must be homologised with the four phallomeres found in other Neoptera. The ontogeny of Pedi cuius suggests that the 'parameres' of Ferris (1951) are true parameres, and the 'V or 'Y'- shaped pseudopenis is a structure formed from the apically-fused mesomeres. In this study the terms paramere and mesomere will be used, and a new temn 'mesomeral arch' coined for the pseudopenis. The homology is extended throughout the Psocodea. It is noted that Pedicuius is atypical in that the division of the primary phallic lobes to form mesomeres and parameres is delayed and limited in extent. The position of the 'penis' during development leads Piotrowski (1961) to suggest that it is produced from elements of the primary phallic lobes, and his assertion that the mesomeral arch is composed solely of parameres indicates that he believes the penis to be composed of the fused mesomeres (i.e. homologous with the aedeagus of other Phalloneoptera). The penis, however, surrounds the gonopore, not

» Figs 27 - 29. Male genitalia of Psocodea (dorsal aspect). (27) Microthoracius cameli (after Kim & Ludwig, 1978a); (28) Enderleinellus longioeps (after Kim & Ludwig, 1978a); (29) Stenopsocus stigmaticus (after Klier, 1956). In each figure the "basal apodeme is indicated "by 'b1, the mesoemere or mesomeral arch by 'm1, and the paramere by 'p1. 183

the endophallus, and it is clear from the identification of the para- meres and mesomeres (see above) that the latter do not form the 'penis'. The 'penis' is, therefore, neither homologous with the aedeagus, nor formed from the primary phallic lobes; it is a secondary sclerotisation developed from the walls of the endophallus surrounding the gonopore. A variously-formed sclerotised structure surrounds the gonopore in many Anoplura and some Ischnocera, and may lie close to the basal apodeme; this structure may be homologous at least within the Anoplura, and the term penis is retained. In Trichodectidae a penis is never present, though the endophallus around the gonopore may be characteristically modified with specialised spicular areas (as in some Lorisicola (Paradoxuroecus) species). In addition to the penis, further sclerites derived from the endophallus (variously termed 'telomeres', 'hypomeres', ' endomeres' etc.) are present in some Phthiraptera. Like the penis, homology of these sclerites is possible only within limited holophyletic groups in the Psocodea. The sclerites vary considerably in form, and it must be assumed that their functions are similarly variable. Keier (1971) suggests that in the Boopiidae some of the sclerites clasp the sperma- tophore and ensure its correct placement in the female. Spermatophores are not found throughout the Phthiraptera, however,- so this function for the sclerites is not universal for the .Order. From Piotrowski's (1961) figure's it is clear that the posterior part of the genital pouch (posterior to a plane passing through the posterior extreme of the basal apodeme and the bases of the primary phallic lobes) becomes the genital chamber. The remaining, anterior, part of the genital pouch differentiates dorsally into the endophallus and ventrally into the basal apodeme. Piotrowski does not indicate any connections for the posterior of the endophallus and the anterior of the genital chamber, but depicts both as terminating abruptly. The mesomeral arch is shown as being connected only to the basal apodeme. Ferris (1951), however, shows the endophallus in Pedicuius to be fused to the mesomeral arch (plus parameres), and Gdnther (1974) indicates in Reuterella (Psocoptera) the fusion of the endophallus with both mesomeral arch and the dorsal faces of the parameres. Schmutz (1955) demonstrates by sections of Bovicola caorae (Trichodectidae) 184

that the wall of the genital chamber is dorsally continuous with the endophallus. Dissections performed during the course of this study confirm this relationship, and show that the v/all of the genital chamber is continuous with the endophallus both dorsally and laterally. Ventrally the connections are more complex (Figs 25 , 30 ). Examination of sagittal sections of the male genitalia (Schmutz, 1955) and of slide-mounted specimens of Trichodectidae reveal that the original two-layered structure of the basal apodeme is retained posteriorly in the separation of a membranous dorsal layer from a more heavily-sclerotised ventral layer; on microscope slides this separation is discernable as a curved line on the heavily-sclerotised lateral margins of the basal apodeme (Fig. 179)* The ventral v/all of the basal apodeme is continuous with the v/all of the genital chamber and the dorsal v/all with the endophallus (Figs 25, 30 ). This indicates that the anterodorsal and anteroventral elements of the genital pouch do not separate completely from each other and from the posterior' element during development, as is maintained by Piotrowski (1961), but retain the posterior connection found in the imago. The fusion of the endophallus to the ventral and dorsal faces of the mesomeres and parameres respectively was not observed by Piotrowski, so no record has been made of the ontogeny of this development. It is suggested, however, that as the bases of the primary phallic lobes move laterally and posteriorly along the sides of the genital pouch, an invagination forms anterior to the lobes and grows posteriad, the outer (distal) face of which fuses to the inner faces of the lobes. The fusion of the endophallus to the inner faces of the phallomeres, the mechanical impossibility of the latter being inverted into the former, and the continuity of the•endophallus with the wall of the genital chamber and the basal apodeme, indicate that the posterior end of the endophallus must be permanently everted (Fig. 25 ). This situation is superficially similar to that found in Thysanoptera, where the endophallus is at least partially fused to the inner faces of the mesomeres and thus permanently partially everted (see Fig. 15 in Heming, 1970). However, the thysanopteran male genitalia are not precisely similar to the psocodean genitalia in ontogeny. Heming (1970) shows that in Hanlothrips verbasci"(Thysanoptera: Phlaeothripidae) tergura VIII

Fig. 30. Sagittal section of psocodean male genitalia (diagrammatic) 186

the mesomeres unite to form an aedeagus (phallotheca of Heming, 1970) which is sclerotised dorsally, ventrally and laterally, with four longitudinal membranous connections between the sclerites (note that four sclerites are present, not three as stated by Matsuda, 1976). The parameres are produced late in development directly from the putative phallobase. The structure termed an 'aedeagus' by Heming (1970) is precisely analogous to the penis of Pedicuius described above ('pseudovirga' auctt. in other Thysanoptera).. In Frankliniella fusca (Thysanoptera: Thripidae) Heming (1970) finds that the fused mesomeres (termed the 'primitive aedeagus') lie ventrally to the 'phallotheca', and a pair of 'parameres' lie laterally. The 'aedeagus' (pseudovirga) is present as in Haplothrips. An alternative interpretation of the structures present in Frankliniella is proposed here: the 'primitive aedeagus' and 'parameres' present in Frankliniella are equivalent to the four sclerites present in Haplothrips - that is, to the partially membranous aedeagus. The parameres do not develop in Frankliniella. but are represented by the pronounced posterolateral angles of the phallobase. By this interpretation a true aedeagus is formed in all Thysanoptera, but is secondarily reduced and membranous. As discussed above the Condylognatha (the Thysanoptera and their sister-group, the Hemiptera) are generally accepted as forming the sister-group of the Psocodea. In both Thysanoptera and'Hemiptera an aedeagus is formed by fusion of the mesomeres. In the sister-group of the Acercaria (Psocodea plus Condylognatha), the Holometabola, an aedeagus formed by fusion of the mesomeres is also present. The most parsimonious explanation of the distribution of the aedeagus would appear to be that it evolved once, in the common ancestor of the Acercaria and Holometabola, and thus the ancestor of the Psocodea had an aedeagus formed by mesomeral fusion. In Haplothrips, as seen above, the posterior of the.endophallus is permanently everted, with four sclerites fused to its outer surface; this is also true of Psocodea. In Haolothrips, however, the four sclerites are.derived from the fused mesomeres, the permanently everted posterior portion of the endophallus is interpreted as membranous joins between the sclerites (i.e. derived from the mesomeres) and the parameres are present and separate. In Psocodea the mesomeres are seen to fuse only apically, the ' everted 187

endophallus' is not mesomeral in origin and the sclerites are interpreted as both mesomeres and parameres. Four alternatives are available to explain these discrepancies: 1) The interpretations given above of the phallomeres in Psocodea are incorrect. Parameres are absent, and a reduced aedeagus is present, as in Haplothrips. The failure of the mesomeres to unite and the consequent differences in the ontogeny of the aedeagus sire apomorphic for the Psocodea, or represent an aberrant species. Further investigations of the ontogeny of more Psocodea may reveal a pattern similar to other Phalloneoptera. 2) The interpretations given above of the phallomeres in Psocodea are correct. The genitalia of the Psocodea represent a partial reversion with the loss of the aedeagu^ and a novel structure, the permanently-everted endophallus, has been developed. 3) The interpretations given above of the phallomeres in Psocodea are correct. The development of the aedeagus by fusion of the mesomeres is a convergence in Condylognatha and Holometabola. 4) The interpretations given above of the phallomeres in Psocodea are correct. The Psocodea are wrongly placed in the Acercaria, but instead form the sister-group to the Condylognatha plus Holometabola. The autapomorphy of the Phalloneoptera is the development of the endophallus, and the autapomorphy of the Condylognatha plus Holometabola is the development of the aedeagus. Psocodea are convergent with the Condylognatha for a number of loss characters (of limited phylogenetic importance), and the modification of the lacinia, the development of the cibarial pump, and the concentration of the nervous system in the thorax. It is noted that the lacinia and the cibarial pump are functionally distinct in the Psocodea and the Condylognatha. Further evidence on which these alternatives may be judged could be derived from more ontogenetic studies of the Psocodea, and a detailed search for apomorphies to clarify the position of the Condylognatha with respect to the Psocodea and the Holometabola. For the purposes of this study it is assumed that the interpretations of the phallomeres of Psocodea are correct, and the aedeagus does not form in this group. Heming (1970) applies the terms phallotheca and endotheca to the male 188

genitalia of Thysanoptera, and Matsuda (1976) states: "the phallotheca type genitalia [of Snodgrass (1935)] has often occurred in the Psocoptera, Thysanoptera, and Mallophaga". It is apparent from the above discussion, however, that in both Psocodea and Thysanoptera the endophallus is connected directly with the mesomeres or the products of the mesomeres and the phallotheca and endotheca as defined by Snodgrass (see above) do not occur.

Summary: homology and terminology of male genitalia in the Psocodea. Examination of a large number of Phthiraptera and the figures of ^socopteran male genitalia in Smithers (1972) and Gilnther (1974) reveals that throughout the Psocodea the following structures of the male genitalia are almost always present and recognisable: a more-or-less sclerotised basal apodeme supporting caudally a pair of parameres which may be fused ventrally and a pair of mesomeres which may be fused dorsally; both parameres and mesomeres may bear setae apically; fused to the parameres and mesomeres is the permanently-everted portion of the eversible endophallus. These structures are considered homologous throughout the- Psocodea. Further sclerites may be present, (penis, 'hypomeres', 'telomeres' etc.) secondarily developed from the endophallus, and may be homologised only within limited holophyletic groups in the Psocodea.

Male genitalia of Trichodectidae. Both parameres and mesomeres are present in most Trichodectidae, the latter usually being apically fused (Figs 180,205 ), the fused portion frequently being extended posteriad (Fig. 234). The parameres may also be fused, forming a median ventral plate (Fig. 259). The parameres and mesomeres may meet the basal apodeme together (Fig. 180) or separately (Fig. 259 )• The full range of variation in the male genitalia of Trichodectidae is discussed in detail in the 'Character Analysis' section and the 'Taxonomy' section below. During copulation the endophallus everts into the female genital chamber. As described above, the interior of the female genital chamber is roughened and lined with scales. The endophallus is like- wise roughened, being covered with small chitinous spicules (Fig. 31) 189

Figs 31 - 32. Trichodectes canis, male genitalia. (31a) ventral, with endophallus extended, showing detail of ornamentation dorsally and ventrally; (31b, 31c) distribution of spicule and scale types on extended endophallus; (32) dorso-lateral aspect, partially extruded from abdomen. 190

or larger sclerites. The probable function of this adaptation is to maintain a firm connection between the male and female genitalia during copulation. The form of the endophallus and the distribution of spicules and sclerites are species-specific and very variable, possibly functioning as pre-zygotic isolating barriers.

Operation of the male genitalia in Trichodectes melis. The male genitalia in Trichodectes melis comprise a spiculate, fully-eversible endophallus, a pair of half-cylindrical, pointed parameres, and a basal apodeme; the mesomeres are absent (Fig. 179). The ventral longitudinal margin of the paramere is broadly convex, whilst the dorsal margin is virtually straight; when at rest the parameres form a tapering cone with the ventral margins overlapping and the dorsal margins not meeting. The endophallus is permanently everted for about half the length ofthe parameres and fused to them, apart from along the dorsal and ventral midlines where fusion ceases submarginally (Fig. 179)• dorsal gap between the parameres is closed by a tongue-shaped sclerite derived from the everted endophallus (Fig. 32 ).. The width of the endophallus membrane free from the parameres is greater apically than basally, allowing divergence of the apices of the para- meres during copulation whilst controlling the degree of that divergence. Copulation has been observed twice in _T. melis during the course of this study. In both cases the insects adopted a '.back to back' position. This was unexpected, the normal position for lice being with the dorsum of the male adjacent to the venter of the female (Wemeck, 1936; Sikora and Eichler, 1941; Schmutz, 1955). The 'back to back' position was assumed in both cases as the insects came into contact through accidental positioning on adjacent hairs. On coming into contact, the male clasped the female around the thorax- abdomen constriction with his antennae. The apex of the male abdomen protruded slightly posterior to that of the female and curved dorsad. The genitalia were then extruded, the apices of the parameres lying together. The male commenced side-to-side ' searching' movements of the tip of the abdomen, and the articulation between the basal apodeme and the parameres flexed so that the parameres v/ere directed anteriad, parallel to the long axis of the male. Searching movements terminated 191

when the tip of the male abdomen came into contact with the female abdomen. It is likely that this contact is sensed by the long setae on the posterior of sternum IX of the male (Fig. 32). As the searching movements ceased the male abdomen pushed anteriad, directing the genitalia into the female subgenital chamber (the space between the subgenital lobe and the postgenita! sterna). The subgenital lobe of jP. melis is fringed with spines, and the postgenital sterna are spinous; it is likely that these spines would 'catch' and perhaps damage the endophallus were it not protected by the conical 'capsule' of the parameres and dorsal endophallus sclerite. The parameral 'capsule' passed between the subgenital lobe and the postgenital sterna; the parameres then opened and the endophallus everted. The precise placement of the endophallus could not be observed, but some conclusions can be drawn following dissection of specimens of both sexes. The endophallus probably everts to a balloon shape, but this cannot be observed in living specimens; in dead specimens the endophallus collapses but when artificially everted it appears as a crumpled cylinder (Fig. 31)• Even accounting for the distortion caused in artificial eversion, the endophallus appears too long to fit completely in the female genital chamber. Thus, when everted, some membrane must lie between the subgenital lobe and the postgenital sterna. The spines of the postgenital sterna are all directed posteriad, so the endophallus, if it presses down on the spines, should not be damaged. The endophallus is covered in chitinous spiculation, with a row of heavy spines dorsally (Fig. 31). The spicules presumably press against the sides of the female genital chamber, providing an anchor as suggested above, but the presence of the row of spines suggests that a similarly specialised structure should be present in the female for these to act against. The most likely structure is the thickened vulval margin, and pressure of the spines against this would provide an efficient anchor. 'Keler (1938) comes to the same conclusion, but it is unclear whether or not this is based on observation of copulation.

2.2.4.5. The tracheal system and spiracles

A posterior commissure joining the two main abdominal tracheal trunks has been demonstrated in Anoplura, Rhyncophthirina, Boopiidae 192

and Trimenoponidae (Amblycera), Philopteridae (Ischnocera) and "some Trichodectidae" (Harrison, 1915; Ferris, 1931). All species of Trichodectidae examined in this study have been found to possess a posterior commissure, which is consequently assumed to have a universal distribution throughout the family. The presence of the posterior commissure is assumed to be plesiomorphic for the Phthiraptera by Clay (1970). Primitively the Psocodea possess two pairs of thoracic spiracles and eight pairs of abdominal spiracles, the latter on abdominal pleura I to VIII. In all Phthiraptera the metathoracic spiracle has been lost, but the mesothoracic spiracle retained. In the Liposcelidae and Phthiraptera abdominal spiracles I and II have been suppressed, so that the first abdominal spiracle is on pleurum III. Clay (1954), however, notes that in Amblycera the characteristic post-spiracular setal complex found in that suborder is present on pleurum II, though there is no indication of a spiracle, and Harrison (1915) notes the presence of a !spiracular scar1 on pleurum II in Philopterus (Ischnocera). Keler (1938) records a single specimen of Felicola subrostratus (Trichodectidae) with rudimentary spiracles on pleurum II, not connected with the tracheal system. Bedford's (1932a) statement that the spiracles may be present on either segments three to eight or two to seven is based on apparent segment numbers, however, and does not reflect a morphological variation in spiracular placement. Many mammal-infesting lice have undergone reduction in the numbers of abdominal spiracles. In all cases except one, this loss has apparently taken place sequentially from the posterior, starting at segment VIII. Thus, within the Trichodectidae, there are species with six, five, four, three, two, one and no abdominal spiracles (see Table VI ), hut in all cases where spiracles are present the first pair is on segment III. If more than one pair of spiracles is present there are no intercalating segments in which spiracles are absent between segments with spiracles. Outside the Trichodectidae the variation and degree of spiracular loss is not so extensive. Trimenoponidae and Gliricolinae (Amblycera) have five pairs of spiracles (on segments III to VII) whilst Eulinognathus (Anoplura) has three pairs (on III to V) or two pairs (on III to IV). The sole exception 193

TAXA NUMBER OF PAIRS

OF SPIRACLES

Geomydoecus, Neotrichodectes. Fellcola (F.) helogale. 0

F. (F.) hopkinsl, F. (S.) fahrenholzi, F. (S.) guinlel.

Lorlslcola (L.) oalayslanus, L. (P.) paralatlceps - mungos clade.

L. (P.) acuticeps - neoafrlcanus clade

Fellcola (S.) bedfordl 1

Fellcola (F.) subrostratus ex Eupleres. 2

-Trlchodectes (S.) potus

Fellcola (F.) all species except helogale & hopkinsl. 3

p. (fi.Vall species except bedfordl. fahrenholzi & gulnlel.

Trlchodectes (S.) fallax. T. IS.) octomaculatus,

T. (S.) PQtus $

Lorlslcola (P.) bengalensls - Juccil clade. L. (P.) aspldorhynchus. U

L. (P.) sumatrensls. L. (P.I lenicornls. L. (P.) wernecki

Trichodectes (n. 5) 5

Trlchodectes (T.). T. (S.) all species except fallax - potus 6

clade. Werneckodectes. Genus n. h, Lutrldia, Protellcola.

Lorlsicola (L.) all species except malayslanus. Dasyonyglnae.

Eutrlchophlllnae. Bovlcollnae

Table VI. Distribution of number of pairs of abdominal spiracles in the Trichodectidae. 194

to the sequential anteriad loss is the anopluran Neolinognathus. in which the single pair of abdominal spiracles is on segment VIII. No other Phthiraptera are known to have undergone spiracular loss. Though spiracular loss in the Trichodectidae is apparently ordered and sequential, there is no evidence that spiracles have necessarily been lost one at a time. Whilst in Procaviphilus (Meganarionoides) angolensis, JP. (M.) colobi and J?. (M.) baculatus the posterior pair of spiracles only is very small and apparently in the process of being lost, in Lorisicola (10 hercynianus and JL. (L.) siamensis the posterior two pairs are extremely small, probably non- functional, and apparently in the process of being lost. In a number of clades, sister-groups exhibit multiple discontinuities in spiracle number. The sister-species Lorisicola (L.) mfjobergi and L. (L. ) malaysianus have six and zero pairs respectively; Felicola viverriculae and an undescribed sister-species have three and zero pairs respectively; the Lorisicola (]?.) lenicornis-wernecki clade and the sister L. (]?.) acuticeps-neoafricanus clade have four and zero pairs respectively. Variation within species can occur, though it is generally erratic. Felicola subrostratus normally has three pairs of abdominal spiracles; the species is widespread and found on many hosts, but on Madagascar, where the host is Eupleres goudoti, there may be three or two pairs, and specimens exhibiting asymmetry are present in the'British Museum (Natural History). Asymmetry has also been noted in Trichodectes (_S.) emeryi, one paratype of which has six spiracles on one side of the abdomen and five on the other. Trichodectes (£>.) potus is unusual in that the female has three pairs of abdominal spiracles and the male only two, the only known example in the lice of sexual dimorphism in spiracle number. Inspection of Table VI reveals that most Trichodectidae have either six, three or no pairs of abdominal spiracles, other numbers being less common. From the cladogram it can be seen that reduction to five, four and one pair has occurred once, reduction to two and three pairs has occurred twice, and reduction to none has taken place eight times. It is plain that the loss of abdominal spiracles cannot be used a priori as a taxonomic character defining (holophyletic) genera, as proposed by Bwing (1936), but equally apparent that the character 195

is not necessarily as variable as suggested by Kdler (1938) and Hopkins (1941), who treat it as a character of specific value only. The selective advantage of this reduction is not known, but it is possible that it is an adaptation to exclude dust from the tracheal system, or to reduce water loss. Webb (1946) regards the latter factor as of limited importance in permanent ectoparasites which can, he argues, obtain water either directly from the blood of the host or indirectly, as a product of breakdown of food material. The presence in non blood- feeding lice of a very efficient mechanism for water uptake from the atmosphere (Williams, 1971) goes some v/ay to refuting Webb1 s contention, but by the same token reduces the likelihood of water loss being a major selective pressure leading to the loss of spiracles. Seduction of water-loss through the spiracles is generally accomplished by means of a spiracular closing mechanism (Richards and Davies, 1977); such a structure is present in Phthiraptera (Harrison, 1915; Webb, 1946). The spiracle opens directly into an enlarged atrium, which may be spherical or elongate. The walls of this atrium are produced into a complex network of cuticular outgrowths, which may reduce water loss (Richards and Davies, 1977) or prevent the entry of dust particles (webb, 1946). Webb (1946) describes the structure of the atrium for a number of lice, and concludes that the size of the atrium and the complexity of the internal outgrowths are directly related to the 'dustiness' of the host. He suggests that birds are less dusty than mammals, and that among mammals the thicker the fur, the lower the dust content of the environment of the lice'. He finds that bird-infesting lice have small simple atria to the spiracles, whilst among the Anoplura the atrial size and the complexity of the internal network increase together with the tendency of the host to have an open coat or to hairlessness. There is considerable variation in size of atrium amongst the Trichodectidae the largest and most complex being that of the thoracic spiracle in the Lorisicola (]?.) p ar al at i c eo s -mungo s clade. No correlation has been attempted with the coat type of the host, however. Ischnocera, Rhyncophthirina and Anoplura all possess a 'spiracular gland', which secretes a sticky fluid onto the outgrowths of the atrial wall (Webb, 1946). Webb (1946) suggests that this is a further adaptation to collect dust from the incoming air.

V 196

2.3. CHARACTER SURVEY OF TRICHODECTIDAE

2.3.1. Introduction

The procedures followed during the character survey and the coding of character states have been discussed in section I.4.3.. The characters selected as potentially of value in analysis of relationships within the Tricliodectidae are listed below, and the distribution of their states is presented in the data matrix in Appendix A. Not all characters recorded were found to convey phyletic information, and the determination of polarity of transformation series is discussed in section 2.4.1. below. The results of the determination of polarity are included in the following list, and those apomorphic character states that are employed in the cladistic analysis are indicated in the list by an asterisk (*). The list of characters is presented in tabular form to show the coding applied. The eight columns of the table are arranged as follows:. Column 1. Number of character Column 2. Character Column 3« Character state Column 4. Coding for phenetic analysis Column 5. Coding for Wagner tree analysis (if applicable) Column 6. Coding for cladistic analysis (if applicable) Column 7. 'Gain1 (g) or 'Loss' (1) characteristic for cladistic analysis (if applicable). On the cladogram gain states are signified by •, loss states by •, and the plesiomorphic states of each, if required, by • and o respectively. Column 8. Phyletic w'eighting category (after Hecht & Edwards) (if applicable) 2.3.2. List of Characters

1 2 3 4 5 6 7 8

1 Basal apodeme symmetric 0 0 0 asymmetric 1 1 1 g IV 2 Posterior of basal apodeme lateral symmetric in vertical plane 0 0 0 struts *asymmetric in vertical plane (Pigs. 207, 215) 1 1 1 g IV 3 Anterior end of basal apodeme sclerotised 0 unsclerotised 1 4 Anterior end of basal apodeme flat or broadly convex (Pigs 213, 259) 0 0 0 0 0 concave (Pig. 106) 1 1 0 0 1 *deeply concave (Pig. 66 ) 2 1 1 0 2 g IV * acuminate (Pig. 187) 3 0 0 1 1' g IV 5 Approach of lateral struts to parallel 0 anterior of basal apodeme convergent 1 divergent 2

6 Posterior expansion of slight or none 0 0 0 basal apodeme lateral struts great (Pig. 183 ) 1 1 1 g IV 7 Posterior bifurcation of basal absent 0 0 0 apoderne lateral struts *present (Pig. 1$7) 1 1 1 g V 8 Posterior of basal apodeme lacking lateral extension 0 0 0 lateral struts 247) *\vith lateral extension to mesomeres (Pig. 1 1 1 g V 1 2 3 4 5 6 7 8

9 Anteposterior spur of basal apoderne absent 0 0 0

cr lateral struts *present (Pig. 97) 1 1 1 0 V 10 Shape of basal apodeme not long and 'waisted' 0 0 0 1 cr *very long, with median 'waist (Pig. 136) 1 1 1 0 IV 11 Posterior of lateral struts of not modified as below 0 0 0 basal apodeme *broad' and obtuse in meeting parameres

IT (Pig.205) 1 1 1 0 IV 12 Posterior of lateral struts of not modified as below 0 0 0 basal apodeme *sharply inturned and convex (Pig. 213) 1 1 1 s IV 13 Posterior of lateral struts of not modified as below 0 0 0 *incurved to parameres (Pigs. 157, 15^) cr basal apodeme 1 1 1 0 V 14 Basiparameral sclerites absent 0 0 0 0 0 *present, fused to parameres 1 0 1 0 1 g IV *present, not fused to parameres 1 1 1 1 2 g IV 15 Basiparameral sclerites separate 0 0 0 fused to each other 1 1 1 g 16 Parameres do not meet mesomeral arch 0 do meet mesomeral arch 1 17 Parameres do not meet basal apodeme 0 do meet basal apodeme 1

vo 00 1 2 3 4 5 6 7 8

18 Paramere length extend to reach basal apodeme 0 0 0 0 do not reach basal apodeme 1 1 0 1 extend anteriorly between lateral struts of basal apodeme 2 0 1 1' 19 Paramere fusion not fused to mesomeral arch 0 0 0 0 *fused to mesomeral arch in part (Pigs 74,85) 1 1 0 1 g IV *completely fused to mesomeres (Pig. 84) 2 1 1 2 or IV . 0 20 Paramere fusion not fused to basal apodeme 0 0 0 0 *fused exteriorly to basal apodeme (Pig.85) 1 1 0 1 0or IV *fused medially to completely to basal apodeme (Pigs 158, 185,186) 2 0 1 1' g IV 21 Paramere fusion not fused together, or as described below 0 0 0 0 ^(faintly fused together 1 1 0 1 g II (clearly fused together 2 1 0 1 g II *not fused but closely associated, with line of division apparent (Pigs 234, 235) 3 0 1 1* g V 22 Parameral apices unfused 0 0 0 *fused 1 1 1 g II 23 Paramere shape not as described below 0 0 0

2 cr *fused to shield-shaped plate (Pigs 247, 5.1) 1 1 1 0 V 24 Paramere shape not as described below 0 0 0 *fused as described in character 23, with antero-median projection (Pig.248) 1 1 1 g IV 1 2 3 4 5 6 7 8

25 Parameral plate shape not as described below 0 0 0 •"•produced apically into incurving points (Pig. 181) 1 1 1 g V 26 Paramere symmetry symmetric 0 0 0 asymmetric 1 1 1 g IV 27 Parameral orientation similar 0 0 0 *oriented at right-angles to each other (Pig. 180) 1 1 1 g V 28 Paramere shape more or less broad, thick 0 0 0 *very thin, deflected asymmetrically (Pig. 186) 1 1 1 g IV 29 Paramere shape more or less broad 0 0 0 "narrow rods (Pigs. 208, 209, 211) cr 1 1 1 0 IV 30 Paramere shape not as described below 0 0 0 *very broad, lanceolate, scoop-shaped (Pig. 185) 1 1 1 cr IV 31 Median internal projection of absent 0 0 0 parameres •"•present 1 1 1 g V 32 Paramere size large or moderate 0 0 0 *small discs (Pig. 99) 1 1 1 1 II 33 Paramere shape not as described below 0 0 0 •cylindrical (Pigs. 179, 212) 1 1 1 g V 1 2 3 4 5 6 7 8

34 Paramere shape not as described below 0 0 0 *basally very narrow, medially broad cr (Fig.205) 1 1 1 0 V 35 Paramere shape not as described below 0 0 0 *with characteristi calfy-different iat ed base and blade (Fig.65 ) 1 1 1 g V 36 Base of parameres not as described below 0 0 0 0 0 cr *broad, club-like (Fig. 203a) 1 1 0 0 1 0 V cr *block-like (Fig. 214) 2 0 1 0 1« 0 V cr *cuboid (Fig. 66 ) 3 0 0 1 1" 0 V 37 Base of parameres lacking flange 0 0 0 *with flange (Fig. 131) 1 1 1 s 38 Paramere and mesomere shape not as described below 0 0 0 *of characteristic asymmetric form (Fig. 258) 1 1 1 g V Reduction of parameres and 39 not reduced as below 0 0 0 0 mesomeres *characteristically reduced (Fig. 66 ) 1 1 0 1 1 II *characteristically greatly reduced (Fig. 67) 2 1 1 2 1 II 40 Mesomeres present 0 0 0 *absent 1 1 1 1 I 1 2 3 4 5 6 7 8

41 Mesomeres do not meet basal apodeme 0 meet basal apodeme 1 42 Mesomeres do not meet basiparameral sclerites 0 meet basiparameral sclerites 1 43 Mesomeral position reach lateral struts of basal apodeme 0 0 0 0 fail to reach lateral struts of basal apodeme 1 1 0 1 IV *extend mesad of lateral struts of basal apodeme 2 0 1 1» g IV 44 Mesomere fusion apically fused to form arch 0 0 0 *not apically fused 1 1 1 1 II 45 Lateral desclerotisations absent 0 0 0 of mesomeral arch *present (Pig. 131) 1 1 1 g IV 46 Lateral flexions of mesomeral absent 0 0 0 arch *present (Pigs 121, 122,236^ 1 1 1 oST IV 47 Me Someral arch not as described below 0 0 0 *modified into tripartite arch (Pig. 105) 1 1 1 g IV 48 Mesomeres basally, between not modified as below 0 0 0 lateral struts of basal apodeme *sharply directed posteriad (Pig.234) 1 1 1 g IV 49 Mesomeral arch medially smooth or with projection 0 0 0

*with two nipples (Pig.232) • 1 1 1 g V 1 2 3 4 5 6 7 8

50 Diameter of mesomeral arch less than half the length of the permanently everted endophallus 0 0 0 0 *more than half the length of the permanently everted endophallus 1 1 0 1 g IV *as great as the length of the permanently everted endophallus 2 1 1 2 g IV 51 Shape of mesomeral arch not modified as below 0 0 0 *with median anteriad deflection (Pigs 203a,203b) 1 1 1 g V 52 Shape of mesomeral arch circular or elliptical 0 0 0 • "•rectangular (Pig.205) 1 1 1 g V 53 Shape of mesomeral arch not modified as below 0 0 0 1 *extension lost, arch 'looped (Pig.257) 1 1 1 g V 54 Shape of mesorneral arch not modified as below 0 0 0 *widely circular (Pig.100) 1 1 1 g V 55 Shape of mesomeral arch not modified as below 0 0 0 *forming pentagonal arch, convex distally (Pig. 76) 1 1 1 g V 56 Shape of mesomeral arch smoothly curved 0 0 0 0 *sharply inturned to parameres (Pig. 95) 1 1 0 1 g IV •"•extending anteriad to posterior end of basal apodeme and sharply recurved

(Pigs. 96 , 230) 2 1 1 2 g IV 1 2 3 4 5 6 7 8

57 Shape of mesomeres not as described below 0 0 0 •very slender, string-like (Fig. 66) 1 1 1 g V 58 Median extension of mesomeral present 0 0 0 arch •absent 1 1 1 1 II 59 Apex of mesomeral arch extension not bifurcate 0 0 0 •bifurcate 1 1 1 g V 60 Shape of mesomeral arch extension not modified as below 0 0 0 •broadly expanded, lanceolate 1 1 1 g V 61 Tongue-like sclerite between absent 0 0 0 paromeres •present 1 1 1 g V 62 Everted part of endophallus not sclerotised 0 0 0 •sclerotised 1 1 1 g V 63 Endophallus spiculation not as described below 0 0 0 •'V'-shaped rods, numerous 1 1 1 g V 64 Endophallus spiculation not as described below 0 0 0 •comprising large hoolc-like spines (Fig. 121) 1 1 1 g V 65 Endophallus spiculation not concentrated about gonopore 0 0 0 0 •concentrated about gonopore 1 1 0 1 g V •with dense V-shaped patch about gonopore 2 1 1 2 g V 66 Endophallus spiculation not as described below 0 0 0 •including median row of hook-like scales 1 1 1 g V 1 2 3. 4 5 6 7 8

67 Endophallus spiculation not as described below 0 0 0 *dense and refringent in part 1 1 1 g V 68 Gonapophysis setae present 0 0 0 *absent 1 1 1 1 I 69 Gonapophysis setae lacking sclerotised tubercles 0 0 0 *some having sclerotised basal tubercles (Pig. 163) 1 1 1 g V 70 Gonapophysis setal tubercles absent, or present and not characteristically fused 0 0 0 *present and characteristically fused (Pig. 125) 1' 1 1 g V 71 Gonapophysis setal tubercles absent, or present and in characteristic pattern (Pig.166 ) 0 0 0 *present and modified from characteristic pattern by loss of apical non-tuberculate seta (Pig. 159) 1 1 1 1 I 72 Gonapophysis shape not as described below 0 0 0 *spoon-shaped (Pig. 168) 1 1 1 g 73 Gonapophysis shape not as described below 0 0 0 0 *hoolc-shaped (Pig. 88 ) 1 1 0 1 g V *hook-shaped with spur (Pig. 89 ) 2 0 1 1» g V 74 Gonapophysis shape not explanate dorsally 0 explanate dorsally 1 1 2 3 4 5 6 7 8

75 Gonapophysis shape not explanate ventrally (other than as discrete lobe) 0 0 0 *thinly explanate ventrally (other than as discrete lobe) 1 1 1 g V 76 Gonapophysis shape with explanations, if present, thick 0 with explanations, if present, thin 1 77 Gonapophysis lobe absent 0 0 0 *present 1 1 1 g V 78 Apex of gonapophysis acute 0 0 0 obtuse 1 0 0 *not projecting beyond lobe (if present) 2 1 1 1 IT 79 Gonapophysis lobe not very thick 0 0 0 *very thick (Fig. 245) 1 1 1 g V 80 Gonapophysis lobe absent, or not as described below 0 round, flat (Figs 188, 221) 1 81 Gonapophysis lobe absent, or not as described below 0 0 0 *broad, formed of fused tubercles (Fig.243a) 1 1 1 g V 82 Gonapophysis lobe absent, or not as described below 0 0 0 *narrow, rectangular, formed of 2 or 3 .190) cr fused tubercles (Fig 1 1 1 0 V 83 Gonapophysis lobe negative for character 81 or, if positive, flattened 0 0 0 0 *positive for 81 and folded anteriorly 1 1 0 1 g IV *positive for 81 and greatly folded 2 1 1 2 g IV 1 2 3 4 5 6 7 8

84 Gonapophysis lobe absent, or not as described below 0 0 0 *produced into spines posteriorly (Pig. 221) 1 1 1 g V 85 fronapophysis junction with ventral acute (Pig. 138) 0 0 0 vulval margin *smoothly continuous (Pig.168) 1 1 1 g 86 Ventral vulval margin % unsclerotised 0 0 0 *sclerotised 1 1 1 g 87 Ventral vulval margin not shorter than length of gonapophyses 0 0 0 *shorter than length of gonapophyses (Pig. 116) 1 1 1 g V 88 Ventral vulval margin shape straight or concave, with chord less than 90° to median, longitudinal axis (Pig. 167) 0 arched or flat, with chord 90° to median longitudinal axis (Pig. 102) 1 89 Ventral vulval margin smooth 0 serrate or toothed 1 90 Ventral vulval margin lacking setal tubercles 0 0 0 *with setal tubercles (Pig. 168) 1 1 1 g V 91 Ventral vulval margin lacking non-tuberculate setae 0 with non-tuberculate setae 1 92 Ventral vulval margin with setae (if present) not limited to sides 0 with setae (if present) limited to sides 1 93 Ventral vulval margin with submarginal setal row 0 lacking submarginal setal row 1 1 2 3 4 5 6 7 8

94 Ventral vulval margin not expanded 0 0 0 •expanded (Figs 109, 138) 1 1 1 g V 95 Ventral vulval margin not as described below 0 0 0 •greatly produced and rounded (Fig. 127) 1 1 1 g V 96 Ventral vulval margin not as described below 0 0 0 •with median narrow projection (Fig. 124) 1 1 1 g V 97 Subgenital lobe absent 0 0 0 •present 1 1 1 g V 98 Subgenital lobe (if present) smoothly continuous with ventral vulval margin 0 arising anterior to margin 1 99 Subgenital lobe margin not serrate (Fig. 166) 0 0 0 0 serrate (Fig.242) 1 1 0 1 g •very serrate (Fig. 188) 2 1 1 2 g V 100 Subgenital lobe margin with serrations (if present), apical only 0 with serrations (if present), apical and lateral 1 101 Subgenital lobe surface smooth dorsally and ventrally 0 0 0 0 0 •with overlapping pointed scales (Fig.219) 1 1 0 0 1 g V •with small spines (Fig. 220) 2 1 1 0 2 g V •with many overlapping spines (Fig. 218) 3 1 1 1 3 g V 1 2 3 4 5 6 7 8

102 Subgenital lobe , basally lacking lateral processes 0 0 0 *with lateral processes 1 1 1 g V 103 Subgenital lobe processes not as described below 0 0 0 (ch. 102) *thinly sclerotised and directe.d posteriorly CP 1 1 1 0 V 104 Subgenital lobe processes not as described below 0 0 0 (ch. 102) *membranous and serrate (Figs 169, 170) 1 1 1 g V 105 Subgenital lobe not bifurcate (Fig. 21$ 0 0 0 *bifurcate (Fig.220) 1 1 1 g 106 Subgenit al lobe dimensions as long as broad 0 not as long as broad 1 longer than broad 2 107 Subgenital lobe dimensions bifurcations (ch. 105) as long as broad 0 bifurcations not as long as broad 1 bifurcations longer than broad 2 108 Subgenital lobe bifurcations not as described below 0 0 0 (ch. 105) *rectangular, widely separate (Fig. 189) 1 1 1 g V 109 Subgenital lobe lacking internal sclerite 0 with internal sclerite 1 110 Subgenital lobe lacking marginal setal patch 0 with marginal setal patch 0 111 Subgenital lobe lacking submarginal setal patch 0 0 0 *with lateral submarginal setal patch 1 1 1 g V 1 2 3 4 5 6 7 8

112 Subgenital lobe lacking setae (other than as in ch. 110,11l) 0 with setae (other than as in ch. 110, 111) 1 113 Sub-vulval area lacking pointed scales 0 0 0 "with pointed scales 1 1 1 e V 114 Dorsal vulval face not spinous 0 0 0 *spinous 1 1 1 s V 115 Post—vulval sclerites present 0 0 0 "absent 1 1 1 1 I 116 Post—vulval sclerites broad, triangular 0 0 0 *long, narrow, oriented parallel to longitudinal axis 1 1 1 s 117 Common oviduct not as described below 0 0 0 "with folded 'collar1 at branching-point 1 1 1 s V 118 Female genital chamber lined with overlapping scales 0 0 0 *lined with sclerotised nodules, some fused 1 1 1 s V

119 Female genital chamber not as described below 0 0 0 "with median dorsal longitudinal fold 1 1 1 s V 120 Female genital chamber not as described below 0 0 0 "with median dorsal area clear of scales 1 1 1 1 II 121 Female genital chamber not as described below 0 0 0 "with median anterior dorsal area clear, thinly sclerotised 1 1 1 e V 1 2 3 4 5 6 7 8

122 Female genital chamber not as described below 0 0 *with spines on dorsal face 1 1 1 g V 123 Female sternum VII lacking processes 0 0 0 *with two long posterior spikes (Fig. 8l) 1 1 1 g V 124 Reproduction sexual 0 0 0 *Parthenogenetic 1 1 1 g V 125 Male tergite VIII absent 0 present, anterior element only- 1 present, posterior element only- 2 present, anterior and posterior elements 3 126 Male tergite VIII (if present) small 0 large (equivalent to tergites II-VIl) 1 127 Male tergite VIII (if present) lacking median division 0 0 0 with median longitudinal division 1 1 1 g II 128 Male tergite VIII (if present) with posterior element absent or, if present, not fused to tergite IX 0 0 0 *with posterior element fused to tergite IX 1 1 1 129 Male segment IX positioned posteriorly- 0 0 0 positioned postero-dorsally or dorsally 1 1 1 g III 130 Male segment IX not overly projecting 0 projecting posteriorly 1 projecting dorsally 2 1 2 3 4 5 6 7 8 « 131 Male genital chamber opening posteriorly or postero-dorsally 0 opening dorsally 1 132 Male segment IX as long as broad 0 not as long as broad 1 longer than broad 2 133 Male tergite IX absent 0 present 1 134 Male tergite IX (if present) small, covering under % total area 0 large, covering at least \ total area 1 135 Male tergite IX (if present) lacking median.division 0 with median longitudinal division 1 136 Posterior margins of male not as described below 0 0 0 tergum IX *greatly expanded (Pig. 93) 1 1 1 8 V 137 Male lateral sternopleural absent 0 sclerite present 1 138 Male lateral stemopleural not fused to tergite IX 0 sclerite (if present) fused to tergite IX 1 139 Male lateral sternopleural not fused to subgenital plate 0 sclerite (if present) fused to subgenital plate 1 140 Male post-genital sclerite absent 0 present 1 141 Male post-genital sclerite lacking median division 0 (if present) with median longitudinal division 1 1 2 3 4 5 6 7 8

142 Male post-genital sclerite not fused to lateral sternopleural sclerite 0 (if present) fused to lateral sternopleural sclerite 1 143 Male post-genital sclerite not fused to sternal sclerite 0 (if present) fused to sternal sclerite 1 144 Male sternal sclerite absent 0 present 1 145 Male sternal sclerite lacking median division 0 (if present) with median longitudinal division 1 146 Male sternal sclerite not fused to subgenital plate 0 (if present) fused to subgenital plate 1 147 Male sternal sclerite not fused to lateral sternopleural sclerite 0 (if present) fused to lateral sternopleural sclerite 1 148 Male segment IX without two longitudinal strengthening sclerites 0 0 0 *with two longitudinal sclerites 1 1 1 g V 149 Pseudostyli absent 0 0 0 "•present 1 1 1 g V 150 Pseudostyli (if present) not as described below 0 0 0 *broad, rounded, long (Fig. 79 ) 1 1 1 g V 151 Pseudostyli (if present) rounded apically 0 0 0 *angular, pointed apically (Fig. 92 ) 1 1 1 g V 1 2 3 4 5 6 7 8

152 Male segment IX lacking single apical projection 0 0 0 •with single apical projection (? fused cr pseudostyli) 1 1 1 o V 153 Male subgenital plate absent 0 present 1 154 Male sternite VII absent 0 present 1 155 Male sternite VIII absent 0 present 1 156 Male sternite VIII not as described below 0 0 0 •characteristically enlarged (Fig.175) 1 1 1 g V 157 Male sternite VIII (if present) not convex posteriorly 0 0 0 •characteristically convex posteriorly (Fig. 229) 1 1 1 g V 158 Male sternite IX absent 0 present 1 159 Male subgenital plate rods absent 0 (s.g.p.r.) present 1 160 Male s.g.p.r. (if present) broad 0 •normal1 1 narrow 2 ro 161 Male sternite VII (if present) not fused to s.g.p.r. 0 -to fused to s.g.p.r. 1 1 2 3. 4 5 6 7 8

162 Male sternite VIII (if present) not fused to s.g.p.r. 0 fused to s.g.p.r. 1 163 Male sternite IX (if present) not fused to s.g.p.r. 0 fused to s.g.p.r. 1 1 164- Male perisetal gap 'normal 0 0 0 0 small 1 1 0 1 1 II *absent 2 1 1 2 1 II 165 Male sternum VI not as described below 0 0 0 *with anterior and posterior sclerites 1 1 1 g V 166 Female flagellomeres unfused (three flagellomeres) 2 0 0 0 fused to form two flagellomeres 1 1 0 1 g *fused to form one flagellomere 0 1 1 2 g 167 Male flagellomeres unfused 0 0 0 •fused 1 1 1 g 168 Male scape greatly expanded 2 0 0 0 slightly expanded 1 1 0 1 1 I •not expanded 0 1 1 2 1 I 169 Male scape setae randomly scattered on posterior face 0 0 0 •in longitudinal row on posterior face 1 1 1 g V 170 Male scape seta! row (if present) numbering at least three setae on posterior 0 0 0 i face •reduced to two setae on posterior face 1 1 1 1 II 1 2 3 4 5 6 7 8

171 Male flagellum *lacking apical 'teeth' 0 1 1 0 0 0 1» 1 I. *with single apical 'tooth' 1 1 0 0 0 0 1 1 I (*)with two apical 'teeth' 2 0 0 0 0 0 0 *with three apical 'teeth' 3 0 0 1 0 0 1'' g IV 1 *with eight apical 'teeth 4 0 0 0 1 0 1"' g IV *with four apical 'teeth' 5 0 0 0 0 1 1"" g IV 172 Male flagellar » teeth1 not on protuberance 0 0 0 *on protuberance 1 1 1 g V 173 PA ale flagellar »teeth 1 articulated basally 0 0 0 *fused to flagellum 1 1 1 g V 174 P/lale flagellum lacking basal projection 0 0 0 0 0 0 with simple basal projection 1 1 0 0 0 1 g V 1 *with basal projection of linked 'teeth 3 0 0 1 0 1' g V *with broad, rough basal projection 2 0 1 1 0 2' g V with other form of basal projection 4 0 0 0 1 1" ocr V 175 PA ale flagellum not as described below 0 0 0 *with simple median and basal projections

only 1 1 1 ocr 176 Inner margin of male flagellum smooth 0 0 0 0 rough 1 1 0 1 g V *serrate 2 1 1 2 g V 177 Inner margin of male pedicel smooth 0 rough 1 1 2 3 4 5 6 7 8

178 Male scape lacking apical projection 0 0 0 *with apical projection 1 1 1 g V 179 Female pedicel lacking projection 0 0 0 "with membranous projection 1 1 1 g V 180 Male flagellum not very long 0 0 0 *very long 1 1 1 g 181 Flagellar sensilla placodea not in pit 0 in pit 1 182 Flagellar sensilla placodea and not in pit 0 0 0 coeloconica *in pit with marginal tongue-like processes 1 1 1 g V 183 Sitophore sclerite not as described below 0 0 0 *with posterior arms extended (Fig. 36) 1 1 1 g V 184 Posterior temple angles lacking projections 0 0 0 0

IT *with projections (Fig. 139) 1 1 0 1 O V *with very long projections (Fig. 149) 2 1 1 2 g V 185 Posterior temple margins not as described below 0 0 0

IT •"produced and convex 1 1 1 O V 186 Osculum not deep 0 0 0

i *deep and with characteristic anterior marginal convexity (Fig. 82 ) 1 1 1 g IV

187 Pretarsus bearing two claws 0 - 0 0 ""bearing one claw 1 1 1 1 II 1 2 .3 4 5 6 7 8

188 Pretarsal claws lacking ventral spines 0 ' 0 0 0 0 *with blunt ventral spines 1 1 1 0 1 g V or • *with sharp ventral spines 2 1 0 1 r» o V 189 Post-coxale of leg three not fused to abdominal pleurum II 0 0 0 *fused to abdominal pleurum II, at least in female 1 1 1 g V 190 Post-coxale of leg three fused to .abdominal pleurum II in female only 0 fused to abdominal pleurum II in both sexes 1 191 Post-coxales of leg three not fused medially 0 0 0 g fused medially 1 1 1 192 Sternum II not as described below 0 0 0 *with sclerotised apophysis articulated to pleurum II 1 1 1 8 V 193 Atrium of thoracic spiracle spherical 0 0 0 *tubular 1 1 1 g V 194 Abdominal spiracles all of similar size 0 0 0 0 *spiracle VIII of male very small 1 1 0 1 1 II *spiracle VII and VIII very small in both sexes 2 0 1 1» 1 II 195 Abdominal spiracle VIII present 0 0 0 *absent 1 1 1 1 I 1 2 3 4 5 6 7 8

196 Abdominal spiracle VII present 0 0 0 •absent 1 1 1 1 I 197 Abdominal spiracle VI present in female 0 0 0 •absent in female 1 1 1 1 I 198 Abdominal spiracle V present 0 0 0 absent 1 1 1 1 I 199 Abdominal spiracle IV present 0 0 0 absent 1 1 1 1 I 200 Abdominal spiracle III present 0 0 0 •absent 1 1 1 1 I 201 Setae of posterior setal row of not stout 0 0 0 0 pleurum II stouter than setae of p.s.r. of PV-VII, though not very stout 1 1 0 1 g very stout, at least ventrally 2 1 1 2 g IV 202 Setae of posterior setal row of not stout 0 0 0 0 pleurum III (stouter than setae of p.s.r. of PV-VII, *ithough not very stout 1 1 0 1 g

(very stout, at least ventrally 2 1 1 2 vLcr> IV 203 Setae of posterior setal row of not stout 0 0 0 0 pleurum IV stouter than setae of p.s.r. of PV-VII, though not very stout 1 1 0 1 g very stout, at least ventrally 2 1 1 2 g IV 1 2 3 4 5 6 7 8

204 Posterior setal row of present 0 0 0 pleura IV-VI absent 1 1 1 1 I 205 Pleurum VIII lacking tuft of very long setae 0 0 0 rr •with tuft of very long setae (Figs 102, 103) 1 1 1 CJ V 206 Abdominal tergal setae •generally very long, obscuring p.l.s. (if present ) (Fig. 171) 0 1 0 1 g V of medium length (Fig..117) 1 0 0 0 very short (Fig.202) 2 0 1 1» 1 V 207 Male tergum II lacking specialised setae as described below 0 0 0 •with long stout setae not found on other

terga (Figs 192, 202) 1 1 1 g V 208 Male tergum III lacking specialised setae as described below 0 0 0 •with isolated pair of long median setae,

longer than those of tergum IV (Fig.200) 1 1 1 g III 209 Male terga II-IV lacking specialised setae as described below 0 0 0 •each with single pair of long median setae

(Fig.193) 1 1 1 g III 210 Female terga I-VIII without specialised setae as described below 0 0 0 •each with single pair of long stout median setae 1 1 1 g III 1 2 3 4 5 6 7 8

211 Female terga I-IV with median setae 0 0 0 *lacking median setae 1 1 1 1 I 212 Male terga II and III lacking specialised setae as described below 0 0 0 *with group of characteristically specialised long setae (Fig.19&) 1 1 1 g V 213 Female terga I-VII without specialised setae as described below 0 0 0 *each with single pair of median setae 1 1 1 g III 214 Male terga II-III without specialised setae as described below 0 0 0 *with long setae arranged in straight rows of four or more(Figs 177,253) 1 1 1 g V 215 Male tergum II without specialised setae as described below 0 0 0 *with long setae arranged in curved rows (Fig. 54) 1 1 1 g V 216 Postero-lateral setae absent 0 0 0 possibly present, but not clearly distinguishable from lateral setal group or unrecognisable 1 0 0 *present on terga II-VT 2 1 1 g V 217 Postero-lateral setae numbering one per site 0 0 0 (if present) numbering two or more per site 1 1 1 g 1 2 3 4 5 6 7 8

218 Male terga II-VI with tv/o median setae the same size as other setae of median group 0 0 0 "with two median setae appreciably smaller cr than other setae of median group 1 1 1 e> V

219 Setae of abdominal sterna II-IV 2iot as described below 0 0 0 rr •short, stout (Fig. 150) 1 1 1 o V 220 Female abdominal setae not as described below 0 0 0

or *very long, fine 1 1 1 o V 221 Setae of male tergum II not arising from modified sclerite 0 0 0 •arising from very long, medially-divided sclerite 1 1 1 g V 222 Abdominal setal bases not enlarged 0 0 0 •enlarged 1 1 1 g V 223 Abdominal pleurum II not modified as described below 0 0 0 0 ST •extending narrowly onto sternum II 1 1 0 1 O V •extending broadly onto sternum II 2 1 1 1« g V 224 Dorsal projection of pleururn II absent . 0 0 0 0 •present, unsclerotised 1 1 0 1 g V •present, sclerotised 2 1 1 2 g V 225 Dorsal projection of pleurum III absent 0 0 0 0 present, unsclerotised 1 1 0 1 g V present, sclerotised 2 1 1 2 g V 1 2 3 ' 4 5 6 7 8

226 Ventral projection of pleurum III absent 0 0 0 present, sclerotised 1 1 1 g V 227 Dorsal projection of pleurum IV absent 0 0 0 0 present, unsclerotised 1 1 0 1 orr V (present, sclerotised 2 1 1 2 g V 228 Ventral projection of pleururn IV absent 0 0 0 0 ^(present, unsclerotised 1 1 0 1 g V (present, sclerotised 2 1 1 2 g V 229 Projections of pleurum IV riot as described below 0 0 0 (if present) *very long (Pig. 150) 1 1 1 s . V 230 Male terga VI-VIII (or VI, if not as described below 0 0 0 .VII-VIII absent) *with anterior sclerite longitudinally divided medially 1 1 1 g V 231 Male terga VI-VII not as described below 0 0 0 *with posterior sclerite longitudinally divided medially 1 1 1 g V 232 Lateral flecks of male abdomen absent 0 0 0 *present 1 1 1 g V 233 Female tergite I absent 0 present 1

f\3 ro OJ 1 2 3 4 5 6 7 8

234 Female tergite II absent 0 present 1 235 Female tergite III absent 0 present 1 236 Female tergite IV absent 0 present 1 237 Female tergite V absent 0 present 1 238 Female tergite VT absent 0 present 1 239 Female tergite VII absent 0 present 1 240 Female tergite •VII I absent 0 present 1 241 Male tergite I absent 0 present 1 242 Male tergum II with anterior sclerite absent 0 with anterior sclerite present 1 243 Male tergum II with posterior sclerite absent 0 With posterior sclerite present 1 244 Male tergum III with anterior sclerite absent 0

v/ith anterior sclerite present 1 1 2 3 4 5 6 7 8

245 Male -tergum III with posterior sclerite absent 0 with posterior sclerite present 1 246 Male tergum IV with anterior sclerite absent 0 with anterior sclerite present 1 247 Male tergum IV with posterior sclerite absent 0 v/ith posterior sclerite present 1 248 Male tergum V with anterior sclerite absent 0 with anterior sclerite present 1 249 Male tergum V v/ith posterior sclerite absent u with posterior sclerite present 1 250 Male tergum VI v/ith anterior sclerite absent 0 v/ith anterior sclerite present 1 251 Male tergum VI v/ith posterior sclerite absent 0 v/ith posterior sclerite present 1 252 Male tergum VII with anterior sclerite absent 0 v/ith anterior sclerite present 1 253 Male tergum VII v/ith posterior sclerite absent 0 with posterior sclerite present 1 254 Male tergum VIII v/ith anterior sclerite absent 0 with anterior sclerite present 1 255 Male tergum VIII with posterior sclerite absent 0 with posterior sclerite present 1 1 2 3 4 5 6 7 8

256 Male and female pleurite II absent 0 present 1 257 Female pleurite III absent 0 present 1 258 Female pleurite IV absent 0 present 1 259 Female pleurite V absent 0 present 1 260 Female pleurite VI absent 0 present 1 261 Female pleurite VII absent 0 present 1 262 Female pleurite VIII absent 0 present 1 263 Male pleurite III absent 0 present 1 264 Male pleurite IV absent 0 present 1 265 Male pleurite V absent 0 present 1 266 Male pleurite VI absent 0

present 1 1 2 3 4 5 6 7 8

267 Male pleurite VII absent 0 present 1 268 Male pleurite VTII absent 0 present 1 269 Female stemite II absent 0 present 1 270 Female stemite III absent 0 present 1 271 Female stemite IV absent 0 present 1 272 Female at emit e V absent 0 present 1 273 Female sternite VI absent 0 present 1 274 Female sternite VII absent 0 present 1 275 Male sternite II absent 0 present 1 276 Male sternite III absent 0 present 1 277 Male sternite IV absent 0 present 1 1 2 3 4 5 6 7 8 278 Male sternite v absent 0

present 1 •

279 Male stcrnite vi absent 0 present 1 229

2.4. CHARACTER ANALYSIS

2.4.1. Introduction

As indicated in section 1, the characters are analysed using a cladistic methodology. The character analysis is performed in two interconnected parts, the determination of character polarity and the construction of the cladogram (described in sections 1.4.2. and 1.3.2. respectively). These are linked through the process of 'reciprocal illumination' already mentioned, and thus, although they are discussed individually in the two sections below, there is some interaction between the two processes which will be manifested in discussion. The .following two sections are intended to explain the reasons for the polarity assigned to the characters listed in section 2.3., and for those groups developed in the cladogram where characters are apparently arranged non-parsimoniously. To increase clarity and conciseness the distributions of characters and character states discussed below are related to taxa and clades developed in the analysis. Clades are referred to by the names of the taxa (species or genera) on the extreme left and right of the clade as depicted on the cladogram, reading from left to right. The cladogram is included below on pages 254 to 262 (Figs 37 to 45) •

2.4.2. Identification of Apomorohic Stales for Phyletic Analysis

In many cases an apomorphic character stale is identified as such by its distinct complexity and very limited distribution, and, to avoid pointless repetition in the following discussion, such instances are not examined individually. The characters are considered under the following headings: 1. Male genitalia, (characters I-67) 2. Female genitalia (characters 68-123) 3. Reproduction (che.ra.cter 124) 4. Male terminal abdominal segments (characters 125-165) 5 . Ant ennae (ch aract er3 16 S -18 2) 6. Head (characters 183-136) 7. Legs (characters 187-188) 230

8. Postcoxale (characters 189-191) 9. Spiracles (characters 193-200) 10. Abdominal setae (characters 201-222) 11. Abdominal pleural projections and modifications (characters 223-229) 12. Abdominal sclerae (characters 192, 230-279)

1. Male genitalia (characters 1-67) Most species of Psocodea have symmetric male genitalia and this stale is consequently assumed to be plesiomorphic for the superorder. In a few species- of Trichodectidae the genitalia are asymmetric, but differences in the form of the asymmetry in different species (Pigs . 180, 186, 258) suggest that several independent autapomorphies have beer- developed. In some cases a characteristic asymmetry is limited to a single species and is therefore of no relevance to phyletic analysis, but the asymmetries described in characters 2, 27, 28 and 38 are more widely distributed and are all employed. The distribution of other apomorphies indicates that character 2, the vertical. asymmetric deflection of the lateral struts of the basal apodeme (1b.a.l.s.1), has been developed twice, once in Pelicola (S.) be'dfordi (Pig.215), ^^ once in the common ancestor of P. (P.) cynictis and p. (P.) setosus (Pig. 207). In most Psocodea the basal apodeme is not' fused to the parameres, but in a few Trichodectidae this fusion, considered to be apomorphic, has taken place (character 20). In the Damalinia (D.) theileri - h-arrisoni clade the posterior ends of the b.a,.l.s. are broad and fused exteriorly to the parameres (Pig. 85 ); this fusion (character 20:1) is unlike that found elsewhere in the family and is believed on that account to be a.ut apomorphic. In the Trichodectes (S>.) mantis - must el ae cla.de, the TM (_S.) fall ax - ootus clade, Genus n. 4 lutrae and Lutridia exilis the parameres are also fused to the b.a.l.s. (character 20:1'),but there are no features in the fusion pattern to indicate whether the apomorphy is homologous or convergent in the four cla.des. There are a number of possible sequences of gains and reversals. The fusion may have talien place three or four tines with no reversals; once, to be lost at least four times; or twice, to be lost at least twice. The genitalia of Lutridia spp. differ from those found in the sister-group 231

(the Trichodectes - Genus n. 4 clade), being more similar to those of Protelicola. Comparison of L. exilis and L. matschiei (Figs . 157, 158) indicates fusion of the parameres and b.a.l.s. in the former species to be associated with the virtual detachment of the basal fused portion of the t>arameres, a unique feature. For this reason the fusion in L. exilis is deemed to be autapomorphic. The Trichodectes (S_.) mart is - must elan clade and the T^. (_S.) fall ax - ootus cla.de a.re placed by character 209 in a trichotomy with _T. (_S.) emeryi, a species in v/hich the parameres are not fused to the b.a.l.s.. If the fusion is homologous in Trichodectes (Stachiella) and Genus n. 4, then it must have been lost in Y7emeckode ct e s, Trichodectes (Trichodectes) and _T. (Subgenus n. 5), and T, (£>.) emeryi. Parsimony suggests that fusion was developed independently in Genus n. 4 and the common ancestor of the two cla.des in Trichodectes (Stachiella) that possess the character, these latter being united as sister-groups. Fusion of the parameres to the b.a.l.s. is thus postulated to have taken place four times in the Trichodectidae, three of those times in the Trichodectes - Lutridia clade.

In most Psocodea the parameres are not fused together, but such- fusion is present, probably apomorphically, in a number of species of Trichodectidae (character 21:1). In some species parameral fusion is difficult to observe, as the portion of the permanent ^-everted endo- phallus lying between the parsmeres is faintly sclerotised, giving the impression that the parameres are fused together; fusion has probably developed in some cases through sclerotisation of the endophallus. The distribution of other apomorphies suggests that parameral fusion exhibits more homoplasy than any other apomorphy in the analysis, being derived 24 times and lost once. In the lorisicola (P.) bengalensis - ,juccii clade the parsmeres are closely associated with each other but a.re not fused (Figs -234? 235) > although fusion has sometimes been assumed ( ,e.g-. •Yemeck, 1948). This proximity is believed to be autapomorphic for the clade (character 21:1*). The form of the fused parameres (parameral plate) may be apomorphic for groups of species (characters 23-25). In a few Trichodectidae the parsmeres and mesomeres are fused, a probable apomorphy (character 19). The distribution of other apornorphies and slight differences in the fusion pattern (Figs 74, 84,:85» 95 ,236) 232

indicate some homoplasy in the character. As described in section 2.2., the mesomeres are frequently fused apically in the Psocodea, -and consequently this fusion, when found in the Trichodectidae, is deemed to be plesiomorphic. Loss of fusion (i.e. reduction to two unfusea oesomeres) is therefore believed to be apomorphic within the Trichodectidae (character 44), and distribution relative to other apomorphies indicates that it has occurred several times in the family (see cladogram). In most species of the Procaviohilus - Eur;/1richodectes clade there is a lateral desclerotisation on each side of the mesomeral arch (Pigs 131,134 145) j restricted distribution of this character within the family indicates its aoomorphy (character 45). The rnesomeral arch may also have lateral flexions. (Figs 121, 122, 236), which can give the arch the appearance of being broken (see V/emeck, 1348). This modification is found in species of the Lorisicola (]?•) lenicornis - neoa.fricanus clade and of Procavicola (Condylocephalus), which on the basis of other apomorphies are widely separated on the cladogram; the lateral flexion of the mesomeral arch (character"46) is consequently believed to be a convergent apomorphy in the two clades named. In most species of the genus Eutrichoohilus the mesomeral arch is divided into three parts by total desclerotisations laterally (Figs .105, 106). This feature is unique in the Phthiraptera and therefore considered apomorphic (character 47). In many Trichodectidae a rod-like sclerite terminating posteriorly in a 'Yf shape or a broad plate is present longitudinally between the b.a.l.s. (Figs 71, 56,161). Although this sclerite (the 'central sclerotisation') is very clear in some species, it is poorly sclerotised in others, and may be either absent or obscured by the sclerotisations of the endophallus in mounted specimens. The uncertainty attached to the observation of this structure has precluded its use in analysis, although it may be of value taxonomically. The central sclerotisation is probably a derivat ive Oi tne oas al aoodeme, developed for muscle attachment. A sclerite or pair of sclerites is present anteriorly to the parameres in some Trichodectidae, and are referred to here as the 'basicaraneral sclerites' (Figs 121, 236). Their presence is believed to be apomorphic 233

because of restricted distribution (character 14:1 + 2). The basi- parsmeral sclerites may be fused to the parameres (pig.236) or to each other (Fig. 105) • They are probably derived from the basal ventral flanges of the parameres found apomorphically (character 37) in a number of Trichodectidae (Figs 131, 134). If they are formed (as suggested here) by deta,chment of the flange from the main body of the paramere (Fig. 33), fusion of the sclerites and the parameres is a stage in the transformation series to the development of free sclerites, but fusion of the basiparamerai sclerites to each other (character 15 ) is a 'terminal' apomorphic state (although it is not used in the cladistic analysis because of its extremely limited distribution). The anterior end of the basal apodeme may be heavily or lightly sclerotised, or apparently not sclerotised at all (character 3). This degree of sclerotisation is very susceptible to modification during preparation of the specimen, and thus is difficult to assess accurately. The character is not used in analysis. The most prominent features of the basal apodeme are the lateral struts (b.a.l.s.), which are generally fairly heavily sclerotised. These struts may approach the .anterior of the basal apodeme in parallel, convergently or divergently (character 5); this character is not used in cladistic analysis because of the difficulty in assigning polarity to the different forms, but is useful taxonomically, and can help in the determination of the form of the anterior margin of the basal apodeme. This anterior margin nay be straight or broadly convex, shallowly concave, very deeply concave (Fig.66 ) or acuminate (Fig. 187) (character 4). Of these forms the last two are almost certainly apomorphic within the Trichodectidae; the elongation of the apodeme and the concomitant parallel-sided concavity is found only in the Bovicola alpinus - tibialis clade (character 4:2), end the acuminate form is found only in three species of Trichodectes, although in fact the apomorphy imparts little useful information for the-construction of the cladogram. The polarity of the character for the other three states is difficult to assess, and they are therefore not used in phyletic analysis. In some species of Damalinia (Tricho 1 ineu.rus) the b.a.l.s. develop a lateral spur before the junction with the parameres, probably at the point at which the dorsal and ventral layers of the basal apodeme 234

33

.anterior setae

posterior setal row All if \])i uv IIM

mi Uv ///; V/-' 34 -7 postero-lateral 1 lateral t v \ seta gap median median setal lateral setal gap group group

Fig- 33. Fostulated evolution of basi-parameral sclerite by detatchment of basal flange of paramere. Fig. 34. Abdominal setal arrangement of Trichodectidae , illustrated by anterior terga and pleura of male. Fig. 35• Sitophore sclerite of Damalinia neotheileri. Fig. 36. Sitophore sclerite of Bovicola hemitragi, showing apomorphic extension of posterior arms. 235

separate; this feature (the * snteposterior spur1, Fig. 97) is not found elsewhere in the Trichodectidae and is considered apomorphic (character 9). The posterior ends of the b.a.l.s. are most frequently not, or only slightly, expanded laterally, but in some Trichodectidae they are greatly broadened and scoop-shaped. This very broad, form (character 6), whilst believed to be apomorphic, is not used as an apomorphy in phyletic analysis because of difficulties in delimiting the state. The posterior forking of the b.a.l.s. (Fig. 187-) is also considered apomorphic (character 7), but was probably developed twice in Tricho&ectes (see cladogram). The basal apodeme probably extends anteriorljr as far as segment VI in the plesiomorphic state, but in a few Trichodectidae it extends up to segment II; in some cases this lengthening has been accompanied.' by a width restriction or * waisting1 medially, and this is believed to be apomorphic (character 10). Other features of the basal apod.eme are, by virtue of their restricted distribution, believed to be apomorphic and are used in phyletic analysis (characters 8, 11, 12, 13). 'Whilst it is not possible to be certain of the plesiomorphic form of the parameres in the Trichodectidae, it is assumed that this is fairly unspecialised., and that the forms of the parameres found in groups of species that are also linked' by other apomorphies are apomorphic (characters 29-36). As noted above, apical fusion of the mesomeres is plesiomorphic for the Trichodectid.s.8, so loss of apical fusion, reduction in siae and complete loss of the mesomeres -are all considered to be aoonorphic within the family (characters 44, 39, 40). The presence of a median longi- tudinal extension to the mesomeral arch is also considered plesiomorphic, as it is present in a number of taxa outside the Trichodectidae. Loss of this extension, or modification of its form from •simple lanceolate1 (Fig. 131), are considered apomorpliic within the family (characters 58-60). In the plesiomorphic state the mesomeres articulate basally with the basal apodeme; articulation of the basal apodeme with amy other part of the mesomeres is considered opomorphic (character 43). Mesad extension of the rnesorneres between the b.a.l.s. (cha.ra.cter 43:1') has apparently arisen twice, once in the Dasyonygina.e, and once in the Lorisicola, and in each case providing .an autaponorphy for the cla.de 236

named. In Lorisicola (£.) bengalensis and philippinensis the parts of the mesomeres between the b.a.l.s, are apomorphic ally defladted posteriad (Pig. 234) (character 48). A similar recurving of the mesomeres occurs elsewhere in the Trichodectidae (character 56), but in this case the mesomeres are exterior to the b.a.l.s., and their recurved portions lie ventrally to the b.a.l.s.. This apomorphy is seen as a transformation series of states in Damalinia (Tricholipeurus) (character 56:1 - 56:2), the most apomorphic of which (56:2) is also exhibited by Lorisicola m.jobergi, although in this case the recurved parts of the mesomeres are very difficult to see (Pig. 230). Other modifications to the mesomeres believed to be apomorphic are present in restricted groups within the Trichodectidae (characters 49-57). The endophallus may be sclerotised in a number of apomorphic ways within the Trichodectidae (characters 62-87).

2. Pemale genitalia (characters 68 - 123) The gonapophyses of most Trichodectidae and many other Psocodea bear at least some setae, which arise directly from the structure and not from tubercles. Absence of setae (character 68) and development of sclerotised setal tubercles (character 69) are therefore both believed to be apomorphic within the Trichodectidae. Setal tubercles are found in Protelicola, Procaviphilus and the Trichodectes - Genus n. 4 clade (T-n4), but as a sister-group relationship between the latter two is not supported by other apomorphies, and the form of the tubercles differs between the two clades (Pigs. 125, 170), the character is probably convergent The relationship between Protelicola and the T-n4 clade is discussed in section 2.4.3. The characteristic pattern taken by the tubercles in each clade is modified by loss (character 71) or fusion (character 70); in both cases these are believed to be apomorphic modifications because of their limited distribution. Tuberculate setae are also found on the ventral vulval margin of most species in Trichodectes (Pig. 170); concordance with other apomorphies suggests the apomorphy of this character (character 90). The plesiomorphic form of the gonapophyses is not certain, but some forms, because of their very restricted distribution, are believed to be apomorphic (characters 72, 73, 75); some convergence in character 73 is indicated by the distribution of other apomorphies. The 237

development of a lobe on the ventral margin of the gonapophysis is restricted to the Trichodectidae and, for this reason, is believed to be apomorphic within the clade (character 77). Distribution of other apomorphies suggests that the lobe developed independently in several different clades, sometimes taking only one form in a clade (character 84), sometimes being apomorphically modified (characters 79, 81, 82, 83). The reduction of the 'spur' - the portion of the gonapophysis distal to the lobe - is considered apomorphic, as it is confined to two- small groups of taxa within the family (character 78). '.'/here present, the gonapophyses in most Phthiraptera meet the ventral vulval margin at an angle (Figs 109, 188), but in some Trichodectidae they meet in a smooth curve (Fig. 167), which may-be sclerotised (characters 85, 86). The ventral vulval margin may extend in a more or less smooth curve between the gonapophyses, as is most frequently the case in lice with gonapophyses, or it may be produced in some manner (characters 94-97). Each of these projections is considered apomorphic because of limited distribution, though some are homoplastically developed in different clades of the family. The distinction between the subgenital lobe (character 97) and the expansion of the ventral vulval margin (character 94, and its apomorphic derivative, character 95), may hot be immediately clear, but whilst the former term is supplied to structures that arise abruptly from the margin, the latter is a more extensive posteriad production of the whole of the margin. Both of these apomorphies occur more than once in the Trichodectidae. The form of the subgenital lobe is variable, though frequently it is. marginally serrate, sometimes with the serrations greatly developed (character 99:2). Several other apomorphies, of restricted distribution 'within the Trichodectidae, are found in the form of the subgenital lobe (characters 101, 102, 103, 104, 108, 111). Two probable apomorphies, the presence of sn internal sclerite in the subgenital lobe (character 109) and the presence of lateral setal patch

(character 110)%, are not used in the cladisiic analysis. The internal sclerite is not readily observable, but this is probably due to the sclerite being rendered undetectable during preparation of the specimens (especially the smaller species), end this likelihood precludes its use. The distribution of the sclerite, where detected, suggests it to be 238

plesiomorphic v/ithin the Trichodectinae,: possibly linked to the development of the subgenital lobe in this clade. The lateral marginal setae appear to share oart of the distribution of the lateral processes of the lobe (character 102),being absent in a few species only. Character 110 is probably closely associated with character 102, and for this reason it is not used. The plesiomorphic form of the genital chamber in the Trichodectida.e is not known, but observations on other Phthiraptera suggest light sclerotisation with a few internal spicules. The development, in some restricted groups of Trichodectidae, of particular patterns of spicules, scales, spines and broad sclerotised areas (characters 118-122) is considered apomorphic.

3. Reproduction (character 124) Parthenogenetic reproduction (character 124)- occurs in a few Trichodectidae, mostly in the Bovicolinae, but also in the species Geomydoecus scleritus. As all other Phthiraptera reproduce bisexually, the character is taken as apomorphic. Parthenogenesis appears to have developed at least four times in the Trichodectidae.

4. Male terminal abdominal segments (characters 125-165) The sclerites of the terminal segments of the male trichodectid abdomen are very variable in presence or absence states, extent, and degree of subdivision. This variability is not, in many cases, readily associated with transformation series of other characters to which polarity has been apolied, -and the plesiomorphic state (ana hence apornorphic states) of the characters of these sclerites is in most cases not known. In a few instances the sclerites are distinctly modified in a restricted group of species, and thus polarity can be assigned (characters 128, 148, 156, 157, 164, 165). Segment IX and the genital opening are apomorphically positioned more or less dorsally in many species of Trichodectidae, as discussed.in section 2.2. above (characters 129, 130, 131). Distortion due to nreuarative processes obscures the position in many of the specimens examined, however, and these characters cannot be used with ars/ confidence. 239

In some species of Trichodectidae the posterior margin of tergum IX is greatly expanded to produce a double convex lobe (character 136); this feature is believed to be apomorphic, though the distribution of other apomorphies indicates that it developed twice. The presence of pseudostyli, discussed in detail in section 2.2. above, is believed to be apomorphic for the Trichodectidae (character 149). The plesiomorphic form of the pseudostyli is not known, as the extant forms are very variable and cannot, in most cases, be resolved into transformation series. In two cases (characters 150, 151), the pseudostyli are of very distinct form and restricted to groups of species believed to be holo- phyletic on other grounds; these character states are believed to be apomorphic. The presence of a single projection posteriorly on segment IX in some species of Trichodectidae, believed to be formed of fused pseudostyli, is also considered apomorphic because of the limited groun of species in which it occurs (character 152).

5. Antennae (characters 166-182) In all Trichodectidae the male flagellomeres are fused together (character 167), a state found elsewhere only in the anopluran families Echinophthiriidae and Hamophthiriidae and therefore considered apomorphic for the Trichodectidae. Fusion of the flagellomeres has also occurred in some female Trichodectidae (character 166), but the distribution of other apomorphies suggests four homoplastic derivations of this apomorphy in the family. The expansion of the scape in the male to house the enlarged musculature is probably plesiomorphic for the Trichodectidae, as similar expansion is found in many Ischnocera and Anoplura. Reduction of this expansion is, however, apomorphic within the family (character 168:1 and 2), and is believed to have taken place three times. The flagellum of most male Trichodectidae bears a number of setae modified into broad pointed 'teeth* (Figs 14a, 238), a feature not found in the same form in amy other Phthiraptera, and therefore considered apomorphic for the family. The plesiomorphic number of 'teeth* is almost certainly two, as this number is the most common in all groups of Trichodectidae; any variation from this number (to zero, one, three, four or eight) is believed, therefore, to be apomorphic (character 171). The loss of the basal 240

articulation of the 'teeth1 (character 173) and. the development of a supporting protuberance (character 172) are both apomorphic. In order that the male antennae should clasp the female with maximum efficiency the 'inner1' (posterior) surface of some or all of the antennal segments may be roughened or bear projections; such developments are considered apomorphic in each form (characters 174-178). The presence of a membranous projection on the female antenna (character 179) is also believed to be apomorphic. In most Phthiraptera the setae of the scape are scattered over its surface in no coherent pattern, and this is true of some Trichodectidae (Pig. 90); in most Trichodectidae, however, the setae of the dorsoposterior surface of the scape are apomorphically arranged in a line along the segment (Pig.14a) (character 169)• In some cases where a row of setae might be expected from the construction of the cladogram, the number of the setae involved in the putative row is only two, and this is believed to represent an apomorphic reduction in number (character 174). The sensilla of the antennae in Trichodectidae and other Phthiraptera have been discussed above; the presence of a fringed pit surrounding the sensilla of the flagellum in two species of Trichodectidae is unique and believed to be apomorphic (character.182).

6. Head (characters 183-186) Although the sitophore sclerite is variable in most phthiraptera (Haub, 1973), it is comparatively uniform in the Trichodectidae. The form found in most Trichodectidae (Pig. 35 ) is believed on this account to be plesiomorphic, and is departed from in Bovicola (Subgen. n. 1), Dasyonyx and Eurytrichodectes, where the posterior arms are extended (Pig.36 ) (character 183). The distribution of other apomorphies indicates that the modification is convergent in Bovicola (Subgen. n. l) and Dasyonyx plus Surytrichodectes. The posterior margins of the temple are generally broadly rounded in Trichodectidae, but in species of the genus Butrichoohilus the convexity is much greater than in the rest of the family (Pig.101); this development is believed to be apomorphic (chara„cter 185). In the three species of the genus "Turytrichodectes (only two of which are described) and the four of Procavicola (Condyloceohalus) the posterior 241

temple angles are developed into pointed projections (character 184), these "being very long in the former genus (character 184:2). This modification is not found elsewhere in the Trichodectidae,although small rounded projections are found in some Dasyonyx spp. and some Damalinia spp.. The presence of pointed projections is believed, on the basis of the distribution of other characters, to be homoplastic in the two genera- mentioned. The form of the osculum has been largely excluded from consideration in the cladistic analysis because of the direct influence of the hair of the host (see above). However, in the Damalinia "(D.) theileri - baxi clade it is quite different from other species of Trichodectidae, (Fig. 82), and is here suggested to be apomorphic (character 186).

7. Legs (characters 187, 188) The loss of one tarsal clew on each leg (character 187) is an apomorphy associated with ectoparasitism on mammals. This character is proposed as an autapomorphy of the Trichodectidae, although it may be- aut apomorphic for a postulated holophyletic group comprising the Anoplura, Rhyncophthirina and Trichodectidae (see section 2.1. for a discussion of the position of the Trichodectidae relative to the two former groups). A number of Psocodea have teeth on the 'inside' face of the tarsal claws, and many Trichodectidae have what appears to be a single basal tooth (Fig.i8b). The occurrence of teeth all along the 'inside' face of the tarsal claws (Figs l8a, 18b) is restricted in the Trichodectidae to Dasyonyx, and is believed to be autapomorphic for the genus (character 185). The two subgenera of Dasyonyx have tarsal claw teeth of different forms: D. (Dasyonyx) have sharp slender teeth (Fig.l8a), whilst I). (ITeodasyonyx) have blunter, broader teeth (Fig. 18b) . These two forms may be co-apomorphies, indicating that the subgenera are sister-groups, or one may be the plesiomorphic state. Ho other characters have been found within Dasyonyx that indicate reliable sister-group relationships within the genus. In this study the two tooth forms are accepted as co-apomorphies and the subgenera retained, but further work on the genus may cause this hypothesis to be challenged. 242

8. Postcoxale (characters 189 - 191) In most species of Trichodectidae the metathoracic postcoxale is either not sclerotised or sclerotised and small, "but the polarity of the transformation series with the extreme states 'sclerotised' and 'not sclerotised' is not known. In Procaviohilus (Iieganarionoides) and some species of Dasyonyx (Dasyonyx)'the postcoxale is greatly enlarged and heavily sclerotised, an apomorphic state not included in the analysis for reasons given below. A further apomorphic condition, the fusion of the postcoxales (character 191), is also found in some members of the same subgenera. The presence of the sclerotised postcoxale is difficult to determine in some of the smaller species of Dasyonyx. but in any case the distribution of other apomorphies indicates convergence of the postcoxale characters in the tv/o subgenera. Neither aoomorphy is used in the analysis.

In

9. Spiracles (characters 193 - 200) For the Phthiraptera (and the Trichodectidae) the plesiomorphic number of spiracles is a single'thoracic pair and six abdominal pairs; further reduction in the number of abdominal spiracles is apomorphic. The numbers of abdominal spiracles in the different species of Tricho- dectidae are summarised in Table VI. Each reduction is considered to be an apomorphy (characters 195-200), though some homoplasy has occurred. Inspection of the distribution of other apomorphies indicates that reduction to five, four and one pair of spiracles has occurred once, reduction to three and two pairs twice, and reduction to none eight times. Because of the sequential pattern of spiracle loss, apomorohy 196 is always associated on the cladogram with apomorphy 195, 197 with 196 and 195, and 200 with 197, 196 and 195.

» 243

In most species of lice each of the abdominal spiracles has atria of roughly the same size; as ha.s been pointed out in earlier discussion, however, some Trichodectidae have posterior spiracles with atria much smaller than those more anterior on the abdomen (character 194). This difference in size is believed to be an apomorphic reduction. In most species of lice the atrium of the thoracic spiracle is as broad of broader than long; species of the genus Cebidicola, however, have a tubular atrium associated with the thoracic spiracle (character 193). This modification of form is believed to be apomorphic.

10. Abdominal setae (characters 201 - 222) It is inferred from the study of abdominal setal patterns throughout the Psocodea that the plesiomorphic pattern is a row of setae running around the abdomen on each of segments I to VIII. On trichodectid abdominal pleura II - VII this row of setae (referred to here as the posterior setal row or 'p.s.r.1) is generally clear, and is absent in only a few species. The distribution of other apomorphies indicates this absence to be apomorphic (character 204). In some Trichodectidae the p.s.r. of pleura II, III and IV comprises setae that are much stouter than those of other pleura (characters 201, 202 and 203 respectively), and the distribution of other apomorphies suggests the apomorpliic.-status of each of these, although each exhibits some homoplasy. Preliminary analysis indicates that characters 201 and 203 (specialisation of the n.s.r. on pleura II and IV) convey little phylogenetic information, and only character 202 (specialisation of the p.s.r. on pleurum III) is used in the final cladistic analysis. The setae on pleura VTIT and IX are frequently longer than the setae of the p.s.r. on anterior pleura.; the extreme length of these setae in Eutrichoohilus is, however, recognised as apomorphic (character 205). The setal row on sterna II and III is usually similar to the row on other sterna, but in Eurytrichodectes the setae of these two rows are short, stout and conical (Fig. 150). The limited distribution of this character indicates its apomorphic status (character 219). The tergal setal row of many Trichodectidae, especially males, is clearly composed of four discrete groups - two lateral and two median - each separated by a. gap (Fig. 34). The positioning of the groups and 244

the number of setae in them are useful taxonomic characters, and may be utilised as landmarks for the identification of particular setae. The groups are, however, difficult to use ir_ phyletic analysis because of the difficulty of assigning polarity to any transformation series. In some Trichodectidae (and in no other Phthiraptera) a seta. - termed here the 'posterolateral seta' or 'p.l.s.' - is present posterolaterally on each side of terga. II - VI (Fig. 34 ) • I*16 restriction of distribution of this seta within the Phthiraptera suggests that its presence is apomorphic (character 216)-. In some cases there is more than one p.l.s. on each side of the tergum (Fig.218); this is believed to be an apomorphy (character 217), but its sporadic occurrence (in terms of clades indicated by other apomorphies) has led to its omission from the cladistic analysis. The presence of the p»l.s. is difficult to assess in some species, either because the lateral group ms.y be reduced in number or because the lateral group is composed of very long setae. In the former case, a single seta in the position of the p.l.s. may be this seta (the lateral group being absent), or it may be the sole remaining seta of the lateral group (the p.l.s. being absent) (Fig. 175) • -n latter case (most Tricho- dectes species and the new subfamily ), the most lateral seta of the lateral group frequently lies slightly posterior to the rest of the row (Fig. 171) and a more differentiated p.l.s. is absent. In both these cases the p.l.s. is postulated to be present, though modified. The distribution of other characters suggests, however, that secondary loss of the p.l.s. has occurred within some taxa. Some setae of the median tergal group, particularly in males, may be specialised. In males of ITeotrichodectes sop. the two central setae of the united median groups, or two setae very near the centre (perhaps separated by one or two unmodified setae) are very much smaller than the other setae of the row (Fig.241). The restriction of the distribution of these 'tergocentral microsetae' suggests that their development is apomorphic (character 218). In Felicola (3.) oygidialis and F. (_S.) mac rums the median group of tergum III is modified in a distinctive manner in the males (Fig. 158) and this modification is assumed to be a.pomorphic for the two species (character 212). Some or all of the setae of the median group on terga II and III of male Trichodectida.e may 245

be enlarged relative to the other tergal setae. This enlargement occurs sporadically both within and outside the family, .and each case is believed to be autaponorphic• 7/ithin the Trichoaectidae, in males of leomydoecus (Thomomydoecus), . (_7 .) concl, Trichodectes (subgenus n. 5) ovalis and ugandensis, and the undescribed sister-subspecies of Trichodectes (T_.) galictidis, the setae of the median row on both terga II and III are enlarged but remain in a straight row (Tigs 177» 253); this arrangement is believed to be apomorphic, but probably convergent in each of the four groups (character 214). In Bovicola (sub.gen n. 1) hemitragi and multispinosa a similar enlargement is confined to some of the setae of tergum II, and the lines are curved (Pig. 54) (character 215). In males of Felicola, the holophyly of which is supported by several apomorphies, the median setal group is reduced to a single, greatly enlarged seta (Fig. 202). The apomorphic status of this character (character 207) is indicated by its restricted distribution .and correlation with other apomorphies. It is notable that the seta.e are single, but of normal size (very small) in the Felicola. (F.) rahmi-viverriculae clade (Fig. 195)» ^^ secondarily increased in number to six in Felicola (_S.) bedfordi and F. (F.) setosus (Figs 192, 20l). On the basis of other apomorphies, the former is believed to be a single autapomorphic reversal, whilst the latter is believed to be a convergent gain. The sclerite from which the pair of setae arise may be long and of characteristic shape (Fig. 194); this feature is found only in conjunction v/ith the enlarged setae (character 207) and is postulated to be apomorphic (character 221). In many males of Felicola the median setal group on terga III - VII is also reduced to a single seta, although in most cases this does not approach the size of the seta of tergum II. In the Felicola (£>.) cooleyi - quadraticens clade this reduction has taken place on terga III - VII, but the setae are similar in length to those of tergum II, the latter being reduced relative to those of other species of Felicola and the former enlarged (Fig.199) (character 208). This aponorohy is convergent on the apomorphic setal pattern of the males of Trichodectes (Stachiella), (character 209), although in this case the setae are all genemllTr long and stout. The median set si grouo of the female tergumi 246

may also be reduced to a single seta (character 210) or lost (character 211). The distribution of these female apomorphies is as follows: The reduction of the median group to a single seta is found only in the •Trichodectes (_3.) fall.ax - o c t o m acul at a s clade; the sister-species, T_.(_3.) pctus, end the sister-group to this clade, the T^. (j>.) mart is - must el ae clade, lack the female median.; group entirely. The sister- species to the whole _T. (J3.) mart is - notus clade, T_.(_S.) emeryi, has the median group unreduced, numbering three setae, on terga I and II, reduced to one seta or absent on tergum III, and absent on terga TV - VIII. It is not certain whether setal loss in the female has taken place only once, the setae being regained in the fall a;: - octomaculatus clade, or-has talc en place independently three times (in emeryi, the martis - mustel ae clade, and ootus). The length of the abdominal setae is difficult to employ in phyletic analysis because of the problem of establishing the polarity of the transformation series 'very short - medium - very long'. The restricted distributions of the two extremes of the series, (concordance with other apomorphies) indicate their apomorphic status, however. The very short, sparse setae (Pig.195) are found in no other Phthiraptera but the Pelicola- Lorisicola clade and some Trichodectes spp., and are probably apomorphic but convergent in the two groups (character 206:1'). The very long refringent seta.e of some Trichodectae (Pig. 171) are considered apomorphic for a similar reason (character 206:1), though in this case similar setae are found in some Philopteridae. These long- setae probabl3r evolved twice in.the Trichodectidae, once in Trichodectes and once in the Heotrichodectinae. The fine, long setae of the females of the Pelicola (J3.) cooleyi - quadratic ens clade are found in no other group and are considered apomorphic (character 220). The setal bases - the circular 'pits' of the setal articulations - are of fairly constant size relative to the setae in'.most Trichodectidae. However, in the Bovicola (13.) ale inns - tibialis clade the bases are noticably large in relation to the setae, and seem to have a double margin• This feature needs to be examined using the scanning electron microscope to elucidate its true structure, but examination using the light microscope is sufficient to detect its presence. This feature is here considered as apomorphic (character 222). 247

11. Abdominal pleural projections and modifications (characters 223 - 229) In many Trichodectidae the dorsoposterior and/or the ventroposterior pleural angles project on pleura II, III or IV (see discussion in section 2.2. above, and Table IV). Projections on these segments of the tyae found here do not occur elsewhere in the 'Phthiraptera, and are therefore considered as aponorphic (characters 224-228). Preliminary analysis reveals that the projections on pleurum III (characters 225 and 226) contribute no useful phyletic information, so the apomorphy is -omitted from the final cladistic analysis. Variation in the degree of development of the dorsal and ventral lobes of the projection on pleurum IV is omitted for the same reason, except for the extreme development in Durytrichodectes (ch.ara.cier 229). Both the presence of a dorsal and a ventral .projection on pleurum IV (characters 227 and 223 respectively) are included in the analysis, though the latter a.pomorphy is reversed in some clades. The presence of a projection on pleurum II (character 224) provides a synapomorphy for Geomydoecus spp., which are also united, as a holophyletic group on other grounds; the projection on this pleurum is found convergently in Trichodectes (Subgen. n. 5) zorillae. The sclerotisation of the dorsal projection on pleurum II (character 224:2) is an autapo- morphy of Geomydoecus (Thomomydoecus) (and some species in Geomydoecus s.str. - see discussion in section 2.4.3« - and T_. zorillae, but . sclerotisation of the projections on the other pleura is very variable, and is not used in cladistic analysis . In Damalinia (Damalinia) pleurum II extends on to sternum II, and the pleurite is expanded at the expense of the stemite (character 223). This extension may be broad (character 223:1') or narrow (character 223:1), but the more plesiomorphic state of these two (should they not be co-apomorphies) is not known. The species with a broa.d ventral extension of pleurite II also possess a more or less extensive dorsal extension, but this is not found in species with a narrow ventral extension. In this treatment the two forms of the ventral extension are used to characterise each of two, sister-groups, but this hypothesis is-.open to challenge, as the group indicated by character 223:1 has no other supporting apomorphy. 248

12. Abdominal sclerae (characters 192, 230 - 279) As noted above, in species of the trichodectid genus Procawicola sternite II is greatly developed as a heavier sclerotised internal apophysis, articulated to pleurum II (character 192). The presence of this unique structure is proposed as apomorphic. The presence of the lateral flecks and their associated small sclerite (see section 2.2. above) is proposed as apomorphic, as the structure occurs in no Phthiraptera other than the Trichodectidae (character 232). The pleura, sterna and terga of the trichodectid abdomen may be sclerotised or not; although some groups (identified on the basis of other apomorphies) may, in general, be more or less sclerotised, the polarity of the transformation series 'sclerotised - not sclerotised' for each segment cannot be determined, and the characters (233 - 279) are not used in the cladistac analysis. The male abdomen may have a characteristic sclerotisation dorsally, in that the terga may have anterior and posterior sclerites (see section 2.2.); this feature is present in some Anoplura, but is probably convergent in this suborder. The presence of doubled tergal sclerites in male Trichodectidae is believed to be apomorphic for the family,but is not used in cladistic- analysis because of the large number of reversals. The tergal sclerites are not further modified in most male Trichodectidae, but in some there is longitudinal division of the anterior or posterior sclerites (characters 230 and 231 respectively), this division being accepted as apomorphic.

2.4.3. Cladistic Analysis

The holophyly of both monotypic and polytypic species is accepted without the need for justification, so species-level autapomorphies have not been indicated unless they are homoplastic with character states elsewhere on the cladogram. Omission of the autapomorphies of species saves both space in the data, matrix and time taken for analysis, and for the same rea.sons many sister-species pairs are justified on the cladogram with fewer autaoornorohieg than are available. Of the 187 apornorphic character states used in the analysis, 86 are postulated to have been developed more than once or to have been 249

secondarily lost, 363 such homoolasies being proposed. V/hen, in the analysis, a choice is available between postulating one reversal or a pair of homoplastic gains (i.e. three clades in a holophyletic group are involved and the topography of the tree is not affected whichever the choice), the latter is chosen (e.g. character 21:1 in the Damalinia theileri - apoendiculata clade). This choice is made so that the distribution of apomorphic character states can more easily be discerned on the cladogram. The number of homoplasies could be slightly reduced without affecting the topology of the tree, because, as explained below, the less parsimonious presentation is sometimes chosen to make the clado- gram more informative and less potentially misleading. The loss of the median extension of the mesomeral arch (character 58) is placed on the cladogram 17 times, frequently in combination with' the loss of apical fusion of the nesomeres (character 44). These apomorphies are not arranged in the most parsimonious manner on the clado- gram, as can be seen by inspection of the Bovicolinae. As presented, the cladogram depicts the loss of the extension 11 times in this subfamily. A more parsimonious arrangement of the apomorphies is achieved b;g postulating characters 44 and 58 as synapomorphic for Damalinia (Damalinia) and Damalinia (Cervicola), and character 58 as synapomorphic for two clades: ".Yemeckiella plus the new genera 2 and 3, and Bovicola (Bovicola) plus 3_. (Leoikentron) and _B. (n.l) . This arrangement reduces the number of proposed homoplasies of character 58 to six within the Bovicolinae, and reduces the number of polychotomies on the cladogram. Alternatively, the loss of the extension might be postulated to have occurred only once, in the common ancestor of the Bovicolinae, and regained six times (_B. crassioes, D. elongata, D. moschatus, D. clayi and the D. in die a - alb irnargin at a clade). The most parsimonious hypothesis is that the structure was lost in the ancestor of Bovicolinae as suggested above, regained twice (_B. crassipes .and the D. (_T.) al b im ar g in at a - elongata clade), and secondarily lost twice (D. (T.) line at a - victoria.e and D. (Tf) oakenhami - bedfordi ) * This last hypothesis, although more parsimonious than the distribution on the tree presented, does not change the topology of the tree. The distribution of character 58 is not as apparent from inspection of the tree in its most parsimonious distribution 250

as it is in the tree presented, as the more scattered distribution of the losses and reversals obscures the alternative possible distributions and implies a spurious confidence in the tree as supported by them. The distribution of character 44 (the loss of nesomer.al fusion) in' 7e m e c k i e 11 a is not presented in the most oarsinonious manner. There is great difficulty in the observation of this character stake in 7erneckiella, and the morphological difference between 'loss of fusion' and 'fusion' is very slight. A detailed examination of the species of this genus for other characters to complete a. full analysis was not made, character 44 only being noted because it occurs elsewhere on the cl ado grain. It is possible but not likely that the distribution of character 44 as observed is supported by other apomorphies, but the proposal of holo- phyletic groups within the genus on 'the basis of the observations made of this single loss character would be -unwise. It is notable that 7 erne ckie 11 a. fulva and _7. neglect a, which differ in the state of character 44, are otherwise very similar, the females apparently being indistin- guishable (Emerson Price, 1979), and it is very likely that they are sister-species. The arrangement of Protelicola, Lutridia. and the Trichodectes - Genus n. 4 clade (T-n4) on the cladogram (Fig. 39) does not accord with the most parsimonious distribution of apomorphies. The cladogram contains four convergencies for 'gain' apomorphies: 20:1' (fusion of parameres and b.a.l.s.) is postulated as homoplastic in Lutridia. Genus n. 4, and Trichodectes (Sta.chiella); 21:1 (fusion of parameres to each other) is postulated as homoplastic in protelicola and Lutridia: and 69 (development of tubercles for the gonaaophysea.l setae) is postulated as homoplastic in Protelicola. and T-n4. Apomorphy 20:1' has been discussed in deta.il in section 2.4.2., and the distribution suggested in the cladogram is believed consistent with the morphological evidence. Apomorphy 69 could be considered in two ways other than that presented: as an autapomorphy supporting the sister-group relationship of Protelicola and T-n4, or as an aut apomorphy of the Tricho dec tin! (the Trichodectes - Protelicola clade), reversed in Lutridia . The first alternative is not sunported by the distribution of any other aromorohies, whereas two altsmative arrangements are each indicated by more than one apomorohy 251

(see "below); the sister-group relationship of Protelicola and T-n4 is therefore rejected. The plesiomorphic arrangement of the gonapo- physis tubercles in T-n4 is clearly distinct from the arrangement in Protelicola. ' If the tubercles are postulated to be homologous in the two clades two further aponorphies (the form of the'tubercles in each clade) would have to be proposed, as neither form appears to be plesio- morphic with respect to the other. This manipulation does not affect the topology of the cladogram (whatever the position of hutridia), and does not clarify the relationships of the clades involved, so the hypothesis of convergence of character 69 in Protelicola and T-n4 is retained. The other two apomorphies nay now be considered together as they both suggest the sister-group relationship of Protelicola .and hutridia. The alternative hypothesis (of the cladogram as presented) is supported by apomorphies 86 (development of a sclerotisation along the ventral vulval margin) and 58 (loss of the median extension of the mesomeral arch). 'Loss' characters are given much less weight than

1 gain1 characters in this analysis, so character 58 should be left out of consideration. The sister-group relationship of Protelicola and Lutridia is therefore supported by two apomorphies and the relationship proposed 011 the cladogram supported by one. As noted in the generic descriptions below, however, an undescribed species of Protelicola has been seen which does not share apomorphy 29. The cladistic position of this species with respect to the other two species in the genus has not been determined because of the poor state of preservation of the specimens, but its existence raises the possibility that character 29 is an apomorphy not of Protelicola but of only two species within the genus (the alternative being a reversal in the undescribed species). Character 29 is also homoplastically developed in Felicola end Bovicola (Lepikentron). If this character is disregarded, apomorphies 80 and 21:1 must be compared for their comparative likelihood of homoplasy. • Apomorphy 21:1 is homoplastically developed at twenty-two other points on the cladogram whilst 86 is found elsewhere only in Bovicola (LeoIk ent ron). Aoomorohy 86 should clearly be xiven much more weight then 21:1 in construction of the cladogram, end 29 is considered of uncertain value in view of the undescribed species of Protelicola. For these reasons the cladonram 252

is retained as proposed, even though it is not maximally parsimonious. The dorsal projection of pleurum IV (character 227:1+2) is lost in the Trichoriectini, but postulated as secondarily regained in "7e ra e ck o de c t e s end Trichodectes (n. 5) zorill as. The form of the projection is different in the two species, however, which indicates the independent development of the structure. V/liilst the genus Cieonydoecus s. lat. is almost certainly holophyletic, this probably does not apply to either-of the two included subgenera (Fig. 38 ). The question of holophyly should be addressed first in the smaller subgenus, Geomydoecus (Thomomydoecus) . All but one of the included species (_G. (_T.) v/ardi) have characteristically asymmetric male genitalia (character 38), and are proposed on this basis to be a:holo- phyletic group (the asymmetricus - zacatecae clade or *a - z cla.de1). Cr. (T_.) warcli and the a. - z clade share the following apomorphies: posterolateral temple margin v/ith single stout seta and associated shorter, finer setae (a character not included in the data matrix); male paraineral plate apically pointed (character 22); gonapophysis smoothly continuous with ventral vulval margin (character 85); male abdominal terga II and III. with median, setal group comprising exceptionally long, stout setae (character 214); and pleural projections sclerotised, especially in females (character 224:2). The possession of a single stout temple seta is unique to these species, but may be a reduction from the two stout setae found in this position in some Geomydoecus (Geom.ydoecus). Apomorphies 22, 85 and 214 are also shared by _G. (_}•) cooel, and this species has the mesomeral arch and oaraneral plate very slender, approaching the shape of the genitalia of the a - z clade more closely than does G_. ) wardi; the posterolateral temple margin lacks any specially-modified setae, but this may be due to secondary loss. Apomorphies 22, 85 and 224:2 are shared by the _G. ((}.) thomomyus - dakotensis clade, but the male genitalia are considerably broader then those of _G. (T.) wardi, and the mesomeral arch lacks a median extension (an autapomorphy of the clade). This clade lias a further autapomorphy in the form of the posterolateral setae of the temple margin, which comprise a single long fine seta and associated shorter fine setae.

As with _G. (_G.) copei, the plesiomorphic form of the temple setae is 253

unknovm, and could have been the form found in _G. (Thomomydoecus) . Other species of Cr. (Geo my do ecus) have a single apex to the paraneral plate (character 22), but do not share any of the other aoomorphies mentioned. jl. (_!.) ward!, the a - z clade, and the _G. (_G.) thomomyus - dakotensis clade are all parasitic on Thonomys sop., whilst -a. (0.) cored is a parasite of Orthogeomys his nidus; both host genera, are para.sitised by other members of Geomydoecus (Geomydoecus). .The a.pomorphies listed a.bove plainly" do not support unequivocally any of the three possible sister-group relationships of the a - z clade without invoking homopla.sy to an unjustifiable extent. It is apparent, however, that Cr. (Geomydoecus) is paraphyletic with respect to G. (Thomomydoecus) and that the latter subgenus is possibly polyphyletic. A full phylogenetic analysis of the 102 species and subspecies of Geomydoecus, which would have been necessary to resolve the problem, was not attempted. The hosts of the genus are all geomyid rodents, the systematic and taxonomic understanding of which is of questionable accuracy (Price, oers. comm.). For the purposes of this study the subgeneric concepts proposed by Price & Emerson (1972) are retained. 254

Fig. 37• Cladogram of Trichodectidae. Clades numbered 1 - 12 are resolved in Figs 38 - 45 below. The Eutrichophilinae (genus Eutrichophilus) is not resolved further. For explanation of numbered apomorphies see text.

V 255

G. (G.) UNRESOLVED SPECIES -» M rsj <0^ ro toas G. (G.) COPEI

G. (G.) THOMDMYUS

••• • G. (G.) DUCHESNENSIS

G. (G.) DAKOTENSIS

G. (T.) WARDI

-0" •G. (T.) A - Z CLADE

•N. (N.) THORACICUS

r-O N. (N.) IHNUTUS

• N. (N.) OS BORN I

N. (N.) MEPHITIDIS

N. (N.) WOLFFHUEGELI

•N. (T.) BARBARAE

•N. (6 ) PALLIDUS

•N. (L.) GASTRODES

•N. (L.) CUMMINGSI

•N. (7 ) INTERFUPTOFASCIATUS

•N. (7 ) CHILENSIS

•N. (7 ) SEMISTRIATUS

•N. (7 ) ARIZONAE

Fig. 38. Cladogram of Subfamily n. (clade 1 of Fig. 37s genera Geomydoecus and Neotrichodectes). For explanation of numbered apomorphies see text. 256

-45k

Fig. 39. Cladogram of Trichodectini (clade 2 of Fig. 37* genera Trichodectes, Werneckodectes, n.4» Lutridia and Protelicola). For explanation of numbered apomorphies see text. 257

F. F. MINIMIS

F. F. RAH MI

F. F. CALOGALEUS

i\) A. tn oi -» rv) F. F. ZEYLONICUS O OO-'•, HG OM F. F. ROHANI MOI U IO - M ffl( O — ri F. F. INAEQ.UALIS

F. F. VIVERRICULAE

F. F. CYNICTIS

F. F. SETOSUS

F. F. LIBERIAE

F. F. ROBERTSI

F. F. HOPKINSr

F. F. SUBRDSTRATUS

F. F. C0NG0ENSIS

F. F. HEL0GAL0IDIS

F. F. HELOGALE

F. F. OCCIDENTALIS 35 2 io°? — r-OWll F. S. BEDFORDI

F. S. ACUTIROSTRIS

F. S. MACHJFUS

F. S. PYGIDIALIS

F. S. COOLETI

F. S. DECIPIENS

F. S. FAHRENHOLZI A iv) rv) ro co oao coro o o ^ F. S. GUINLEI

F. s. VULPIS

F. s. QUADRATICEPS

Pig. 40. Cladogram of Felicola (clade 3 of Fig. 37)• For explanation of numbered apomorphies see text. 258

L. p. BENGALENSIS

L. P. PHILIPPINENSIS

L. P. JUCCII

L. P. ASPI DO RH YNCHUS —DO*- L. P. SUMATRENSIS

L. P. LENICORNIS

L. P. WERNECKI

L. P. ACUTICEPS

L. P. AFRICANUS

L. P. NEOAFRICANUS

L. P. PARALATICEPS

•atMii L. P. LATICEPS

L. P. TTJNGOS

L. L. - - M N MALAYSIANUS r-DOHH- * ^ ® — — — L. L. MJOBERGI Ul 01 - M en vi -J ro L. L. AMERICANUS

L. L. BRAZILIENSIS

L. L. CAFFRA

L. L. FELIS

L. L. NEOFELIS

L. L. SIMILIS

L. L. SPENCERI

L. L. SUDAMERICANUS

L. L. HERCYNIANUS

L. L. SIAMENSIS

Fig. 41. Cladogram of Lorisicola (clade 4 of Fig. 37)* For explanation of numbered apomorphies see text. 259

C. EXTRARIUS

-a+m-MOI- C. SEMIAFMATOS ® s t C. ARMATUS

PROCAVICOLA (P.) 16 SPECIES

P. C.) LINDFIELDI

p. C.) HOPKINSI a) A. g p. C.) DISSIMILIS

p. c.) BEDFORDI

p. C.) UN IVIRGATUS

PROCAVIPHILUS (P.) ROB

P. P.) HARRISI

P. P.) F. GRANULOIDES

P. P.) F. HINDEI

P. P.) F. FERRISI

P. P.) DUBIUS

P. P.) GRANULANS

P. M.) S. SCLEROTIS

P. M.) S. MAJOR

P. M.) SERRATICUS

P. M.) SCUTIFER

P. M.) JORDAN I

P. M.) N. NEUMANNI

P. M.) N. BACULATUS

P. M.) TENDEIRQI

P. M.) IUESEBECKI

P. M.) CONGOENSIS

P. M.) COLOBI

P. M.) ANGOLENSIS

P. M.) AFRICANUS

m- DASYONYX (D.) 9 SPECIES

NUN CO - -4 -k ® D. (N.) 6 SPECIES fr ® t fro E. PARADOXUS

OB — — IO to E. MACHADOI

Fig. 42. Cladogram of Dasyonyginae (clade 5 of Fig. 37: genera Cebidicola, Procavicola, Procaviphilus, Dasyonyx and Eurytrichodectes) For explanation of numbered apomorphies see text. 260

B. (B.) ALPINUS

B. (B.) BOVIS

B. (B.) LIMBATUS in u in ro as r, in -«j — M B. (B.) CAPRAE

B. (B.) OREAttUDIS

B. (B.) OVIS

w * r-CH> B. (B.) JELLISONI

<0 ^ B. (B.) LONGICORNIS ro

B. (B.) TARANDI

B. (B.) TIBIALIS

B. (H.) CRASS I PES

B. (L.) BREVICEPS * N O „ _ A 00 (O o> 15? S B. (1 ) HEMITRAGI

•A• 0 1 ••no —• 10 B. (1 ) MJLTISPINOSA

2 S. BISON

2 S. SEDECIMDECEMBRII

3 TRAGULI

W. ASPILOPYGA

W. EQUI

W. OCELLATA

W. FTJLVA

W. NEGLECTA

W. ZEBRAE

W. ZULUENSIS

Fig. 43. Cladogram of part of Bovicolinae (clades 6 - 9 of Fig. 37s genera Bovicola, n.2, n.3 and Merneckiella). For explanation of numbered apomorphies see text. 261

D. D.) THEILERI

D. D.) NEOTHEILERI

D. D.) SEMITHEILERI

io ni m M. kJl D. D.) CRENELATUS

D. D.) BAXI

D. D.) HARRISONI

D. D.) CHORLEYI

D.- D.) 0 FN ATA

D. D.) TH0WS0NI

D. D.) ORIENTALIS r®- D. D.) DIH5RPHA

to ^ D. D.) PELEA

D. D.) ADENOTA

D. D.) HILLI

D. D.) FAHRENHOLZI

D. D.) APPENDICULATA

D. C.) HENDRICKXI

D. C.) MARTINAGLIA

D. C.) MAAI

D. C.) PORFICULA •••• D. C.) METERI

D. C.) MJNTIACOS

D. C.) REDUNCAE

0. C.) UGANDAE

•am-ro D. C.) TRABECULAE to — D. C.) LEROUXI

D. C.) ANNECTENS

D. C.) HOPKINSI

M MO ? • D. C.) NATALENSIS

Fig. 44. Cladogram of D. (Damalinia) and D. (Cervicola) (clades 10 and 11 of Fig. 37). For explanation of numbered apomorphies see text. 262

D. (T.) AEPYCERUS

D. (T.) ANTIDORCUS

D. (T.) C. COPNUTA

D. (T.) C. OUREBIAE

D. (T.) PARKERI

D. (T.) LGNGICEPS

D. (T.) SPINIFER

D. (T.) ALBIMARGINATA

D. (T.) DORCEPHALI

D. (T.) LIPEUROIDES

D. (T.) PARALLELA

D. (T.) INDICA

D. (T.) LINEATA

D. (T.) VICTORIAE

D. (T.) CONECTENS

D. (T.) PAKENHAMI

D. (T.) BEDFORDI

D. (T.) CLAYI

D. (T.) MOSCHATUS

D. (T.) ELONGATA -t 01

Fig. 45. Cladogram of Damalinia (Triehoiipeurus) (clade 12 of Fig. 37)• For explanation of numbered apomorphies see text. 263

TAXONOMY, PHYLCGENY AND HOST RELATIONSHIPS

CP THE TRICHGDECTIDAE (PHTHIRAPTERA: ISCHNOCERA)

CHRISTOPHER HENRY COUTTS LYAL, B.Sc.

VOLUME 2

October 1983

A thesis submitted for the degree of Doctor of Philosophy of the University of London and for the Diploma of Imperial College.

Department of Pure and Applied BioTogy, Imperial College, London Stf7

and

Department of Entomology, British Museum (Natural History), London SIT7 SECTION 3

CLASSIFICATION AND TAXONOMY

OF TRICHODECTIDAE 265

3.1 TAXONOMIC HISTORY OF TRICHODECTIDAE

Buzmeister (1838) divided the Mallophaga into two families, Liotheidae and Philopteridae, the latter comprising the two genera Philopterus and Trichodectes. Kellogg (1896) proposed the suborders Amblycera and Ischnocera for Liotheidae and Philopteridae (sensu ^ Burmeister) respectively, and erected the family Trichodectidae for the genus Trichodectes.

Mjdberg (1910) described Damalinia and Butrichophilus, the second and third genera of Trichodectidae, and Stobbe (1913a) described a fourth genus, Burytrichodectes. Stobbe (1913b) revised the family for the first time. Ewing (1929) described four further genera and provided a key to all eight, though Ferris (1929) regarded Ewing's new genera as of "most dubious value". Bedford (1929, 1932a, 1932b, 1936) described a further ten genera, two of which were junior synonyms of genera proposed by Bwing (1929), thus bringing the total to sixteen genera; Bwing (1936) provided a key to' fourteen of these. Keler (1938) recognised twenty-four genera, ten of them new (although one of these had been published previously by Keler, 1934 as a nomen nudum). The two genera omitted by Keler (1938) were the same two .previously omitted by Swing (Cebidicola Bedford, 1936 and Lorisicola Bedford, 1936); the three species included in these genera were placed by Keler with two others, also from primates, in his new genus Meganarion (although with the proviso "without, of course, intending to establish the congeneric status of these species"). Keler (1938) provided a key to most of the genera described in his paper (with the sole exception of Meganarion) and to many of the species. V/erneck (1941) introduced the subgenus concept to the taxonomy of Trichodectidae, describing three new subgenera in two of the four genera of the family parasitic on hyraxes. During the decade following Keler* s (1938) review of the family a number of new genera were described, bringing the total number of available names in the genus-group to 43 by the end of 1948. Werneck (1948, 1950) reviewed all the Trichodectidae, and recognised twenty genera (though one of these doubtfully) and three subgenera. There have been no genera and only ".one subgenus described since 1948. The most recent name to be added (L ak shminar a.y an el 1 a Eichler, 1982) was published as a replacement name 266

for a junior homonym, and brings the total of available names to 45- There have been no revisionary works of the family since those of Werneck (1948, 1950), although Hopkins & Clay (1952) when cataloguing the 'Mallophaga1 accepted thirteen genera, some of these doubtfully, and Eichler (1963) recognised thirty-eight genera (with no subgenera). The problem of establishing criteria by which taxa can be distinguished at the generic level received early attention. An attempt to identify morphological characters for this purpose was initiated by Bedford (1929) • Bedford (1932a) provided a more thorough discussion, and concluded that v/hilst the shape of the head, the presence or absence of abdominal sclerites and the form of the female gonapophyses were of value, the form of the male genitalia and the number of abdominal spiracles provided no useful guide. Swing (1936) came to a quite different conclusion, regarding abdominal spiracle number as "the most important generic character". Bedford (1939) realised the unworkability of any system involving a. priori assessment of morphological'characters for generic discrimination, although he still felt that abdominal spiracle number was not of value at the generic level, and noted that (morphological) generic characters "may not be very striking". To supplement or replace morphological characters Bedford (1939) made use of host data, the possibility of which was first discussed by Kellogg (1913, 1914) and Harrison (1916). Bedford (1939) wrote: "Before placing a species in a new genus one should ask oneself: would it be possible to say from what kind of host the parasite was taken off had it not been recorded? If it is impossible to answer the question, then one should be justified in placing it in a new genus.". Hopkins (1941) used this principle to a certain extent in his discussion of Felicola. He also discussed the morphological characters used for the discrimination of the genus from others, and pointed out that "the singling out of one character [on which to base genera] only tends to obscure natural relationships". Werneck (1936) perceived and treated the problem of generic discrimination in a rather different way from those described above. He noted that whilst the type-species (and sometimes a few species similar to the type) of each of the described genera were quite distinct, other species showed intermediate characters. The existence of these 'transitional forms1 convinced V/erneck that there 267

was 110 validity in the separate genera, so he synonymised them all (with the exception of some genera not found in South America, which were outside the scope of the paper). Bedford (1939) regarded this action as "unwarrantable" and re-erected all the genera. Hopkins (1942, 1943) reviewed the characters used to separate the genera of Trichodectidae parasitic on carnivores and antelopes respectively. In each case he found annectent species as described by Werneck (1936), and took similar action, though modified by the belief that louse genera should somehow reflect host taxa. Hopkins (1942) therefore accepted three genera of Trichodectidae parasitic on carnivores and Hopkins (1943) accepted one genus parasitic on antelopes. Werneck (1948, 1930) though less influenced by the host data, recognised more genera than had Hopkins. He accepted the morphologically 'distinct1 species and species groups as genera, and placed annectent species in the genus which they most' clos.ely resembled. Werneck provided diagrams to illustrate the position of these species (Fig. 46 ). Hopkins (1949 "conceded subgeneric status to many groups which seem likely to be accepted by.systematists whose views .... differ from mine" and recognised fourteen genera and twenty subgenera. Hopkins

8c Clay (1952) synonymised some of the subgenera accepted by Hopkins (1949), but raised others to generic status. Ledger (1980) held views similar to Hopkins 8c Clay (1952), though in some cases followed the views of Hopkins (1949); the resultant generic arrangement still involves fewer genera than accepted by Werneck (1948, 1950) and many more subgenera.

Emerson 8c Price (1981), however, can "find no basis for rejecting the classification of Trichodectidae given by Werneck (1948, 1950)", and suggest that the "question of genera vs subgenera will perhaps continue until Mallophaga have been described from all likely hosts". A number of other workers also follow these views. A third group of taxonomists (e.g. Eichler, 1963; Zlotorzycka, 1972) have accepted not only all the genera recognised by Weraeck (1948, 1950), but also a number of genera that Werneck considered as junior synonyms; the subgenus category is not used, however. The present generic placement of most of the Trichodectidae is thus a matter of some contention and a review of the variations in status of some genera and subgenera is presented in Tables VTI-IX.

Hopkins (1941) presented an explanation for the presence of so many / T. mal ays i anus ca ca F. subrostratus £ -H F. intermedius •H t) 0)H / P« pygidialis 0 0s iH rt F. minimus P 4 ci +> F. felis F. macrurus ca £ bfl-p o 8 o 0 0 F. acutirostris •P A H fi Cj F. bedfordi 0 O rt

PH PH PH PH PH PH PH / \

ca '"s •H \ ca £ 0 rH rt W) £ 0

CU

ca ca .H &o 0 p4«n £ £ -H 0 o® o fao .£ £ •H 0 O £ 0 -H £ -H £ £ £ rd "P O 0 0 •h rt 0 £ rH Pi B 0 £ rt 0 Ph PH CU • •

W Fig. 46. Werneck's (1948) schematic representation of the relationships of Felicola, Parafelicola and Neofelicolai

ON 00 u 0 bo & rH S A c A bo 01 —' •H iH O u a B o xi HI) 0) o <0 a) Presen t stud y Hopkin s (1949 ) a w Hopkin s (1943 ) te a a 1. Damalinia Mjttberg, 1910 1 1 1 1 1 1 1 1 1 1 1 2. Bovicola lining, 1929 2 2 2 = 1 1(2) 2 1(2) 2 1(2) •2 2 3. Tricholipeurus Bedford, 1929 3 3 3 = 1 1(3) 3 1(3) 3 1(3) 3 1(3) 4. Cervicola Kdler, 1938 4 4 - 1(4) = 1 = 1 & s:2 4 s1 s.l. =2 . 1(4) 5. Holakartikos Kdler, 1938 5 5 - 5 =2 =2 5 =1 S.l. =2 1(5) 6. Lepikentron Kdler, 1938 6 6 - 6 «=2 1(6) 6 =1 S.l. =2 1(6) 7. Rhabdopedilon Kdler, 1938 7 7 - 1(7) =2 =6 7 =1 s.l. =2 =2 8. Vlerneckiella Eichler, 1940 8 - 1(8) =2 =2 8 »1 s.l. =2 8 9. Subgenus n. 1 2(9) [from 2] 10. Genus n. 2 10 [from 2] 11. Genus n. 3 11 [ from 1]

Table VII. Generic concepts in the Bovicolinae. The genera included in the table are represented by numbers 1-11; '2* indicates thgit the genus (Bovicola) is given full generic status, 11(2)' indicates that the genus (Bovicola) is considered as a subgenus (of Damalinia) and '=1' indicates that the genus is considered as a junior synonym (of Damalinia)» The generic name Bovidoecus Bedford, 1929 is omitted, as it was reoognised as a junior Bynonym of Bovicola by Bedford (1932a) and has not since been used. There is no reference in the table to Werneck (1936), who treated all genera of ro ON Trichodectidae as synonyms of Trichodect es. ^ Q> O X— CM CO ON & rn •H nJ iH vo rt CO ON CN ON ON O ON £ X- X— X— rn PH ^ ON ON X— CM rt (fl 1a m JO rt & 00 rt M0 rt a> 0) rt ON rt •H •H 0) •H •a c rH 0 «- d) XX rt M xi W -— 1—1 0 £ & Pi 0 XD O 0 % 0) rt W A tz & iz w0 il Presen t stud y

1. Trichodectes Nitzsch, 1818 1 1 1 1 1 1 1 1 1 1 2. Neotrichodectes Ewing, 1929 2 2 = 1' 2 1(2) 1(2) 2 1(2) 2 2[removed to Subfam. n.] 3. Galictobius K61er, 1938b 3 3 = 1 = 1 1(3) = 1 3 = 1 = 1 = 1 4. Lutridia K61er, 1938a 4 4 = 1 4 1(4) 4 4 1(4) 4 4 5. Stachiella KSler, 1938a 5 5 = 1 5 1(5) 1(5) 5 1(5) 5 1(5) 6. Ursodectes Kdler, 1938a 6 6 = 1 =1 1(6) =1 6 = 1 = 1 =1 7. Potusdia Conci, 1942 1(7) =1 7 = 1 = 1 = 1 8. Trigonodectes Kdler, 1944 =1 1(8) =1 8 = 1 = 1 2(8) [removed to Subfam. n.] 9. VJerneckodectes Conci, 1946 =1 1(9) =1 9 = 1 = 1 9 10. Genus n. 4 10 [from 4] 11. Subgenus n. 5 1011) [from 5]

Table VIII. Generic concepts in the Trichodectini (plus Neotrichodectes and Trigonodectes). The genera included in the table are represented by numbers 1-11; •1• indicates that the genus (Trichodectes) is given full generic status, 11(3)' indicates that the genus (Galictohius) is considered as a subgenus (of Trichodectes) and 'si1 indicates that the genus is considered as a junior synonym (of Trichodeotes). Protelioola is included in Table IX. The generic name Grisonia Kdler, 1938a (nec Gray, 1843) is omitted and its replacement name^ Galiotobius Kdler, 19381} used throughout. There is no reference in the table to Werneck (1936), who treated all genera as synonyms of Trichodectes. 1 1 genera ofTrichodectidae assynonym s ofTrichodectes. the descriptionofFelicol a s.strbelowThereinoreferenc e inthtabltoWerneck(1936) , whotreatedall respectively) haveno t beenincludedastheirstatuhano t varied;theyarediscussedinthcomment s following genus (Protelicola)i s consideredasubgenu(ofFelicola ) and'elindicatesthatthegenu isconsidereda by numbers1-9»indicatethatth e genus(Pelicola)igivenfullgenericstatus,•1(2)•indicatethatth Table IX.GenericconceptsinthFelicolin i (plusProtelioola).Thegeneraincludedn junior synonym(ofFelicola) . ThegenericnamesBedfordia andFelicinia(juniorhomonym n absoluteBynonym 9. 6. 8. 7. 5. 4. 2. 3. 1. Parafelicola Werneck,1948 Neofelicola Vferneck,1948 Paradoxuroecus Conci,1942 Fastigatosculum Kdler,1939 Eichlerella Conci,1942 Protelicola Felicola Ewing,1929 Lorisicola Bedford,1936 Suricatoecus Bedford,1932 Bedford,1932 = 1 1 1 Bedford (1936) 0 - 2 W 3 rH ro ON K> x— •» x 2 4 5 3 W XI i-H •H ON 1 O fo = 1 = 1 - 2 1 W * •H ON O P CQ =3 = 1 =3 . = 0 9 8 IS * * 4 4 4 4 4 3 ON •p- 1 1 1 1 U P Q) d> 1 1(9) 1(8) 1(3) 1(2) - - - w % % •H ON •sl" ON Pi O P CQ w 1(9) # 1(8) •H - "O U P CQ >— O Q) = 1 =3 =3 =3 ON 9 8 4 1 0 thetablarrepresented 4(7) 2 1(3) =7 =7 =3 = [removedtoTrichodectini] 4 1 1 Present study 272

annectent species in the Trichodectidae and a justification for synonymising many of the genera, writing: "I believe the explanation to be that the Trichodectidae are in the process of dividing up into genera; in some cases the divergence has proceeded far enough for us to recognise the segregates as generically distinct, but in a much greater number of cases extreme members of a group may have become strikingly distinct whilst the others remain as connecting-links which entirely undo our attempts to find characters peculiar to the group;". Hopkins' statement implies that the species in a genus are somehow evolving as a unit, and is linked to the typological approach to taxonomy. Relationships between three or more taxa, if assessed by a simple count of character .states (i.e. not distinguishing between plesiomorphies, apomorphies and homoplasies) are frequently reticulate in aspect (Simpson, 1961; Hennig, 1966; Mayr, 1969) (Fig. 47 )• The greater the number of homoplasies the more complex the reticulum is likely to be, and the more difficult it is to combine the taxa into groups. If genera are constructed on this principle some morphologically distinct species and species-groups will be distinguished, 'linked' by annectent species, with the concomitant absence of 'gaps' between genera - precisely the problem v/ith traditional groupings of the Trichodectidae. The 'problem' of annectent species is therefore engendered by the typological approach; the difficulties in distinguishing supra-specific groups are also a result of this, but combined in the Trichodectidae with a high degree of homoplasy.

The difficulties discussed above have discouraged authors from attempting to produce keys to genera of the Trichodectidae. Since the key to genera published by Keler (1938) very fev; such keys have been published, and none that included all the genera. Keler (1944) produced a key to some genera, slightly emended from Keler (1938), but a promised second half to the paper containing the rest of the key v/as never published (Keler, 1960b). Werneck (1948, 1950), despite describing all of the genera, did not attempt to produce a key. A few keys have since been published in faunistic works, for example Toulechkoff (1955) produced a key in Russian to the genera found in Bulgaria, and Zlotorzycka (1972) published a rather inaccurate key in Polish to the genera, found in Poland. 273

Fig. 47- Reticulate pattern of phenetic relationships of hypothetical taxa A, 3, C, D, E and F. The numbers denote number of character states in common. 274

Although most authors follow Kellogg (1896) in their conception of the Trichodectidae, and retain familial rank for the group (e.g. Hopkins & Clay, 1952; Hopkins, 1960; Ledger, 1980; Emerson & Price, 1983), the rank of the group has been raised by others. Keler (1938) raised the Trichodectidae to superfamily level and included three families: Tricho- dectidae (Trichodectinae nov. [sic], Felicolinae nov., Eurytrichodectinae nov. and, dubiously placed, in this family, Eutrichophilinae nov.), Bovicolidae nov. and Dasyonygidae nov. (Fig. 48). Eichler (1940) described two further subfamilies in the Trichodectidae (sensu Keler, 1938), Lymeoninae and Cebidicolinae. Eichler (1941) described the Damaliniinae nov. in the Bovicolidae, attributing the nominate subfamily to himself. He also transferred the Eurytrichodectinae and the Eutrichophilinae to the Dasyonygidae, again attributing the newly-defined nominate subfamily to himself. Eichler (1941) considered the rank of the whole group to be not superf amily but 'family group* and termed it the Trichodectiformia (attributed to Keler, 1938); the subfamily Trichodectinae is also attributed to Keler (1938) but the nominate family is attributed to Burmeister (1838). The classification proposed by Eichler (1941) is depicted in Fig. 49. Keler (1944) retained the Eurytrichodectinae in Trichodectidae and moved the Eutrichophilinae to Bovicolidae; the subfamilies described by Eichler (1940, 1941) Were not mentioned (Fig. 50). The Trichodectidae were attributed by Keler (1944) to Kellogg (1896) but the nominate subfamily to Keler (1938). The superfamily rank was retained and the 'family group' not mentioned. Hopkins (1949) regarded the families proposed by Keler (1938) as subfamilies and the subfamilies as at most tribes. Hopkins also, following his synonymy of Bovicola with Damalinia, considered that "Bovicolinae must be known as Damaliniinae". Eichler (1963) retained the higher ranks and included • interfamilia Trichodectiformia' within the superfamily Trichodectoidea. He moved the Lymeoninae to Dasyonygidae (wherein he retained Eurytricho- dectinae), but accepted the move of Eutrichophilinae to Bovicolidae proposed by Keler (1944). Eichler (1963) also proposed the division of the Trichodect(oidea) into tribes, indicating their existence and composition by variations in typography in the list of genera presented (Eichler, 1963: 159, lines 31-37). Eichler (1963) aid not publish any of the tribal names, but Lakshminarayana (1976) listed all of them. Neither Eichler (1963) uor Lakshminarayana (1976), however, gave any statement that purported "to give characters differentiating the tax(a); 275

DASYONYX • • DASTOHYGIDAE

EUKYTRICHODECTES • EXJRYTRICHODECTINAE

EUTRICHOPHILUS

PROCAVICOLA EUTRICHOPHILINAE

PROCAVIPHIUJS

BEDPORDIA

FELICOLA FELICOLIHAE PROTELICOLA

SURICATOECUS TRICHODECTIDAE

GEOMYDOECUS

GRISONIA

LUTRIDIA

NEOTRICHODBCTES

STACHI ELLA TRICHODBCTINAE

TRICHODBCTES

URSODECTES

BOVICOLA

CERVICOLA

DAMALINIA

HOLAKARTIKOS BOVICOLIDAE

LEPIKENTRDN

RHABDOPEDILOH

TRICHOLIPEUFUS

MEGANARION • INCERTAE SEDIS

Fig. 48. Classification of Trichodectoidea according to Kdler (1S38). 276

DASYONYX 3 DASYOHTGIHAE

EIJRYTRICHQDBCTES • EURYTRICHODECTINAE

EUTRICHOPHILUS DASYONYGIDAE

PROCAVICOLA EUTRICHOPHILINAE

PRDCAVIPHILUS

BEDPORDIA

FELICOLA FELICOLINAE PRDTELICOLA

SURICATOECUS

GALICTOBIUS

GEOMYDOECUS

LUTRIDIA

NEOT RICHODECTES TRICHODBCTINAE TRICHODECTIDAE

STACHIELLA

TRICHODECTES

URSODECTES

LYMEON • LYfEOHIHAE

CEBIDICOLA

LORISICOLA CEBIDICOLINAE

MEGANARIONOIDES

CERVICGLA

DAMALINIA DAMALINIINAE

TRICHOLIPEUFUS

BOVI COLA BOVICOLIDAE HOLAKARTIKOS

LEPIKEUTROH BOVICOLINAE

RHABDOPEDILON

WERNECKIELLA

TRICHOPHILOPTEFOS TRICHOPHILOPTERIDAE

Fig. 49* Classification of Trichodectiformia according to Eichler (1940 277

DASTONYX DASYONYGIDAE EURYTRICHODBCTES EURYTRICHODECTINAE l FASTIOATOSCUUJM1

FELICGLA1

PRDTELICOLA1

1 SURICATOECUS FELICOLINAE

CEBIDICOU2

LORISICOLA2

JEG ANA RIONO IDES2

GALICTOBIUS TRI CHO DECT I DAE GEOMYDOECUS

LUTRIDIA

NEOTRICHODECTES TRICHODECTINAE STACHIELLA

TRICHODBCTES

TRIGONODBCTES

URSODECTES

EUTRICHOPHILUS

LYMEON2 EUTRICHOPHILINAE PRDCAVICOLA1 ^

PRDCAVIPHIUJS1 ^

BOVICOLA1

1 CERVICOLA BOVICOLIDAE DAMALINIA1

1 HOLAKARTIKOS BOVICOLINAE

LEPIKENTRDN1

RHABDOPEDILON1

TRICHOLIPEURUS1

TRICHOPHILOPTEFUS TRICHOPHILOPTERIDAE

Fig. 50» Classification of Trichodectoidea according to KSler (1944)• 1 Position of genus inferred from K41er (1938).

2 Position of genus inferred from key in Kdler (1944). 278

•TRIBES'

DASTONTX NEODASTONTX ] DASYONYGINAE PROCAVIPHILUS • • EU RTTRICHODECTES DASYONYGIDAE PROC AVI COLA EURYTRICHODECTIHAE CONDYLOCEPHALUS

tCGANARIONOICES

LYWEON 3 • LYMEONINAE FELICOLA • SURICATOECUS

EICHLERELLA

FASTIGATOSCULUM

FELICOM3RPHA FELICOLINAE PROTELICOLA

NEOFELICOLA

PARAFELICOLA

PARADOXUROBCUS

TRICHODECTES

GALICTOBIUS TRICHODECTIDAE POTUSDIA

TRIGONODECTES

URSODECTES TRICHODECTINAE WEFNECKODECTES STACHIE1LA ZJ • NEOTRICHODBCTES LUTRIDIA • GEOMYDOECUS

CEBIDICOLA • CEBIDICOLINAE LORISICOLA ] BOVICOLA ] HOLAKARTIKOS ] LEPIKENTRON BOVICOLINAE RHABDOPEDILON ] WEFNBCKIELLA BOVICOLIDAE

DAMALINIA ] CERVICOLA • DAMALINIINAE TRICHOLIPEURUS

EUTRICHOPHILUS 3 EUTRICHOPHILINAE

Fig. 51* Classification of Trichodectiformia according to Eichler ( 1963) • The 'tribes' are indicated by square brackets in the appropriate column, the first genus listed in 'each 'tribe' being intended by Eichler (19^3) as the type-genus of that 'tribe'. 279

DASYONYX

NEODASYONYX DASYONYOINAE DASYONYGIDAE

P ROC AVI PHILUS

EURTTRICHODECTES EURYTRICHODBCTINAE PROCAVICOLA ] PROTELICOLA

FELICOLA

SURICATOECUS

FELICOLINAE NEO FELICOLA

CEBIDICOLA TRICHODBCTIDAE

LORISICOLA

GEOMYDOECOS

STACHIELLA

TRICHODBCTES TRICHODECTINAE

TRIGONODECTE3

URSODEETES

EUTRICHOPHILUS EUTRICHOPHILINAE

BOVICOLA

CERVICOLA

DAMALINIA BOVICOLIDAE BOVICOLINAE LEPIKENTRON

TRICHOLIPEUFUS

WERNECKIELLA

Fig. 52. Classification of Trichodectoidea accord-ing to KSler (19^9)• 280

or a definite bibliographic reference to such a statement" as is required by the International Code of Zoological Nomenclature for any name published after 1930 (Article 13). None of the (eleven) names, therefore, are available for taxonomic use. Eichler (1963) attributed Trichodectoidea, Trichodectifoimia, Trichodectidae and Trichodectinae to Burmeister (1838) but, whilst (correctly) attributing Bovicolidae and Dasyonygidae to Keler (1938) he attributed the nominate subfamilies of both to Eichler (1941). The classification proposed by Eichler (1963) is depicted in Fig. 51• Keler (1969) proposed a classification similar to that proposed by Keler (1944), but omitting a number of genera (Pig. 52). Article 36 of the International Code of Zoological Nomenclature (1964) states that all categories in the family-group (tribe, subfamily, family, superfamily and any supplementary categories, according to Article 35a) are co-ordinate, and a name established for any category within the group is available with its original date and author for a taxon with the same type genus in each of the categories. The Tricho- dectinae, Trichodectidae, Trichodectiformia and Trichodectoidea should therefore all have the same date and author. The first use of a family-group name based on the type genus Trichodectes was by Kellogg (1896), who ranked "the Nitzchian families as suborders, the Nitzschian genera as families, and the Nitzschian subgenera, the genera of present- day writers, as genera." (Kellogg, 1896). Kellogg (1908) attributed the Trichodectinae (the only subfamily included in the Trichodectidae, which bore no attribution) to "Burmeister (?)". As explained above, subsequent attribution has frequently been to Burmeister (1838) and, in the case of some co-ordinate names, to Keler (1938) . Burmeister (1838) did not mention any taxon in the family-group with the type genus Trichodectes. All family-group names with the type genus Trichodectes should therefore be attributed to Kellogg (1896). Trichodectidae Kellogg, 1896 has been placed on the Official List of Family-group Names in Zoology (Opinion 627, 3ull.Zool.Nom. (1962), 19: 91-96). The names Bovicolinae, Bovicolidae, Dasyonyginae ana Dasyonygidae should all be attributed to Keler (1938), not Eichler (1941). The action of Hopkins 281

(1949) in synonymising the senior family-group name (Bovicolinae) into the junior (Damaliniinae) was taken because be believed Bovicola and Damalinia (the type genera) to be synonyms, and Damalinia is senior to Bovicola. This action is incorrect under Article 40 of the Code, however, which states that in the case of type-genus synonymy the senior family-group name is to be used for the family-group' taxon that contains both senior and junior synonyms. This Rule can be set aside for such an action if taken before 1961, if the name as "won general acceptance" (Article 40a). The subfamily Damaliniinae sensu Hopkins has rarely; if ever been used since Hopkins (1949), whilst the name Bovicolidae (= Damaliniinae sensu Hopkins) has been employed by Eichler (1963) and Keler (19&9) • action of Hopkins (1949) is therefore rejects Keler (1944) included the Trichophilopteridae - a family containing a single genus, parasitic on Lemurs - within the Trichodectoidea, though Keler (1969) referred this family to the Philopteroidea. Eichler (1963) retained the Trichophilopteridae -in the Trichodectoidea, but distinguished it as 1interfamily Trichophilopteriformia1 as opposed to 'interfamily Trichodectiformia'. Stobbe (1913a), Ferris (1933) and 7/erneck (1948) all considered the affinities of Trichoohilooterus to lie v/ith the 'Philopteridae1 rather than with the Trichodectidae. In this study no apomorphies were found to indicate a sister-group relationship between Trichoohilooterus and all or part of the Trichodectidae. 282

3.2 PROPOSED CLASSIFICATION

The proposed classification is derived from the results of a cladistic analysis of the Trichodectidae (Trichodectiformia sensu Sichler, 19&3) at the species level (Figs 37—45)* species are grouped on the basis of criteria i - iv below, and ranked according to the principles of phyletic sequencing (see section 1.3.4. above).

Criterion i. Holophyly The classification includes, as far as possible, only holophyletic groups (see section 1.3.4. above). Some genera and subgenera, however, may be found not to follow this criterion (see discussion of Damalinia s. str., Dasyonyx and Geomydoecus below).

Criterion ii. Utility Genera are ideally of 'moderate* size and relative morphological unifoimity. If a genus is large and diverse, recognition is difficult and useful discussion on many aspects of biology or distribution prohibited; if genera are too small, identification is time-consuming and discussion again impeded. Ho ' absolute* size can be recommended, however, as the most satisfactory size will depend on a number of properties of the species, and must (in this study) conform to such limitations as are imposed by the criterion of holophyly. Subfamilies are chosen in this study to aid discussion by prbviding names for holophyletic groups of genera, and to fulfil the logic of the phyletic sequencing convention.

Criterion iii. Stability To ensure that the classification has maximal stability the generic concepts accepted in this study conflict as little as possible with established usage (though see discussion in section 3.1. above).

Criterion iv. Distinctness To facilitate identification, taxa in the genus-group should be as distinct from one another as possible. The requirements of Mayr (1969) that genera must be separated by a decided gap, and that the * size' of the gap should be inversely proportional to the size of the taxon, are not necessarily compatible with the criterion of holophyly followed here, 283

however, and the problem of annectent species (that 'fill1 any such gap) has been discussed above. Despite the apparent drawback of adherence to holophyletic groups at the expense of inter-generic 'gaps', it has been possible in this study to produce a key to the genera of Trichodectidae (see discussion of keys to genera of Trichodectidae in section 3.1 above). The formation of the genera of Trichodectidae is discussed below, to give an indication of the rationale behind each decision. The genera are discussed by subfamily, and the division into subfamilies is discussed last.

Subfamily Bovicolinae (Figs 43,, 44, . 45, 53) This clade was not resolved fully in the analysis, and a primary pentachotomy was obtained. The clade has been treated as a single genus (see Table VTI), but the diversity of morphology and of hosts indicates that this concept is too broad to be of great value. Subdivision of the clade into smaller holophyletic groups increases the value of the classification for information retrieval, and leads to the acceptance of genera that approach the concepts of Werneck (1950). To obtain a measure of conformity each of the five branches of the clade has been accorded generic status. The monobasic genera Nov.. 2 and Nov. 3 require no comment, and the genus Werneckiella was revised by lloreby (1978). Of the two remaining clades, one corresponds approximately to a restricted concept of Bovicola, the other to Damalinia plus Tricholipeurus (sensu Werneck, 1950). The Bovicola clade (genus Bovicola) has a primary tetrachotomy, with most of the species belonging to only one of the four resultant clades (Fig. 43) • The species in this large clade are morphologically more similar to one another than they are to any of the species in the other three clades. To recognise this morphological divergence (and thus facilitate identification), and to demonstrate in the classification the extent of the phylogenetic knowledge, the four branches are each 'accorded subgeneric status. The Damalinia plus Tricholioeurus clade (genus Damalinia) has a primary trichotomy (Fig. 37), snJ it is clear that discussion of the genus will be facilitated by the recognition of each of these branches as a subgenus. Damalinia s. str. comprises two major clades, each 284

characterised by the form of an apomorphic development of abdominal pleurum II onto the sternum. As discussed above (section 2.4.1.) these two forms may be co-apomorphies, or may represent two states in a trans- formation series. If the latter interpretation is correct, one of the clades is probably paraphyletic with respect to the other.

Subfainil?^ Butrichophilinae Only the single genus Butrichonhilus is included, with no change in generic concept.

Subfamily Dasyonyginae (Fig. 42) The previously-accepted generic concepts in this subfamily remain essentially unchanged at the subgenus level. The only change is the transfer of the subgenus Meganarionoides from Procavicola to Procaviohilus, and the inclusion of Procaviphilus sclerotis and P^ serraticus in (Meganarionoides). Subgenera are used (as in V/erneck, 1941, 1950; Ledger, 1980) as no advantage accrues from regarding each of the clades so recognised as a full genus, and application of the principles of phyletic sequencing allows retention of all the currently-used generic and subgeneric names with no higher taxa required, wh'ereas recognition of all these as genera would require the description of a number of intercalating family-group taxa. It is notable that one of the two subgenera of Dasyonyx may be paraphyletic with respect to the other, as the subgenera are characterised by apomorphic developments of the teeth of the tarsal claws. These may be co-apomorphies or two states in a transformation series (see section 2.4.1. above). If the latter interpretation is correct, one of the subgenera is probably paraphyletic with respect to the other.

Subfamily Trichodectinae (Figs 37, 39, 40, 41, 53) The first dichotomy in this clade splits it roughly into Felicola (sensu Ledger, 1980, but without Protelicola. and with Lorisicola) on one side and Trichodectes (sensu Ledger, 1980, but without Neotrichodectes and Trigonoaectes, and with Protelicola) on the other. The diversity of morphology of the lice, and the variety of hosts infested, indicates that the very broad generic concepts endorsed by Ledger (1980) are too inclusive to be of great value in data-retrieval and discussion. For 285

this reason the genera proposed here are smaller than those of Ledger (1980), and, in some cases, approach the concepts held by 'Verneck (1948). Most of the species in the Trichodectes side of the initial dichotomy arise from the three branches of an apical trichotomy (Fig.39)* The branch of this trichotomy comprising the oinguis - galictidis clade corresponds roughly to the concept of Trichodectes held by Werneck (1948), whilst the other two branches (the ovalis-zorillae clade and the emeryi- potus clade) correspond roughly to Stachiella sensu V/erneck (1948) (though fall ax, octomaculatus and potus were placed in Trichodectes by Werneck, 1948). However, placing two of the three clades of the trichotomy in a taxon Stachiella and excluding the third results in a group that is not holophyletic. Recognising each of the three branches of the trichotomy as a separate genus is undesirable, as the three inter- grade phenetically. The course followed here is to recognise the genus Trichodectes comprising all three branches, each of these being considered a subgenus (Fig.53). Using the principle of phyletic sequencing the sister-group of Trichodectes is also considered a genus, for which the name Werneckodectes is available. Likewise the next three branches of this clade are also considered genera. This process necessitates dividing the genus Lutridia into two genera, but retention of the genus as it stood calls for recognition of a paraphyletic group in the classification, and, although the species in the two clades comprising Lutridia (sensu Werneck, 1948) are superficially similar, some of these similarities may be homoplastic.

The other branch of the initial dichotomy of the Trichodectinae clade comprises, as noted above, most of the species consigned to Feljcola by Ledger (1980) plus the single species of the genus Lorisicola (sensu V/erneck, 1950). The two branches of this clade (Fig. 37) are each considered as genera which, taking the most senior available names, are known as Felicola and Lorisicola. For reasons of utility, each genus is divided into two holophyletic subgenera. Hone of the genera or subgenera coincides with any previous generic concept, as such concepts relied heavily on head shape and abdominal spiracle number, both of which characters have proved to be subject to a considerable degree of homoplasy. 286

In order to maintain the logic of phyletic sequencing, if the Felicola - Lorisicola clade is to be considered as comprising two genera, the rank of this clade and of the Trichodectes - Protelicola clade must be equal and formally recognised. Use of the tribal category permits this, and the family-group names Trichodectini and Felicolini are available (see full classification below). It must be stressed that these tribes aire inserted to maintain the formal structure of the classification, and are not intended (or believed) to have any other significance.

Subfamily n. (Fig. 38) The first dichotomy in this subfamily divides the clade into those species previously assigned to the genus Geomydoecus on one side, and species from Heotrichodectes, Lahshminarayanella and Trichodectes (sensu \7erneck, 1948) on the other. The two branches will be discussed separately. The genus Geomydoecus as previously recognised is fairly uniform in morphology, distribution and host species, and may be identified readily. To divide this genus into others would inhibit rather than encourage discussion, and the genus is retained in its present form. The two subgenera as proposed by Price & Emerson (1972) are also retained though, as indicated in section 2.4.3•» neither are holophyletic groups. The other branch of the primary dichotomy comprises the ten species previously assigned to the genus ITeotrichodectes (considered a subgenus of Trichodectes by Hopkins, 1949 and Ledger, 1980), the two species previously assigned to the genus Lakshminarayanella (formerly Lymeon), and a single species formerly placed in Trichodectes by most authors (£. barbarae). The clade is plainly close to the established concept of Heotrichodectes, and it is preferable that this name is applied to as much of the group as possible. The ten species of Neotrichodectes auctt. do not form a holophyletic group, however, though morphologically they are quite uniform. Inclusion of JT. barbarae is unlikely to create problems, but Lakshminarayanella (as Lymeon) has been placed by some authors in a subfamily of its own (Eichler, 1940, 19&3), considered close to the hyrax lice (Keler, 1944; Hopkins, 1949; Eichler, 1963). If Lak shminar a:/anell a is sunk into Heotrichoaect es and given no formal recognition it is likely to be raised from synonomy by future workers 287

because of its distinctive morphology, leaving Heotrichodectes paraphyletic. The course taken here is to recognise Lakshminarayanella as a subgenus of Heotrichodectes, which necessitates recognition of four other (holo- phyletic) subgenera, names already being available for tv/o of these. Application of the principles of phyletic sequencing permits equal ranking of the subgenera within the genus.

Subfamilies (Fig.53) To divide the family into 'manageable* holophyletic groups for the purposes of discussion and to maintain the logic of phyletic sequencing, supra-generic groupings had to be employed. Use of the principles of phyletic sequencing permitted the use of the subfamily category throughout (with the addition of the tribes mentioned above). The limits of the subfamilies were chosen for maximum utility, modified by the dictates of the sequencing convention. It would be surprising, given the high degree of homoplasy of structures in the Trichodectidae, if the subfamilies fulfilled criterion iv above and were readily distinguishable. A key to subfamilies is provided, however, largely to satisfy the requirements of the International Code of Zoological nomenclature (Article 13) for a description to accompany any new name for, although names were available for most of the subfamilies, a single new name is required and will be published in due course. A complete classification of the Trichodectidae to generic level is set out in below in phyletic sequence (as recommended by Wiley, 1979b, 1981).

Sequenced Cla.ssifica.tion of the Trichodectidae

Family Trichodectidae

Subfamily Bovicolinae (all genera sedis mutabilis) Genus Bovicola (all subgenera sedis mutabilis) Subgenus Bovicola Subgenus Holakartikos Subgenus Leoikentron

Subgenus n. 1 Genus n. 2 Genus 7/erne cki ell a Genus n, 3 Genus Damalinia (all subgenera sedis mutabilis) Subgenus Damalinia Subgenus Cervicola Subgenus Tricholiueurus Subfamily Eutrichophilinae Genus Eutrichophilus Subfamily Dasyonyginae Genus Cebidicola Genus Procavicola Subgenus Procavicola Subgenus Condylocephalus Genus Procaviohilus Subgenus Procaviohilus Subgenus Meganarionoides Genus Dasyonyx Subgenus Dasyonyx Subgenus IIeodasyonyx Genus Burytrichodectes Subfamily Trichodectinae Tribe Trichodectini Genus Protelicola Genus Lutridia Genus n. 4 Genus V/erneckodectes Genus Trichodectes (all subgenera sedis mutabilis) Subgenus Trichodectes Subgenus n. 5 Subgenus Stachiella Tribe Pelicolini Genus Felicola Subgenus Felicola Subgenus Suricatoecus Genus Lorisicola Subgenus Lorisicola Subgenus Paradoxuroecus Subfamily n. Genus ffeotrichodectes Subgenus Ileotrichodectes Subgenus Trigonodectes Subgenus n. 6 Subgenus Lakshminarayanella Subgenus n. 7 Genus Geomydoecus Subgenus Geomydoecus (paraphyletic) Subgenus Thomomydoecus (? polyphylet 290

GEOMYDOECUS (G.) E G. (THOMOMYDOECUS) NEOTRICHODECTES (N.)

N. (TRIGONODECTES) SU3FAMILY n,

N. (n. 6) N. (LAKSHMINARAYANELLA) N. (n. 7) TRICHODECTES (T.) T. (n. 5) T. (STACHIELLA) WE RNEC KODECTES n. U

LUTRIDIA • TRICHODECTINAE KELLOGG, 1896

PROTELICOLA

FELICOLA (F.)

HI F. (SURICATOECOS)

LORISICOLA (L.)

KI L. (PARADOXUROECUS)

EURYTRICHODECTES

DASYONYX (D.) "-E D. (NEODASYONYX) PROCAVIPHIUJS (P.) DASYONYGINAE KfiLER, 1938 KI P. (ICGANARIONOIEES) PRDCAVICOLA (P.)

-E P. (CONDYLOCEPHALUS)

CEBIDICOLA

EUTRICHOPHILUS EUTRICHOPHILINAE KfiLER, 1938 BO VICOLA (B.)

B. (HOLAKARTIKOS)

B. (LEPIKENTRON)

B. (n. i)

n. 2 • BOVICOLINAE KfiLER, 1938 WERNECKIELLA

n. 3

DAMALINIA (D.)

D. (CERVICOLA)

D. (TRICHOLIPEURUS)

Fig. 53 . Cladogram of the genera of Trichodectidae, with subfamily assignments. 29 T

3.3. DESCRIPTIONS OF GENERA AND SUBGENERA

3.3.1. Introduction

The generic and subgeneric descriptions below are arranged by sub- family in the order of the sequenced classification of the Trichodectidae (see section 3«2. above). Descriptions are set out in the following order: paragraph one - head, both sexes, with details of sexually-dimorphic features of antennae, if present; paragraph two - thorax, both sexes, omitting mention of the anterior setae (on the post-temporal margin) which are present in all species; paragraph three - abdomen, both sexes, with details of sexually- dimorphic features of the setae, sclerites or shape, if present; fourth paragraph - female terminalia and genitalia; fifth paragraph - male subgenital plate, terminalia and genitalia. Descriptions are given of each genus as a whole, even where subgenera are present. The descriptions of subgenera (if any are present) follow that of the genus in which they are placed, and give only subgeneric characters, so that some of the paragraphs listed above may be omitted. Characters that vary between subgenera, if mentioned in the generic description, are indicated by an asterisk (*).

Each description is followed by an indication of the host group or groups parasitised, and by any pertinent comments on the taxonomy, morphology or biology of some or all of the included species. A check- list of all species included in each genus or subgenus is also given, the names being placed in alphabetical order. Following each species name in the check-lists is an indication of the number of specimens of each sex examined in the study. Two species have not been placed, and are considered incertae sedis: Trichodectes baculus SchUmmer, 1913* Type-host: Capra hireus Linnaeus Trichodectes tigris Ponton, 1870. Type-host: Felis tigris Linnaeus These species are discussed by '.Verneck (1950). The subfamilies, genera and subgenera are keyed in section 3*4. below. 292

3.3.2. Bovicolinae

3.3.2.1. Bovicola Swing

The genus 3ovicola comprises four subgenera.

Description

Anterior of head with osculum absent or, if present, broad and shallow*; pulvinus of normal length or short and not attaining anterior margin of head; dorsal preantennal sulcus present or absent*; clypeal marginal carina not or only slightly broadened medially, or broadened to variable degree into bar with posterior and anterior margins roughly parallel, bar either straight and at right angles to long axis of head or curved and parallel to anterior margin of head*; anterolateral margin of head smoothly rounded; preantennal portion of head short, outline broadly rounded or trapezoid*. Temple margin smoothly convex or with posterior projection*, sometimes convexly produced posteriad*. Male scape expanded or not expanded*, with setal row apparently present or setae randomly scattered; flagellomeres fused in males and females; male flagellum with two b as ally-articulated 'teeth1 and interior face . not 'roughened* • Dorsum of head with more or less abundant setae, short, long or of moderate length*. Sitophore sclerite unmodified or with posterior arms extended*. Thorax with more or less abundant setae, short, long, or of moderate length, frequently longest on postero-lateral margin of pterothorax*. Abdomen oval or elongate, frequently tapering posteriorly more in the male than in the female*. Abdominal spiracles present on segments III - VIII. Abdominal setae variable*; anterior setae always present on pleura, sometimes on sterna and terga; postero-lateral setae absent. Abdominal pleural projections absent. Sclerites present at least on sterna III - VII (males) and III - VIII (females), terga II - VII (males) and III - IX (females) and pleura II - VIII; male terga v/ith posterior sclerites present or absent*. Gonapophyses with marginal setae; ventral lobe present, though sometimes not pronounced*. Gonapophyses meet ventral vulval margin acutely, not linked by sclerotised band. Ventral vulval margin not sclerotised, or sclerotised only medially; subgenital lobe absent, 293

though small median membranous projection may be present (Fig* 55)*. Genital chamber sometimes with median antero-dorsal area lacking scales or spicules*. Male subgenital plate -variable*. Pseudostyli present or absent*. Male genital opening dorsal or postero-dorsal. Male genitalia variable*.

Hosts

Bovidae, Cervidae and Camelidae (Artiodactyla).

Comments

Some species of Bovicola are parthenogenetic, males being rare or unknown. A summary of the varying taxonomic treatments of Bovicola, its sub- genera and synonyms, is presented in Table VII.

Subgenus Bovicola Swing

Bovicola Ewing, 1929: 193- Type-species: Trichodectes caorae Gurlt, by original designation. Bovidoecus Bedford, 1929: 518. Type-species: Pedicuius bovis Linnaeus, by original designation. [ Synonymy by Bedford, 1932a: 356] Rhabdooedilon Keler, 1938: 453* Type-species: Trichodectes longicornis Hitzsch, by original designation. [Synonymy by Y/erneck, 1950: 59]

Description Clypeal marginal carina not broadened medially, or more or less broadened into bar with posterior margin straight or matching curvature of osculum. Temple margin smoothly convex, lacking projection on postero-lateral angle, not convexly produced posteriad to great extent. Male scape not expanded or only slightly expanded. Dorsum of head with setae short or of moderate length, of greater abundance anteriorly than posteriorly. Sitophore sclerite unmodified.

Thorax with lateral and dorsal setae long and of moderate length, sometimes abundant and numerous on disc of prothorax and pterothorax, otherwise less abundant and sparsely scattered on disc of pterothorax with only two setae present on disc of prothorax; setae present along lateral margins and posteriorly (dorsally) on prothorax and pterothorax; posterior setal row of prothorax marginal, with median gap present or absent; posterior setal row of pterothorax submar.ginal, with median gap

V 294

Fig. 54. Bovicola (n.l) hemitragi male: abdominal terga I and II, Fig. 55• Bovicola (B. ) .jellisoni female: terminalia, ventral. Fig. 56. Bovicola (L.) breviceps female: terminalia, ventral. Fig* 57* Bovicola (H.) crassipes female: gonapophysis, ventral. Fig. 58. Bovicola (B.) caprae female: gonapophysis, ventral. 295

absent, row incorporating two very long setae between postero-lateral and postero-median angles or, if setae generally abundant on thorax, postero- lateral setae of pterothorax longer than others. Abdominal setae short, long or of medium length; setal bases, at least of setae of posterior setal row on sterna and terga, enlarged, clearly with doubled margins. Pregenital sclerites present on sterna and terga (v/here present) of all segments, except sometimes tergum I and (independently) tergum VTII of males; terga of males, at least of segments IV - VI, with both anterior and posterior sclerites. Gonapophyses with lobe rectangular, acute, rounded or not pronounced; marginal setae confined to lobe, long. Ventral vulval margin not sclerotised; convex, biconvex with median indentation, or convex with small median membranous projection (Fig. 55) J margin smooth or spinose. Postgenital area lacking spinose patch. Genital chamber with antero- median dorsal area lacking spicules, scales or other decoration, either very narrow and strongly-defined or wide and ill-defined, or with very narrow longitudinal fold. Male subgenital plate variable; stemites 1711 and IX present, fused to s.g.p.r., sternite VIII absent or, if present, fused or not fused to s.g.p.r. (Figs 59, 61). Pseudostyli absent (Fig. 62) or, if present, setose and lobulate (Figs . 60, 61 ). Basal apodeme very concave anteriorly, the sides of the concavity frequently being parallel, though sometimes obscure. Parameres with broad basal flange or block; sometimes very reduced. Basiparameral sclerites present and fused, or absent. Meso- meres, if fused apically, forming very narrow arch lacking median extension; otherwise mesomeres not fused, sometimes very reduced and obscure. Male genitalia depicted in Figs 65, 66, 67, 68.

Hosts Bovidae and Cervidae (Artiodactyla).

Comments Some of the species in this subgenus are parthenogenetic•

Species included alpinus Keler, 1942 (5c, 3?) bovis (Linnaeus, 1758) (76, 137?) canrae (Gurlt, 1843) (c50c, c50$) 296 Figs 59 - 64 (facing): Fig. 59. Bovicola (B. caprae male: abdomen. Fig. 60. Bovicola (B. caprae male: terminalia. Fig. 61. Bovicola (B. bovis male: terminalia. Fig. 62. Bovicola (B. concavifrons male: terminalia. Fig. 63. Bovicola (H. crassipes male: terminalia. Fig. 64. Bovicola (n.1) hemitragi male: subgenital plate (setae omitted)• 297 298

concavifrons (Hopkins, 1960) [Raised from syn. with longicornis (ITitzsch) ] (22, 98?) .jellisoni Emerson, 19&2 (102, 10$) limb at us (Gervais, 1844) (c506, c65$) longicornis (ITitzsch, 1818) (44$) oreamnidis (Hopkins, 1960) (Holotype 2) ovis (Schrank, 1781) (592, 64$) tarandi (Mj8berg, 1910) (2$, 7 nymphs) tibialis (Piaget, 1880) (clOO?)

Subgenus Holakartikos Keler Gen. rev., Stat. n.

Holakartikos Keler, 1938: 461. Type-species: Trichodectes oilosus Piaget (nec Giebel) [= Trichodectes crassipes Rudow] , by original designation.

Description Anterior of head with osculum absent; pulvinus very short, not attaining anterior margin of head; dorsal preantennal sulcus absent, though ventral preantennal sulcus sometimes present; clypeal marginal carina not always pronounced and not, or only slightly, broadened medially; preantennal portion of head very short, outline smoothly and shallowly rounded. Temple margin smoothly convex, lacking projection on postero- lateral angle, convexly produced posteriad. Male scape very slightly expanded, with setae randomly scattered. Dorsum of head with abundant setae of moderate length; temple with long postero-lateral marginal setae. Sitophore sclerite unmodified. Thorax with abundant setae, long and of moderate length, present on margins and disc of prothorax and pterothorax; setae longest on the rounded posterolateral angles of prothorax and pterothorax. Abdomen with long setae of posterior setal row, and shorter anterior setae, present on sterna, terga and pleura (where present) of all segments (Fig. 63). Pre-genital sclerites sometimes very faint, present on sterna and terga (where present) of all segments except tergum I and sternum II; male terga lacking posterior sclerites. Gonapopliyses with broadly rounded lobe smoothly continuous with ventral margin; marginal setae long, present all along ventral margin, 71 65 70

Figs 65 - 71. Bovicola spp.: male genitalia. (65) B. (B.) bovis; (66) B. (B.) caprae; (67) B. (B.) concavifrons; (68) B. (B.) concavifrons, detail of right paramere (p) and mesomere (m); (69) B. (H.) crassipes; (70) B. (n.1.) ro vo bemitragi; (71) B- (L-) breviceps. vo 300

including lobe. Ventral vulval margin not sclerotised; produced into three weakly-developed lobes. Postgenital pleural area with patch of short, spine-like setae. Genital chamber lacking dorsal non-ornamented area or fold. Male sub genital plate with s.g.p.r. not joining sternites VII and VIII, and sometimes failing to contact either or both; sternites sometimes very faintly sclerotised, obscure; s.g.p.r. v/ith broad lateral flange - on VIII and IX (Fig. 63 ). Pseudostyli absent. Male genital opening dorsal. Basal apodeme long, not concave anteriorly. Parameres long, slender, v/ith basal block and flange. Basiparameral sclerites absent. Mesomeres fused apically, with median extension present (see comments below). Male genitalia depicted in Fig. 69.

Hosts

Bovidae (Artiodactyla).

Comments

The only included species is not known to be parthenogenetic. Y/erneck (1950) failed to depict the median extension of the mesomeral arch, and considered it absent. Holakartikos was considered a synonym of Bovicola by Werneck (1950) and Emerson & Price (1981); a more extensive history of the variations in status of this subgenus is presented in Table VII. Species included crassipes (Rudow, 1866) (246, 31?)

Subgenus Lepikentron Keler Gen. rev., St at. n.

Lepikentron Keler, 1938: 452. Type-species: Trichodectes breviceos Rudow, by original designation.

Description Anterior of head with osculum absent; pulvinus very short, not attaining anterior of margin of head; dorsal preantennal sulcus absent; clypeal marginal carina not pronounced, not broadened medially; preantennal portion of head shorter in male than female, outline broadly and smoothly rounded. Temple margin smoothly convex, lacking projection on postero- lateral angle, not convexly produced posteriad to great extent. Male 301

scape expanded, with setal row apparently present, though setae may be scattered randomly. Dorsum of head with setae of moderate length, slender; setae sparsely distributed, more abundant anteriorly than posteriorly. Sitophore sclerite unmodified. Thorax with lateral and dorsal setae slender, long and of moderate length; setae present postero-laterally and posteriorly on prothorax and along lateral margins and posteriorly (dorsally) on pterothorax; posterior setal row of prothorax submarginal, sparse, with large median gap; posterior setal row of pterothorax submarginal, with small median gap, setae shorter medially than laterally, with two long setae laterally; pair of setae, widely spaced, present on disc of prothorax dorsally; setae not present on disc of pterothorax.

Abdominal setae of moderate length, slender; anterior setae never present on sterna and terga. Pre-genital sclerites present on terga II - VII (males) and terga III - IX (females) and sterna III - VII (males) and III - VIII (females); male terga III - VII with both anterior and posterior sclerites, though the posterior elements may be very faintly sclerotised and difficult to see. Gonapophyses with small, pronounced lobe and broad tapering.spur (Fig. 56); marginal setae confined to lobe. Ventral vulval margin sclerotised medially; shallowly convex or biconvex (Fig. 56). Post- genital pleural area lacking spinous patch. Genital chamber lacking median non-ornamented area or fold.

Male sub genital plate with stemite VTI fused to s.g.p.r. and sternites VIII and IX absent; s.g.p.r. broad. Pseudostyli absent. Male genital opening postero-dorsal. Basal apoderne slightly longer than parameres, convex anteriorly. Parameres long, slender. Basi- parameral sclerites absent. Mesomeres not apically fused; each with median desclerotisation, and appearing as tv/o rods (Fig. 71) •

Hosts

Camelidae (Artiodactyla).

Comments

Only one male of the single included species is known and the :species ipay be parthenogenetic. The subgenus was treated as a synonym of Bovicola by '.Vemeck (1950) 302

and Emerson & Price (1981); a more extensive history of the variations in status of this subgenus is presented in Table VTI.

Species included breviceos (Rudow, 1866) (Id, 24?)

Subgenus n. 1

Type-species: Trichodectes hemitragi Cummings

Description

Anterior of head with osculum absent; pulvinus very short, but attaining anterior margin of head; dorsal preantennal sulcus absent; clypeal.-marginal carina broadened medially into straight bar with posterior margin slightly irregular; anterolateral margin of head smoothly rounded; preantennal portion short, with outline rounded, almost straight anteriorly. Temple margin convex laterally, straight posteriorly, with short posteriorly-directed projection on postero-lateral angle bearing two setae. Male scape slightly expanded, with setae randomly scattered. Dorsum of head with abundant setae of moderate length. Sitophore •sclerite with posterior arms extended (Pig. 36). Thorax with dorsal and lateral setae abundant, long or of moderate length, present marginally and on disc of prothorax and pterothorax; longest setae present postero-laterally on pterothorax. Abdomen tapering more acutely in male than in female. Abdomen with posterior setal row comprising long setae on sterna, terga and pleura, anterior setae shorter; anterior setae present on sterna, terga and pleura (where present) of all abdominal segments except sometimes tergum I; male tergum II with curved row of 3 - 4 long stout setae on each side, modified from posterior setal row, these setae being linked by a curved sclerite (modified tergite) (Pig. 54). Pre-genital sclerites present on terga II - VTI or VIII (males) and II - IX (females) and sterna II - VTI (males) and II - VIII (females); male terga lacking posterior sclerites. Gonapophyses with broadly rounded lobe smoothly continuous with ventral margin; marginal setae long, confined to lobe. Ventral vulval margin not sclerotised; convex. Postgenital pleural area lacking spinose patch. Genital chamber lacking dorsal non-ornamented area or fold. 303

Male segment IX produced posteriorly into narrow, sclerotised extension; subgenital plate tapering characteristically, comprising stemites VT and VTI linked by s.g.p.r. (Fig. 64). Pseudostyli absent. Ivlale genital opening dorsal. Basal apodeme not as long as pa,rameres, not concave anteriorly. Parameres fused basally, long and tapering to pointed apices. Basiparameral sclerites absent. Mesomeres absent, or represented by very short sclerites, not apically fused. Male genitalia depicted in Fig. 7&

Hosts

Bovidae (Artiodactyla).

Comments Neither of the two included species is known to be parthenogenetic.

Species included hemitragi (Cummings, 1916) (20d, 26$) multisninosus Emerson & Price, 1979 (80, 11$)

3.3.2.2. -Genus n. 2

Type-species: Bovicola sedecimdecembrii Eichler

Description

Anterior of head v/ith osculum absent, though pulvinus attaining margin; margin anteriorly to pulvinus membranous, hyaline; dorsal preantennal sulcus present; clypeal marginal carina slightly broadened medially; anterolateral margin of hea,d smoothly convex; preantennal portion of head short or longer, but not as long as postantennal portion, outline broadly rounded, though slightly truncate anteromedially. Temple margin broadly and smoothly convex. Male scape expanded, with setae randomly scattered; flagellomeres fused in males and females; male flagellum with two or three basally-articulated 'teeth1; male flagellum uot 'roughened' on interior face. Dorsum of head with numerous setae of medium length. Sitophore sclerite unmodified. Thorax v/ith lateral and dorsal setae long and of moderate length; setae present along lateral margins and posteriorly (dorsally) on prothorax and pterothorax; posterior setal row on prothorax marginal, with median gap; posterior setal row of pterothorax submarginal, with median gap 304

Figs 72 - 74» n.2. sedecimdecembrii. (72) female terminalia, ventral; (73) male subgenital plate (setae omitted); (74) male genitalia. Figs 75 - 76. ITerneckiella ecrni. (75) male terminalia; (76) male genitalia. Figs 77 - 78. n.3« traguli. (77) female head, dorsal; (78) male genitalia 305

absent; pterothorax with posterior setal row incorporating two very long setae between postero-lateral and postero-median angles; pair of setae, widely spaced, present on disc of prothorax dorsally; scattered setae sometimes present posteriorly on disc of pterothorax dorsally. Atria of thoracic spiracles very large. Abdomen oval. Abdominal spiracles present on segments III - VIII. Abdomen v/ith setae short and of medium length; anterior setae present on sterna, terga and pleura; postero-lateral setae absent. Abdominal pleural projections absent. Sclerites present on sterna, terga and pleura (where present) of all pre-genital abdominal segments except tergum I; male terga V, VI and VII with anterior and posterior sclerites. Gonapophyses broad, truncate; setae present along postero-median margin; ventral lobe absent (Fig. 72)* Gonapophyses meet ventral vulval margin acutely, not linked by sclerotised band. Ventral vulval margin not sclerotised; very short, more or less straight; subgenital lobe absent. Male subgenital plate with sternites VII and IX fused to s.g.p.r., sternite VIII present but not fused to s.g.p.r.; s.g.p.r. heavily sclerotised, widest on sternum VTII (Fig. 73). Pseudostyli present, large, broadly triangular (Fig. 73). Male genital opening dorsal. Parameres separate, rod-like, fused to mesomeral arch. Basiparsmeral sclerites absent. Mesomeres fused apically; median extension absent. Male genitalia depicted in Fig. 74.

Hosts Bovidae (Artioaactyla).

Species included sedecimdecembrii sedecimdecembrii (Eichler, 1946) Comb. n. from Bovicola. (5

3.3.2.3. ".Verne cki ell a Eichler Gen. rev.

'Verneckiella Eichler, 1940: 160. Type-species: Trichodectes eaui, by original designation. 306

Description Anterior of head with osculum absent; pulvinus not attaining margin; dorsal preantennal sulcus present; clypeal marginal carina slightly broader medially than laterally, or not broadened; anterolateral margin of head smoothly convex; preantennal portion of head not long, outline broadly rounded, sometimes slightly flattened anteriorly. Temple margin convex or rectangular. Male scape expanded, with setae randomly scattered; flagellomeres fused in males and females; male flagellum with two basally-articulated 'teeth'; male flagellum not 'roughened' on interior face. Dorsum of head with abundant short setae. Sitophore sclerite unmodified.

Thorax with lateral and dorsal setae short and of medium length; setae present along lateral margins and posteriorly (dorsally) on prothorax and pterothorax; posterior setal row on prothorax marginal, with median gap; posterior setal row on pterothorax marginal or submarginal, with no median gap; posterior setal row of pterothorax incorporating two very long setae with intervening shorter setae between postero-lateral and postero-median angles; prothorax with seta or setae on disc laterally (dorsally).

Abdomen elongate-oval. Abdominal spiracles present on segments III - VIII. Abdomen with setae short and of moderate length; anterior setae frequently present on sterna and terga, always present on pleura; postero-lateral setae absent. Abdominal pleural projections absent. Sclerites present on abdominal pleura II VII and sterna and terga of at least abdominal segments II - VTI; male terga with single sclerites only.

Gonapophyses broad, truncate, with median faces almost parallel to one another; marginal setae of moderate length; ventral lobe absent. Gonapophyses meet ventral vulval margin acutely, not linked by sclerotised band. Ventral vulval margin not sclerotised, very short, straight; sub genital lobe absent.

Male subgenital plate with stemite VTI present and fused to s.g.p.r., stemites VIII and IX absent; s.g.p.r. sinuate and broadest on sternum VIII (Fig. 75). Pseudostyli present, small, simple setose lobes. Male genital opening postero-dorsal. Parameres long, straight or 307

flared and twisted medially, sometimes fused basally. Basiparameral sclerites absent. Mesomeres fused or almost fused into pentagonal mesomeral arch with median extension absent; mesomeres broadest basally (external to b.a.l.s.) and more or less broad distally; mesomeres projecting basally between b.a.l.s. to contact pararneres. Male genitalia depicted in Pig. 76.

Hosts

Eauidae (Perissodactyla) and Bovidae (Artiodactyla).

Comments

Some species are parthenogenetic, the males being unknown. 7/erneckiella was considered a subgenus of Damalinia by Hopkins (1949) and a synonym of Bovicola by 7/emeck (1950); it is here raised from synonomy with Bovicola. A more extensive history of the variations in status of this genus is presented in Table VII. The genus was revised by Moreby (1978).

Species included aspilooyga (Y7erneck, 1956) Comb. n. from,Bovicola (9^, 11?) equi (Denny, 1842) Comb. n. from Bovicola (3<5, clOO?) fulva (Emerson & Price, 1979) Comb. n. from Bovicola (4d, ?1?) neglecta (Kgler, 1942) Comb. n. from Bovicola (5d, 6?) ocellata (Piaget, 1880) Comb. n. from Bovicola (17?) zebrae Moreby, 1978 Comb. rev. from Bovicola (15* 2?) zuluensis ('.Yerneck, 1950) Comb. n. from Bovicola (13c, 13$)

3.3.2.4. Genus n. 3

Type-species: Damalinia traguli Xlerneck

Description

Anterior of head with osculum present, broad; dorsal preantennal sulcus present; clypeal marginal carina broadened medially into less heavily sclerotised dorsal sclerite, which is broad, posteriorly convex and with median posterior projection (Pig. 77); anterolateral margin of head straight or slightly sinuate; preantennal portion of head as long as its maximum width, outline trapezoid. Temple margin convex or 308

rectangular. Male scape expanded, with setal row present and comprising at least four setae; flagellomeres fused in males and females; male flagellum with two basally-articulated 'teeth1. Dorsum of head with setae short posteriorly and of moderate length anteriorly. Sitophore sclerite unmodified.

Thorax with dorsal setae short or of moderate length; prothorax and pterothorax with marginal or submarginal posterior setal row$ the longest setae being posirerolaterally except in the male, which has a pair of long setae medially on the posterior row of the pterothorax; male with setal patch centrally on disc of pronotum, hut no other setae present on disc of either sex. Abdomen elongate, with male tapering to more acute posterior angle than female. Abdominal spiracles present on segments III - VIII; atria oblate-spheroids, very large. Abdomen with setae of moderate length, the longest being those comprising the pleural posterior setal row,- particularly of the posterior pleura; anterior setae present on all pleura, but not sterna or terga; postero-lateral setae absent. Abdominal pleural projections absent. Sclerites present on sterna, terga and pleura (where present) of all abdominal segments except tergum I, which is reduced and obscure; male terga III - VI with anterior and posterior sclerites. Gonapophyses broad medially, tapering smoothly distally; ventral margin with long, abundant setae; ventral lobe absent. Gonapophyses meet ventral vulval margin acutely, not linked by sclerotised band. Ventral vulval margin not sclerotised; convex; subgenital lobe absent; marginal spines present, though difficult to see. Female genital chamber v/ith dorsal wall lacking spicules over narrowly triangular area antero- medially. Male subgenital plate v/ith sternites IX and VIII linked by s.g.p.r., but sternite VTI not attached. Pseudostyli present, short, conical; median ventro—posterior projection also present, longer than pseudostyli. Male genital opening postero-dorsal. Basal apodeme acuminate apically. Parameres broad, triangular, poorly-sclerotised, asymmetrically deflected (may be artifact of preparation, though deflected the same way in all specimens seen). Basiparameral sclerites absent. Mesomeres fused 309

apically, median extension absent; mesomeral arch fused to b.a.l.s. about one-third the length of the basal apodeme anteriad from the posterior end. Male genitalia depicted in Fig. 78 •

Hosts Tragulidae (Artiodactyla).

Species included traguli (Jerneck, 1950) Comb. n. from Damalinia (706, 75?)

3.3.2.5. Damalinia Mjflberg

The genus Damalinia comprises three subgenera.

Description

Anterior of head v/ith osculum present, narrow or broad, deep or shallow, or osculuin absent, in which case head as described below for D.(T.) conectens*; dorsal preantennal sulcus present or absent*; clypeal marginal carina more or less broadened medially and of variable form*; anterolateral margin of head straight, slightly concave, slightly sinuate or convex; preantennal portion of head of variable length, outline triangular, trapezoid, rectangular or rounded*. Temple margin smoothly convex, sometimes convexly produced posteriad*, with posterolateral angle sometimes developed laterally or v/ith small posterior rounded projection*. Male scape expanded, with setal row present or setae randomly scattered*; flagellomeres fused in males and females; male flagellun v/ith two or three basally-articulated 'teeth1* and interior face serrate or 'roughened'. Dorsum of head v/ith setae sparse or more or less abundant, short or of moderate length, frequently longer along the anterolateral margins and across the clypeus than elsewhere. Sitophore sclerite unmodified.

Thorax with dorsal and marginal setae short, long or of moderate length, frequently longest on posterolateral margin of pterothorax. Prothorax with setae sparse or absent on anterolateral margin; posterior setal row marginal, though directed onto disc medially and median setal pair sometimes isolated, row more or less sparse, with median gap between setae (other than isolated median pair) present, sometimes wide*; single seta frequently present on dorsal disc anterolaterally. Pterothorax 310

with posterior setal row marginal or submarginal, sometimes irregular or 'doubled1*, median gap present or absent, posterior setal row incorporating two very long setae with intervening shorter setae between postero-lateral and postero-median angles; setae absent from disc. Abdomen oval, elongate, or very elongate and narrow*. Abdominal spiracles present on segments III - VIII. Abdominal setae short or of moderate length, frequently longer on pleura than on sterna and terga; anterior setae present on all-pleura except, occasionally, pleurum II, rarely on sterna and terga; postero-lateral setae absent. Abdominal pleural projections absent. Pregenital sclerites present on sterna, terga and pleura (where present) of all segments except tergum I, sometimes absent or very small on pleura (Fig.. 54)*; male terga with or without posterior sclerites*. Gonapophyses variable; ventral margin without rounded lobe, but sometimes with hook-shaped projection*; marginal setae present. Gonapo- physes meet ventral vulval margin acutely, not joined by sclerotised band. Ventral vulval margin sclerotised or not sclerotised; subgenital lobe present or absent*. Dorsal margin of vulva and post-vulval area with or without pointed scales*. Common oviduct, at branching point, with or without collar (see subgenus Cervicola)*. Male subgenital plate variable, sternites VII and VIII always being present and fused to s.g.p.r., sternite IX sometimes modified. Pseudostyli absent or, if present, of variable form*. Posterior margins of tergum IX sometimes greatly expanded*. Male genital opening postero-dorsal or dorsal. Male genitalia variable*.

Hosts

Bovidae and Cervidae (Artiodactyla).

Subgenus Damalinia Mjflberg Damalinia MjSberg, 1910: 69. Type-species: Trichodectes crenelatus Piaget, by monotypy.

Description Anterior of head with osenium present, narrow or broad, deep or

shallow; dorsal preantennal sulcus present; clypeal marginal carina 311

Fig. 79 Damalinia (D.) crenelata male: abdomen. Fig. 80 Damalinia (D.) theileri male: terminalia. 312

broadened medially either into more or less developed simple bar with posterior margin straight or concave, or into more or less broad •U'- shaped sclerite, or into broad, heavily-sclerotised margin of deep osculum; anterolateral margin of head convex or slightly sinuate, in the latter case slightly concave at junction of margin and clypeofrontal sulcus and convex anteriorly; preantennal portion longer or shorter than posterior portion, outline triangular, trapezoid or rounded, sometimes with slight protuberances on either side of osculum. Temple margin smoothly convex, slightly produced posteriad, sometimes with postero- lateral angle developed laterally, or with small posterior rounded projection. Male scape with setae randomly scattered; male flagellum with two basally-articulated 1 teeth1 . Dorsum of head with setae more or less abundant. Pterothorax with posterior setal row sometimes irregular or doubled*. Abdomen oval or elongate, sometimes very narrowly elongate. Abdominal setae present anteriorly on all pleura, occasionally on terga and sterna, but only laterally and as irregularity or 'doubling1 of posterior setal row. Pleurum II with sclerite extending broadly or narrowly onto sternum II and sometimes tergum II, frequently 1 crowding1 sternite or tergite II (Pig.79 ); pleurites not reduced in size or absent; tergum I lacking sclerite; male terga V and VI (at least) with both anterior and posterior sclerites. Gonapophyses variable, sometimes hook-shaped, though lacking distal spur, more frequently obtuse, sometimes with ventral (median) margin concave or convex; ventral margin with setae long or of moderate length, abundant, setae sometimes also present on anterior margin. Ventral vulval margin not sclerotised, sometimes short, straight or convex; subgenital lobe absent, though posterior margin of sternum VTI sometimes developed into two spikes (Pig. 8l). Dorsal margin of vulva and post- vulval area usually without pointed scales. Common oviduct without ' collar1 . Male subgenital plate with sternites VTI, VIII and IX fused to s.g-.p.r., sternite IX and postgenital sclerite sometimes fused; s.g.p.r. more heavily sclerotised than stemites (Pig.79 )• Pseudostyli absent or, if present, long and broad, parallel-sided or with basal constriction 313

Pigs 81 - 86. Damalinia (Damalinia) spp. (81) D. theileri female: terminalia, ventral (setae omitted); (82) D. baxi female: head, dorsal; (83) D. app endi culat a male: terminalia; (84) D. orientalis male genitalia; (85) D. neotheileri male genitalia; (86) D. crenelata male genitalia. 314

(Figs , 79, 80), or long and narrow (Fig. 83)* Posterior margin of tergum IX not greatly expanded. Parameres more or less broad, sometimes fused together. Basiparameral sclerites absent. Mesomeres unfused apically, sometimes fused to b.a.l.s. in characteristic manner (Fig. 85 ) or to parameres, basally or in entirety, in the latter case apparently absent. Sndophallus lacking spicular patch (c.f. subgenus Tricholioeurus). Iiale genitalia depicted in Figs , 84 , 85, 86.

Hosts

Bovidae (Artiodactyla).

Comments

Emerson & Price (1982) distinguish their new species orientalis (described in Bovicola) from the very similar species thompsoni Bedford on the following grounds: "The female of 3. orientalis is smaller than that of B. thompsoni and the lateral margins of the forehead are even [sic] so slightly indented for _B. orientalis and always even for j3. thomosoni; the median plates on tergites II - VIII are of different shapes for the two species; the chaetotaxy of terminal abdominal segments is different, with each gonapophysis having at least 20 median and anterior setae for 13. thomosoni•<[ orientalis having, according to the preceeding description, 12-17 setae] ; and the posterior margin pf the temple of _B. thompsoni has small projections that are not present for 13. orientalis. The male of 13. thompsoni is unknown.". Five male and seven female paratypes of 13. orientalis were examined in this study, together with a further eight males and eight females from the same host (Capricornis crispus swinhoei) not examined by Emerson 8c Price when they prepared their description of orientalis, and three females of thompsoni, including the holotype. Treating the supposed distinguishing characters in order, the three specimens of thompsoni are larger than any of the females from _C. jc. swinhoei; whilst none of the specimens of thompsoni has an indentation on the forehead (at the junction of the clypeo-frontal sulcus v/ith the margin), not all of the specimens from Capricornis swinhoei have either; the shapes of abdominal tergites II - VII are not significantly different in the two species; the chaetotaxy of the terminal abdominal segments is not different, and no specimen of thompsoni has more than 17 setae on the median and anterior margins of the gonapophyses; the small 315

projections of the posterior temple margins are present in all specimens of orientalis. It seems, therefore, that the differences between the two species are twofold: host (orientalis being described from Capricornis crisous swinhoei and thompsoni being known only from Capricornis sumatrensis sumatrensis), and size. The biological significance of the latter character is not clear, and the two species may be found to differ in other characters not so far discovered. Until a larger sample can be examined, collected from more localities, no taxonomic action is taken to reduce the rank or synonymise orientalis, though on the basis of the information so far available the species probably should not stand.

Species included adenota (Bedford, 1936) Comb. n. from Bovicola. (39c, 35$) app endi culat a (Piaget, 1880) (19b, 25?) baxi Hopkins, 1947 (l6d, 24?) chorleyi (Hopkins, 1941) (21o, 19?) crenel at a (Piaget, 1880) (27d, 20$) dimorpha (Bedford, 1939) Comb. n. from Bovicola. ( $ of type series) fahrenholzi (Eichler, 1949) Comb. n. from Tricho1ipeurus harrisoni (Cummings, 1916) (3 c, 3?) hilli (Bedford, 1934) Comb. n. from Bovicola. (42d, 27?)

neotheileri Emerson 8c Price, 1971 (Id, 6$) orientalis (Emerson 8c Price, 1982) Comb. n. from Bovicola (108b, 135?) ornata 7/emeck, 1957 (Holotyped ) pelea (Bedford, 1934) Comb. n. from Bovicola

semitheileri Emerson 8c Price, 1971 (Holotype c , allotype? ) theileri Bedford, 1928 (26, 4?) thompsoni (Bedford, 1936) . Comb. n. from Bovicola (3?)

Subgenus Cervicola Keler Gen. rev. Stat. n.

Cervicola Keler, 1934: 263. Norn. nud.

Cervicola Kller, 1938: 460. Type-species: Trichodectes tibialis Keler (nec Piaget) [ = Trichodectes meyeri T aschenber.-], by original designation. 316

87

88

89

Figs 87 - 90. Damalinia spp. (87) D. (T.) elongata female: terminalia, ventral; (88) D. (C.) me.yeri female: gonapophysis, ventral; (89) D. (C.) hendrickxi female: gonapophysis, ventral; (90) D. (C.) martinaglia male: scape. 317

Description Anterior of head with osculum present, narrow or broad, deep or shallow; dorsal preantennal sulcus present; clypeal marginal carina broadened medially, either into simple bar with posterior margin straight or convex but occasionally with median posterior projection, or into longer posteriorly-developed sclerite with posterolateral angles more or less convex and more or less pronounced median posterior projection; anterolateral margin of head straight or slightly sinuate, in the latter case slightly concave at junction of margin, and clypeofrontal sulcus and convex anteriorly; preantennal portion of head as long as or shorter than post antennal portion, outline triangular, trapezoid or rounded. Temple margin smoothly convex, slightly produced posteriad. Male scape with setae randomly scattered; male flagellum with two basally—articulated 'teeth1. Dorsum of head with setae more or less abundant, sometimes less so posteriorly than anteriorly. Pterothorax with posterior setal row single. Abdomen oval or elongate. Abdominal setae present anteriorly on all pleura except, occasionally, pleurum II, but may be very short, fine and difficult to see; anterior setae never on sterna and terga. Pleurum II never with sclerite extending onto sternum II; pleurites not reduced in size or absent; male terga V and VT (at least) with both anterior and posterior sclerites. Gonapophyses hook-shaped, apex of curved portion acute or rounded, sometimes with distal (dorsal) spur (Pigs. 88, 89). Gonapophyses with setae long or of moderate length on posterior margin and sometimes on apex of 'hook', smaller setae sometimes present on anterior margin of 'hook'. Ventral vulval margin not sclerotised. Subgenital lobe absent. Dorsal margin of vulva and post-vulval area with pointed scales. Common oviduct at branching-point with folded and more or less apparent »collar', sometimes partially sclerotised and refracting light when viewed in phase-contrast or bright field transmitted light. Male subgenital plate with sternites VII , VIII and IX fused to s.g.p.r., though sternite IX sometimes not complete, s.g.p.r. sometimes not attaining posterior margin of segment IX, perisetal gaps sometimes absent; s.g.p.r. more heavily sclerotised than sternites. Pseudostyli 318

Figs 91 - 94- Damalinia male terminalia* (91) (£•) natalensis (ventral, setae omitted); (92) D. (C.) martinaglia (ventral, setae omitted); (93) D. (T.) indica (setae omitted); (94) D. (T.) aepycerus. 319

absent or, if present, apically pointed and more or less broad (Pig. 92) or apically rounded and very narrow (Pig. 91)* Posterior margins of male tergum IX not greatly expanded. Parameres broad or narrow, larger or smaller than mesomeres, may be reduced to small discs, in which case mesomeres absent; parameres fused or unfused; apices sometimes widely divergent. Basiparameral sclerites present or absent. Mesomeres absent or, if present, not fused apically, nor fused to parameres or b.a.l.s.. Endophallus lacking spicular patch (c.f. subgenus Tricholioeurus Male genitalia depicted in Pigs 98, 99, 100.

Hosts

Bovidae and Cervidae (Artiodactyla).

Comments

Cervicola was treated as a synonym of Damalinia by 7/erneck (1950), and of both Damalinia and Bovicola by Hopkins & Clay (1952, pp 102 and 67 respectively). The history of the variations in status of Cervicola is presented in Table "VTI.

Species included annectens Hopkins, 1943 Comb. rev. from Tricholioeurus. (212, 25?) forficula (Piaget, 1880) (52, 8?) hendrickxi Hopkins, 1947 (42, 6?) hopkinsi Bedford, 1936 (152, 28$) lerouxi (3edford, 1930) Comb. n. from Tricholioeurus. (9$) maai Emerson & Price, 1973 (Kolotypeo, allotype$) martinaglia (Bedford, 1936) (342, 27$) meyeri me:/eri (Taschenberg, 1882) (52, 54$) meyeri hydropotis (Dobroruka, 1975) Comb. n. from Cervicola meyeri sika (Dobroruka, 1975) Comb. n. from Cervicola muntiacus (Seguy, 1948) (122, 12$) natal ens is Emerson, 1964 (22, 2$) reduneae (Bedford, 1929) Comb. n. from Tricholioeunis. [Raised from subsp. of D. trabeculael (282, 29$) trabeculae (Bedford, 1929) Comb. n. from Tricholipeurus. (102, 10$) ugandae (7/emeck, 1950) Comb. n. from Tricholioeurus. [Raised from subsp. of _D. trabeculae] (142 , 13 $ ) Unless otherwise stated, all species were previously considered as placed in Damalinia s. str.. 96 97

100 98

Figs 95 ~ 100* Damalinia male genitalia. (95) b. (t.) victoriae (96) D. (t.) indica; (97) D. (t.) aep.yceras; (98) D. (C.) reduncae (99.) (£•) hopkinsi; (100) D. (c.) meyeri. 321

Subgenus Tricholioeurus Bedford Stat. n.

Tricholineurus Bedford, 1929: 514. Type-species Tricho 1 it>eurus aeuycerus Bedford, by original designation.

Description Anterior of Jiea-d variable, one of two types: a) Osculum absent; pulvinus short, not attaining anterior margin of head; dorsal preantennal sulcus absent; clypeal marginal carina insignificant, not, or only slightly, broadened medially; anterolateral margin of head straight posteriorly, convex anteriorly; preantennal portion of head as long as postantennal portion, outline rounded anteriorly (D.(T.) conectens only) b) Osculum present, rarely deep; dorsal preantennal sulcus present; clypeal marginal carina broadened medially either into simple bar with posterior margin straight or convex, or into longer posteriorly-dveloped 'U'- or 'W-shaped sclerite, with posterolateral angles more or less acutely convex, and frequently with more or less pronounced median posterior projection; anterolateral margin straight, slightly sinuate, or slightly concave; preantennal portion of head longer or shorter than postantennal portion but not short, outline trapezoid or rectangular. Temple margin smoothly convex, more or less convexly produced posteriorly. Male scape with setal row frequently present, comprising.four or five setae, though setae sometimes more or less randomly scattered; male flagellum with two or three basally-articulated 'teeth'. Dorsum of head with setae sometimes sparse. Prothorax with median gap of posterior setal row sometimes almost the width of the posterior margin. Pterothorax with posterior setal row single. Abdomen usually very elongate, narrow. Abdominal setae present anteriorly on all pleura except, occasionally, pleurum II, but may be very short, fine and difficult to see; anterior setae never on sterna and rarely on terga. Pleurum II never with sclerite extending onto sternum II; pleural sclerites frequently reduced to small anterior plate or absent; tergum I sometimes lacking sclerite; male terga lacking posterior sclerites, or with both anterior and posterior sclerites present 322

on at least segments V and VI (though may be present on any segments up

to II - VIII). Gonapophyses variable, sometimes hook-shaped with distal spur and marginal setae on posterior (dorsal) margin of 'hook1, or long v/ith ventral margin convex, straight or sinuate (though not v/ith lobe), apically pointed or rounded; marginal setae of moderate length. Ventral vulval margin sometimes sclerotised; subgenital lobe present, variable, not marginally serrate, ventrally smooth or scaled. Dorsal margin of vulva and post-vulval area without pointed scales. Common oviduct v/ithout 'collar' . Male sub genital plate v/ith sternites VII, VIII and IX fused to s.g.p.r. perisetal gaps sometimes large, occasionally absent; s.g.p.r. heavily sclerotised and sternites sometimes very lightly scerlotised; s.g.p.r. sometimes curved or sinuate (Fig. 94)* Pseudostyli absent or, if present, variable, short and rounded or apically angular, posteriorly or medially directed, narrow or broad, sometimes fused to form single caudal projection. Posterior margins of male tergum IX frequently greatly expanded (Fig. 93). Lateral struts of basal apodeme sometimes v/ith anteposterior spur (Fig. 97). Parameres unfused or, if fused, plate apically pointed or bifurcate. Basiparameral sclerites present or absent. Mesomeres unfused apically or, if fused, symmetric or asymmetric, median extension absent or present; mesomeral aroh frequently recurved abruptly at base to contact parameres, sometimes extended between b.a.l.s.; mesomeres not fused to parameres or b.a.l.s.. Eudophallus v/ith patch of regularly-arranged and numerous spicules sometimes very apparent. Male genitalia depicted in Figs 95 , 56 , 97.

Hosts

Bovidae and Cervidae (Artiodactyla).

Comments

Damalinia (_T.) longiceps is included following the statement of Clay & Hopkins (1955) that it resembles D. (£.) soinifer Hopkins "most closely among known species". Tricholipeurus has been treated as a full genus and synonym (Hopkins, 1943) or subgenus (Hopkins, 1949) of Damalinia; the history of the variations in status of Tricholioeurus is presented in Table VII. 323

Species included aeoycerus (3edford, 1929) Comb. n. from Tricholipeurus (lb, l?) albimarginata (Y/erneck, 1936) Comb. n. from Tricholipeurus (9b, 13?) antidorcus (Bedford, 1931) Comb. n. from Tricholipeurus (lib, 18?) bedfordi (Hill, 1922) Comb. n. from Tricholipeurus (2b, 2?) clayi (Y/erneck, 1938) Comb. n. from Tricholipeurus. (14b, 12?) conectens Hopkins, 1943 . Comb. rev. from Tricholipeurus. (18b, 16?) cornuta cornuta (Gervais, 1844). Comb. n. from Tricholipeurus (20b, 27?) cornuta ourebiae Hopkins, 1943- Comb. rev. from Tricholioeurus (18 b, 17?) dorcephali (Y/erneck, 1936) Comb. n. from Tricholipeurus. (2b, 2?) elongata (3edford, 1934) Comb. n. from Tricholipeurus. (10b, 10?) indi ca (Yerneck, 1950) Comb. n. from Tricholipeurus. (65b, 60?) lineata (Bedford, 1920). Comb. n. from Tricholipeurus. (75b, 83?) lipeuroides (Megnin, 1884) Comb. n. from Tricholipeurus. (106b, 91?) longiceps (Rudow, 1866) moschatus (Emerson & Price, 1971) , Comb. n. from Tricholipeurus. (Holotypeo, dissacociated b head) pakenhami (v/erneck, 1947) .. Comb. n. from Tricholipeurus. (21b, 20?) parallela (Osborn, 1896) . Comb. n. from Tricholipeurus. (50b, 10.6?) parkeri (Hopkins, 1941) Comb. n. from Tricholipeurus. (10b, 10?) spinifer Hopkins, 1943 Comb. rev. from Tricholipeurus.(17o. 16?) victoriae Hopkins-', 1943 Comb. rev. from Tricholipeurus. (33b, 35?)

3.3.3. Eutrichophilinae

3.3.3.1. Eutrichophilus MjBberg

Eutrichoohilus MjiJberg, 1910: 71. Type-species: Eutrichophilus cercolabes M.jOberg by subsequent designation (Harrison, 1916: 21).

Description Anterior of head with osculum present or absent, but pulvinus always attaining margin; dorsal preantennal sulcus absent; clypeal marginal carina with median expansion absent or slight, or present as broad or narrow parallel-sided bar with transverse margins convex, dorsal; (102) E. maximus female: terminalia; (103) E. setosus male abdomen• 325

straight, or concave; anterolateral margin of head straight or convex; preantennal portion of head long or short; outline triangular, rounded or broadly trapezoid. Temple margin convex or with posterolateral angle apparent; temples greatly expanded posteriad (Fig. 101). Male scape expanded, with longitudinal setal row present and comprising two setae; male flagellomeres fused; female flagellomeres fused or flagellum comprising two flagellomeres; male flagellum very long, with two basally- articulated 'teeth'. Dorsum of head with setae short or long, sometimes longer anteriorly than posteriorly. Sitophore sclerite unmodified.

Thorax with dorsal setae shorter of moderate length marginally or submarginally on posterior of prothorax and pterothorax, absent from disc of both; one or two long setae on posterolateral margins of pt erothorax.

Abdomen oval and elongate. Abdominal spiracles- present on segments III - VIII. Abdominal setae short or of moderate length, with tufts of long setae on at least pleurum VIII, sometimes also pleurum VH (males) or IX (females) (Fig. 103); anterior setae present on all pleura but absent from sterna and terga; postero-lateral setae present. Abdominal pleura lacking projections dorsally or ventrally. Sclerites present on sterna, terga and .pleura (where present) of all pre-genital abdominal segments except, sometimes, tergum I, which may be very small; male terga, at least of abdominal segments V and VI, with anterior and posterior sclerites.

Gonapophyses frequently la.rge, broadly triangular or rounded, ventral margin lacking lobe but with more or less dense marginal setae which are long or of moderate length. Gonapophyses meet ventral vulval margin acutely, not linked by sclerotised band. Ventral vulval margin not sclerotised; smoothly convex, with or without median indentation or setose-projection; subgenital lobe absent. Female terminalia depicted in Fig. 102.

Male subgenital plate with sternites VII and VIII present and fused to s.g.p.r., IX absent or, if present, fused to s.g.p.r. (Fig. 103). Pseudostyli absent. Male genital opening dorsal, male segment IX posterior. Parameres long or short, narrow or broad; with basi- parameral sclerite or flange sometimes present and fused medially, thus 104

(105) E. moojeni: (106) E. guyanensis; (107) E. guyanensis, detail 327

linking parameres, but otherwise unfused. Mesomeres present, fused apically to form arch with no median extension; arch smoothly rounded, or with lateral desclerotisations, in which case median portion is straight . and at right angles to lateral portions, very poorly sclerotised and thin, or absent. Male genitalia depicted in Figs 104 , 105, 106, 107.

Hosts Erethizontidae (Rodentia).

Species included cercolabes Mjdberg, 1910. (296, 25$) comitans V/erneck, 1950 (6 c, 2$) cordiceps MjBberg, 1910 (23d, 27$) exiguus Werneck, 1950 (Holotyped , allotype $ ) guyanensis Werneck, 1950 (80, 7$) lobatus Ewing, 1936 (5d, 8$) maximus Bedford, 1939 (lid, 11$) mexicanus (Rudow, 1866) (50d, 50$) minor Mjdberg, 1910 (34d, 27?) moojeni Werneck, 1945 (3d, 3?) setosus (Giebel, 1874) (102d, 102$)

3.3-4. Dasyonyginae

3.3.4.1. Cebidicola Bedford

Cebidicola 3edford, 1936: 52. Type-species: Trichodectes armatus

Neumann, by original designation. Meganarion Keler, 1938: 465. Type-species: Trichodectes armatus Neumann, by original designation. [ Synonymy by Eichler, 1941]

Description Anterior of head with osculum present, deep; dorsal preantennal sulcus present or absent; clypeal marginal carina broadened medially into dorsal, posteriorly convex, sclerite; anterolateral margin straight, slightly convex, concave or sinuate anteriorly, more or less abruptly concave at junction with clypeofrontal sulcus, v/ith or without anterior sclerotised projection on either side of osculum (Figs 110, 111); 328

Nf-A//'^

YUZMRRFRY n

' N\ Z-' S

^YY///

Tt i

108

Figs 108, 109. Cebidicola armatus. (108) male: abdomen; (109) female: terminalia. 329

preantennal outline broadly triangular. Temple margin convex.or slightly acute and angular posterolaterally, with eyes more or less prominent (Pigs 110, 111). Male scape expanded, with setal row present and comprising two or more setae; male flagellomeres fused, with "two basally-articulated 'teeth1; female flagellomeres fused or unfused. Dorsum of head with setae short or of moderate length, sparse. Sitophore sclerite unmodified. Tarsal claws lacking ventral spines or teeth. Postcoxale absent or present, not greatly developed. Thoracic setae present dorsally only along posterior and posterolateral margins of pterothorax; setae short medially, longer laterally. Atrium of thoracic spiracle tubular or conical. Abdomen oval, sometimes tapering posteriorly more in male than female. Abdominal spiracles present on segments III - VIII. Abdominal setae short or of moderate length, longest on pleura VT - VIII; anterior setae sometimes present on pleura and laterally on sterna and terga; postero-lateral setae present, sometimes numbering more than one per site. Pleural projection present ventrally on abdominal pleurum IV, large, sclerotised. Sclerites present on all abdominal pleura, on at least abdominal terga II to VIII and at least abdominal sterna V to VTI; male terga, at least on segments V - VTI, with anterior and posterior sclerites (Fig.108). Gonapophyses broad, especially medially, though lobe absent; marginal setae long, densely crowded. Gonapophyses meet ventral vulval margin acutely, not linked by sclerotised band. Ventral vulval margin not sclerotised; sometimes expanded, otherwise smoothly convex (Fig.109); bilobed median spinose projection may be present, but subgenital lobe absent. Median longitudinal sclerite sometimes present on female sternum VIII (Fig. 109). Male subgenital plate with sternite VII fused to s.g.p.r., VTII fused or not fused to s.g.p.r., and IX absent of, if present, not fused to s.g.p.r.. Pseudostyli absent or, if present, small, slender, incurved (Fig. 108) . Male genital opening postero-dorsal or dorsal, Parameres fused or unfused; basiparameral sclerites present, fused. Mesomeres present, fused or unfused and, if fused, median extension 330

Figs 110 - 113. Cebidicola spp. (110) C. semiarmatus female: head, dorsal; (111) C. armatus female: head, dorsal; (112) C. extrarius male genitalia; (113) C. armatus male genitalia. 331

absent; mesomeral arch produced basally between b.a.l.s. to contact parameres, which do not meet b.a.l.s.. Male genitalia depicted in

Figs 112,113-

Hosts Cebidae (Primates)

Species included armatus (Neumann, 1913) (3 d, 4?) extrarius Werneck, 1950 (2ld, 13$) semi armatus (Neumann, 1913) (126, 12$)

3.3.4.2. Procavicola Bedford

The genus Procavicola comprises two subgenera.

Description

Anterior of head with osculum present, semicircular; dorsal preantennal sulcus present; clypeal marginal carina broadened medially into dorsal, posteriorly convex, sclerite (Fig.119); conus large, as long as female scape; anterolateral margin of head straight or convex anteriorly, more or less abruptly concave at junction with clypeofrontal sulcus; preantennal outline broadly triangular. Temple margin smoothly convex, sometimes produced posteriad, or with posterolateral angle developed into posteriorly-projecting triangular or rounded process* Male scape expanded, with setal row represented by two setae only; male flagellomeres fused, v/ith two basally-articulated 'teeth'; female flagellomeres unfused and closed associated, fused to tv/o closely associated annulations, or completely fused*. Dorsum of head with setae short, sparse; anterior margin of head with setae longer than on disc. Sitophore sclerite unmodified. Tarsal claws lacking ventral spines or teeth. Postcoxale of metathoracic leg absent or present; if present, may be well-developed, but not to the same degree as described for Procaviohilus (Meganarionoides) and not fused to abdominal pleurite II. Thoracic setae present dorsally only along posterior and lateroposterior margins of prothorax and pterothorax; setae short, except for laterally on pterothorax, where of moderate length. 332

Figs 114 - 116. Procavicola female terminalia (ventral). (114)

(p.) natalensis; (115) P. (G.) lindfieldi; (116) P. (C.) dissimilis. 333

Abdomen oval, more or less elongate (Figs . 117, 120). Abdominal spiracles present on segments III - VTII, all approximately the same size. Abdominal setae short or of moderate length; anterior setae present on pleura only; postero-lateral setae present, sometimes numbering more than one per site*. Pleural projections present dorsally and ventrally on abdominal pleurum IV, sclerotised. Sclerites present on sterna, terga and pleura of all abdominal segments except, occasionally, tergum I*; male terga, at least of segments IV - VI, with anterior and posterior sclerites; second abdominal sternum with broad, heavily- sclerotised apophysis underlying sternite, articulating with median extensions of abdominal pleurite II (Figs. 117, 120).

Gonapophyses broad, lacking lobe; marginal setae lacking tubercles, occasionaly on small conical protuberances (Figs 115, 116). Gonapophyses meet ventral vulval margin acutely, not linked by sclerotised band. Ventral vulval margin not sclerotised; expanded, sometimes 'V/1 -shaped medially (Fig. 114), sometimes broadened posteriorly (Fig.115), sometimes contracted, shorter than length of gonapophyses (Fig. 116)*.

Male subgenital plate with sternites VII and VIII present and fused to s.g.p.r., sternite IX absent or, if present, fused to s.g.p.r. and perisetal gaps small*; Sig.p.r. not always attaining posterior margin of segment IX*. Pseudostyli absent (Figs 117,118, 120). Male genital opening postero-dorsal. Male genitalia very variable*; parameres not fused, basiparameral sclerites present or absent*, mesomeres fused or unfused*.

Ho st s Procaviidae (Hyracoidea).

Subgenus Procavicoli . _—-....a Bedford

Procavicola Bedford, 1932: 711. Type-species: Trichodectes stematus Bedford, by original designation.

Description Temple margin smoothly convex, sometimes projecting posteriorly, but never with posterior membranous or lightly-sclerotised process. Postero-lateral setae present, single at each site. Sclerites present on sterna, terga and pleura of all abdominal segments exceot 334

Figs 117 - 120. Procavicola spp. (117) P- (P.) eichleri male: abdomen; (118) P. (P.) vicinus male: terminalia; (119) P. (P.) natalensis female: head, dorsal; (120) P. (C.) dissimilis male: abdomen. 335

segment I, where tergal sclerite absent.

Ventral- vulval margin expanded as described, much broader than length of gonapophyses (Fig.114).

Male subgenital plate with stemites VTI and VIII present arid fused to s.g.p.r., stemite IX absent; s.g.p.r. may be very slender, may not attain posterior margin of segment IX ^Figs 117, 118). Parameres unfused, more or less narrow, rod-like, sometimes asymmetrically curved. Basi- parameral sclerites frequently present, fused or separate. Mesomeres not fused, short. Endophallus lacking large, hook-like sclerites. Male genitalia depicted in Fig. 123.

Hosts

Procaviidae (Hyracoidea).

Species included affinis Werneck, 1941 (102, 13$) brucei Werneck, 1941 (43

Figs 121 - 123. Frocavicola male genitalia. (121) P. (C.) dissimilis. (122) P. (C.) dissimilis detail of right pramere and mesomere at junction with basal apodeme; (123) P. (P.) pretoriensis. 337

Subgenus Condyloceohalus Werneck

Condylocechalus Werneck, 1941: 497 [as subgenus of Procavicola Bedford], Type-species: Procavicola (Condylocephalus) bedfordi Werneck, by original designation.

Description Temple margin convex, v/ith posteriorly-projecting membranous or lightly-sclerotised process, more apparent in male than female, triangular or as small rounded bump (linfieldi females). Female flagellomeres unfused, but closely associated. Abdomen v/ith postero-lateral setae present, frequently doubled, trebled or numerous at each site. Sclerites present on sterna, terga and pleura of all abdominal segments. Ventral vulval margin expanded, sometimes broadened posteriorly and broader than length of gonapophyses (Fig. 115), otherwise narrov/er, width less than length of gonapophyses (Fig.116). Male subgenital plate v/ith sternites VII, VIII and IX present and fused to s.g.p.r., v/ith perisetal gaps small (Fig. 120). Parameres unfused, curved, with anterolateral projections, not asymmetric. Basi- parameral sclerites present, fused or unfused. Mesomeres fused apically; mesomeral arch with median extension and lateral double flexion (Figs 121, 122). Endophallus ornamented with large, hook-like sclerites (Fig. 121).

Hosts

Procaviidae (Hyracoidea).

Comments

Though Condylo ce-phalus has been treated by most authors as a subgenus of Procavioola, Eichler (1963) considered it to have full generic status.

Species included bedfordi Werneck, 1941 (26, 1$) dissimilis 'Werneck, 1941 (646, 58?) hookinsi Werneck, 1941 (246, 29?) lindfieldi (Hill, 1922) (776, 63?) univirgatus (Neumann, 1913) (336, 32$) 338

3.3.4.3. Procaviphilus Bedford

The genus Procaviohilus comprises two subgenera.

Description

Surface of head, thorax and abdomen frequently covered with clearly- visible scales or sclerotised nodules. Anterior of head variable, one of two types: 1procaviphilus1 or 1 procavicola' *. a) 1procaviphilus1 type: Osculum absent 'or, if present, slightly concave only; dorsal preantennal sulcus absent; clypeal marginal carina broadened medially into straight or slightly curved bar (Pig.126); conus small, not as long as female scape; anterolateral margin of head straight or convex anteriorly, no abrupt concavity at junction with clypeofrontal sulcus; preantennal outline trapezoid (Pig.126). b) 'procavicola1 type: Osculum present, semicircular; dorsal preantennal sulcus present or absent; clypeal marginal•carina broadened medially into dorsal, posteriorly convex, sclerite (Fig.119); conus large, as long as female scape; anterolateral margin of head convex anteriorly, more or less abruptly concave at junction with clypeofrontal sulcus; pre- antennal outline broadly triangular. Temple margin smoothly convex, more or less projecting posteriorly. Male scape expanded, with setal row represented by two setae only; male flagellomeres fused, with two basally-articulated 'teeth'; female flagellomeres unfused, though sometimes very closely associated. Dorsum of head with setae short, sparse. Sitophore sclerite unmodified.

Tarsal claws lacking ventral spines or teeth. Postcoxale of leg III absent or, if present, frequently enlarged, heavily sclerotised, displaced posteriad to occupy abdominal sternum II, and fused to sclerite of abdominal pleurum II, in which case gap between postcoxales sometimes obscured by sternite II*. Thoracic setae present dorsally only along posterior and lateroposterior margins of prothorax and pterothorax; setae short, except for laterally on pterothorax, where of moderate length.

Abdomen oval-elongate. Abdominal spiracles present on segments III to VIII, though sometimes very small and possibly non-functional on 339

Figs 124 - 127. Procaviphilus spp. (124) P. (M.) n. baculatus female: terminalia, ventral; (125) P. (M.) n. baculatus female: gonapophysis, ventral; (126) P. (P.) f. granuloides female: head, dorsal; (127) P. (M.) scutifer female: terminalia, ventral. 340

VTII*. Abdominal setae short or of moderate length; anterior setae present on pleura only, sparse; postero-lateral seta present. Pleural projections present ventrally and dorsally on abdominal pleurum IV, sclerotised. Sclerites, frequently faint, present on sterna, terga and pleura of all abdominal segments except I; male terga, especially tergum VI, frequently v/ith anterior and posterior sclerites.

Gonapophyses v/ith setae non-tuberculate and, frequently, tuberculate; setal tubercles, if present, sometimes fused characteristically*; lobe absent or, if apparently present, formed of fused tubercles and thick, v/ith submarginal setae (Fig.125). Gonapophyses meet ventral vulval margin acutely, not linked by sclerotised band. Ventral vulval margin not sclerotised; expanded, frequently v/ith posterior broadening (Fig. 124')', sometimes lengthened (Fig. 127) or v/ith median lobulate process (Fig.124)*. Male subgenital plate v/ith sternites VTI, VTII and IX present and fused to s.g.p.r., perisetal gaps very small or absent, rarely large (Figs 128, 12$). Pseudostyli absent. Male genital opening postero- dorsal. Male genitalia very variable: basal apodeme short or long v/ith median constriction*; parameres unfused, frequently v/ith basal flange, or fused, faintly or clearly, and parameral plate apically bifurcate. Basiparameral sclerites absent. Mesomeres fused, basally extending betv/een b.a.l.s. to contact parameres, v/hich may also contact b.a.l.s. (Figs 130, 132, 133, 135,136); mesomeral arch v/ith median extension more or less broad, lateral desclerotisations sometimes apparent (Figs

131,134)*.

Ho st s

Procaviidae (Hyracoidea) and Cercopithecidae (Primates) .

Subgenus Procaviphilus Bedford

Procaviohilus Bedford, 1932: 725. Type-species: Procaviphilus ferrisi Bedford, by original designation Description Anterior of head of 1procaviphilus' type. Postcoxale absent or, if present, not greatly developed and not fused to abdominal pleurite II. Abdominal spiracles all the same size. 341 342

Gonapophyses with setal tubercles present but not fused; gonapophyses not apically truncate. Ventral vulval margin expanded, sometimes tT.7» - shaped medially, not broadened posterly.

Male genitalia with basal apodeme short, attaining abdominal segment VTI or VT, lacking median constriction; mesomeral arch with lateral desclerotisations.

Hosts Procaviidae (Hyracoidea).

Species included dubius '.Verneck, 1941 (10c, 10$) ferrisi ferrisi Bedford, 1932 (122, 5$) ferrisi granuloides Bedford, 1939 (262, 19$) ferrisi hindei Werneck, 1946 (3°, 4$) granulatus (Ferris, 1930) (13^, 17?) harrisi Werneck. 1946 (152, 19$) robertsi (Bedford, 1928) (162, 18$)

Subgenus Lleganarionoides Sichler Comb. n.

Mevanarionoides Eichler, 1940: 159. Type-species: Trichodectes colobi Kellogg, by original designation. Acondyloceohalus 7/erneck, 1941: 478 [as subgenus of Procavicola Bedford], Type-species: Trichodectes congoensis Ferris, by original designation. [Synonymy by Vemeck, 1946: 85] .

Description Anterior of head of 'procaviphilus1 type or, more frequently, of 'procavicola' type. Postcoxale of leg III enlarged as described in description of Procaviphilus s. lat., and fused to abdominal pleurite II, at least in female. Gonapophyses with setal tubercles absent (in which case gonapophyses characteristically broad and vulval margin produced posteriad as in Fig. 127), or present basally, marginally and submarginally, and fused characteristically to form basal process (Fig. 125); gonapophyses more or less truncate. Ventral vulval margin expanded, as described for 343

Figs 130 - 13o. Frocaviphilus male genitalia. (130) F. (M.) angolensis; (131) P. (P.) dubius; (132) F.(M.) .jordani; (133) P. (M.) .jordani detail of junction of mesomere, paramere and basol apodeme; C134) P- (M.) serraticus; (135) (£•) granuloides; (136) P. (M.) n. neumanni. 344

Procaviphilus s. str. or, more frequently, broadened posteriorly, some- times produced posteriad (Fig. 127) or with median lobulate process (Fig. 124). Male genitalia with basal apodeme attaining abdominal segment VII or VI or, more frequently, long, attaining segment III or II, with median constriction (Figs 130,136);- mesomeral arch with or without lateral desclerotisation.

Hosts

Procaviidae (Hyracoidea) and Cercopithecidae (Primates).

Comments

There has been some disagreement in the literature over the correct host of one species in this subgenus. Most species included in P. (Meganarionoides) are parasites of Procaviidae, as are .all other species in the subfamily Dasyonyginae (other than the three speices of Cebidicola. which are included in the subfamily for the first time in this study)• One species, however, P. (M.) colobi (Kellogg, 1910), was described from the monkey Colobus guereza caudatus Thomas. Keler (1938) included this species with the others described from Primates in his genus Meganarion (an objective synonym of Cebidicola), although realising that the species were not truly congeneric. Eichler (1940) described the new genus Meganarionoides for colobi, and placed it with Cebidicola and Lorisicola in the new subfamily Cebidicolinae. Werneck (1946) recognised the identity of colobi v/ith the hy rax lice, and synonymised M e ganari ono ides with Procavicola (Acondylocephalus) Werneck, 1941, the subgenus thus taking the name Procavicola (Meganarionoides). Werneck (1946) also suggested that Colobus was not the true host of P. colobi, but that the louse v/as probably a parasite of Dendrohydrax validus subsp.. He suggested that the host record of the type specimens was erroneous and due to mislabelling (the collection having included both Colobus and Dendrohyrax), and that a second record was due to contamination (other hyraxrlice having been associated v/ith the specimens of colobi) . Hopkins (1949) reported having examined twenty-five skins of Colobus oolykomos, which he identified as the 'supposed host1, without having found any Trichodectidae, and agreed with Wemeck (1946) that Dendrohyrax validus 345

subsp. was the correct host. Hopkins & Clay (1952) also identified D. validus subsp. as the host,the record from 'Colobus caudatus1 being termed an 'error1. Sichler (1963) agreed, and removed Ivleganarionoides from the Cebidicolinae and placed it in the Dasyonygidae with Procavicola (Fig. 51 )• Emerson 8c Price (1981) include _P. colobi, without comment, as a parasite of Dendrohyrax validus validus, although the association with the nominate subspecies of this animal has not appeared elsewhere in the literature. Kuhn & Ludwig (1964), however, reported a specimen of Colobus guereza with "hundreds of eggs and adult and larval Procavicola on it, all clasping the hairs tightly; most of them on the back and on the throat", and were able to state that the monkey had not been in contact with a Dendrohyrax or any other Procaviidae after its death. They concluded "There is no doubt .... that Colobus guereza is a natural host of Procavicola (IvIeganarionoides) colobi." . In view of the fact that there are now three records of the species from Colobus guereza and none from any member of the Procaviidae, this conclusion seems fully justified.

Meganarionoides was, as described above, treated as a subgenus of Procavicola by Y/erneck (1946). In this he has been followed by most authors, although Eichler (1963) considered it to be a full genus. Before the present study Meganarionoides had not been placed as a subgenus of Procavjphilus.

Species included africanus ('.Yerneck, 1941) .. Comb. n. from Procavicola. (56, 5?) angolensis (Bedford, 1936) Comb. n. from Procavicola. (86, 11 $ ) colobi (Kellogg, 1910) Comb. n. from Procavicola. (16, 1?) congoensis (Ferris, 1930). Comb. n. from Procavicola. (226, 25?) jordani (Bedford, 1936) Comb. n. from Procavicola. (2 6, 2$) muesebecki (Emerson & Price, 1969) Comb. n. from Procavicola (106,10?) neumanni neumarni (Stobbe, 1913) Comb. n. from Procavicola. (26, 1?) neumanni baculatus (Ferris, 1930) . Comb. n. from Procavicola. (136,14$) sclerotis sclerotis Bedford, 1932 [treated as Proca/yiphilus s. str.

by previous authors] (1C6, 17?) sclerotis major Maltbaek, 1937 [treated as Procavishilus s. str. by previous authors] 346

scut if er (Werneck, 1941). Comb. 11. from Procavicola. (146, 19?) serraticus (Hill, 1922) [treated as Procaviphilus s. str. by previous authors] (506, 70$) tendeiroi (Emerson, 19&5) Comb. n. from Procavicola. (36, 2$)

3.3.4.4. Dasyonyx Bedford

The genus Dasyonyx comprises two subgenera.

Description

Anterior of head v/ith osculum present, variable in degree of excavation*; dorsal preantennal sulcus absent; clypeal marginal carina broadened medially, v/ith posterior curvature of broadened portion similar to curvature of osculum* (Pig.137); conus not large; anterolateral margin of head straight, convex or concave, though not very concave at junction with clypeofrontal suture; preantennal portion of head short or long, outline broadly triangular, trapezoidal or rounded*. Temple margin shallowly convex, sometimes v/ith small rounded projection postero- laterally*. Male scape expanded, with setal row present; male flagello- meres fused, with two basally-articulated 'teeth1; female flagellomeres unfused. Dorsum of head v/ith setae of moderate length. Sitophore sclerite v/ith posterior arms extended (cf. Pig. 36), though sclerite difficult to see. Tarsal claws v/ith ventral spines or teeth (Pigs 18a, 18b)*. Post- coxale of metathoracic leg absent or, if present* enlarged, though not as described for Procaviohilus (Meganarionoides) and not fused to abdominal pleurite II. Thorax v/ith dorsal setae present only posteriorly on prothoracic margin and posteriorly and posterolaterally on pterothoracic margin; setae short anteriorly, longer posteriorly v/ith the longest setae on the posterolateral margins of the pterothorax; setae generally sparse. Abdomen broadly oval, v/ith male segment IX not projecting greatly (Pig. 139)• Abdominal spiracles present on segments III to VTII, all approximately the same size, frequently inconspicuous. Abdominal setae of moderate length; anterior setae present on pleura only; postero- lateral setae present. Pleural projections present dorsally and ventrally 347

Figs 137 - 139* Dasyonyx spp. (137) B. (D.) dendrohyracis female: head, dorsal; (138) D. (D.) v. ugandensis female: terminalia, ventral; (139) B. (D.) v. validus male: abdomen. 348

on abdominal pleurum IV, sclerotised. Sclerites present on sterna, terga and pleura (where present) of all abdominal segments except I; male terga, at least of abdominal segment VI, with anterior and posterior sclerites. Gonapophyses with sparse marginal setae and variably-developed lobe ventrally, the lobe bearing two apical or sub-apical setae and frequently being serrate along dorsal (posterior) margin (Fig.138). Gonapophyses meet ventral vulval margin acutely, not linked by sclerotised band. Ventral vulval margin not sclerotised; greatly expanded, sometimes with posterolateral angular projections (Fig. 138); subgenital lobe absent. Male subgenital plate with sternites VII, nil and IX present and fused to s.g.p.r., but variably modified (Figs 139,140, 141, 142, 143), frequently lacking perisetal gap. Pseudostyli absent. Male genital opening postero-dorsal or dorsal. Parameres fused or, if unfused, then with basal flanges (Figs 145*146,147,148). Basiparameral sclerites absent. Mesomeres fused; 'mesomeral arch with median extension and lateral desclerotisations; mesomeres more or less produced basally between b.a.l.s. to meet parameres, which sometimes do not contact b.a.l.s. (Figs 144, 145,146,147,148 ).

Ho st s

Procaviidae (Hyracoidea).

Subgenus Dasyonyx Bedford

Dasyonyx Bedford, 1932b: 720. Type-species: Dasyonyx validus Bedford, by original designation.

Description Osculum deeply concave; preantennal outline of male head sub- triangular or subtra.pezoidal. Temple margin frequently with small rounded projection posterolaterally (Fig. 137) • Tarsal claws with ventral spines slender and sharp ( Fig. 18a). Postcoxale of metathoracic leg generally present and enlarged, though not as described for Procaviphilus (Keganarionoides).

Hosts Procaviidae (Hyracoidea). 349

m £ •H M a o .£

& f=5|

CM

0 •H 0 £ 0 •H ,0 O £ •H rt £

& Ol

0 •H 0 £ 0 0 £ •H £B£

2

*=!i o

•H rH rt £ •H s £ 0 -P

S£I o >: 0 • rt 0 ft •H rH a$ > • o rO /—^ • P! 1 o • Pll x 0 rn tjfl •H T~ PH 350

Species included bedfordi 7/erneck, 1945 (llo, 16?) dendrohydracis (Ferris, 1930) (126, 15?) guineensis Werneck, 1941 (56, 3$) hopkinsi Y/erneck, 1941 (336, 32?) minor Bedford, 1939 (3 6, 1?) oculatus Bedford, 1928 ovalis Bedford, 1932 (36 6, 35?) smallwoodae Emerson & Price, 1969 (7 4 3 3) validus validus Bedford, 1932 (226, 22?) validus ugandensis Y/erneck. 1941 (296, 26?)

Subgenus ITeodasyonyx 7/erneck

Heodasyonyx 7/erneck, 1941: 543 [as subgenus of Dasyon:^x Bedford] . Type-species: Dasyonyx transvaalensis Bedford, by original designation.

Description Osculum shallowly concave; preantennal region of male head short, outline rounded. Temple margin lacking posterolateral projection. Tarsal claws with broad ventral teeth (Fig. 18b). Postcoxale of metathoracic leg absent.

Hosts

Procaviidae (Hyracoidea).

Species included capensis Emerson, 19&5 (Holotype6, allotype ?) diacanthus (Ehrenberg, 1828) (96, 8?) nairobiensis Bedford, 1936 (896, 92?) ruficeps Emerson, 1964 (15^, 13?) transvaalensis Bedford, 1932 (24^, 31?) waterburgensis Bedford, 1932 (56, 6?) 351

Figs 144 - 148. Dasyonyx male genitalia. (144) B. (N.) ruficeps; (145) B. (D.) minor; (146) D. (D.) guineensis; (147) (£•) validus, with detail (a) of endophallus sclerotisation; (148) B. (D.) ovalis. 352

3.3.4.5. Eurytrichodectes Stobbe

Burytrichodectes Stobbe, 1913a: 111. Type-species: Burytrichodectes paradoxus Stobbe, by monotypy.

Description Anterior of head with osculum absent or, if present, very shallowly concave; dorsal preantennal sulcus absent; clypeal marginal carina slender, not greatly developed medially or, if developed, in the form of a median posteriorly-directed narrow-based triangle; anterolateral margin of head slightly sinuate; preantennal portion of head very short, outline broadly triangular or trapezoid. Temple margin produced posteriorly into broad triangular spike, almost as long as prothorax or, if shorter, attaining front of pronotum (Fig. 149)• Male scape expanded, with longi- tudinal setal row comprising only two setae; male flagellomeres fused, though semicircular sclerite at apex may be vestige of terminal flagellomere; two basally-fused 'teeth1 present on male flagellum; female flagellomeres unfused; membranous projection present on female pedicel (Fig. 149)• Dorsum of head with setae short, sparse. Sitophore sclerite with posterior arms extended (cf. Fig.36 ), though sclerite difficult to see.

Tarsal claws ridged ventrally, lacking teeth or spines (Fig. 18c). Postcoxale of metathoracic leg absent. Thorax with dorsal setae present on posterior margin of prothorax and on posterior margin of pterothorax; setae short, sparse.

Abdomen broad and oval, sometimes with male terminal segments tapering and projecting slightly. Abdominal spiracles present on segments III to VIII, all approximately of the same size. Abdominal setae short on sterna IV to IX, terga and pleura, some tergal setae very short; sterna II and III with stout, conical setae (Fig. 150); anterior setae present, sparse on pleura; setal row on terga and sterna may be irregular; postero-lateral setae present. Pleural projections present dorsally and ventrally on abdominal pleurum IV, sclerotised, very long, reaching or almost reaching posterior margin of pleurum V. Sclerites present on sterna, terga and pleura, (where present) of all abdominal segments except, sometimes, tergum I; male terga, at least of abdominal segments II to VI, with anterior and posterior sclerites. 353

Figs 149 ~ 151. Eurl7trichodectes paradoxus. (l4S>)female head, dorsal; (150) male abdomen; (151) male genitalia. 354

Gonapophyses with ventral marginal setae present, each with a small, conical tubercle; ventral lobe absent. Gonapophyses meet ventral vulval margin smoothly, not linked by sclerotised band. Ventral vulval margin not sclerotised; medially expanded and trapezoid; subgenital lobe absent. Male subgenital plate with sternites VTI, VIII and IX present and fused to s.g.p.r., with setal gaps very small or absent (Fig. 150). Pseudostyli absent. Male genital opening dorsal. Parameres not fused, sometimes apically bifurcate. Basiparameral sclerites absent. Meso- meres fused; mesomeral arch with median extension (see comments below); mesomeral arch lacking lateral desclerotisations; mesomeres sometimes produced basally between b.a.l.s. to contact parameres. Male genitalia depicted in Fig. 151.

Hosts

Procaviidae (Kyracoidea).

Comment s

The mesomeral arch,., of E. paradoxus is printed upside-down in Werneck

(1941: 452).

Species included machadoi Y7erneck, 1958 (Holotypec, allotype?) paradoxus Stobbe, 1913 (292, 28?)

3.3.5. Trichodectinae

3.3.5.1. Protelicola Bedford Gen. rev.

Protelicola Bedford, 1932a: 355. Type-species Protelicola intermedia Bedford, by monotypy.

Description Anterior of head with osculum present; clypeal marginal carina broadened just laterally to osculum, tapering medially and interrupted by dorsal preantennal sulcus; anterolateral margin of head straight or convex; preantennal outline of head triangular or rounded. Temple margin convex. Male scape slightly expanded, with longitudinal setal row present and comprising three or four setae; flagellomeres fused in males and females; male flagellum with two basally-articulated 'teeth1. 355

Dorsal setae of head short or of moderate length, longest and most abundant anteriorly. Sitophore sclerite unmodified. Thorax with setae long or of moderate length laterally and dorsally, with setae on postero-lateral angles of pterothorax shorter, more spine- like. Prothorax with setae present sparsely on lateral and posterior margins; median gap present, wide; single seta present antero-laterally on disc. Pterothorax with setae present postero-laterally and sub- marginally posteriorly; median gap present, wide; no setae present on disc. Abdomen rounded, similar in shape in males and females. Abdominal spiracles present on segments III - VIII. Abdominal setae of moderate length; anterior setae present on pleura only; postero-lateral setae absent. Abdominal pleura lacking projections. Abdominal segments with tergal sclerites absent except tergite IX in female; pleural sclerites present on pleura II and III; sternites IV - VII present in male and V - VII present in female, very slender and difficult to see in both sexes. Gonapophyses with basal setae and rectangular lobe on ventral margin, lobe formed from more or less fused setal tubercles. Gonapophyses meet ventral vulval margin smoothly, not linked by sclerotised band. Ventral vulval margin not sclerotised. Subgenital lobe present, not marginally serrate, but sometimes with apical papillae. Female genital chamber with small lapped scales on walls, dorsal wall with median longitudinal anterior 'slit' where scales are absent. Male subgenital plate comprising very slender sternite VII and s.g.p.r. only, s.g.p.r. not reaching posterior of sternum IX. Male genital opening dorsal or postero-dorsal. Pseudostyli absent. Basal apodeme with b.a.l.s. widely divergent anteriorly. Parameres long, slender, fused basally, projecting anteriorly between b.a.l.s. (but see second paragraph of 'comments' below). Basiparameral sclerites absent. Mesomeres fused apically to form arch, with median projection present, broad (see second and third paragraphs of 'comments' below). Male genitalia depicted in Fig. 152. Alimentary canal with numerous small spines in crop.

Ho sts Hyaenidae and Protelidae (Canaivora). 356

Figs 152 - 153• Protelicola male genitalia. (152) hyaenae; (153) undescribed species. 357

Comments

Protelicola was treated as a subgenus and a junior synonym of Felicola by Hopkins (1949) and Werneck (1948) respectively; its most recent placement (Emerson & Price, 1981) was as a synonym of Felicola. A more detailed history of the variations in status of Protelicola is presented in Table IX. In the B.M.H.H. collection there is a slide bearing 1 nymphal, 3 female and 2 male (one of which is teneral) lice from Proteles cristatus termes. Hopkins has identified the lice as Protelicola intermedius. The females are indistinguishable from females of JP. intermedius s. str., but the male genitalia are very distinct, with the parameres completely fused to form a broad plate, the mesomeral arch wide, and the b.a.l.s. with a small postero-lateral projection contacting the mesomeres (Fig. 153). In all other respects the males resemble _P. intermedius s. str.. It seems that the males represent a new species, differing from Protelicola intermedius by the structure of the genitalia, but the identity of the females is doubtful. In view of the limited number of specimens available, the species is not formally described. Bedford (1932) described JP. intermedius from Proteles cristatus, the first louse known from a hyaena. Hopkins (1960) described JP. intermedius hyaenae (in Felicola) from Hyaena brunnea, distinguishing it from the nominate subspecies by the size, the outline of the preantennal- portion of the head, and the male genitalia. Ledger (1980) raised JP. jL. hyaenae to specific rank although Emerson & Price (1981) retained' its subspecific status. Hopkins (1960) indicated three features of the male genitalia in which the two taxa differ: the greater anterior

divergence of the b.a.l.s. in

variably divergent anteriorly, the parameres fused only basally and the mesomeral arch with a median broad extension (although this is very thinly sclerotised and difficult to see in both taxa). The male genitalia of .P. hyaenae are depicted in Fig. 152. The two species may be distinguished by the smaller size and shorter preantennal region of P. hyaenae (see photographs in Hopkins, 1960) .

Species included hyaenae (Hopkins, 1960) Comb. n. from Felicola Stat n. (Holotype 0' , allotype ?)

intermedius Bedford, 1932 Comb. rev. from Felicola (192, 48?)

3.3.5.2. Lutridia Keler Lutridia Keler, 1938: 433. Type-species: Trichodectes exilis Giebel, by original designation.

Description Anterior of head with osculum absent; dorsal preantennal sulcus absent; clypeal marginal carina broadened medially to form dorsal sclerite with three posteriorly-directed projections (Fig.155); preantennal portion of head short, outline smoothly rounded. Temple margin convex or rectangular. Male scape not expanded, longitudinal setal row comprising two setae positioned distally on segment; flagellomeres fused in males and females; male flagellum lacking 'teeth' . Dorsum of head with setae short or of moderate length, sparse. Sitophore sclerite unmodified.

Thorax with dorsal setae long or of moderate length, limited to posterior and postero-lateral margins of prothorax and pterothorax. Abdomen oval or slightly elongate-oval, with male segment IX projecting slightly posteriorly (Fig. 154). Abdominal spiracles present on segments III - VIII. Abdominal setae as follows: pleurum II with setae sparse, stout and short, anterior setae and p.s.r. present; pleurum III with setae.short and of moderate length, stout, very sparse (exilis) or with p.s.r. present (matschiei); pleura IV - VII or VIII lacking setae, those on VIII, if present, very small and posteriorly positioned; sternal setae short and stout or longer (about two-thirds length of segment); sternum II with median gap of posterior setal row

» 359

Figs 154 - 156. Lutridia spp. (154) i* exilis male: abdomen; (155) L. matschiei female: head, dorsal; (156) i- matschiei female: terminalia. 360

small or absent; sterna III - IV and VII - VIII (matschiei) or III - VTII (exilis) with setae very sparse, setae absent from sterna V - VT of L. matschiei; terga I - IV (males) or I - III (females) with median setal group including one or two setae as long as segment, other terga with median group absent or comprising shorter setae; terga with lateral seta or setae generally present, of moderate length on terga II - III, shorter on more posterior segments (these setae may represent either lateral setal group or postero-lateral seta); anterior setae present only on pleurum II. Abdominal pleura lacking projections. Sclerites absent from abdominal pleura, present on terga IV - VIII (males) or terga IV - IX (females) and sterna III - VIII (males) or sterna IV - VIII (females); sclerites frequently very faint, may not be seen; male terga lacking posterior sclerites. Gonapophyses with non-tuberculate setae on ventral margin; ventral lobe not present. Gonapophyses meet ventral vulval margin smoothly, linked by sclerotised band. Ventral vulval margin sclerotised, with chord at 90 degrees to long axis of abdomen, submarginal non-tuberculate setae present. Subgenital lobe present, small, rectangular, sometimes serrate along posterior margin (Fig. 156) • Male subgenital plate comprising s.g.p.r. only or with s.g.p.r. linked by broad sternite VIII (and possibly VII); in either case s.g.p.r. not att aining segment IX. Pseudostyli absent. Male genital opening dorsal. Basal apodeme slender, long, attaining at least abdominal segment III. Parameres long, slender, basally fused; basal fused portion may be partially detached from rest of parameres; parameres sometimes fused to b.a.l.s.. Mesomeres absent. Male genitalia depicted in Figs .157, 158.

Hosts

Lutrinae (Carnivora: Mustelidae).

Comments

Lutridia has been treated as a synonym and a subgenus of Trichodectes (Hopkins, 1942 and Hopkins, 1949 respectively), though the most recent treatment (Emerson & Price, 1981) considered Lutridia as a full genus. The history of the variations in status of Lutridia is presented in Table VIII. 361

Figs 157, 158* Lutridia male genitalia. (157) B* matschiei; (158) L. exilis. 362

Species included

exilis (Nitzsch, 1961) (36, 10$)

matschiei (Stobbe, 1913) (366, 38$)

3.3.5.3. Genus n. 4

Type-species: Trichodectes lutrae Y/erneck

Description

Anterior of head with osculum absent, though pulvinus attaining margin; dorsal preantennal sulcus absent; clypeal marginal carina slightly broadened medially at junction with pulvinus; preantennal outline broadly and smoothly rounded. Temple margin rectangular. Male scape not expanded; longitudinal setal row present, comprising four setae; flagellomeres fused in males and females; male flagellum lacking 'teeth'. Dorsum of head with setae short or of moderate length, sparse; temple margin with two or three longer setae. Sitophore sclerite unmodified. Prothorax v/ith two setae of medium length on posterior margin; pterothorax v/ith one or two short, spine-like setae anterolaterally and six to ten long setae dorsally on posterior margin. Abdomen oval, v/ith male segment IX projecting posteriorly. Abdominal spiracles present on segments III - nil. Abdominal setae as follows: pleurum III with short, stout setae anteriorly and posteriorly; pleura III - VTII lacking setae; terga I - VI (males) or I - IV (females) v/ith central seta of median groups as long as segment, setae otherwise short; terga VII - VIII (males) or V - IX (females) with short setae, sparse; sterna with stout, short setae, sparse; anterior setae present only on pleurum II; postero-lateral setae absent. Abdominal pleura lacking projections. Sclerites absent from abdominal pleura and sterna, but present, slender, on at least terga III - VII (males) or V - IX (females), though may, be very faint arid not seen; male terga lacking posterior sclerites. Gonapophyses with small ventral lobes formed from fused setal tubercles. Gonapophyses meet ventral vulval margin smoothly, linked by sclerotised band. Ventral vulval margin sclerotised, v/ith chord at 363

Fig. 159- n.4 lutrae female: terminalia, ventral. Fig. 160. n.4 lutrae male genitalia (after VIerneck). Fig. 161. VIerneckodectes ferrisi male genitalia (after Kerneck). 364

90 degrees to long axis of abdomen; sub-marginal non-tuberculate setae present. Subgenital lobe present, large, with lateral spine-like projections and associated setae present basally (Fig. 159)* Male subgenital plate represented by sternite VTII with lateral arms extending anteriad. Pseudostyli absent. Male genital opening dorsal. Basal apodeme attaining abdominal segment III, not slender. Parameres broad, scoop-shaped, not fused together, but fused to b.a.l.s.. Mesomeres absent. Male genitalia depicted in Fig. 160.

Hosts Lutrinae (Camivora: Mustelidae) .

Species included lutrae (7/erneck, 1937) Comb. n. from Lutridia. (16, 1$)

3^3^5^4•^ 7/erneckodectes Conci Gen. rev.

7/erneckodectes Conci, 1946: 59- Type-species: Trichodectes ferrisi 7/erneck, by original designation.

Description Osculum absent, though pulvinus attains anterior margin of head; dorsal preantennal sulcus present; clypeal marginal carina broadened slightly medially; preantermal portion of head short, outline smoothly and broadly rounded. Temple margin convex. Male scape expanded; flagellomeres fused in males and females; . male flagellum with two basally-articulatea 'teeth' and basal projection. Dorsum of head with setae of moderate length.

Thorax with dorsal setae longest postero-laterally on pterothorax and along posterior margin on pterothorax; shorter setae present sub- marginally along posterior of pterothorax and on postero-lateral angles of prothorax, where there is a small setal patch; disc and posterior margin of pronotum each with pair of small setae.

Abdomen oval, tapering posteriorly rather more in male than female. Abdominal spiracles present on segments III - VIII. Abdominal setae numerous, anterior setae being present on stema, terga and possibly pleura of all segments; anterior setae smaller than setae of posterior setal rows on each segment; postero-lateral setae, if present,obscured 365

by large number of other setae. Abdominal pleurum IV with ventral projection in male, possibly with dorsal projection in female; pleurum III possibly v/ith dorsal projection in female. Sclerites present on all abdominal pleura, on terga V - IX (males) or terga VII - IX (females) and on sterna III - VI (males only - sclerites absent on female sterna); male terga without posterior tergites, but -anterior tergites, v/here present, v/ith median longitudinal division. Gonapophyses v/ith non-tuberculate setae on ventral margin; ventral lobe present, small; setae on lobe stout, short, whilst setae distally to lobe longer, more slender. Gonapophyses meet ventral vulval margin smoothly, linked by broad sclerotised band. Ventral vulval margin sclerotised, v/ith chord at 90 degrees to long axis of abdomen; marginal non-tuberculate setae present, stout, short. Subgenital lobe present, broad, with lateral rounded projections and associated setae present basally. Male subgenital plate with stemite vii fused to s.g.p.r., sternites VIII and IX absent; s.g.p.r. v/ith sinuate margins. Pseudostyli absent. Male genital opening dorsal. Parameres not as long as basal apodeme, not fused together. Basiparameral sclerites absent. Mesomeres not apically fused, abutting parameres and b.a.l.s. basally. Male genitalia depicted in Fig.-|6l.

Hosts Ursidae (Carnivora).

C omment s Werneckodectes has been treated as a synonym and a subgenus of Trichodectes (by Hopkins, 1942 and Hopkins, 1949 respectively); its most recent placement was as a synonym of Trichodectes. A more comprehensive history of the variations in status of Werneckodectes is .given in Table VIII. Species included ferrisi (".Yerneck, 1944) Comb. n. from Trichodectes. 366

3.3.5.5. Trichodectes ITitzsch

The genus Trichodectes comprises three subgenera.

Description

Anterior of head with osculum present or absent*, but always with pulvinus attaining margin; dorsal preantennal sulcus present or absent; clypeal marginal carina broadened to variable extent medially to form simple bar with posterior margin straight or concave, or carina broadened into dorsal sclerite which is heavily-sclerotised laterally (dorsal to margin of clypeus and pulvinus) and lightly on very lightly-sclerotised medially (posterior to the osculum), more or less convex posteriorly or 'U'-shaped with median posterior process*; anterolateral margin of head straight, convex or sinuate*; .preantennal portion of head long or short*, outline broadly rounded, broadly triangular, trapezoid or only slightly produced anteriad between coni*. Temple margin convex, rectangular or produced laterally*. Male scape expanded or not expanded*; longitudinal setal row present, comprising at least four setae; flagellomeres fused in males and females; male flagellum with one, two or four basally- articulated 'teeth' or 'teeth' absent*. Dorsum of head with setae short, of moderate length or long, longest setae generally present along posterior temple margin; setae sometimes sparse. Sitophore sclerite unmodified.

Thorax with prothoracic dorsal setae sparse, short or of moderate length* posteriorly and posterolaterally, absent from disc; pterothorax with setae on oosterolateral .angles short and so me — lik e or of moderate length, dorsal setae otherwise present on posterior margin only, long or short, numerous, sparse or absent*. Abdomen oval, male segment IX sometimes slightly projecting posteriad, but usually positioned dorsally on the abdomen (Figs 171, 173, 174, 175, 178) Abdominal spiracles present on segments III - IV, III - V, III- VTI or III - VTII*; spiracle on segment VIII, if present, sometimes much smaller than those on segments III - VTI*. Abdomen with at least some tergal and sternal setae as long as segment, or setae very short, sparse and absent from pleura V and VI*; terga with lateral and median groups of setae frequently distinct; tergal setae numerous, or median group reduced 367

Figs 162 - 165. Trichodectes (T.) female -terminalia. (162) T. (T.) galictidis; (163) T. (T.) canis; (164) T. (T.) emersoni; (165) T. (T.) pinguis, ventral aspect. 368

to a single seta or absent*; male terga II and III sometimes with median group comprising exceptionally long, stout setae (Fig. 177)*; anterior setae present only on pleurum II; postero-lateral setae presumed absent, or presence obscured by numerous long setae or reduction (or absence) of lateral setal group. Abdominal pleura lacking projections, or projections present dorsally on pleura II, III and IV* (Fig. 173). Abdominal sclerites variable, present or absent*; male terga v/ith or without anterior and posterior sclerites*. Gonapophyses with separate tuberculate setae and s'ingle apical non- tuberculate seta on ventral margin (Figs .162,163 , 164,166 , 168), or tubercles more closely associated (Fig.169) or all setae non-tuberculate (Figs 165, 167)*; ventral lobe absent. Gonapophyses meet ventral vulval margin smoothly, linked by sclerotised band, or band absent. Ventral vulval margin sclerotised or, rarely, not sclerotised; v/ith chord at 90 degrees to long axis of abdomen; marginal setae present, tuberculate or non-tuberculate. Subgenital lobe present, usually with marginal serrations and lateral basal projections, the latter sometimes v/ith associated setae. Male subgenital plate absent (Figs 171,172,173, 177, 178), represented only by s.g.p.r. (Fig. 176) or enlarged sternite VIII (Fig. 175), with sternites VII, VIII and IX present and fused to s.g.p.r. (Fig.174), or of the latter form but with sternite VIII divided medially. Pseudostyli absent. • Male genital opening dorsal or postero-dorsal. Parameres fused or separate, fused tob.a.l.s. or free*. Basiparameral sclerites absent. Mesomeres absent or present, fused or unfused; if mesomeres fused apically, median extension absent.

Hosts Canidae, Mustelidae, Procyonidae, Ursidae and Viverridae (Carnivora).

Comments The different concepts of the extent of the genus Trichodectes held by various workers are summarised in Table VIII. 369

Figs 166, 167. Trichodectes female terminalia. (166) T. (S.) emeryi; (167) T. (n.5) zorillae. 370

Figs 168 - 170. Trichodectes (Stachiella) female terminalia.• (168)

T. (S.) erminiae; f 160) T. '(S.) octomaculatus; (170) T. (S.) £otus. 371

Subgenus Trichodectes ITitzsch

Trichodectes Hitzsch, 1818: 294. Type-species: "Trichodectes canis Degeer (syn. _T. latus II.)", by subsequent designation [johnston & Harrison, 1911: 326] . Ursodectes Keler, 1938: 435. Type-species: Trichodectes oinguis

Burmeister, by original designation. [Synonymy by Hopkins, 1942: 444]. Grisonia Keler, 1938 [ nec Gray, 1825: 339 nom. nud.; Gray, 1843: 6s] .

no type-species designated. Galictobius Keler, 1938a: 228. Horn. nov. for Grisonia Keler nec Gray. Type-species: Trichodectes galictidis Werneck, by original designation. [Synonymy by Hopkins, 1942: 444].

Description Male scape expanded; male flagellum with two or four basally- articulated 'teeth1. Abdominal spiracles present on segments III - VTII-, spiracle on segment VIII not smaller than those on anterior segments. Abdomen v/ith at least some tergal and sternal setae as long as segment, or setae very short, sparse and absent from pleura V and VI (kuntzi and emersoni); male terga II and III sometimes v/ith median group comprising exceptionally long, stout setae (undescribed sister-species to T. galictidis), tergal setae never with median group reduced to single seta or absent (except sometimes on tergum I). Abdominal pleura lacking projections. Abdominal sternal sclerites absent, or present on sterna V - VIII only; abdominal tergal sclerites present or, more usually, absent; if tergites present on male, then not with anterior and posterior sclerites on each segment. Female genital chamber with ventral wall frequently obscure, dorsal wall bearing sclerotised nodules, sometimes fused together. Male subgenital plate absent, represented by s.g.p.r. only, or by enlarged stemite VTII only. Parameres fused to form plate or unfused; symmetric or asymmetric; not fused to b.a.l.s.. Faintly-sclerotised tongue-like sclerite of uncertain homology sometimes present dorsally between parameres if mesomeres absent (Figs 32, 179). LIesomeres-absent or, if present, fused or unfused. Male genitalia depicted in Figs 32, 179, 180, 181. 372

Fig. 171. Trichodectes (T.) canis male: abdomen. Fig. 172. Trichodectes (T.) pinguis euarctidos male: terminalia. 373

Fig. 173. Trichodectes (n.5) zorillae male: abdomen. Fig. 174* Trichodectes (S.) erminiae male: terminalia. 374

Fig. 175* Trichodectes (T.) emersoni male: abdomen.

Fig. 176. Trichodectes (T.) ga,lictidis male: terminalia. 375

Fig. 177• Trichodectes (n.5) ovalis male: abdomen. Fig. 178* Trichodectes (S.) emeryi male: -terminalia. 376

Hosts

Canidae, Mustelidae, Ursidae and Viverridae (Carnivora).

Comments

Trichodectes Hitzsch, 1818 was placed on the Official List of Generic ITames in Zoology, v/ith the type-species Trichodectes canis DeGeer, by Opinion 627 of the International Commission of Zoological Nomenclature (3ulletin of Zoological Nomenclature, 19 (1962): 91).

Species included canis ( DeGeer, 1778) (clOOd, clOO?) . emersoni Hopkins, 1960 (15 d, 12?) galictidis Werneck. 1934 (15d, 16$; also 2d, 1? of an undescribed sister-species) kuntzi Emerson, 1964 (15d, 14?) melis (Pabricius, 1805) (c60d, c60?) pin.guis oinguis Burmeister, 1838 (Id, 5?) pinguis euarctidos Hopkins. 1954 (20d, 20?) vosseleri Stobbe, 1913 (2d, 7?)

Subgenus n. 5

Type-species: Trichodectes ovalis Bedford

Description

Anterior of head v/ith osculum present; clypeal marginal carina broadened medially into dorsal sclerite which is heavily-sclerotised laterally and lightly-sclerotised medially, more or less convex posteriorly or TU' -shaped with median posterior process; anterolateral margin of head convex or sinuate; preantennal portion of head not long, outline rounded or subtriangular. Temple margin convex or rectangular. Male scape not greatly expanded; male flagellum v/ith two basally-articulated 'teeth1 . Thorax with prothoracic dorsal setae sparse, of moderate length posteriorly and posterolaterally, absent from disc; pterothorax v/ith setae on posterolateral angle long dorsally, short and spine-like ventrally; dorsal posterior pterothoracic setae submarginal, long, comprising two 377

179

Figs 179 - 181• Trichodectes (Trichodectes) male genitalia, (179) T. canis; (180) T. galictidis; (181) T. emersoni. 378

pairs with wide median gap. Abdominal spiracles present on segments III - "VTI. Abdomen with at least some tergal and sternal setae as long as segment; setae present on all pleura; tergal setae numerous, median group not reduced to a. single seta or absent except sometimes on tergum I or on posterior terga of males only, if male terga II and III with median setal group comprising exceptionally long, stout setae; postero-lateral setae presumed absent, though may be present as the most lateral seta of lateral group, which is frequently situated more posteriorly than other setae. Abdominal sterna and terga with or without sclerites; male terga, if sclerites present, with anterior sclerites only. Female genital chamber with dorsal wall not" bearing sclerotised nodules. Male subgenital plate unsclerotised. Parameres separate or thinly fused to each other; symmetric or asymmetric; not fused to b.a.l.s.. Tongue-like slcerite not present. Mesomeres absent. Male genitalia, depicted in Figs .182, 183, 184.

Hosts Mustelinae (Carnivora: Mustelidae).

Species included ovalis Bedford, 1928 Comb. rev. from Stachiella' (152, 12?) ugandensis Bedford, 1936 Comb. rev. from Stachiella (442, 47? ) zorillae Stobbe, 1913 Comb. rev. from Stachiella. (17 2, 26?)

Subgenus Stachiella Keler Stat. n.

Stachiella Keler, 1938: 428. Type-species: Trichodectes pusillus ITitzsch [ = pedicuius mustelae Schrank] , by original designation. Potusdia Conci, 1942: 141. Type-species: Trichodectes potus Werneck, by original designation. [Synonymy with Trichodectes by Verneck,'1948: 110; Syn. n. with Stachiellal

Description Clypeal marginal carina broadened medially into dorsal sclerite which is heavily-sclerotised laterally and lightly or very lightly- sclerotised medially, more or less convex posteriorly or 'U'-shaped with 379

Figs 182 - 184. Trichodectes (n.5) male genitalia. (182) T. ovalis; (183) T. ugandensis; (184) T. zorillae. 380

median posterior process; anterolateral margin of head smoothly convex; preantennal portion of head long or short, outline broadly rounded, sometimes only slightly produced anteriad between coni. Male scape not, or only slightly, expanded; male flagellum with one or two basally- articulated 'teeth' or 'teeth' absent.

Thorax with prothoracic dorsal setae sparse, of moderate length posteriorly and posterolaterally, absent from disc; pterothorax with setae on posterolateral angles short and spine-like or of moderate length, dorsal posterior setae submarginal, long, comprising one or more pairs with wide median gap, or absent. Abdominal spiracles present on segments III - IV, III - V, or III - VIII; spiracle on segment VTII, if present, sometimes much smaller than those on segments III - VII. Abdomen with at least some tergal and sternal setae aalong as segment; setae present on all pleura; terga with median setal group of male reduced to a single seta on most segments, lateral group small; female tergal setae of similar arrangement or with median group absent (see discussion on p.246above); male terga, II and III never with median group comprising exceptionally long, stout, setae; postero-lateral setae presumed absent, though perhaps present as the most lateral seta of lateral .setal group, which is frequently situated more posteriorly than other- setae of the group. Abdominal pleura lacking projections. Abdominal sterna variably sclerotised, with sternites, if present, most commonly on posterior segments; abdominal terga with sclerites present on segments III - VTII or III - IX, sometimes on I and II; male terga frequently with both anterior and posterior sclerites, though posterior sclerites may be faintly-sclerotised or absent; abdominal pleurum II frequently sclerotised, otherwise pleura unsclerotised. Gonapophyses with separate or closely-associated tuberculate setae and single apical non-tuberculate seta on ventral margin. Gonapophyses meet ventral vulval margin smoothly, linked by sclerotised band. Ventral vulval margin sclerotised. Female genital chamber with dorsal wall not bearing sclerotised nodules. Male subgenital plate absent, or with stemites VII, VIII and IX present and fused to s.g.p.r., or of this form but with sternite VTII divided medially. Parameres separate or thinly fused to one another; 381

Figs 185 - 187. Trichodec-bes (stachiella) male genitalia. (185) T. octomaculatus; (186) T. erminiae; (187) T. emeryi. 382

symmetric, asymmetric or asymmetrically deflected; fused or not fused to b.a.l.s.. Tongue-like sclerite not present. Mesomeres absent, present, small and unfused,or present and fused. LI ale genitalia depicted in.Figs 185, 186, 187.

Hosts

Llustelidae; LIustelinae and Procyonidae (Camivora).

Comments

Stachiella has been considered a synonym and a subgenus of Trichodectes (Hopki ns, 1942 and Hopkins, 1949 respectively), although its most recent treatment (Emerson 2c Price, 1981) was as a full genus. Potusdia has also been considered as a full genus, subgenus of Trichodectes and synonym of Trichodectes but has not, before this study, been considered a synonym of Stachiella. A more comprehensive history of the variations in status of Stachiella ana Potusdia is given in Table "7TII.

Species included divaricatus Harrison, 1915- Comb. rev. from Stachiella. emeryi Emerson 2c Price, 1974 > [Treated by Emerson 2c Price (1974, 1981) as Trichodectes s. str.] (4d, 17?) eiminiae (Hopkins, 1941) -. Comb. n. from Stachiella. (92d, 100?) fall ax Wemeck, 1948. [Treated by 7/erneck (1948)' andall subsequent authors as Trichodectes s. str.] (2

s. str.] (48d, 58?)

ootus Y/emeck, 1934 [ Treated by Verneck (1948) and Emerson 2c Price (1981) as Trichodectes s. str.] (19d, 25$) retusus retusus Buraeister, 1838 Comb. rev.from Stachiella. (Id, 1?) retusus mart is (7/erneck, 1948) Comb. n. from Stachiella. (Id) retusus saYLfii (Conci, 1940) Comb.n. from Stachiella..

1 383

3.3.5.6. Felicola Swing

The genus Felicola comprises two subgenera.

Description

Anterior of head with osculum present or absent; dorsal preantennal sulcus present or absent; clypeal marginal carina, if osculum absent,, v/ith very slight median broadening, or, if osculum present, carina broadened medially into dorsal sclerite which is heavily-sclerotised laterally (dorsal to margin of clypeus and pulvinus) and very lightly-sclerotised medially (posterior to osculum); anterolateral margin of head straight, slightly sinuate or convex; preantennal portion of-head long or short, outline triangular or broadly rounded*. Temple margin rectangular or convex. Male scape variably expanded.or not expanded*, with longitudinal setal row present and comprising at least three setae; flagellomeres fused in males and females; male flagellum v/ith one or three basally- articulated 'teeth', or variable number of 'teeth' fused to flagellum, or 'teeth' absent*. Dorsum of head v/ith setae short, of moderate length or long, but in any case frequently longer than abdominal tergal setae; setae sparse,_though most numerous anteriorly and along lateral margins. Sitophore sclerite unmodified.

Thorax v/ith dorsal setae long or of moderate length though frequently short and spine-like on posterolateral angle's of pterothorax, not- present on disc of prothorax or pterothorax. Abdomen oval or elongate, frequently terminating in more or less acute projection of segment IX in the male (Figs 19V 194-197, 199-202). Abdominal spiracles absent, or present on segment III, segments III - IV, or segments II - V*. Abdominal setae short, of moderate length, or occasionally long and fine* (Figs 193, 199, 202); male tergum II frequently with 2-6 very long setae medially (Figs 191,192, 20Q 201); abdominal pleurum III frequently with -posterior setal row comprising stout, conical setae (Figs 193 , 194, 202); anterior setae, if present, only on pleurum II; postero-lateral setae present or absent. Abdominal pleurum III with projections absent, or, if present, dorsal or ventral,sclerotised or unsclerotised; abdominal pleurum IV v/ith projections absent or, if present, dorsal and occasionally ventral, sclerotised or unsclerotised. 384

Fig. 188. Felicola (F.) ze.yl onions female: terminalia. Fig. 189. Felicola (S.) vtilpis female: terminalia, ventral. Fig. 190. Felicola (S.) decipiens female: gonapophysis, ventral. 385

Fig. 191. Felicola (F.) cynictis male: abdomen. Fig. 192. Felicola (F.) setosus male: abdomen. /' T "I t-

J I \ \ \

193

Figs 193, 194. Felicola (p.) male abdomens. (193) F. (F.) minimus; (194) p. (p.) congoensis. o>

\ I 387

Figs 195 - 198. Felicola spp. ( 195) (F.) calogaleus male: abdomen; (196) F. (F.) hopkinsi male: terminalia; (197) (F.) helogale male: terminalia; (198) F. (S.) pygidialis male: abdominal terga I - III. 388

Abdominal" sclerotisation variable; sternal, tergal and pleural sclerites, if present, generally on anterior segments, becoming less clear on posterior segments; male terga. sometimes with anterior and posterior sclerites, at least on terguin VI*.

Gonapophyses with iion-tuberculate setae and rounded or rectangular lobe present on ventral margin; apical spur present or absent*. Gona- pophyses meet ventral vulval margin acutely, not linked by sclerotised band. Ventral vulval margin not sclerotised; straight or concave, with chord less than 90 degrees to long axis of abdomen; subgenital lobe present, apically single or bifurcate, v/ith margins serrate, at least posteriorly (Fig. 188) (see comment below).

Male subgenital plate not present, though sternite VTII sometimes v/ith posterolateral projections probably homologous v/ith lateral rods of subgenital plate (Fig.194). Pseudostyli absent. Male genital opening postero-dorsal or dorsal; segment IX frequently produced posteriad. Parameres generally long and slender (Figs 203e,,203b,209,214^ occasionally broader (Fig. 205); frequently fused completely or,more usually, basally only. Basiparameral sclerites absent. Mesomeres present or absent; if present, unfused (Fig.213) or, if fused, median extension absent*. Male genitalia depicted in Figs . 203 - 217.

Hosts

Kerpestidae, Viverridae, Felidae and Canidae (Carnivora).

Comment s

Emerson £c Price (1980) distinguish the females of their new species Suricatoecus OccidentaJLis (transferred to Fslicola in this -study) from other species in the 'helogale Group1 (equivalent to the con.goensis - occidentalis clade) by the presumed absence of the subgenital lobe in occidentalis. Examination of the type series of this species, however, reveals that the subgenital lobe, although very fine, is present in all females of the species. A second species of this clade, close to F. helogaloidis, has been taken from skins of Crossarchus obscurus (the host of F. occidentalis) and specimens are in the collections of the B.K. (E.K.). 389

Fig. 199. Felicola (S.) fahrenholzi male: abdomen. Fig. 200. Felicola (S.) decipiens male: abdomen. Fig. 201. Felicola (S.) bedfordi males abdomen. Fig. 202. Felicola (S.) acutirostris male: abdomen. 391

Sub genus Felicola Swing

Felicola Ewing, l.vi.1929: 121, 122, 192. Type-species: Trichodectes subrostratus Burmeister [attributed to Kitzsch] , by original designation. Felicinia Bedford, -.x.1929: 519. Type-species: Trichodectes subrostratus Burmeister [attributed to ITitzsch] , by original designation. [Synonymy by Bedford, 1932a: 356]. Bedfordia Keler, 1938: 463 [nec Fahrenholz, 1936: 55]. Type-species: Felicola helogale Bedford, by original designation. Fastigatosculum Keler, 1939: 11 • Horn. nov. for Bedfordia Keler nec Fahrenholz [Synonymy v/ith Felicola by Hopkins, 1941: 36 (as Bedfordia); otherwise synonymised with Suricatoecus by 'Werneck, 1948]

Description Preantennal portion of head v/ith outline narrov/ly or broadly triangular or rounded. Hale flagellum v/ith 'teeth* absent or, if present, numbering one, two, three or four and fused to flagellum, not basally articulated. Abdominal spiracles absent, or present on segments III to IV or III to V. Abdominal setae very short or of moderate length. Gonapophyses with rounded lobe on ventral margin; spur present or absent. Everted portion of endophallus frequently thinly sclerotised (Figs 204,208,210). Mesomeres present and fused or absent.

Hosts

Felidae, Herpestidae and Viverridae (Carnivora).

Comments

The species F. genettae (Fresca) is included on the basis of the figures and description of Fresca (1924) which, although poor, suggest an affinity v/ith the calogaleus - viverri.culae clade. If this is a correct placement, the host record of Genetta genetta rhodanica is anomalous. The various treatments of the junior synonyms of Felicola are summarised in Table IX; omitted from the Table is Conci (1946) who, like Sichler (1941, 1963) regarded Fastigatosculum as a good genus. 392

Figs 203 - 206. Felicola (Felicola) male genitalia. (203a) F. inaecrualis, endophallus not everted; (203b) F. inaecrualis, endophallus partially everted; (204) F. calogaleus; (205) F. congoensis; (206) F. helogale. 393

Species included

calogaleus (Bedford, 1928) (145', 24?)

congoensis (Emerson & Price, 1967). Comb. n. from Suricatoecus.

(23

rahrai Emerson & Stojanovitch, 1966 (95, 16$)

robertsi Hopkins, 1944 (85, 15$)

rohani Werneck, 1956 (686, 62$) setosus Bedford, 1932 (146, 18$)

subrostratus (3urmeister, 1838) (IO36, 120$)

viverriculae (Stobbe, 1913) Comb. n. from Parafelicola. (2l6, 26?; 6 6, 6$ of undescribed sister-species) zeylonicus Bedford, 1936 (66,

Subgenus Suricateocus Bedford Stat. n.

Suricatoecus Bedford, 1932a: 354. Type-species: Trichodectes cooleyi

Bedford, by monotypy. Sichlerella Conci, 1942: 140. Type-species: Trichodectes vulois Denny, by original designation. [Synonymy by 'Werneck, 1948: 172].

Description

Preantennal portion of head with outline narrowly triangular or broadly rounded. Male scape not, or very slightly, expanded; male flagellum with 'teeth' absent or, if present, numbering one or three and basally articulated.

Abdominal spiracles absent, or present on segment III or segments

III to V. Abdominal setae very short, of moderate length, or long 394

Figs 207 - 212. Felicola male genitalia. (207) F« (F.) setosus; (208) F. (F.) robertsi; (209) F. (F.) hopkinsi; (210) F. (F.) subrostratus; (211) F. (F.) liberiae; (212) F. (S.) decipiens, with detail of paramere. 395

and fine. Tergal and sternal sclerites generally present on abdomen, though less clear on posterior segments; male terga never v/ith posterior sclerites.

Gonapophyses v/ith rounded or rectangular lobe on ventral margin; spur present. Everted portion of endophallus never sclerotised.

Hosts Canidae and Herpestidae (Carnivora).

Comment s

Suricatoecus has been treated not only as a full genus, but also as a synonym and a subgenus of Felicola (Bedford, 1936 and Hopkins, 1949 respectively); a more complete history of the variations in status of Suricatoecus and of Eichlerella is provided in Table IX. Sichler (1963) ihcluded the hitherto unpublished name Felicomorpha in his catalogue, without providing any further details, making the name a nomen nudum. In an earlier, unpublished work, Eichler had attributed this name to Keler m.s., and noted the type species, v/hich is a junior synonym of

T>. vulpis Denny. Felicomorpha is, however, not an available name.

Species included acutirostris (Stobbe, 1913) [Treated as Felicola s. str.by previous authors.] (26, 2$) bedfordi Hopkins, 1942 [Treated as Felicola s. str. by previous authors.] (4d, 11$) cooleyi (Bedford, 1929) Comb. n. from Suricatoecus. (3Qc, 28?) decipiens Hopkins, 1941 Comb. rev. from Suricatoecus. (9°, 9?) fahrenholzi (Y/erneck, 1948) , Comb. n. from Suricatoecus (l6d, 16$) fennecus (Emerson 2c Price, 1981). Comb. n. from Suricatoecus guinlei (Werneck, 1948). Comb. n. from Suricatoecus. (4°, 10?) macrurus Werneck, 1948 . [Treated as Felicola s. str. by previous

authors.] (23d, 19?) oygidialis Y/erneck, 1948 [Treated as Felicola s. str. by previous authors.] (366, 41?) ouadraticeps (Chapman, 1897) Comb. n. from Suricatoecu.5. (5d, 9?) vulpis (Denny, 1842) Comb. n. from Suricatoecus. (18°, 25?) 396

Figs 213 - 217. Felicola male genitalia. (213) F. (S.) acutirostris; (214) F. (S.) vulpis; (215) F. (s.) bedfordi; (216) F. (F.) miniimis (parameres displaced slightly apart); (217) F. (£>.) fahrenholzi. 397

3-3.5.7. Lorisicola Bedford

The genus Lorisicola comprises two subgenera.

Description

.interior of head with osculum present or absent*; dorsal preantennal sulcus cresent or absent; clypeal marginal carina broadened medially, median sclerite variable*; anterolateral margin of head sinuate or convex; preantennal portion of head of variable length, outline more or less broadly triangular or rounded, sometimes, if osculum absent, convexly produced anteriorly*. Temple margin convex, rectangular, or slightly produced laterally*. Male scape expanded or only slightly expanded, with longitudinal setal row present and comprising at least two setae*; male and female flagellomeres fused; male flagellum with one or two basally-articulated teeth*. Dorsum of head with setae short or of moderate length, sparse. Sitophore sclerite unmodified. Thorax with dorsal setae short or of moderate length* though frequently short and stout on posterolateral angles of pterothorax, not present on disc or medially posteriorly on prothorax or pterothorax. Abdomen oval or elongate, male segment IX not produced greatly. Abdominal spiracles absent, or present on segments III - VI.or III - VIII; posterior two pairs of spiracles, if six pairs present, sometimes very small and possibly non-functional*. Abdominal setae short or very short, frequently sparse dorsally; abdominal pleurum III frequently with posterior setal row comprising stout, conical setae (Fig.229); anterior setae absent except on pleurum II; postero-lateral setae present or absent, sometimes numbering more than one per site* (Fig.2l8). Pleural projections present dorsally and sometimes ventrally on pleurum IV, sclerotised or unsclerotised. Abdominal sclerotisation variable; sterna with sclerites absent except for subgenital plate (in male), or present on more posterior segments (VII, VI + VII-, V - VTI, IV - VII or

III - VII); terga with sclerites on I - VIII, II - VIII, III - VIII or IV - VIII; pleura with sclerites on at least II, sometimes also on III and IV; male terga with anterior and posterior sclerites present on at least terga IV - VTI, or posterior sclerites not present*. Gonapophyses with ventral marginal non-tuberculate setae; rounded 398

Figs 218 - 222. Lorisicola spp. (218) L. (L.) m.jobergi female: abdomen; (219) L. (L«) spenceri female: terminalia, ventral (setae omitted apart from on gonapophyses); (220) L. (P.) bengalensis female: subgenital lobe, ventral; (221) L. (L.) felis female: gonapophysis, ventral; (222) L. (P.) africanus female: head, dorsal. 399

or rectangular lobe present or absent on ventral margin*; apical spur present or absent*. Gonapophyses meet ventral vulval margin acutely, not linked by sclerotised band. Ventral vulval margin not sclerotised; straight or concave, v/ith chord less than 90 degrees to long axis of abdomen; subgenital lobe present, ventral surface more or less covered in overlapping scales*.

Male subgenital plate v/ith sternites VII, VIII and IX present and fused to s.g.p.r., VII and VIII fused to s.g.p.r. and IX absent, or VII fused to s.g.p.r. and VIII and IX absent or present but not fused to s.g.p.r.*. Pseudostyli absent. Hale genital opening postero-dorsal or dorsal; male segment IX lying dorsally on abdomen. Paraineres short, broad, sometimes fused. Basiparameral sclerites present or absent*.

I.Iesomeres present, fused; median extension present or absent*; mesomeres extending basally between b.a.l.s. to contact parameres, or terminating exteriorly to b.a.l.s.*. Male genitalia depicted in Figs .230 — 236.

Hosts Felidae, Herpestfdae and Viverridae (Carnivora) and Lorisidae (Primates).

Subgenus Lorisicola Bedford

Lorisicola Bedford, 1936: 51. Type-species: Trichodectes m,jobergi Stobbe, by original designation.

Description Anterior of head with osculum present; clypeal marginal carina broadened medially to form rectangular or 'W' -shaped sclerite, or broadened slightly to either side of osculum, very lightly sclerotised posterior -to osculum. Temple margin rectangular or slightly produced laterally. Male scape expanded or only slightly expanded, with setal row comprising two setae; male flagellum v/ith basally-articulated 'teeth1 on projection.

Abdominal spiracles absent, or present on segments III to VTII; posterior two pairs of spiracles, if six pairs present, sometimes very small and possibly non-functional. Males with posterior tergal sclerites absent. 400

Figs 223 - 225. Lorisicola (L.) male terminalia. (223) L* mjobergi; (224) L. similis; (225) L. spenceri. 401

Gonapophyses with rounded lobe, or lobe absent (Fig.218); apical spur present or absent (Fig.221). Subgenital lobe covered in overlapping pointed scales (Fig.215) or spines (Fig.218). Male subgenital plate with sternites VII, VIII and IX present and fused to s.g.p.r. (Fig.223 ), or with sternite VII fused to s.g.p.r. and VIII and IX present but not fused to s.g.p.r. (Fig.224), or of the latter form but lacking sternite IX (Fig.225 ), or lacking VIII and IX but VTI very broad. Basiparameral sclerites absent, Mesomeres fused, with median extension. Mesomeres extending basally between b.a.l.s. to contact parameres (Fig.233 ), terminating exteriorly to b.a.l.s. (Fig .23or extending anteriad to posterior end of b.a.l.s. and abruptly reversing, lying ventrally to b.a.l.s., though reversed portion is difficult to see (Fig .230).

Hosts Felidae and Viverridae (Carnivora) and Lorisidae (Primates).

Species included americanus (Emerson Sc Price, 1983) Comb. n. from Felicola. braziliensis (Emerson & Price, 1983)' Comb. n. from Felicola. caffra (Bedford, 1919) Comb. n. from Felicola. (16, 1$) felis (Werneck, 1934) Comb. n. from Felicola. (12, 1?) hgrcynianus (K£ler, 1957) • Comb. n. from Felicola. (66, 6?) malaysianus (Y/eraeck, 1948) Comb. n. from Trichodectes. (46, 5$) mjobergi (Stobbe, 1913) (cl002, clOO?)

neofelis (Emerson & Price, 1983) Comb. n. from Felicola. siamensis (Emerson, 1964) - Comb. n. from Felicola. (52, 2?) similis (Smers on Sc Price, 1983) Comb. n. Afrom Felicola. (12, 1?) soenceri (Hopkins, 1960) Comb. n. from Felicola. (82, 5$ ) sudamericanus (Emerson & Price, 1983) Comb. n. from Felicola.

Subgenus Paradoxuroecus Conci Gen. rev., Stat. n.

Paradoxu.roecus Conci, 1942: 141. Type-species: Paradoxuroecus juccii . Conci,. by original designation. Pa.rafelicola 7/emeek, 1948: 226. Type-species: Trichodectes acuticeos Neumann, by original designation. Syn. n. 402

Figs 226, 227. Lorisicola (Paradoxuroecus) spp. (226) L. (P.) acuticeps male: abdomen; (227) L. (?•) laticeps male: terminalia. 403 404

ITeofelicola YTeraeck, 1948: 235. Type-species: Neofelicola asoidorhynchus

Y/erneck, "by original designation. Syn. n.

Description Anterior of head with osculum present, in which case clypeal marginal carina broadened slightly to either side of osculum arid very lightly scler.otised posterior to osculum, or osculum absent, in which case clypeal marginal carina broadened medially to form posteriorly-convex or straight heavily sclerotised bar (Fig.222); outline of preantennal portion of head more or less broadly rounded or triangular, sometimes, if osculum .absent, convexly produced anteriorly (Fig.222). Temple margin convex or rectangular. Hale scape expanded or only slightly expanded, with longitudinal setal row comprising at least three setae; male flagellum with two basally-articulated 'teeth', only rarely on projection. Thorax with dorsal setae short. Abdominal spiracles absent, or present on segments III - VI. Postero-lateral setae present or absent, but never numbering more'than one per site if present. Males with posterior tergal sclerites absent, or anterior and posterior sclerites present on at least terga IV - VII. Gonapophyses with lobe present, rounded or rectangular; apical spur present. Subgenital lobe bilobate, with scales modified into short spines in some cases, though spines may be sparse (Fig.220).

Male subgenital plate with sternites VTI, VIII and IX present and fused to s.g.p.r. (Fig.226 ), VII and IX present and fused to s.g.p.r. but ViII absent (Fig.227), or VII and VTII present and fused to s.g.p.r. but IX absent (Figs 228,229). Parameres short, broad, not fused together (see second paragraph under 'comments' below), but may be vary closely associated (Figs 234, 235). Basiparameral sclerites present (Fig*. 236) or absent. Mesomeres fused apically with median extension present or, if absent, arch with two apical nipples (Fig* 232); mesomeral arch sometimes with lateral double flexion (Fig.236); mesomeres extending basally between b.a.l.s. to contact parameres, sometimes sharply recurved

posteriorly between b.a.l.s. (Fig.234)- Endophallus frequently with spicular collar, sometimes 'V'-shaped, around gonopore (Fig. 236a). 405

(231) L. (L.) malaysiaims; (232) L. (P.) laticeps; (233) L. (L.) spenceri. 406

Figs 234 - 236. Lorisicola (Paradoxuroecus) male genitalia. (234) L* (P.) bengalensis; (235) (£•) juccii; (236) L. (P.) acuticeps, with endophallus (a) not everted, and (b) everted. 407

Hosts

Herpestidae and Viverridae (Carnivora).

Comments

Pqxadoxuroecus has been considered by most authors to be a synonym of Felicola (synonymised by Y/emeck, 1948), and is raised here from synonymy to be a subgenus of Lorisicola. Neofelicola and Parafelicola were both considered by Hopkins (1949) to be subgenera of Felicola. A more complete history of the variations in the status of P arado xuro e cu a, II eof elicola and Paraf elicola is presented in Table IX. Werneck (1948) figured the parameres of the species aspidorhynchus, sumatrensis and juccii as fused together; examination of the type material of the former two species and of numerous specimens of the latter has revealed that this is not the case, although in JL. ,iuccii the parameres are very closely associated with each other. Lorisicola (Paradoxuroecus) bengalensis (Werneck, 1948) was described in the genus Neofelicola from three females, the male being unknown. These females were taken from a museum skin of Paradoxurus hermaphrodytus canus Miller, which was itself collected in Thailand. Female lice subsequently collected from_P. he imaphrodytus subspp. in Thailand agree with ' Yemeck' s (1948) description of II. bengalensis. Emerson (1965) describes the male of bengalensis, collected from the type host in Thailand, and distinguishes a new species, II. philippinensis, collected from Paradoxurus ohilippinensis. Numerous specimens of lice from a number of subspecies of Paradoxurus hermaphrodytus have been examined during the course of this study, and it has become apparent that lice of two clades are present: the Felicola (F.) ze:/lonicus - viverriculae clade and the Lorisicola (P.) philippinensis - ju-ccii clade. These clades may be distinguished by the following characters: F_. (F.) zeylonicus - viverriculae clade - Three pairs of abdominal spiracles; male flagellum v/ith 'teeth' not basally articulated; parameres long, slender, extending anteriorly between b.a.l.s.;

mesomeres not basally extending between b.a.Ls.; mesomeral arch lacking median extension; female subgenital lobe broad, smooth ventrally, with long, flattened marginal spines. 408

L. (]?•) ohilinuinensis - juccii clade - Four pairs of abdominal spiracles; male flagellum with 'teeth' basally articulated; parameres short, broad, closely-associated with one another but not fused, not extending between b.a.l.s., if reaching them; mesomeres basally extending between b.a.l.s. to contact parameres; mesomeral arch with median extension; female subgenital lobe narrow, apically bilobate, covered

ventrally with small pointed scales, lacking marginal spines. The female described as N. bengalensis by Werneck (1948) is, by the structure of the subgenital lobe and the number of abdominal spiracles, a member of the L. (.P.) •ohiliooinensis - juccii clade, as are both mal'e and female of IT. philippinensis as described by Emerson (1965) • The male described by Emerson (1965) as N. bengalensis is, however, a member of the (_F.) zeylonicus - viverriculae clade and therefore not the true male of II. bengalensis. Female lice of the latter clade are now known from Paradoxurus hermaohrodytus subspp., as are males of the former.

The male of L. bengalensis has genitalia of the same type as (in fact indistinguishable from) that of L. ohiliooinensis. Emerson (1965) distinguishes _L. ohiliupinensis from L. bengalensis by the male genitalia, the greater number of sternal and tergal setae in both sexes of the former, and the greater number of vulval setae in the foimer. The characteristics of the male genitalia, as stated above, are the same in the two species. Study of the large sample of specimens now available indicates that the vulval setal number of L. ohiliuoinensis is not outside the range of L. bengalensis. The tergal and sternal setae in the males are the same, as are the sternal setae in the females. The tergal setae of the female paratype of L. uhiliooinensis in the B.LI.(N.H.) collection are the same as those of jL. bengalensis, but do not agree with the figure in Emerson (1965), where far more setae are depicted. The host species

Paradoxurus philiooinensis is at best a subspecies of P. hermaohrodytus% and all other subspecies appear to harbour L. bengalensis. L. philippinensis (Emerson) is consequently provisionally synonomised with L. bengalensis (V/emeck), subject to examination of the female allotype of ohilioninensis.

Species included acuticeos (Neumann, 1902) , Comb. n. from Parafelicola. (2G5, 34?) africanus (Emerson c: Price, 1966) Comb. n. from Parafelicola. (156, 7?)

V 409

asoidorhynchus (Y/erneck, 1948) Comb. n. from ITeof elicola. (62, 7$) bengal ensis (Y/erneck, 1948) . Comb. n. from ITeof elicola. (262, 59$) juccii Conci, 1942 Comb. n. from Felicola. (822, 81?) laticeos (Y/erneck, 1942) Comb. n. from Suricatoecus. (102, 13?) lenicorais (Y/erneck, 1948) Comb. n. from Parafelicola. (92, 22?) mungos (Stobbe, 1913) Comb. n. from Suricatoecus. (12, 3$) neoafricanus (Emerson & Price, 1968) Comb. n. from Parafelicola. (Holotype 6,4?)

oaralaticeos (7/erneck, 1948) , Comb. n. from Suricatoecus. (12, 4?) nhilinpinensis (Emerson, 1965) Comb. n. from ITeof elicola. (52, 5?) sumatrensis (TYemeck, 1948) Comb. n. from ITeofelicola. (46, 4?) wernecki (Hopkins, 1941) Comb. n. from Parafelicola. (72, 11?)

3.3.6. Subfamily n.

3.3.6.1. Heotrichodectes Swing

The genus ITeotrichodectes comprises five subgenera.

Description • • Anterior of head with osculum present, sometimes very shallow*; dorsal preantennal sulcus present, not always clear; clypeal marginal carina broadened medially into dorsal sclerite of variable form* which is always more heavily sclerotised laterally (dorsal to margin of clypeus and pulvinus) than medially; anterolateral margin of head convex, straight or sinuate*; preantennal portion o'f head short or longer, some- times as long as postantennal portion*, outline broadly rounded or triangular*. Temple margin broadly convex. Hale scape expanded, with longitudinal setal row present and comprising at least four setae;, flagellomeres fused in males and females; male flagellum with two or more basally-articulated 'teeth1 distally, and with basal toothed or rough projection sometimes present* (Fig.238); female pedicel lacking membranous projection. Dorsum of head with most setae short or of moderate length, more or less sparse, sometimes with long seta on posterior margin of temple* (Fig.237 ) . Sitophore sclerite unmodified. Thorax with dorsal setae long or of moderate length* though, frequently short and spine-like on posterolateral angles of orothorax 410

Figs 237 - 241. Ueotrichodectes spp. (237) N. (T.) barbarae male: temple margin; (238) N. (n.7) semistriatus male: flagellum; (239) N. (N.) mephitidis female: abdominal pleura II and III, dorsal; (240) N. (T.) barbarae male: anterior abdominal segments, dorsal; (241) N. (TT.) mephitidis male: terminalia. 411

and pt ero thorax; posterior margin of pronotum with four setae and wide median gap, posterior of pterothorax dorsally with setae more abundant, marginal or submarginal, median gap present or absent. Abdomen broadly rounded, not greatly projecting posteriad in male (Fig.241). Abdominal spiracles absent. Abdominal seta.e generally abundant, as long as segment or sparse, shorter, with long setae present only on posterior pleura*; terga, especially of males, with median and lateral setal groups distinct, though median groups generally united; male terga II - VI with median two setae much smaller than other setae in the row, sometimes separated by one or more longer setae (Fig.241); anterior setae present on pleura II and III only; postero-lateral setae presumed absent, though, possibly present as the most lateral seta of lateral group, which is frequently situated more posteriorly than other setae. Abdominal pleura lacking projections, except for small sclerotised or unsclerotised projection dorsally on pleurum III of female IT. meuhitidis. Abdominal terga and sterna lacking .sclerites, except for lateral rods of subgenital plate in male and, sometimes, tergite IX in female; abdominal pleura- usually unsclerotised, sometimes with sclerites on pleura II, III,

IV and, in females, "'/III*. Gonapophyses v/ith or without lobe on ventral margin, very variable*. Gonapophyses meet ventral vulval margin smoothly-or acutely, but not linked by sclerotised band. Ventral vulval margin not sclerotised; generally convex medially*, subgenital lobe present or absent*. Common oviduct not notably striate (cf. Geomydoecus). Male subgenital plate with only s.g.p.r. present (Fig.241). Pseudo- styli absent. Male genital opening dorsal. Parsmeres fused to form single plate with apex pointed or bifid*. Basiparameral sclerites absent. Mesomeres fused apically; mesomeral arch with median extension*; mesomeres basally abut postero-lateral projections of b.a.l.s. but do not contact parameral plate (Fig.251), or contact neither b.a.l.s. nor parameral plate (Fig. 250).

Hosts Mustelidae and Procyonidae (Carnivora) and Bradypodidae (Edentata). 412

Figs 242 - 244. Neotrichodectes female terminalia. (242) N. (H.) mephitidis, ventral aspect; (243a) N. (n.7) chilensis, showing setal distribution; (243b) N. chilensis. showing sclerites; (244) N. (T.) barbarae. 413

Comments ITeotrichodectes has been considered a synonym and a subgenus of Trichodectes (by Hopkins, 1942 and Hopkins, 1949 respectively); a more complete history of the variations in status of ITeotrichodectes is presented in Table VIII.

Subgenus ITeotrichodectes Swing

ITeotrichodectes Swing, 1929: 194. Type-species: Goniodes meohitidis Packard, by original designation.

Description Anterior of head with clypeal marginal carina broadened into dorsal sclerite which is more or less convex posteriorly, sometimes almost circular; preantennal portion of head sometimes as long as postantennal portion. Male flagellum with two basally-articulated 'teeth' distally, and with toothed projection sometimes present basally (Fig.238). Temple with long seta present on posterior margin. Thoracic and abdominal setae long, abundant. Abdominal pleura, lacking projections, except for small sclerotised or unsclerotised projection dorsally on pleurum III of female H. meohitidis (Fig.239). Abdomen lacking sternal, tergal and pleural sclerites, except for lateral rods of male subgenital plate. Gonapophyses broad, membranous, with ventral marginal setae absent or, if present, basal only (Fig.242); ventral lobe absent. Ventral vulval margin with lobe present though difficult to see; lobe serrate, at least along posterior margin (Fig.242). Female genital chamber with dorsal wall bearing slanting scales laterally, lightly sclerotised and lacking scales or other decoration medially (anteriorly). Parameral plate slender, apically bifid, with median basal extension reaching anteriorly between b.a.l.s.. Mesomeral arch with median extension pointed; mesomeres basally abut postero-lateral extensions of b.a.l.s.. Male genitalia depicted in Fig. 248*

Hosts Hustelidae and Procyonidae (Camivora) . 414

Figs 245, 246. Neotrichodectes female terminalia. (245) N. (L.) gastrodes; (246) N. (n.6) pallidas. 415

Comments IT e o t r i cho d e c t e s s. str. is most readily distinguished from other subgenera of Neotrichodectes by characters. of the female terminalia: the retention of the plesiomorphic form of the gonapophyses (found also in Geomydoecus) and the apomorphic development of a membranous subgenital lobe (cf. the very different structure in IT. (Trigonodectes)). Assignment of male insects to the subgenus relies on absence of the basal flagellar projection (a character-state reversal not undergone by the type species of the subgenus), the presence of an anterior development of the parameral plate between the b.a.l.s., and the slenaerness of that parameral plate in relation to its length. ITeotrichodectes wolffhuegeli* (Herneck) is known only from the male, although Y/erneck (1948) predicted that the female would be very similar to that of N. chilensis (placed in subgen. n. 7 in this study), and so must be assigned to subgenus on the basis of male characters. IT. wolffhuegeli does have a toothed projection on the base of the male flagellum, although, as indicated above, cannot be eliminated from IT. s. str. on that basis. The form of the parameral plate of IT. wolffhuegeli is. much the same as members of IT. s. str. and on this evidence the species is placed in the subgenus. Emerson (oers. comm.), however, suggests that the male genitalia of N. wolffheugeli lie within the limits of permissible variation of IT. chilensis, which he therefore considers as a junior synonym; IT. wolffheugeli and IT. chilensis are found on the same host, although IT. chilensis is also found on several other species of the host genus. All specimens identified e.s IT. wolffheugeli Coy V/erneck and in the present study) have a. much narrower parameral plate than those identified as IT. chilensis. However, the parameral plate, whilst not extending anteriad between the b.a.l.s. in most specimens of IT. chilensis, does do so in some. In those specimens of N. chilensis where there is no anteriad extension of the parameral plate the endophallus lacks large, heavily-sclerotised teeth, or such - teeth are few in number; in those where the extension is present the teeth are correspondingly more developed. The degree of development of the endophallus teeth seems to be proportional to the degree of development of the anterior margin of the parameral plate in IT. chilensis and such

*The original spelling ''wolffhtlgeli' is modified here to 'wolffhuegeli' in accordance v/ith Article 32(c)(i) of the International Code of Zoological Nomenclature (1964). 416

Figs 247 - 251. Neotrichodectes male genitalia. (247) N* (n.6) pallidus; (248) N. (IT.) mephitidis; (249) N- (L-) gastrodes; (250) N. (T.) barbarae; (251) N. (n.7) chilensis. 417

teeth are present and well-developed in IT. wolffhuegeli. The sample of specimens of both species was too small to permit any correlations of genetalia type with host species or geographical distribution, though within IT. chilensis specimens exhibiting both extremes were found from the same host in the same area. For the purposes of this study, the two species are treated as separate, N. wolffheugeli is assigned to

Neotrichodectes s. str, , and IT. chilensis is assigned to ITeotrichodectes (Subgen. n. 7). This conclusion is regarded as the most satisfactory " for the data presently available, but further collecting from species of the host genus (Conepatus) is needed to clarify the situation.

Species included mephitidis (Packard, 1873) (c50d, clOO?) minutus (Paine, 1912) (646, 76?) osborni Keler, 1944 (162, 21?) thoracicus (Osborn, 1902) (Il6, 9?) wolffhuegeli (7/erneck, 1936) (12)

Subgenus Trigonodectes Keler Gen. rev., Stat. n.

Trigonodectes Keler, 1944: 179, 185. Type-species: Trichodectes barbarae Neumann, by original designation.

Description Anterior of head with osculum very shallow; clypeal marginal carina broadened medially into dorsad 'U'-shaped sclerite with median posterior process; anterolateral margin of head convex; preantennal portion of head with outline rounded. Hale flagellum with two basally- articulated 'teeth' distally, and with toothed projection present basally. Temple with long seta present on posterior margin. Thoracic and abdominal setae long, abundant. Abdominal pleura lacking projections. Abdomen lacking sternal, tergal and pleural sclerites, except for lateral rods of male subgenital plate.

Gonapophyses slender, apically acute, sclerotised, with strong setae present along ventral margin; ventral lobe absent. Gonapophyses meet ventral vulval margin acutely. Ventral vulval margin with median 418

lobulate projection, with margin not serrate, and submarginal setal row present (Fig.244). Female genital chamber lacking lateral slanting scales and anterior sclerotised area on dorsal wall, but both ventral and dorsal walls bearing numerous scales with posterior spinules. Parameral plate triangular or shield-shaped, pointed apically. Mesomeral arch with median extension rounded, covered in small tubercles; mesomeres extend basally anterior to ends of b.a.l.s., and do not contact b.a.l.s. or parameral plate. 3asal apodeme lacking postero-lateral projections on b.a.l.s.. Male genitalia depicted in Fig.250.

Hosts Mustelinae (Carnivora: Mustelidae).

Comments

Trigonodectes has been treated as a synonym end a subgenus of Trichodectes (by Y/erneck, 1S48 and Hopkins, 1949 respectively); in this study it is raised from synonymy with Trichodectes and barbarae is placed for the first time in Neotrichodectes. A more complete history of the varying status accorded to Trigonodectes is presented in Table VIII.

Species included

• barbarae (Neumann, 1913) - Comb. n. from Trichodectes. (136, 13?)

Subgenus n. 6

Type-species: Trichodectes oallidus Piaget

Description

Anterior of head with osculum shallow; clypeal marginal carina broadened into dorsal rectangular sclerite; anterolateral margin of head convex; preantennal portion of head not as long as postantennal portion, outline broadly rounded. Male flagellum with two basaily- articulated 'teeth1 distally, -and with toothed projection present basally. Temple with long seta present on posterior margin. Thoracic and abdominal setae long, abundant. Abdominal pleura lacking projections. Abdomen lacking sternal, tergal and pleural sclerites, except for lateral rods of male subgenital plate. 419

Gonapophyses with rounded ventral lobe with submarginal setae; spur distal to lobe very short. Gonapophyses meet ventral vulval margin acutely. Ventral vulval margin convex, but subgenital lobe or lobulate process not present. Female terminalia depicted in Fig. 246. Female genital chamber with dorsal wall bearing slanting scales laterally, spines medially (spines most apparent anteriorly, though may be obscured). Parameral plate apically bifid, lacking median basal extension reaching anteriorly between b . a. 1. s •. I'.I esomeral arch v/ith median extension pointed; mesomeres basally abut postero-lateral extensions of b.a.l.s.. Male genitalia depicted in Fig.247-

Hosts Procyonidae (Camivora).

Species included oallidus (Piaget, 1880) . ['Treated by previous authors as ITeotrichodectes s. str.] . (65d, 75?)

Subgenus Lakshminarayanella Eichler Stat. n.

Lymeon Eichler, 1940: 158 [ Nee Foerster, 1868: 176] Type-species: Trichodectes gastrodes Cummings, by monotypy. Lakshminarayane11a Bichler, 1982: 83* Horn.-nov. for Lymeon Eichler nec Foerster.

Description Anterior of head with osculum present, deep; clypeal marginal carina broadened medially into posteriorly convex bar, parallel to curvature of osculum; anterolateral margin of head straight or slightly sinuate; preantennal portion of head short in male, longer in female; outline broadly triangular. Make flagelluin with seven basally-articulated 'teeth' distally, and with roughened projection present basally. Temple v/ith no long setae present on posterior margin.

Thorax with dorsal setae of moderate length, longest on postero- lateral angle of pterothorax; pterothorax with setae sparse along posterior dorsal margin. Abdominal setae of moderate length, not as long as segment except on posterior pleura; terga with lateral and median setal groups not 420

clearly distinct, median gap sometimes pronounced. Abdominal pleura lacking projections. Abdomen lacking tergal and sternal sclerites except for lateral rods of male subgenital plate and tergite IX of female; abdominal pleura II, III, IV and, in femal^ VIII, with sclerites, though that of IV sometimes very small.

Gonapophyses broad, very thick; ventral lobe present, thick, with setae along posterior margin; spur distal to lobe not short, rounded apically (Fig.245 ) • Gonapophyses meet ventral vulval margin acutely. Ventral vulval margin convex, but subgenital lobe or lobulate process not present. Female genital chamber with dorsal wall bearing slanting scales laterally, spines medially (spines most apparent anteriorly). Parameral plate apically bifid, sometimes projecting slightly anteriad between b.a.l.s.. Mesomeral arch with median extension pointed, broad basally; mesomeres basally abut posterolateral extensions of b.a.l.s.. Male genitalia depicted in Fig.249.

Hosts 3radypodidae (Edentata).

Species included cummingsi (Eichler, 1943) Comb. n. from Lakshminarayanella. gastrodes (Cummings, 1916) Comb.n. from Lakshminarayanella. (46, 6$, 4 nymphs)

Subgenus n. 7 Type-species: Heotrichodectes semistriatus Emerson and Price

Description As generic description, but: Anterior of head with clypeal marginal carina broadened medially into dorsal 'U'-shaped sclerite with median posterior process; antero- lateral margin of head convex or sinuate; preantennal portion oihead not as long as post ant ennal portion, outline broadly rounded. Male • flagellum with two basally-articulated 'teeth' distally, and with'.toothed projection basally. Temple with long seta present on posterior margin.

Thoracic and abdominal setae long, abundant. Abdominal oleura lacking projections. Abdomen lacking sternal and tergal sclerites, 421

except for sternal rods of male subgenital plate andtergite IX of female; abdominal pleura II, III and VTII in female sometimes with sclerites, though that of III may be very small. Gonapophyses with ventral lobe present, large, apparently comprising fused setal tubercles, with setae along posterior margin and anterior margin, the latter frequently directed posteriad; spur distal to lobe present, not short, frequently obtuse apically. Gonapophyses meet ventral vulval margin acutely. Ventral vulval margin convex, but subgenital lobe or lobulate process hot present. Female terminalia depicted in Figs 243a, 243b* Female genital chamber v/ith dorsal wall bearing slanting scales laterally, spines medially, sometimes lightly sclerotised and lacking scales,' spines or other decoration anteromedially.

Parameral plate apically bifid, sometimes projecting slightly anteriad between b.-a.l.s.. Mesomeral arch v/ith median extension pointed; mesomeres basally abut posterolateral extension of b.a.l.s.. Male genitalia depicted in Fig.251.

Hosts Mustelinae and Mephitinae (Carnivora: Mustelidae).

Species included arizonae Werneck, 1948 (186, 39?) chilensis 7/erneck, 1948 (c506, c50?)

interruptofasciatus (Kellogg & Ferris, 1915) (27^, 46?) semistriatus Emerson & Price, 1976 (56, 5?) All the above species have been treated previously as ITeotrichodectes s. str.

3.3.6.2. Geomydoecus Ewing

The genus Geomydoecus comprises tv/o subgenera.

Description

Anterior of head v/ith osculum present; dorsal preantennal sulcus present; clypeal marginal carina broadened medially into posteriorly convex bar; .anterolateral margin of head convex or sinuate; preantennal portion of head not long, outline broadly triangular or rounded. Temple 422

margin broadly convex. Male scape expanded; longitudinal setal row present and comprising at least three setae; male scape sometimes with median posterior projection; flagellomeres fused in males and females; male flagellum "with, two b as ally-articulated 'teeth'; female pedicel with membranous postero-ventral projection (Fig. 252), sometimes obscure. Dorsum of head with setae short or of moderate length, more or less sparse; temple margin sometimes with specialised long, fine or short and stout lateroposterior setae.* Sitophore sclerite unmodified. Thorax with dorsal setae long or of moderate length; posterior margin of pronotum with four setae and wide median.gap, posterior margin of pterothorax dorsally with varying number of marginal or submarginal setae. Abdomen broadly rounded or more elongate and tapered, particularly in male. Abdominal spiracles absent. Abdomen with at least some setae as long as segment; setae generally abundant.; terga, especially, of males, with median and lateral setal groups distinct, though median groups generally united; male terga II - VI without median setae shorter than others; male terga II and III sometimes v/ith median group comprising exceptionally long, stout setae* (Fig.253); anterior setae present on pleura II and III only; postero-lateral setae sometimes clearly present (Fig.253),.otherwise obscured, though may be present as most lateral seta of lateral group, which is frequently situated more posteriorly than other setae. Abdominal pleura v/ith projections dorsally on pleura II, III, TV and, at least in female, ventrally on IV, sclerotised (Fig.255) or unsclerotised*; projections generally more apparent in females than males. Abdominal terga and sterna lacking sclerites, except for lateral rods of male subgenital plate and, sometimes, terga II - IV of male*; abdominal pleura II and sometimes III and IV sclerotised, at least in female; other pleura unsclerotised.

Oonapophyses broad, membranous, with ventral marginal setae, if present, generally basal only; ventral lobe absent (Figs 254 , 255)• Gonapophyses meet ventral vulval margin smoothly or acutely, but not linked by sclerotised band. Ventral vulval margin not sclerotised; generally convex or very convex medially; subgenitaJ lobe not present. Female genital chamber v/ith dorsal wall bearing slanting scales' laterally 423

(254) G. (G«) californicus female: terminalia. 424

lightly sclerotised and lacking scales or other decoration antero-medially. Common oviduct generally with distinct striae of species-characteristic form. Male subgenital plate v/ith only s.g.p.r. present. Pseudostyli absent. Male genital opening dorsal. Parameres fused to form single plate v/ith apex pointed or bifid*. Basiparameral sclerites absent. • Mesomeres fused apically; mesomeral arch v/ith or without median extension; mesomeres basally abut b.a.l.s., which sometimes have posterolateral extensions; mesomeres do not contact parameral plate. Male genitalia depicted in Figs 256 - 259.

Hosts

Geomyidae (Rodentia).

Comments

A few of the:-species are partheno genetic. ITo more than 25 specimens of most species were examined during the course of this study, though in many cases large numbers v/ere available. Detailed descriptions of all species of Geomydoecus s. lat. and a phenetic treatment of the genus may be -, found in Kellenthal and price (1976, 1980), Price, R.D. (1974, 1975), Price and Emerson (1971, 1972), Price and Hellenthal(1975a, 1975b, 1976, 1979, 1980a, 1980b, 1980c, 1981a, 1981b), Price and Timm (1979), Timerand Price (1979, 1980).

Subgenus Geomydoecus.- Swing

Geomydoecus Swing, 1929: 193- Type-species: Trichodectes geomydis Osbom, by original designation.

Description Temple margin sometimes v/ith two short, stout setae latero-posteriorly, or single long, fine seta latero-posteriorly. Male abdominal terga II and III only rarely v/ith median setal group comprising exceptionally long, stout setae (J7-. conei). Pleural projections rarely sclerotised. Hale terga II - IV not sclerotised. Hale genitalia, not a.synmetric. Parameral plate apically pointed or bifid. Male genitalia depicted in Figs 256, 257. 425

Hosts Geomyidae (Rodentia).

Species included actuosi Price & Hellenthal, 1961. (256, 25$) albati Price 2c Hellenthal. 1981. (25d, 25?) alcorni' Price & Emerson, 1971. (66, 6$) alleni Price 2c Emerson, 1971. (2

guadaluoensis Hellenthal 8c Price, 1980. (252, 25?) heaneyi Timm Sc Price, 1980. (252, 25?) hoffmanniPrice Cc Hellenthal. 1976. (252, 25?) hueyi Price 1 Hellenthal, 1980. (25c, 25?) idahoensis Price & Emerson, 1971. (25 2, 25?) illinoensis Price 8c Emerson, 1971. (25c, 25?) jaliscoensis Price 8c Hellenthal, 1981. (252, 25?) jonesi Price S: Emerson, 1971. (42, 5?) limitaris limitaris Price & Hellenthal, 1981. (252, 25?) limitaris bakeri Price 8c Hellenthal. 1981. (252, 25?) limitaris halli Price & Hellenthal. 1981. (252, 25?) limitaris tolteci Price & Hellenthal, 1981. (25 2, 25?) martini Price & Hellenthal, 1975. (25c, 25?) mcgregori Price & Emerson, 1971. (25c, 25?) merriami Price & Emerson, 1971. (25 4 25?) mexicanus Price 8c Emerson, 1971. (252, 25?) mobilensis Price, 1975. (25?) musculi Price 8c Hellenthal-, 1981. (252, 25?) nayaritensis Price 8c Hellenthal, 1981. (162, 25?) nebrathkensis Timm 8c Price, 1980. (252, 25-?) oklahomensis Price 8c Emerson, 1971. (25 2, 25?) oregonus Price 8c Emerson, 1971. (252, 25?) panamensis Price 8c Emerson, 1971. (23c, 22?) oattoni Price & Hellenthal, 1979. (12 2 , 8?) perotensis perotensis Price 8c Emerson, 1971. (25o, 25?) oerotensis irolonis Price 8Z Emerson, 1971. (25.2, 25?) oolydentatus Price 8c Emerson, 1971. (25 2, 25?) auadridentatus Price & Emerson, 1971. (25 2, 25?) scleritus (McGregor, 1917) (36, 25?) setzeri Price, 1974. (66, 8$) shastensis Price & Hellenthal, 1980. (256, 25$) sinaloae Price 8c Hellenthal, 1981. (256, 25$) spickai Timm & Price, 1980. (252,25?) sub c al if o mi cu s Price 8c Emerson, 1971. (25 c, 25?) subgeomydis Price 8c Emerson, 1971. (25 2, 25?) 427

subnubili Price Zz Hellenthal, 1975. (256, 25?) tamaulipensis Price & Hellenthal, 1975. (3 6, 25$) texanus texsnus Ewing. 1936. (256, 25$) texanus trooicalis Price & Hellenthal. 1975. (256, 25$) thomomyus (McGregor, 1917). (25c, 25$ ) tolucae Price & Emerson, 1971. (256, 25$) traubi Price Zz Emerson, 1971. (25 6 , 25$) trichopi Price Zz Emerson 1971. (256, 25$)

t rune at us Werneck. 1950. (256, 25$) umbrini Price Zz Emerson, 1971. (256, 25$) ustulati ustulati Price & Hellenthal, 1975. (256, 25$) ustulati clarkii Price Zz Hellenthal, 1975. (256J 25$) veracruzensis Price Zz Emerson, 1971. (25 6, 25$) v/armanae Price Zz Hellenthal, 1981. (25c, 25$) v/ellerl v/elleri Price Zz Hellenthal, 1981* (256, 25$) v/elleri multilineatus Price Zz Hellenthal, 1981. (256, 25$) wernecki Price & Emerson, 1971. (256, 25$) yucatanensis Price Zz Emerson, 1971. (256, 25$)

Subgenus Thomomydoecus Price Zz Emerson

Thomomydoecus Price & Emerson, 1972: 464 [as subgenus of Geomydoecus Ewing] . xjnpe-species: Geomydoecus (E homo my do ecus) j amesbeeri Price Zz Emerson, by original designation.

Description Temple margin with single stout seta and finer, shorter adjacent setae latero-posteriorly. Male abdominal terga II and III v/ith median setal group comprising exceptionally long, stout setae. Pleural projections sclerotised, at least in female (Fig. 233) • Male terga II - IV sometimes v/ith sclerites. Gonapophyses meet vulval margin smoothly. Male genitalia symmetric (Fig.259) or asymmetric (Fig.258). Parameral plate apically pointed.

Eosts Thomomvus spp. (Rodentia: Geomyidae). 428

Fig. 255* Geomydoecus (T.) minor female: abdomen. Hatched areas indicate damage to specimen. Figs 256 - 259. Geomydoecus male genitalia. (256) G. (G.) actuosi; (257) G. (G.) thomomyus; (258) G. (T.) minor; (259) G. (T.) wardi. Species included

asymmetricus Price Zc Hellenthal, 1980. (256, 25$) birneyi Price & Hellenthal, 1980. (256, 25$) dickermani Price Zz Emerson, 1972. (256, 25$) genowaysi Price & Emerson, 1972. (256, 25$) greeri Price Zz Eellenthal, 1980. (26) jamesbeeri Price Zc Emerson, 1972. (86, 10$) jolinhafneri Price Zc Hellenthal, 1980. . (25 6, 25$) markhafneri Price Zz Hellenthal, 1980. (25 6, 25$) minor Wsrneck, 1950 (25 6,25$) neocooei Price Zc Emerson, 1971. (26, 1$) orizabae Price Zc Hellenthal, 1980. (106, 26$) peregrini Price & Hellenthal, 1980. (46, 4$) ootteri Price & Hellenthal, 1980. (166, 25$) timmi Price Zc Hellenthal. 1980. (256, 25 $ wardi Price Zc Emerson, 1971. (256, 25$) williamsi Price Zc Hellenthal, 1980. (8d, 16 $ zacatecae Price Zc Hellenthal, 1980. (25d, 25?) 431

3.4. KEYS TO TRICHODBCTIDAE

3.4.1. Introduction Two keys are provided: a key to subfamilies and a key to'genera and subgenera. The subfamily key is derived in part from the generic key, but is not as usable, and should not be employed as a supplement or introduction to the latter. As noted in section 3.2. above, it is a criterion of availability that any new taxon name be accompanied by some form of description, and it is in fulfilment of this formal require- ment that the subfamily key is presented.

3.4.2. Key to Subfamilies of Trichodectidae

1. No abdominal spiracles present; majority of tergal and sternal

setae at least two-thirds the length of the segment or, if not, median setal group on tergum II comprising at least three setae (and, frequently, median groups running together). New World. Subfamily n. At least one pair of abdominal spiracles present or, if not, majority of abdominal sternal and tergal setae less than two-thirds the length of the segment or median setal group on tergum II comprising only one seta. Old and New World 2

2. Female subgenital lobe present, frequently v/ith serrate margin, at lea.st posteriorly; if margin of subgenital lobe smooth, gonapophyses meet vulval margin smoothly (Fig.166); female flsgellomeres fused; abdominal spiracles numbering six or fewer pairs. Parasitic on Carnivora and Primates. Old and New World Trichodectinae Kellogg, 1896 Female subgenits.l lobe absent or, if present, not marginally serrate and gonapophyses meet vulval, margin acutely; female flagellomeres fused or unfused; abdominal spiracles numbering six pairs,though spiracles on segment VIII may be very small and inconspicuous, possibly non-functional (some species- of Procaviohilus (ileganarionoides)). Not parasitic on Carnivora. Old and Hew World 3 432

Dorsal or ventral projection present on abdominal pleurum IV; mesomeral arch generally produced basally between b.a.l.s.; female antennal flagellomeres generally not fused, or only partially fused; parasitic on hyraxes

and primates. Old and New World Dasyonyginae Keler, 1938 Pleurum IV lacking -any projection; mesomeral arch rarely produced basally between b.a.l.s.; female flagellomeres generally fused or, if not, then female with long setal tufts on abdominal pleura VIII and IK; not parasitic on hyraxes or primates 4

Posterior margin of temple generally produced, very convex; (Fig. 101); very long setae present on at least pleurum VIII, sometimes also on pleura VII (male) or IX (female); basi- parameral sclerites present; mesomeral arch lacking extension if complete, otherwise tripartite, median part sometimes-.obscure (Figs 105, 106); if mesomeral arch entire, male genitalia as in Fig. 104, temple margins not greatly produced, and female with two fla,gellomeres; parasitic on- New World porcupines (Erethizontidae)

Eutrichophilinae Keler, 1938 Temples not so developed; setae on pleurum VIII not exceptionally long; basiparameral sclerite present or absent; nesomeral arch, if present, with or without extension, but arch never tripartite; female flagellomeres fused; pseudostyli frequently present; parasitic on Artiodactyla and Perissodactyla Old and ITew World Bovicolinae Keler, 1938

.3. Key to Genera and Subgenera of Trichodectidae

No abdominal spiracles present; majority of tergal and sternal setae at least two-^thirds the length of the segment or, if not, median setal group on tergum II comprising at least three setae (and, frequently, median groups running together). 433

At least one pair of abdominal spiracles present or, if not, majority of abdominal sternal and tergal setae less than two-thirds the length of the segment or median setal group on tergum II comprising only one seta. Old and Hew World 8

2. Abdominal pleura II-IV with dorsal projections (Fig.255), though most apparent in females and sometimes very inconspicuous; male lacking tergocentral microsetae; 1ateroposterior corner of temple margin frequently with single long fine seta or one or two shorter, stout setae; female pedicel with dorsal membranous projection (Fig.252) (sometimes obscure). [Geomyiaae] Geomydoecus s. lat .... 3 Abdominal pleura lacking dorsal projections, or single membranous projection present on pleurnm IV only (EIg.239); male with tergocentral riiicrosetae on abdominal terga II - VI (Fig. 240): long seta frequently present on temple margin but shorter stout seta not developed; female pedicel lacking any projection Heotrichodectes s. lat .... 4

3. Pleural projections on pleurum II sclerotised; temple margin with single stout seta and associated smaller finer setae; male abdominal terga II and III v/ith rows of enlarged setae (Fig. 253); parameral plate with single apical point; male genitalia symmetric or asymmetric....Geomydoecus (Thomomydoecus) Pleural projection on pleurum II unsclerotised or, if sclerotised, posterolateral temple margin v/ith single long fine seta and associated smaller setae; temple margin with or without specialised setae but not v/ith single stout seta; male abdominal terga II and in rarely v/ith rows of specialised setae (G. cooei only); parameral plate v/ith single apical point or apically bifid; male genitalia symmetric ...... Geomydoecus (G eomydoe cus) 434

4. Female sub genital lobe present, v/ith serrate margins; female genital chamber v/ith clear, flat dorsal region but lacking single scattered spines; gonapophyses broad, membranous, lacking lobe; parameral plate slender, v/ith basal projection between b.a.l.s. (Fig.248). [ivlustelidae and Procyonidae] .... ITeotrichodectes (I'eotrichodectes) Female subgenital lobe absent or, if present, lobe v/ith smooth margins and longitudinal setal rov/s (Fig.244); female genital chamber, if with clear flat dorsal area, then v/ith single spines scattered over it; gonapophyses not broad and membranous, frequently v/ith lobe; parameral plate broad, with very limited projection between b.a.l.s 5

5. Ventral vulval margin v/ith lobulate process v/ith smooth margins and longitudinal rov/s of setae (Fig.244); gonapophyses slender, sclerotised,lacking lobe; parameral plate v/ith single apical point; mesomeral arch extension broad, clubbed (Fig.250). [i.Iustelidae] ITeotrichodectes (Trigonodectes) Ventral vulval margin convex, but not produced; gonapophyses not slender, lobe present; parameral plate v/ith apex bifid; mesomeral arch with pointed extension 6

6. Large species, over 2.75 mm. long; male flagellum v/ith 7 articulated 'teeth1; female gonapophyses thick, with lobe and spur (Fig. 245) J female pleurum VIII sclerotised; abdominal setae relatively small, not attaining following setal row. f Bradyoodidae] .. .ITeotrichodectes (Lakshminarayanella) Smaller species, under 2.25 mim long; male flagellum v/ith 2 articulated 'teeth'; female gonapophyses otherwise; female pleurum VIII not sclerotised; abdominal setae long, attaining or nearly attaining setal bases of following setal row. [Camivora] 7

7. Gonapophyses with flat lobe and small spur (Fig. 246); male mesomeral arch extension attaining end of parameral plate (Fig. 247) [ Procyonida.eJ ITeotrichodectes (n. 6) Gonapophyses with lobe comprising fused setal tubercles, long spur present (Fig.243); male mesomeral arch extension reaching beyond apex of parameral plate (Fig.251). [l.Iustelidae] ITeotrichodectes (n. 7) 435

8. Five pairs of abdominal spiracles present; vulval margin sclerotised, with or without setal tubercles, and meeting gonapophyses smoothly; subgenital lobe present; parameres not fused to b.a.l.s.; mesomeres absent; postcoxale absent; abdominal segments II - V with median setal group present, comprising at least three

setae.[ I.Iustelidae] Trichodectes (n. 5) Other than five pairs of abdominal spiracles present, though spiracles on segment "'/III may be very small, inconspicuous and possibly non-functional (some species of Pro caviohilus (Meganarionoides) as described in key couplet 26, and some species of Trichodectes (Stachiella), as described in key couplet 14) 9

9 Abdominal pleura V -VI (at least) lacking setae 10 Abdominal pleura III - VIII (at least) with posterior setal row and, sometimes, anterior setae,... 12

10. Abdominal tergal setae on segments I - VI less than half the length of the segment, shorter than the postero-lateral setae; pleura V - VI lacking setae (Fig. 175); male flagellum'with two basally-articulated 'teeth'; mesomeres present, unfused; parameres fused, with distinct inturned

apices arising from plate (Fig.181); subgenital lobe bifurcate, with long basal lateral processes (Fig.164). [llust elidao] Trichodectes (Trichodectes)(in part) At least some setae on abdominal terga I - III as long or longer than segment and postero-lateral seta, and postero- lateral setae sometimes absent; pleura IV - VII (at least) lacking setae (Fig. 154); male flagellum lacking 'teeth'; mesomeres absent; parameres unfused or united at ba.se only; subgenital lobe not bifurcate or only slightly so; basal processes-of subgenital lobe absent or, if present, not long (Fig. 159). [Lutrinae ] 11 436

11. Posterior setal row present on pleurum III; parameres slender, rod-like, fused basally (Figs 157 » 158); subgenital lobe lacking basal lateral processes; gonapophyses lacking setal tubercles (Fig. 156) Lutridia Posterior setal row hot present on pleurum III; parameres broad, not fused to each other (Fig.160); subgenital lobe with basal lateral processes; gonapophyses with setal tubercles (Fig-. 152) Genus n. 4

12. Ventral vulval margin meets gonapophyses smoothly, joined by sclerotised band; subgenital lobe present, frequently v/ith basal lateral processes; sternal setae on at least segments III - VT attaining or nearly attaining base of following setal row; dorsum of head v/ith setae sparse; male scape expanded or, if not, parameres fused to b.a.l.s 13 Ventral vulval margin meets gonapophyses acutely or, if meeting smoothly, not joined by sclerotised band; subgenital lobe present' or absent, but, if present, never with basal lateral processes (except Damalinia (Tricholioeurus) elongata; see Fig. 87 ); sternal setae on segments III - VI not attaining base of following setal row, usually less than three-quarters length of segment or, if longer, either female genitalia not as described and dorsum of head with dense setal covering (Bovicola (Holakartikos) and 3, (Subgenus n. 1)) or male scape not expanded and parameres not fused to b.a.l.s 15

13. Pleurum IV with dorsal projection; anterior setae present on abdominal terga and sterna, [ursidae] Wemeckodectes Pleurum IV without dorsal projection; anterior setae not present on abdominal terga and sterna 14 437

14. Male abdominal terga II - IV (at least) with median setal group reduced to one seta (Fig. 178); parameres fused to b.a.l.s. (Figs 185, 186) .or characteristically asymmetric (Fig. 187); female abdominal terga III - VII (at least) with median setal group reduced to one seta, or absent. [l.Iustelidae ana Procyonidae] Trichodectes (Stachiella) Tergal setae of both sexes more abundant, with at least two setae in median setal group; parameres not fused to b.a.l.s. [ Canidae, Viverridae, Ursidae and Mustelidae] Trichodectes (Trichodectes)(in part)

15. Posterior setal row of pleurum III with setae stouter than those of p.s.r. of pleurum V (Figs. 193, 202) or, if not, species with four pairs of abdominal spiracles; otherwise species with 0, 1, 2, 3, 4 or 6 pairs of abdominal spiracles; subgenital lobe present; gonapophyses with lobe present 16 Posterior setal row of pleurum III with setae not stouter than those of p.s.r. of pleurum V; six pairs of abdominal spiracles present, if.gonapophyses with lobe, then subgenital lobe absent 19

16. Abdominal spiracles numbering 6, 4 or 0 pairs; if no abdominal spiracles present, then female with gonopore surrounded by spicular refringent patch, or gonapophysis lobe comprising two fused tubercles, or antenna! sensilla in pit with peripheral tongue-like projections; female

subgenital lobe frequently with overlapping scales or spines; male mesomeral arch always present, with median extension or two- apical nipples; mesomeres produced basally between b.a.l.s. or, if not, antennal sensilla as described above; parameres usually broad, contacting mesomeres only, not b.a.l.s.; male abdominal tergum II lacking specialised setae of median group Lorisicola s. lat 17 438

Abdominal spiracles numbering 3, 2, 1 or 0 pairs; if no abdominal spiracles present, then female gonopore not surrounded by spicular refringent patch, gonapophysis lobe not comprising two fused setal tubercles; antennal sensilla of male and female never in pit with peripheral tongue-like projections; female subgenital lobe never with overlapping scales or spines; male mesomeres fused, unfused or absent; if mesomeres fused, mesomeral arch never with median extension or apical nipples; parameres frequently narrow, rod-like, contacting mesomeres, b.a.l.s. or both; male abdominal tergum II frequently with long, specialised setae (Figs.191* 152, 200, 201, 202) .Felicola s. lat 18

17- Male antenna! flagellum with 'teeth' on projection; six pairs of abdominal spiracles present or abdominal spiracles absent, in which case antennal sensilla in pit v/ith peripheral tongue-like projections. [?elidae, Viverridae and Lorisidae].• Lorisicola (Lorisicola.) Male antennal flagellum with 'teeth' not on projection, or, if projection present, mesomeral arch lacking extension; four pairs of abdominal spiracles.present or abdominal spiracles absent, in which case male gonopore surrounded by spicular patch (Fig. 236). [viverridae and Herpestidae] Lorisicola (Parado:iuroecus)

18. Male antennal flagellum with one or three basally-articulated 'teeth', or 'teeth' absent, in which case male abdominal tergum III with median setal group reduced to one seta of similar size to those on tergum II, which are not greatly enlarged, and parameres not fvised; female gonapo- physis v/ith lobe and spur present, lobe rounded or rectangular and formed of fused tubercles; subgenital lobe bifid or not; if subgenital lobe bifid, lobes pointed, rounded or with rectangularly obtuse posterior margins (Fig. 189); everted portion of male endophallus never sclerotised; abdominal spiracles numbering 0, 1 or 3 pairs. [Verpestidae and Canida.e] Felicola (Suricatoecus) 439

Male antennal fl age Hum v/ith one, two, three or four non- articulated 'teeth', or 'teeth' absent, i n v/hich case male abdominal tergum III v/ith median setal group reduced to one seta much smaller than those on tergum II, v/hich are greatly enlarged, and parameres fused, at least basally; female v/ith gonapophysis lobe rounded, v/ith or v/ithour spur, but lobe never rectangular; subgenital lobe bifid or not; if subgenital lobe bifid, lobes of various shapes, but never v/ith rectangularly obtuse posterior margins; everted portion of male endophallus frequently thinly sclerotised (Figs 204 208); abdominal spiracles numbering 0, 2 or 3

pairs. [llerpestidae, Viverridae and Felidae]

Felicola (Felicola)

19. Dorsal or ventral projection present on abdominal pleurum IV (Fig. 117); mesomeral arch generally produced basally between b.a.l.s.; female antennal flagellomeres generally not fused, or only partially fused. [ Procaviidae and Primates] 20 Pleurum IV lacking any projection; mesomeral arch rarely produced basally between b.a.l.s.; female flagellomeres generally fused or, if not, then female with long setal tufts on abdominal pleura VIII and IX (See couplet 27); not parasitic on hyraxes or primates 27

20. Sitophore sclerite modified, v/ith posterior aims extended (Fig. 36 ) (sclerite difficult to see) 21 Sitophore sclerite unmodified (Fig. 35) (not, generally, difficult to see) 23

21. Tarsal claws with ventral teeth or spines; temple margin v/ith

or without small rounded projection; -pleural projection on abdominal pleurum IV not elongate. [ Procaviidae] Dasyonyx s. lat 22 Tarsal claws lacking ventral teeth or spines; tenrole margin with long, broad, triangular projection (Fig. 149); pleural projection on abdominal pleurum IV long (Fig. 150); [Procaviidae] Eurytrichodectes 440

22. Tarsal claws with sharp, fine spines (Fig. 18a) Dasyonyx (Dasyonyx) Tarsal claws with broad, saw-like teeth (Fig.18b) Dasyonyx (ITeo dasyonyx)

23. Abdominal sternum II with broad, heavily-sclerotised band articulated with abdominal pleurum II (Fig.117); setal row of male scape comprising only two setae; basipara- meral sclerites present, [procaviidae] Procavicola. s. lat 24 Abdominal sternum II lacking sclerotised band or, if sclerotised band present, this is fused to abdominal pleurum II or medially broken; setal row of male scape numbering more than two setae; basiparameral sclerites absent or, if present, thoracic spiracle with tubular atrium and female flagellomeres fused 25

24. Posterior angle of temple with small projection; mesomeral arch with lateral double flexion and median extension; endophallus with large hook-like spines (Fig.-|2l) Procavicola (Condyloceohalus) Posterior angle of temple lacking projection; mesomeres unfused and lacking lateral double flexion and median extension; endophallus lacking large hook-like spines (Figs. 123) Procavicola (Procavicola)

25. Atrium of thoracic spiracle spherical; mesomeral aj?ch with median extension and lateral desclerotisations; gonapophyses with setal tubercles, or, if not, postcoxale greatly developed and fused to abdominal pleurum II. [Procaviidae and Cercopithecidae] Procaviuhilus s. lat 26 Atrium of thoracic spiracle tubular; mesomeral anch lacking median extension and not desclerotised laterally; gonapo- physes lacking setal tubercles; postcoxale not greatly developed and fused to abdominal plcurum II. [Cebida.e] ..... Cebidicola 441

26. Parameres with "basal flange, sometimes fused faintly; perisetal gap of male subgenital plate absent; postcoxale not fused to abdominal pleurum II; setal tubercles of gonapophyses not fused characteristically. [Procaviidae] Procaviphilus (Procaviohilus) Parameres usually lacking basal flange; perisetal gap of male subgenital plate present or, if absent, parameres fused together and articulated with mesomeral arch as in Fig. 132; and mesomeral arch produced has ally along b.a.l.s. (Fig. 133); postcoxale fused to abdominal pleurum II, at least in females; setal tubercles of gonapophyses fused characteristically (Fig. 125) or, if not, ventral vulval margin as in Fig. 127 . [ Procaviidae and Cercopithecidae] . Procaviphilus (ITeganarionoides)

27. Posterior margins of temple generally produced, very convex (Fig.101); very long setae present on at least pleurum VIII (Figs 102 , 103), sometimes also on pleurum VTI (males) or IX (female); basiparameral sclerites present; mesomeral arch lacking extension if complete, otherwise tripartite, median part sometimes obscure (Figs . 105,106); if mesomeral arch entire, male genitalia as in Fig.104, temples not greatly produced, and female with two flagellomeres,

otherwise female flagellomeres fused. [Erethizontidae] Eutrichouhilus Temples not so developed; setae on pleurum VIII.not exceptionally long; basiparameral sclerites present or absent; mesomeral arch, if present, with or without

extension, but never tripartite; female flagellomeres fused 28

28. Parameres narrow, rod-like and fused basally; mesomeral arch with broad lobe-like extension; b.a.l.s. widely divergent anteriorly (Fig. 152); gonapophyses with setal tubercles; pseudostyli absent. [protelidae and TTyaenidae] .. Protelicola 442

Parameres not narrow and fused basally or, if so, then b.a.l.s. not widely divergent anteriorly; mesomeral arch without broad lobulate extension; gonapophyses lacking setal tubercles; pseudostyli frequently present 29

Subgenital lobe present; endophallus with dense patch of regularly-arranged spicules or, if not, parameral plate with single apex (Fig. 95); mesomeral arch entire, v/ith abrupt bend to enable bases to meet parameres (Figs ,95> 96 ), or mesomeres unfused and b.a.l.s. v/ith anteposterior spur (Fig. 97 );. interior face of male flagellum serrate (Fig.14a); abdominal sterna never with anterior setae; long, slender species. [Bovidae and Cervidae] Damalinia (Tricholioeurus) Subgenital lobe absent or, if present, as small flap (Fig. 55) and species v/ith anterior setae on abdominal sterna (Bovicola jellisoni); sternum VII sometimes developed posteriorly into two projecting spikes (Damalinia theileri, neotheileri and semitheileri, Fig. 81 ); endophallus lacking spicular patch; parameres with apices free; mesomeres apically fused, unfused or absent, but abrupt bend not present; b.a.l.s. lacking anteposterior spur; interior face of male flagellum without serrations; broader species 30

Dorsal face of vulva with pointed scales; gonapophyses hook-shaped (Fig. 88); common oviduct, at branching point, with folded and more or less apparent collar, sometimes partially sclerotised and refracting transmitted light; mesomeres unfused; abdominal" pleurum never extending ventrally onto abdominal sternum II; interior face of male flagellum serrate. [ Bovidae and Cervidae] . .Damalinia (Cervicola) Dorsal face of vulva larking pointed scales; gonapophyses not hook-shaped or, if they are, then abdominal pleurum II extending onto sternum II (Fig. 79); common oviduct lacking 'collar' as described above; mesomeres fused, unfused or absent; interior face of male flagellum with or without serrations 31 443

31. Abdominal pleurum II v/ith sclerite extending onto sternum II and occasionally tergum II, sometimes at the expense of sternite or tergite (Fig. 79 ); mesomeres unfused, may be fused to parameres and apparently absent; pseudostyli

absent or, if present, broad or narrow (Figs 79 , y 83 ); interior face of male flagellum serrate. [ 3ovidae] Damalinia (Damalinia) Abdominal.pleurum II not extending onto sternum II; mesomeres fused, unfused or absent; pseudostyli, if present, not as figured above; interior face of male ' flagellum lacking serrations 32

32. Atria of abdominal spiracles large, clear; mesomeral arch fused to b.a.l.s.; parameres broad, asymmetrieally deflected (Fig. 78); thorax v/ith setae sparse dorsally in female, but male with median patch of setae on prothorax; head elongate, trapezoid, with deep osculum present (Fig. 77 ) [Tra.gulidae] Genus n. 3

Atria of abdominal spiracles not large; mesomeral arch not fused to b.a.l.s. or, if it is, parameres and mesomeres also fused (Fig. 74); parameres not broad or asymmetrically deflected; thoracic setae less sparse but male thorax lacking central setal patch; head not elongate but rounded, osculum absent or, if present, not deep (Fig. 12a) 33

33. Parameres fused to mesomeres (Fig. 74); pseudostyli present, apically angular (Fig. 73 ); gonapophyses broad, truncate (Fig. 72 ); osculum absent, but anterior margin of head slightly flattened or concave medially, v/ith hyaline border where pulvinus attains margin. [Bovidae] Genus n. 2 Parameres not fused to mesomeres; pseudostyli, if present, apically rounded; gonapophyses not broad, and with lobe variably apparent, or, if gonapophyses broad and truncate (7/emeckiella), then pulvinus not attaining anterior margin of head, which is smoothly rounded and lacks a median hyaline border 34 444

34.- Gonapophyses broad, truncate; mesomeres of characteristic pentagonal form (Fig. 76). [Equidae and Bovidae] ... Werneckiella Gonapophyses with more or less discrete lobe (Fi^s 55, 56); mesomeres, if present, not pentagonal; base of parameres frequently heavily block-like. [Bovidae, Cervidae and Camelidae] Bovicola s. lat 35

35. Species with more or less dense covering of long setae; anterior setae present on abdominal terga, sterna and pleura, slightly shorter than setae of posterior setal rows on these elements (Fig. 63); gonapophyses with very limited lobe formation (Fig. 57 ) 36 Species with shorter setae or, if setae long, then sparsely distributed and anterior setae not present on abdominal terga and sterna; gonapophyses generally with more developed lobe (Figs 56, 58) 37

36. Sitophore sclerite with posterior aims extended (Fig. 36 ); male v/ith specialised setae on abdominal tergum II (Fig. 54); mesomeres absent (Fig. 70 ); female lacks spinose patch on post genital pleural area Bovicola (n. 1) Sitophore sclerite v/ith posterior arms not extended; male lacking specialised setae on abdominal tergum II; mesomeres present, fused apically, v/ith median extension (Fig. 69); female v/ith spinose setal patch on postgenital pleural .area Bovicola (H 01 ale art iko s)

37. Setae on head long, fine; osculum absent; preanteimal sulcus absent; gonapophyses v/ith very distinct lobe of characteristic form (Fig. 56 ); male genitalia with mesoneres not fused, bipartite (Fig. 71 ) Bovicola (Leoikentron) Setae on head not long and fine; osculurn present or absent; preanteimal sulcus present or absent; gonapophyses v/ith less distinct lobe, of different form (Figs . 55 , 58 ); male genitalia of different form, mesomeres never bipartite ... Bovicola (Bovicola) SECTION 4

HOST-PARASITE RELATIONSHIPS 446

4.1. INTRODUCTION

Lice are obligate ectoparasites and cannot survive for long away from the environment provided by the skin and fur or feathers of the host (the 1dermecos'). They have neither a cyclical change of host taxa nor a free-living stage. These restrictions may lead to a louse spending all its life on a single host individual or, because opportunities for transfer between hosts are mainly intra-specific, on host individuals of only one taxon (or population). Such a relationship might be ejected to lead to some loss of evolutionary independence for the lice, and thus favour co-evolution and perhaps co-speciation of louse and host. This in turn would result in phylogenetic relationships between host taxa being reflected in similar phylogenetic relationships of their lice. The latter proposal, which is supported in general terms by many observations, has led to the formulation of a number of * phylogenetic rules1, of which the most important is 'Fahrenholz' Rule* • Fahrenholz* Rule was initially proposed by Eichler (1940a), but has been stated in a number of forms (Eichler, 1940a, 1948; Lakshminarayana, 1977; Brooks, 1979), the most simple of which is: 'Parasite phylogeny mirrors host phylogeny1 (Brooks, 1979). This has been accepted as axiomatic by a number of louse taxonomists, and it has been employed (see section 3.2.) in the construction of both parasite and host classifications. In some cases the argument has been reduced to complete circularity, the postulated relationships of the hosts being used to claim louse relation- ship, which is in turn used to support the initial hypothesis of host relationship (e.g. Traub, 1980). A number of theoretical and practical reservations are discussed by Hopkins (1957), Clay (1949a, 1957) and Hennig (1966), the last of whom dismisses absolutely the applicability of parasitological evidence in reconstruction of host phylogenies (and, by implication, vice versa). Despite these objections Fahrenholz* Rule is still considered to be generally applicable by many taxonomists and systematists, and is widely employed (e.g. Emerson and Price, 1981). As noted in the Introduction (p. 18), one objective of the present study is to test Fahrenholz1 Rule and therefore determine whether dependence on host-parasite associations in the process of systematic 447

decision-making as described above is justified. Fahrenholz' Rule, as an expression of the evolutionary relationships between parasite and host, is based primarily on the ecological relation- ships of the two as summarised in the phenomenon of host specificity. Besides the obvious test of the Rule by comparison of hast and parasite phylogenies (which is undertaken in section 4.4.), therefore, the Rule may be tested by study of the predictions that can be made from it regarding host specificity. Such tests are made in section 4.3. as part of an examination of host specificity in lice. Before further consideration of host-specificity or other aspects of co-evolution, a species concept for lice must be developed for use in discussion. 448

4.2. SPECIES CONCEPT

Sexually-reproducing species are usually defined as 'groups of actually or potentially interbreeding natural populations, reproductively isolated from other such groups* (Mayr,,1942, 1969). This definition is operative only in terms of data unavailable to the majority of taxonomists, who must make assumptions about- reproductive relationships on the basis of morphological data garnered from dead specimens, supplemented with whatever other data are available. It is not a necessary premise that all species differ morphologically nor that every morphologically-distinct group constitutes a species, and examples in the mosquitoes and the poly- morphic butterflies provide evidence against the universal applicability of these assumptions. These two guidelines do, however, govern the description of the vast majority of insect species,' simply because no other data are available. Most specimens of Fhthiraptera are associated with one piece of non- morphological data, the identity of the host. The majority of problems of species-level taxonomy in the lice are directly attributable to the differences in weight attached to this information. Some taxonomists attach great importance to host identity and propose that lice on different host species must be reproductively isolated from one another and therefore may (or must) be considered as taxonomically distinct, even if morphological criteria for differentiating them are not available (Zlotorzycka,1964; Eichler, 1966). These views have led to the description of a number of species and subspecies that can be distinguished from others only on the basis of the host identity. However, Mayr's definition of the biological species includes populations that are potentially interbreeding (i.e. that have no intrinsic reproductive isolation), and populations separated only by an extrinsic barrier are not debarred by that from being members of the same species. Any other conclusion is, as pointed out by B. Nelson (1972), anticipating rather than documenting the course of evolution. Unless it is certain that there is no gene flow between lice on different hosts (and there is no a. priori reason to suppose this, as will be discussed below) louse species must not be distinguished on the basis of host identity alone, but should be demonstrably distinct on morphological or other characters. 449

This principle is followed in the present study, and species are only recognised as such on the basis of morphological distinctness. A further difficulty arises in the use of the subspecies category. Hellenthal & Price (1980) distinguish qualitative and quantitative differences between louse populations. If qualitative differences are noted, they generally regard the populations as distinct species; if the differences are quantitative only, the populations are treated as different subspecies. Clay (1962) does not recognise size or size-related qualitative differences as worthy of taxonomic recognition, maintaining that the absence of data for lice indicating whether size differences are host-related or genetic in origin precludes the use of these criteria. It is notable that a number of subspecies described by Price and his associates do span different host taxa, perhaps indicating a genetic basis for qualitative differences. Whether this basis is of such a nature that subspecies can be erected is,another matter. There is much disagreement over the justification of the subspecies category in.taxonomy and its significance (Lane & Marshall, 1981), but the two criteria most frequently applied are some degree of morphological individuality and geographical distinctiveness (Mayr et al.. 1953; Mayr, 1969). These criteria appear to be rarely used in the Phthiraptera, and the subspecies category is frequently reserved for morphologically similar lice from closely-related hosts (Johnson, 1960). The value of the subspecies concept in studies of Phthiraptera is questionable, and its significance plainly variable, but in this study no changes have been made to the status of infra- specific taxa without taxonomic study. Retention of subspecies therefore reflects the lack of such study. In the following discussion subspecies are treated as if they were full species.

V 450

4.3. HOST SPECIFICITY

4.3.1. Introduction

The term 'host specificity' has been used to refer to the occurrence of a single species of parasite on a particular region of the host's body (site specificity), on a single host species, on a (presumed) holophyletic higjier taxon or on a non-holophyletic group of species, and to the occurrence of a higher parasite taxon on a species or higher taxon of host. In the following discussion the emphasis will be on the factors governing the host range of ah individual parasite species, and the implications of non-specificity at this level. Before commencing the discussion several tezms must be defined. A 'primary host' is a host taxon that is associated with its parasites because of inheritance (adherence to Fahrenholz* Rule), whereas a •secondary host' has acquired its parasites from another taxon in the host class by the process of*•secondary infestation' • The association of parasite and secondary host is also referred to as a * secondary infestation'. 'Secondary absence' refers to the absence of a parasite taxon from a host taxon that might, from its presumed ancestry, be expected to be associated with that parasite. are- referred to as being of the 'host class' if they are.potentially suitable for parasitism by a member of the parasite group under discussion; for example, the host class for lice comprises mammals and birds, and for Streblidae (Diptera), bats. The phenomenon of host specificity is examined in two ways, below. In section 4.3.2. it is considered in the light of predictions made by Fahrenholz' Rule, and observations of host-parasite associations in the lice used to test the rule. In section 4.3.3. an attempt is made to determine the causal factors of host specificity, with particular reference to lice.

4.3.2. Predictions of Fahrenholz' Rule

4.3.2.1. Introduction As noted above, Fahrenholz* Rule makes necessary several predictions about host specificity. The first, which is discussed below under the heading •phylogenetic component of specificity', is that cladogenesis of 451

the parasite only takes place in response to host cladogenesis, and invariably does so. The second, which is discussed under the heading 1 ecological component of specificity', is that a host is only parasitised by lice of a taxon that it has inherited. These aspects of specificity are only treated in these sections to determine if they are valid predictions; the nature of specificity is treated more comprehensively in section 4.3.3. • The phylogenetic prediction is in fact only made necessary by some formulations of the Rule (e.g. Eichler, 1940a; Brooks, 1979); it is not implicit in the 'explanation' of the Rule given by Eichler (1948) "The ancestors of extant parasites must have been parasites on the ancestors of extant hosts". Both predictions'will be examined with particular reference to lice.

4.3.2.2. Phylogenetic Component of Specificity

The prediction of Fahrenholz' Rule to be examined in this section is that speciation of the parasite only takes place in response to host speciation, and invariably does so (•co-speciation' of Brooks, 1979)- "If the above prediction is correct and no other factors have modified the host-parasite associations, the number of parasite species and of their infested hosts will be the same, no host will be parasitised by more than one louse and no louse will parasitise more than one host. A bias in the taxonomic basis of this comparison has already been noted (section 4.2.), in that some taxonomists have named louse taxa only on the basis of their host associations; this, however, is not considered to be a problem in the Trichodectidae. A further complication is caused by the existence of subspecies, both of host and louse. In the Tricho- dectidae, 337 species parasitise 244 species of mammalian host, and 350 species and subspecies parasitise 577 species and subspecies of host. Much of the numerical difference in the host numbers is due to the subdivision into 306 subspecies of the 31 species of Geomyidae known to be parasitised. Although the relationship between numbers of host and louse species is obscured, it can be seen that they differ, contrary to the prediction of Fahrenholz' Rule. Considering the number of lice parasitic on each host, slightly over half of the host taxa are parasitised 452

NUMBER OP NUMBER OP HOST SPECIES AND SUBSPECIES INFESTED TRICHODECTIDAE PER HOST TAXON GEOMYIDAE HYRACOIDEA OTHER HOSTS TOTAL

1 192 7 179 378 2 99 7 21 127 3 30 5 6 41 4 11 5 1 17 5 1 1 0 2 6 2 3 0 5 7 1 0 1 2 8 • 1 1 2

16 1 1

TOTAL 337 30 208 575

Table X. Numbers of trichodectid species and subspecies parasitising individual host taxa. 453

number of hosts host taxa per trichodectid gecmyidae hyraccidea other hosts total taxcn sspp .* spp. sspp.1. sfp. sspp. sfp. sspp. spp.

1 42.2 85.3 67.7 88.1 77.5 78.0 65.5 81.8 2 15.7 11.8 11.9 10.2 17-6 17.1 16.1 14.4 3 4.9 1.9 8.5 1.7 2.7 2.7 4.3 2.3 4 5.9 1.0 3.4 1.1 1.1 2.9 0.9 5 5.9 3.4 0 0 2.3 0 6 2.9 3.4 0 0 1.4 0 7 3.9 1.7 1.1 1.1 2.0 0.6 8 1.0 0.3 9 4.9 1.4 10-19 7.9 2.3 20-29 1.9 0.6 30-39 0 0 40-49 1.9 0.6 50-59 0 0 .60-69 1.0 0.3

number of 102 102 59 59 187 187 348 348 parasite species

and subspecies

Table XI. Percentage of species and subspecies of Trichodectidae parasitising various host groups in different classes of number of hosts attacked (species and subspecies). * 16 host species with Trichodectidae recorded from only one subspecies. * 3 host species with Trichodectidae recorded from only one subspecies. 454

by more than one taxon of trichodectid, the maximum number of louse species and subspecies found on one host species being 16 (Table X). Similarly, although most Trichodectidae are found on only one host taxon, up to 7 species or 60 subspecies may be parasitised (Table XI). Because this distribution may be caused not only by independent speciation but also by secondary infestation (see below), the phylogenetic specificity prediction must be examined in more detail, and in such a way that secondary infestation is not a complicating factor. The prediction will be examined in two parts. The first component of the statement denies the possibility of parasite speciation without concomitant host speciation. Although such independent speciation (Fig. 260) would falsify Fahreriholz' Rule, it does not of itself preclude the use of parasite distributions in phylogenetic studies of the hosts as in no case does the possession of one or more parasite taxa characterise a non-holophyletic assemblage. Hennig (1966) may have realised this, for he did not consider this possible exception to Fahrenholz' Rule, although he dealt with other problems (see below). Brooks (1979) criticises Hennig for this omission, but himself fails to see the validity of parasite associations as apomorphies for use in host phyletic analysis if independent parasite speciation takes-place (Fig. 1, cladograms IIB of Brooks, 1979, but not legend for IXB). Evidence for independent speciation of the parasite may be found in sister-species pairs of parasites on single host species. The most familiar case of sister-species of lice on a single host is the head and body lice of man (Pediculus capitis and >p. hum anus respectively) . Whilst these have been thought to be transient types of a polytypic species (Ferris, 1935, 1951), they are now believed to be either two subspecies (Busvine, 1948; Clay, 1973) or two species (Busvine, 1978; Schaefer, 1978; Lyal, 1982). The cladistic analysis of the Trichodectidae suggests seven speciation events without concomitant host speciation (Table XII). The first of the instances listed on Table XII concerns the water mongoose Atilax paludinosus; a further species (as yet undescribed) belonging to the F. acutirostris - pygidialis clade is also associated with this host, suggesting three speciation events for 455

/ host parasite phylogeny phylogeny

Fig. 260. Phylogenies of hosts and associated parasites, showing speciation of the parasite independent of host speciation. HOST SPECIES LOUSE SPECIES CLADISTIC RELATIONSHIP OF LICE

Atilax paludinosus Felicola (S.) acutirostris macrurus pygidialis

Prionodon linsang LoriBicola (P.) aspidorhynchus sumatrensis

Connochaetes taurinus Damalinia (L.) theileri neotheileri semitheileri

Cervus 61aphus Bovicola (B.) concavifrons longicornis

Ammotragus lervia Yferneckiella fulva neglecta

Table XII. Sister-species of Triohodectidae parasitising the same primary host species. The sister-species relationship of Werneckiella fulva and W. negleota is not established firmly. VJ1 ON 457

the parasite. It is notable that no two of the three described species in the clade have been collected from the same host individual. It is not known whether specimens of any two (or all three) of Damai ima (D.) theileri, neotheileri and semitheileri have been collected from a single host specimen, but it is known that specimens of each of the other three sister-species pairs in Table XII have been found on single host individuals. In addition to these cases, a number of other instances on the cladogram suggest, more equivocally, louse speciation without host speciation (Table XIII). The results of this study therefore indicate that speciation of lice can and does take place without host speciation, thus falsifying Fahrenholz* Rule. Clay (1949a, 1957, 1976) and Brooks (1979) suggest that the develop- ment of sister-species of parasite on a single host species can be accounted for by the allopatric speciation model, and postulate the temporary isolation of a population of hosts, during which the parasites speciate but the hosts do not. After the host populations are re-united, the two sister-species of parasite would be free to spread throughout the full range of the host species, giving rise to two sympatric parasite species on the same host (Clay, 1949a, 1957, 1976). P.W. Price (1977) considers the environmental constraints experienced by parasites and suggests that these lead to the fractioning of gene pools and favour inbreeding and asexual reproduction, which in turn promote predominantly homozygous populations. Such populations, he points out, have all their genes exposed to selection in each generation, allowing selection pressure to affect a population rapidly. Two further attributes of parasites (including lice) are short generation time (in comparison to the host) and the ability to increase population .-size rapidly. Both of these favour rapid differentiation of populations under different selection pressures (P.W. Price, 1977). Price suggests that the founder effect and genetic drift are very important in speciation of parasites; his conclusion is that evolutionary and speciation rates of parasites can be high compared to those of the hosts, a factor which will encourage the development of sister-species of lice on single host species in the manner described above. It is notable that one of the sister-species pairs 458

HOST SPECIES LOUSE SPECIES CLADISTIC RELATIONSHIP OF LICE

CaPra hircus Bovicola limbata Capra hircus caprae

Cephalophus monticola Damalinia bedfordi Cepahlophus monticola, adersi,] pakenhami Sylvicapra grimmia |

Odocoileus virginianus, hemionus Damalinia lipeuroides Odocoileus virginianus, hemionus parallel^

Atilax paludinosus Lorisicola paralaticeps Atilax paludinosus laticeps

Genetta tigrina, victoriae Lorisicola lenicornis G. tigrina, thierryi wernecki G. tigrina, genetta, abyssinica acuticeps G. tigrina neoafricanus G. genetta africanus

Table XIII, Possible cases in the Trichodectidae of louse speciation without concomitant host speciation. 459

(Bovicola concavifrons and B. longicornis) comprises two parthenogenetic species. Price* s suggestion that parasites have high evolutionary rates, and the observation of sister-species pairs of lice on single hosts, both contradicts another-parasitdlogical 'rule*, that parasites evolve more slowly than their hosts (Manter, 1955, 1966; "Manter*s first rule", according to Brooks, 1979) • This rule is based on the supposition that the environment of the parasite is provided entirely-by the host, but is unlikely to change as rapidly as does the host itself (as evolution of the host is likely to concern features other than those directly concerning its parasites). This rule, although not cited by name, has been considered applicable to most groups of lice, those parasitising hyraxes and tinamous being excluded (Hopkins, 1949, 1949a; Clay, 1957; Hennache, 1981). The second part of the prediction of phylogenetic specificity demands speciation of the parasite whenever the host speciates. Should this fail to occur, either both daughter species of host will have the same parasite, or one of them will not be parasitised. In both cases Fahreriholz* Rule is falsified. In the first case the parasite taxon would by its presence indicate a holophyletic group of hosts (Fig.26li), but any subsequent speciation of the parasite (Fig.26lii) would cause a paraphyletic group of hosts to be recognised by the possession of a single parasite taxon (Hennig, 1966). Such a failure of co-speciation may be recognised in the association of a single louse species with two or more host species that themselves form a monophyletic (i.e. holophyletic or paraphyletic) group. Such evidence cannot be considered conclusive if it is believed that lice can establish populations on more than one host taxon by secondary infestation (a possibility which is established as fact in section 4.3.2.30* This is particularly relevant since the more closely related two host taxa are, the more similar the environments they offer to parasites are likely to be, and the more likely it is that any colonisation of the secondary host will result in successful establishment of the lice. There are a number of cases in the Trichodectidae of single louse species parasitising two or more congeneric species of host (i.e. plausible monophyletic groups), but the unavailability of host phylogenies and the possibility 460

phylogeny phylogeny

Fig. 261. Phylogenies of hosts and. associated parasites, showing host speciation without parasite speciation; (i) leading to a holophyletic group of hosts (B + C) being parasitised by a single parasite taxon (i"b); and (ii) leading to a paraphyletic group of hosts (B + C) being parasitised by a single parasite taxon (b). 461

of secondary infestations make this line'of evidence difficult to pursue. The second alternative can be recognised in the- absence of a parasite from a host taxon though present on related hosts and thus probably present on the ancestor of the host taxon (Fig. 262) (Hennig, 1966). This is difficult to demonstrate in practice, as a presumed absence may be the result of inadequate collecting. Hopkins (1949) suggests that such 'secondary absence' is quite frequent in the Trichodectidae, but this proposal is based on the assumption that secondary infestation is not a factor, and therefore all observed host-parasite associations are primary. For example, Hopkins (1949, 1957) notes that Trichodectidae are present on rodents of the families Geomyidae and Erethizodontidae and primates of the families Lorisidae and Cebidae, but on no other families of these orders; for both orders he maintains that Trichodectidae were present but have become extinct on all other families or their ancestors. All four of these associations are believed here to result from secondary infestations. A more acceptable instance in the Anoplura is described by Hopkins (1949, 1957). He points out that most species of Sciuridae are parasitised by-one species of Enderleinellus and one of Neohaematopinus, and suggests that the absence of Neohaematopinus from Xerus erythropus and Enderleinellus from X. inauris are secondary. Ward (1958) notes the absence of two ischnoceran genera (Cuclotogaster and Lipeurus) from New World , although both are represented by many species on Old World birds of this order. He suggests that the lice were not present on the Galliformes that originally crossed the Behring land bridge from Asia. The absence of Piagetiella (Amblycera) from some of the species of Phalacrocorax is possibly secondary (Clay, 1976). There are two possible reasons for the secondary absence of a louse taxon. The first is that lice were originally present, but have become extinct throughout the range of the host. This is not subject to test, as extinction of a louse species without concomitant extinction of the host has not been observed. The second possibility is that of primary absence, lice having been absent from the ancestral host population when it became isolated. This 462

N 'A a\ B /'C b\ ZD e r'F d \ ,'G e \

sr * 0

host parasite phylogeny phylogeny

Fig. 262. Phylogenies of hosts and associated parasites showing secondary absence of parasites, leading to a paraphyletic group of hosts (a+c+d+f+g) being parasitised by a single (higher) parasite taxon (holophyletic group of parasites). 463

alternative may be tested through observation of modern louse-host associations. If such primary absence is not to occur, lice must be present throughout the host range, as a new (daughter) host population may theoretically become isolated at any place in that range. The assumption that lice are so distributed is not always simple to test; Kloet & Hinks (1964) in their check-list of the insects of the British Isles, for example, state that the list -of lice has been complied from the distribution of the hosts alone. A further difficulty is the paucity of records; many species are represented in collections by only a few specimens, and collecting sites rarely span the full range of the host or hosts. Some examples are available, however, that indicate that lice do not invariably extend throughout the host range. Clay (1949a) notes that not all British populations of the chaffinch are parasitised by Ricinus (Amblycera). She also mentions that Laemobothrion (Amblycera) has not been recorded in the British Isles from the coot or moorhen, although present on these birds in North Africa, and Ricinus may be absent from the crested lark (Galerida crist at a) over part of its range. These apparent absences need to be -confiimed by more collecting, as Laemobothrion has since been collected from a British coot, and a further reported absence (Picicola from the green woodpecker in Britain) has also been found to be an artifact of collecting.. Boyd (1951) finds that only three of the four species- of lice found on the starling in Europe are present on this species in the United States, the bird having been introduced there in 1890. Timm & Price (1980) indicate that Geomydoecus nebrathkensis (Trichodectidae) is found in only part of the range of its host. Geomys bursarius lutescens (Pig.263). Plomley & Thompson (1937) and Thompson (1940) note that Heterodoxus spiniger (Amblycera), parasitic on the domestic dog and some wild canids and felids in many parts of the World, is found almost entirely between 40°S and 40°N. Trichodectes canis (Trichodectidae), also parasitic on the domestic dog and wild canids, has not yet been recorded from the Ethiopian Region although widespread elsewhere (Ledger, 1980). Hopkins (1945) states that the hyrax Procavia capensis coombsi is parasitised by the trichodectid Procavicola pretoriensis over most of its range, but where it is 464

Fig. 263 • Distribution of the eight species of Geomydoecus that parasitise pocket of the Geomys bursarius complex in relation to the distribution of their hosts. Records of lice are indicated by symbols, a key to v/hich is given in the inset box. The inner boundary lines on the map indicate the boundaries between subspecies of gophers. After Timm & Price (1980). 465

sympatric with Procavia £ • letabae this louse is replaced by the louse of the latter hyrax species (Procavicola mokeetsi). Ledger (1980) expresses doubts on the identification of the host species in this case. Hopkins (1949), as already noted, states that no two of the three Tricho- dectidae Felicola acutirostris. macrurus and pygidialis. all parasitic on the mongoose Atilax paludinosus. have been recorded from the same locality. The herpestid Crossarchus obscurus is parasitised by Felicola occidentalis in the northern part of its range but by an undescribed member of the F. congo ensis - occidentalis clade in the southern part. Clay (1964) shows that the two species Sula leucogaster and £5. sula are both hosts to the ischnoceran Pectinopygus garbei in the Atlantic, and to the related sulae in the Indian Ocean. Clay (1966) notes that Stigiphilus aitkeni (Ischnocera) is parasitic on Tyto alba in the Nearctic, Neotropical, Australasian and Oriental Regions, whereas 3• rostratus is parasitic on the same host in the Palaearctic and Ethiopian Regions. Similarly, two species of Myrsidea (Amblycera) are found parasitising Corvus orru in different parts of its range (Klockenhoff, 1972). Clay (1976) provides a number of other examples in the bird lice. These examples indicate that the distribution of a .louse species is not necessarily identical to that of the host, and therefore that there is a possibility lice will not be on a species even though they sire on its sister-species and where on the common ancestor of both. Apart from technically falsifying Fahrenholz* Rule, this demonstrates that it is not justifiable to use the presence of a holophyletic group of lice to postulate the holophyly of the hosts concerned, as the host group may, because of secondary absence of the lice, be paraphyletic (Fig.262). The causes of lack of congruence between louse and host distributions are probably numerous, but five will be mentioned here, i. Speciation. A species of louse may have been distributed over all of the range

of the host but undergone speciation in part of that range, leaving the host with two allopatric sister-species of lice. 466

ii. Extinction. A species of louse may have been distributed over all of the host range but become extinct in part of it • iii. Climate. The effect of climatic factors on louse distribution has not been investigated, but is probably responsible for the restriction in distribution of Heterodoxus spiniger. iv. Population structure. Parasite populations frequently have an •overdispersed' structure (Kennedy, 1975), so that the majority of parasites are found on comparativel few host individuals, and most host individuals of a population have relatively few parasites. This structure has been recorded for several lice (Ward, 1957; Cook & Beer, 1958; Eveleigh & Threlfall, 1976). This population structure allows a new host population, if founded by few individuals, to lack parasites present on the parent population. However, very few studies on louse population dynamics have been made, and the overdispersed structure may not be typical; Price (pers. comm.) has found that most Geomyidae have numerous lice. v. Secondary establishment. If lice are able to transfer between host animals of different taxa (as will be demonstrated in the next section), it is possible that lice present in only part of a host's range colonised the host (and perhaps underwent speciation) from another host taxon after the dispersal of the •new' host over its full range.

4*3.2.3« Ecological component of specificity

The prediction of Fahrenholz' Rule to be examined in this section is that the ancestors of extant parasites must have parasitised the ancestors of the extant hosts, i.e. that secondary infestation does not take place. Exceptions to this rule have been recognised by phthirapterist for many years, but the extent to which these have influenced present louse-host associations has been believed to be minimal. Hopkins (1949) considered that there were only six or eight cases of secondary infestation in all the lice and, whilst Hopkins (1957) accepted that he 467

had probably underestimated the importance of inter-specific transfer in his earlier works, he still believed it to have been very rare, and cases of anomalous distribution of lice best explained by hypotheses of secondary absence. Clay (1949a, 1957) accepted the possibility of secondary infestation, but believed it to have been most common at an early stage in the evolution of lice, "before [they] had developed extreme host specificity" and before the habitat (plumage of birds) had developed much from a simple uniform cover not offering different niches. Traub (1980) suggests that any secondary infestations by fleas and lice must be limited to hosts of a similar or higher phylogenetic or evolutionary level than the primary host, although he does not indicate how such levels can be deteimined. A number of clear cases of secondary infestation have been reported. Price & Beer (1963) found the amblyceran Colpocephalum turbinatum to be parasitic on 35 species of falconiform, some "presumably distantly related" and on the domestic pigeon (Columbiformes). The multiple infestation of may be a result of non-speciation of the louse with the host coupled with secondary absence, but the presence of the species on the pigeon can hardly be accounted for in this way. Price (1975a) records Menacanthus eurysternus (Amblycera) from 118 species of Passeriformes (in 70 genera and 20 families) and 5 species in the . Tarry (1967) records the dog louse Linognathus setosus (Anoplura) in very large numbers on domestic poultry, the only known record of Anoplura established on a bird. Another dog louse, Heterodoxus spiniger (Amblycera: Boopiidae) is found not only on the domestic dog and five species of wild Canidae, but also on a wallaby (Keler, 1971) • As all other members of the Boopiidae (except one) are found on marsupials, secondary infestation of a canid, probably the dingo, must have taken place from the wallaby, and this has been followed by further infestations of other Canidae (and some Felidae) around the World (Hopkins, 1949, 1957; Murray & Calaby, 1971). The lagomorph Sylivlagus brasiliensis is parasitised by three species of Amblycera that are primary parasites of the guinea-pig Cavia (Rodentia); these species have presumably become associated with Sylvilagus by its sympatry with Cavia. as heavily-infested hares have been found in the open grassland far from human dwellings but 468

in the same area as Cavia (Hopkins, 1949). Boyd (1951) records that since the introduction of the starling Sturaus vulgaris to North America from Europe, it has acquired the ischnoceran species Brueelia omatissima (= 3. illustris) from a native icterid, whilst the native species Turdus migratorius has become host to the starling louse Brueelia nebulosa. The above examples are all of secondary infestation without divergence of the lice on the secondary host, but divergence and eventually speciation may occur if these lice are isolated from those on the primary host for sufficient time (i.e. the secondary infestation resulted from a rare colonisation, not repeated frequently enough for a significant level of gene flow between the lice on the two hosts). Clay (1953) reports that a species of the ischnoceran Rallicola. primarily parasitic on Porohyrio sp. (Rallidae), is established on a corvid on Guam Island. There is no species of Porphyrio on Guam (Mayr, 1945), and the lice on the two hosts are 1 subspecifically' distinct (Clay, 1957). The Rhyncophthirina comprises only two species: one parasitic on both Indian and African elephants and the other on the (Phacochoerus aethiopicus). Clay (1963) suggests that transfer of an ancestral stock took place between one of the Elephantidae and Phacochoerus, to be followed by speciation. Clay (1971) describes a species of Boopiidae parasitic on the cassowary (Aves). • As noted above, other members of the Boopiidae, except one on Canidae, are parasitic on Australasian marsupials; a transfer has plainly taken place between members of different Classes. The Boopiidae them- selves are almost certainly derived directly from bird lice (Clay, 1970). Trichophilopterus babakotophilus (Ischnocera: Trichophilopteridae) is fouiid on at least four species of lemurs,but is almost certainly derived from bird-infesting Philopteridae.

The above examples are distributed throughout the Phthiraptera and, while they indicate the occurrence of secondary infestation, they do not give any idea of its frequency. For comparative purposes, therefore, the cases of secondary infest at ion known in the Trichodectidae will be considered. There are many trichodectid species parasitising more than one host species (see Table XI) but, as a large proportion of these instances may follow failure of co-speciation, only those involving 469

apparently distantly-related hosts (i.e. non-congeners) are discussed (Table XTV)• Eight louse species and sixteen cases of secondary infestation are listed, but of these only the first three are unlikely to have resulted in some way from the actions of man. i) Damalinia pakenhami. Sylvicapra grimmia is identified as the secondary host because the other hosts of.the louse and of its sister- species are in the genus Cephalophus. The possibility of the association with both Sylvicapra and Cephalophus being primary cannot be entirely dismissed, and is discussed below in section 4.4.2.1.. ii) Damal ipia coxnuta. Gazella dorcas is most likely to be the primary host of this species because other lice of the same clade are parasitic on Gazella but none are on Antilope. iii) Bovicola alpinus. The identity of the primary host of this species is not certain, although several samples of the species have been recorded from Rupicapra but only one from Capra. It is notable that most of the other species of Bovicola on Capra are all associated with the domestic goat Capra hircus, and thus secondary infestation of the domestic goat from an unknown wild host may be likely. iv) Werneckiella equi. Eichler (1954) described specimens of this species (as W. tramp el) from a batrachian camel in a circus. The primary . host is undoubtedly the domestic horse, Equus caballus (or another member of the Equidae - see below). v) Bovicola concavifrons. This species is found on the red deer and the wapiti with some frequency, but has also been taken from Pere David* s deer. The latter host is not as closely related to the red deer as are other species and, unless a number .of secondary absences are postulated, must be assumed to have obtained the lice by secondary infestation. All specimens of Pere David* s deer are in captivity, and have been for many years. vi) Bovicola tibialis. This infestation is documented by Westrom, Nelson &'Connolly (1976). Three specimens of Odocoileus had been kept in an enclosure with fallow deer, and transfer of the lice probably took place during this period of enforced association. 470

LOUSE SPECIES PRIMART HOST SECONDARY HOST CI RCUMSTANCES

DAMALINIA PAKENHAMI CEPHALOPHUS ADERSI SYLVICAPRA GRIMMIA IN WILD

and/ore. MONTI COLA (Cephalophlnae)

(Cephalophinae)

DAMALINIA CORNUTA GAZELLA DORCA§ ANTILOPE CERVICAPRA IN WILD

(Antlloplnae) (Antlloplnae)

BOVICOLA ALPINUS FUPICAPRA FUPICAPRA CAPRA HIRCUS PY REN I AC A IN WILD or (Caprlnae) (Caprlnae)

WERNECKIELLA EQUI EQUUS CABALLUS CAMELUS BACTRIANUS IN CIRCUS

(Bquldae) (CAMELIDAE)

BOVICOLA CONCAVIFRDNS CERVUS ELAPHUS ELAPHUFUS DAVIDIANUS IN CAPTIVITY

(Cervidae) (cervidae)

BOVICOLA TIBIALIS DAMA DAMA ODOCOILEUS HEMIONUS IN CAPTIVITY

(Cervidae) (Cervidae)

EUTRICHOPHILUS SETOSUS ERETHIZON DORS ATOM MACACA MJLATTA IN ZOO

• (Erethlzontldae) (Cercoplthecldae)

FELICOLA SUBROSTRATOS 7ICHNEUMIA ALBICAUDA SALANOIA CONCOLOR UNKNOWN

(Herpestldae) (Herpestldae)

CIVETTICTIS CIVETTA IN CAPTIVITY

(Vlverrldae)

ELJPLERES GOUDOTI UNKNOWN

(Vlverrldae)

FELIS CATOS ? IN CAPTIVITY

(Felldae)

FELIS LYBICA UNKNOWN

FELIS SYLVESTPIS UNKNOWN

TRICHODECTES CANIS CANIS SP. VULPES VULPES EXPERIMENTAL

(Canldae) (Canldae)

VULPES BENGAL ENSIS UNKNOWN

CIVETTICTIS CIVETTA IN CAPTIVITY

(Vlverrldae)

Table XIV. Secondary infestations in the Trichodectidae. 471

vii) Butrichophilus setosus. The infestation of a monkey with porcupine lice occurred in a zoo, the two host species being kept in nearby cages (Fenstermacher & Jellison, 1932) • viii) Felicola subrostratus. The primary host of this species is identified in section 4.4.5«3. as Ichneumia albicauda. and the suggestion made that all other hosts of the louse are parasitised because of secondary infestations. The opportunity for so many transfers may have arisen from the practice of keeping Ichneumia as a pet, this almost certainly providing the link that allowed infestation of the domestic cat. Felis

lybica and

ix) Trichodectes canis. The identity of the primary host is uncertain, but it is almost certainly a-canid, the record from a viverrid (Civettictis civetta) being the result of a secondary infestation. The association with the domestic dog, whether primary or secondary, undoubtedly enabled infestation of Civettictis and Vulpes bengalensis, and possibly other Canidae, due to the enhanced distribution of the dog following its association with man. Only a few specimens of T .canis have been recorded from V. bengalensis, but the males of these exhibit slight differences from other specimens of JP. canis in the form of the genitalia. The record from Vulpes vulpes is of a single experimental infestation by Eichler (1940a), who introduced 200 specimens of T.. canis on to one individual of this host; eleven weeks later the fox was found to be carrying 14,000 lice. The host associations of other species of Tricho- dectes suggest that JP. canis is derived from a louse parasitic on a mustelid.

There are consequently only three known cases in the modern Tricho- dectidae where secondary infestation has occurred naturally, although in twelve other cases infestation may have occurred following human-induced 472

proximity of the hosts. Examination of the cladogram of the Trichodectidae permits an estimate to be made of the number of secondary infestations in the history of the family. These have been discussed in detail in section 4.4., and, assuming that all parasitism among the hyraxes and pocket gophers can be explained in phylogenetic teims (which it almost certainly cannot), a mini mum of 41 secondary infestations are necessary to explain the distribution of primary host associations on the cladogram. One hundred and ninety-eight cladogeneses have been postulated and depicted on the cladogram, of which 41 are associated with secondary infestation; the implication of this is that a minimum of 20,7% of speciation events in the history of the Trichodectidae have been associated with secondary infestation. The difference between the level of secondary infestation expected by analogy with modern records and that estimated from the cladogram may be explained by two factors: a high rate of parasite evolution and inadequate collecting. The rate of evolution of parasites, as already noted, may be very rapid, and it is likely that lice on a •new' - host, if isolated from the lice on the primary host, may diverge and speciate in a comparatively short time. If this is the case, multiple (secondary) infestations may be of comparatively short duration and thus rare, whilst secondary infestations 1 preserved* by speciation would be more common. In addition, the practice of dismissing records of lice from the 'wrong' hosts as * stragglers' and in some cases discarding the specimens (Clay, pers. comm.) must have lowered the number of recognised secondary infestations. Finally, there are still a large number of mammals that have, not been adequately (or ever) examined for lice (Emerson & Price, 1981), and some of these may provide more instances of *modern* secondary infestation.

4.3.2.4. Summary

The predictions of phylogenetic and ecological specificity made by Fahrenholz* Rule have been falsified. Speciation of the parasite does not necessarily take place with host speciation but may take place 473

independently. Parasites may be absent from a host taxon despite having been present on an ancestor of that host. Single parasite species, or holophyletic parasite groups, may be associated with holophyletic or paraphyletic host groups without any secondary infestation taking place. Lice (and presumably other parasites) can and do become permanently associated with host taxa following secondary infestation from other host taxa. Single parasite species, or holophyletic parasite groups, may therefore be associated with polyphyletic host assemblages. A minimum of 20.7% of speciation events in the phylogeny of the Trichodectidae are associated with secondary infestations, a proportion far higher than would be expected from perusal of the literature. Departures from the co-speciation and co-evolution predicted by Fahrenholz' Rule are unlikely to follow the same patterns for different species. Equating the probability of a postulated host relationship being correct with the number of parasite taxa supporting the relationship ('Hopkins' Principle') may, therefore, be justifiable, if (a) the phylo- genies of all the parasites concerned have been established using cladistic principles (Brooks, 1981) and all show the same distribution of host associations, and (b) any non-phylogenetic factors which may explain the association of some parasites do not explain the distribution of the majority of them (i.e. the hypothesis of adherence to Fahrenholz' Rule is the most parsimonious explanation for the observed host-parasite associations). No predictions of the phylogenetic relationships of the hosts of Trichodectidae are made as a result of the present study. The failure of Fahrenholz' Rule to explain all aspects of host- parasite phylogenetic associations, the ways in which" the Rule may be departed from, and the lack of rigorous cladistic studies of any lice other than the Trichodectidae, indicate that attempts to postulate host relationships on the basis of presumed louse relationships ( e.g. Webb, 1949; Timmermann, 1957; Kettle, 1977) are not justified.. Falsification of Fahrenholz' Rule in detail, and its rejection for the purposes of phylogenetic postulation, are not intended as an absolute refutation of the Rule as a description of evolutionary events. It is possible that a large proportion of host-parasite associations can be described in terms of the Rule, although such descriptions could only be postulated following a cladistic analysis of both host and parasite groups. 474

4.3*3. Factors Influencing Host Specificity

4.3.3.I. Introduction

Host specificity in the Phthiraptera is frequently considered to have developed at some stage after the initial association with birds and mammals, and subsequently to have become 'fixed1 (Hopkins, 1949; Keler, 1957a; Clay, 1949a, 1957; KBnigsmann, 1960; Ledger, 1980). This follows from the identification of host specificity as a stable feature of a parasite taxon, with attendant advantages and disadvantages (e.g. Hopkins, 1949; Ludwig, 1982). This approach, although perhaps sometimes adopted as a 'shorthand' for more detailed discussion, encourages a teleological view of the factors which influence host specificity, whereas the phenomenon should be considered as the result of the interaction of these factors through time. The purposes of this Section are: (a) to identify and briefly discuss the factors controlling host specificity in the Phthiraptera, (b) to determine-the reasons for the high level of monoxenia exhibited by the Phthiraptera compared to other ectoparasites of vertebrates, and (c) to discuss further the phenomenon of secondary infestation. The two latter objectives are met within the framework of the discussion resulting from objective (a). The factors controlling host specificity that are identified in this study, together with their possible interactions, are depicted in Fig. 264. It can be seen from Fig.264 that there is some positive feedback, so that the continuation of factors which initially promoted the restriction of a parasite to a given host becomes advantageous to the parasite as it adapts to and becomes more dependent upon the host. Despite this, in the following discussion the factors associated with host specificity will be considered primarily as pressures limiting the hosts available to a parasite rather than as adaptations of the parasite to maintain or achieve a 'chosen' level of host specificity, as the former approach is less teleological. Two processes are involved in the development of an association between a parasite and a (novel) host individual or taxon. The first is 'colonisation', which is the detection .(if applicable) and invasion Fig. 264 Diagrammatic representation of the factors influencing host specificity of parasites of birds and mammals. 476

of the host by the parasite, and the second is •establishment', which is the successful reproduction (for at least two generations) of the parasite on the 'new' host. Colonisation may refer to the arrival of any number of parasites, but establishment requires prior colonisation by a fertile female or a pair; the factors governing colonisation by an individual or a pair of parasites do not differ, except perhaps in probability. Application below of the term 'host specificity' to families or other higher taxa of parasite refers to the degree to which species within the taxon are restricted in the number of host species parasitised.

4.3*3.2. Independence of dispersal

The observation that host specificity is related to the restriction of the parasite to the host has been made several times (Hopkins, 1949, 1957; Theodor, 1957; Oldroyd, 1964; Askew, 1971; P.W. Price, 1980; Marshall, 1981). P.W. Price (1980) attempts to illustrate this relationship by comparing specificity and dependence on the host in Philopteridae, Streblidae, Oestridae, Hystrichopsyllidae and Hippoboscidae, all of which are insect parasites of birds or mammals. These five families are reconsidered here with the addition of Trichodectidae, Nycteribiidae, Polyctenidae and Cimicidae. The percentage of parasite species in each family (or sample of each family) associated with different numbers of host species is summarised in Table XV, the families being arranged in order of decreasing specificity (as indicated by the percentage of parasite species associated with only one host species and the maximum number of hosts parasitised). The dependence of the parasite on the host (i.e. the degree to which the location of the parasite is determined by the location of the host) is summarised in Table XVT. The biology of each family as it relates to dependence is briefly reviewed by Order.

Phthiraptera Philopteridae and Trichodectidae, like other Phthiraptera, stay on the host body throughout their life-cycle, generally leaving the host only when it is possible to walk directly onto another (see below). They are oviparous, with three nymphal instars and no pupal stage. The difference in host specificity between the two families is probably an artifact, caused primarily by the much greater amount of taxonomic 477

CO ft M a ft 3 CO ft M Ei a ft ft CO M ft <3 ft ft ft O s ft ft H <15 ft <3? M ft CO ft ft EH H ft ft ft w ft 0 M ft ft 0 p o g ft <5 ft CO Eh H M ft ft p 0 M 1 fi EH ft H M H ft CO ft O 0 ft ft o EH a §5 M g ft § M M P ft 03 CO ft ft o g ft § H ft ft ft CO O P u • ft

f 87 82 70 64 56 49 43 37 17 2 9 14 28 17 22 19 38 20 9 3 2 2 2 7 13 7 4 9 15 4 1 1 4 5 6 4 5 13 5 1 2 6 3 5 7 6 1 4 3 6 4 7 1 4 2 4 8 4 3 4 9 4 3 2 10-19 2 9 4 20-29 1 7 30-39 1 2 40-49 50-59 6O-69 70-79 80-89 NUMBER OF 122 348 40 28 153 53 73 172 46 PARASITES

Table XV. Percentage of species of families of insects parasitic on mammals and. birds in different classes of number of hosts attacked (modified from P.Vf. Price, 1980). From analysis of host-parasite lists for Philopteridae (Theodor & Costa, 1967), Trichodectidae (Emerson & Price, 1281 ; original data), Nycteribiidae (Marshall, 1980),

Polyctenidae (Maa, 1964; Ueshima» 1572), Streblidae (tfenzel, Tipton & Kiewlicz, 1966; Wenzel, 1976), Oestridae (Zumpt, 1965)» Cimicidae (Usinger, 1966), Hystrichopsyllidae (Hopkins & Rothschild, 1962, 1966) and Hippoboscidae (Bequaert, 1956). LARVAL INSTARS ADULT TOTAL

STAGE OVUM I II III IV V PUPA PRE-REPRODUCTIVE •REPRODUCTIVE' (•INDEPENDENCE SCORE')

PARASITE FAMILY

POLYCTENIDAE - 1 1 1 - -- 1 1 5

PHILOPTERIDAE 1 1 1 1 - - - 1 1 6

TRICHODBCTI DAE 1 1 1 1 - - - 1 1 6

NYCTERIBIIDAE - - - - - 2 2 2 1 7

STREBLIDAE - - - -- 2 2 3 1 8

HIPPO BOSCI DAE - -- - - 3 3 • 3 1 10

OESTRIDAE 1 1 1 1 - - 3 3 3 13

HYSTRICHOPSYLLIDAE 2 2 2 2 - - 2 3 1 Ik

CIMICIDAE 2 2 2 2 2 2 - 2 2 16

Table XVI. Degree to which parasite location is determined by host location in nine families of parasites of birds and mammals. The dependence of each parasite instar is scored as follows: * 1• — location of parasite totally determined by location of host; '2* - location of parasite determined by host-created or local environment (e.g. nest, bat—cave); '3' - location of parasite independent of location of host.

00 479

•splitting* in the Philopteridae, and the difference in the geographic area sampled. Whereas the host numbers for Trichodectidae include all known hosts, those for Philopteridae include only those hosts occurring in a restricted geographical area (Israel), some additional hosts thereby being excluded. A further source of error in the host numbers is that some Trichodectidae are specific to subspecies rather than species, particularly those lice parasitic on the hyraxes and pocket gophers. If the specificity is determined by subspecies rather than species-, ..a much greater spread of host numbers is obtained (Table XI). The uncertainties attached to the .taxonomy of the hyraxes and pocket gophers limits the value of subspecies-specific figures, and they are not used in this study. However, the indication that further collecting and taxonomic work may change the apparent host specificity recognised in Table XV is clear.

Hemiptera The Polyctenidae are haematophagous ectoparasites of bats, and are believed to stay on the host throughout their life-cycle. It is the only family of Heteroptera in which the nymphs are produced vivi- parously, but in general the biology is very poorly known. There are three nymphal instars and no pupal stage. The Cimicidae are related to the Polyctenidae and are also chiefly associated with bats, although birds and man are also attacked. Members of the family are haematophagous, but spend little or no time on the body of the host, living in crevices of caves or in bird nests. The method whereby the insects are distributed between caves or nests is not certain, as cimicids have only rarely been recorded from their hosts away from their cave or nest (Usinger, 1966). The number of unpublished records of cimicids on their hosts in such circumstances (particularly on bats) is increasing, however (Hutson, pers. comm.), and phoresy, perhaps of the first nymphal instar (which would be both difficult to see and difficult to identify) is still the most likely dispersal method (Dolling, pers. comm.). Cimicidae are oviparous, with five nymphal instars and no pupal stage. The host data presented in Table XV for Cimicidae are abstracted from Usinger (1966). The confidence possible in the host range of species recorded numerous times from identified hosts is clearly greater than that possible for species

1 480

recorded only once or a few times, and for those recorded only from unidentified hosts; unfortunately, many of the records available fall into the latter two categories. It is possible, therefore, that the number of hosts utilised by many species of cimicid is greater than that recorded by Usinger (1966) and that the percentage of monoxenous species recorded in Table XV is too high.

Dipt era Streblidae and Nycteribiidae are also haematophagous ecto- parasites of bats; the females either deposit fully-developed larvae on the surface of the bat roost (all of the Nycteribiidae recorded here and probably most of the Streblidae) or drop the larvae on the floor of the roost (Askew, 1971; Marshall, 1981). The larvae pupate very shortly after deposition by the adult. P.W. Price (1980) suggests that the" pupae and emergent adults of the Streblidae remain in close proximity to the bat host population and isolated from populations of other roosts, but he considers neither the species which drop their larvae freely, nor the possibility of mixed bat roosts. Streblidae are here considered rather more independent of their hosts than Nycteribiidae because many species of the former family are winged whereas species of the latter are apterous.

Hippoboscidae are ectoparasitic on birds and some mammals; the female produces fully-developed larvae which drop to.the ground and rapidly pupate; the emergent adult lacks immediate contact with the host. The adult flies are mostly winged, although sometimes lose their wings on contact with the host. The percentage of monoxenous species of this family presented in Table XV is taken from P.W. Price (1980), who in turn took it from Bequaert (1956). For the purposes of the discussion below a different figure, obtained from Maa (1963) is used, as Maa lists all species in the family, not just the American ones. Maa (1963) cannot be used as a source for figures in Table XV, as he employs the general terms ' oligoxenous1 and 1 polyxenous* but does not give numbers of host species. It is notable that the percentage of monoxenous species calculated from Maa (1963) (33.1%) is higher than that calculated from Bequaert (1956) (17%). 481

Oestridae are, as larvae, endoparasitic on large mammals, either living in the nasal cavities or burrowing in the flesh. The adult females are free-flying, depositing eggs or first-instar larvae on the hosts. There are three larval instars and a pupa.

Siphonaptera Adult Hystrichopsyllidae are ectoparasites of small mammals; the larvae live freely in the nesting material, while the adults are long-lived and mobile, frequently changing host (P.V7. Price, 1980). They are oviparous, with three larval and a pupal stage.

The score ('independence score') obtained in Table XVI is plotted against the percentage of mpnoxenous species of each family (from Table XV) in Fig. 265. Despite the crudity of the independence score it is apparent that there is a broad (negative) correlation between the two (regression described by y = 92.917-3.709x, with a probability that the slope s 0.0 and the regression is not significant of < 0.05). Two parasite families, the Philopteridae and the Hippoboscidae, depart very clearly from the expected level of host specificity as indicated by the regression line. The bias in the sampling of the Philopteridae has already been discussed, and an overestimate of the monoxenia of this family was expected. Any bias in the sampling of the Hippoboscidae, however, would also be expected to raise the apparent level of monoxenia, for reasons similar to those outlined for the Cimicidae. Despite this probable bias, the Hippoboscidae are much less host specific than would be predicted from their host independence score. There are three factors that might cause this apparent anomaly; a high level of vagility of the fly, a high level of vagility of the host, and a low level of specificity in the stimulus used by the fly to detect the host. These will be discussed in section 4.3.3.3.• The relationship between the percentage of monoxenous parasites in a family and the degree to which the parasites are necessarily associated with the hosts and their immediate environment may be viewed in two ways: (1) the less time the parasite spends away from the host the lower the chances of it encountering more than one host species and therefore the greater the probability of it developing a dependence relationship v/ith 482

Percentage of species in Family restricted to one host species

Fig.265 • Relationship between percentage of monoxenous species in families of parasites (from Table XV) and the extent to which location of parasite is determined by location of host ('independence score' from Table XVl). Family names are abbreviated as follows: C - Cimicidae; H - Hippoboscidae; Hy - Hystrichopsyllidae; N - Nycteribiidae; 0 - Oestridae; P - Philopteridae; Po - Polyctenidae; S - Streblidae; T - Trichodectidae* 483

that host; or (2) the most host specific a parasite the less it can afford to leave the host because of the concomitant lowering of probability that it will encounter a host to which it is adapted. A third alternative, that both might be correlated with an as yet unidentified factor, is not considered here. Of these alternatives only the first offers an explanation (though not necessarily complete) of host specificity, and for this reason it is preferable to the second. If the second alternative is correct, the restrictions on independent movement of the parasite would post-date the assumption of parasitism, and host specificity would be determined solely by other factors. This prediction may be investigated. It is believed that the Phthiraptera, Siphonaptera and perhaps Hippoboscidae evolved from nest-dwelling ancestors that developed a dependence on the bird or mammal inhabitant of that nest, possibly through a phoretic association (to'-ensure location of new nests) (Waage, 1979). Jobling (1949) suggests a similar pathway for the Streblidae and Nycteri- biidae, although Oldroyd (1964) believes that the association with cave environments is secondary for these flies, and that they developed from glossinid-like ancestors (see below). The ancestor of the Phthiraptera was almost certainly apterous, just as the modern Liposcelidae. The morphological (and hence ecological) similarity of adult and nymphal lice suggests that a shift from detritus feeding to parasitism, with its attendant constraints, would affect both stages. The evolution of oviposition on the host rather than in the nest removed any need to leave the dermecos, and the -restriction of independent dispersal of the louse (as adult or nymph) would inevitably follow. The evolutionary processes postulated do not require the development of •host specificity1 by any other means. Waage (1979) suggests that the ancestor of the Siphonaptera was apterous and probably capable of leaping, just as the modern (non-parasitic) Boreidae (Mecoptera). The distinctness of the larval and adult stages of fleas permits different habits for each, and this also almost certainly antedates the development of parasitism. The degree to which the dispersal of the flea would be necessarily constrained by its dependence on the host was therefore determined to some extent before 484

the development of parasitism in the order, although some families of fleas show a greater degree of host specificity (and dependence on the host) than do others. The effect on host specificity of adaptation to nest type is discussed briefly below. The different vagilities of Hippoboscidae, Streblidae and Nycteribiidae may pre-date or post-date the development of parasitism. Nycteribiidae and Streblidae are generally agreed to be sister-groups (Wenzel et al., 1966; Schlein, 1970; Griffiths, 1972). If this is so, the association with bats is likely to be ancestral, and the reduction in vagility in Nycteribiidae (occasioned by the loss of wings) post-dates parasitism. These two families may have an origin independent to that of the Hippo- boscidae (Wenzel et al., 1966; Schlein, 1970) or be the sister-group of this family (Theodor, 1957; Griffiths, 1972). If the latter suggestion is correct, parasitism is likely to be plesiomorphic for the clade, and the differences in vagilities of the three families are all subsequent developments. It is notable that although most of the Hippoboscidae parasitic on birds axe fairly strong fliers, some, and some of the mammal parasites, are apterous, or with wings reduced or caducous. The relatively low independence score of the Hippoboscidae, Streblidae and Nycteribiidae is due in part to the development of pupipary, a modification they share with the haematophagous Glossinidae. The Glossinidae, which are placed by Griffiths (1972) as sister-group to the Hippoboscidae plus Streblidae and Nycteribiidae, are not dependent on the host for dispersal at all, and exhibit a very low host specificity. Pupipary is probably a k - selective strategy developed in response to a high-protein adult diet. If this is the case and the relationship of the flies is as suggested by Griffiths (1972), pupipary probably post-dates the association of the flies with vertebrates, but antedates parasitism. The development of endoparasitism in the Oestridae was probably via facultative myiasis, with the development of obligate myiasis (obligate endoparasitism) ensuring dependence of the larva on the host. It has been shown that there is a strong correlation between host specificity and the degree to which the parasite is dependent on the host for dispersal, and that this restriction on independent dispersal 485

occurred (at least in some groups) concomitant with the development of obligate parasitism and not as a result of host specificity. These observations indicate that the observed host specificity of the groups concerned (including lice) is, at least in part, an inevitable consequence of parasitism. This is not to say that selection does not favour mechanisms to maintain the level of independent dispersal of the parasites. In the lice the development of positive thermotaxis (Murray, 1957a) and negative phototaxis (Wigglesworth, 1941) is clearly in response to selection pressure against the lice moving from close to the skin of the .host to the inhospitable outer regions of the pelage where they might be displaced from the host. The hypothesis that restriction of independent dispersal determines the degree of host specificity does not explain all aspects of host specificity, however, and other factors are discussed below.

4*3*3*3. Host suitability

If host specificity is due only to the restriction of distribution of a parasite as described above, any parasite if placed on a 'new* host will be able to survive as well as if it were on its original host. This patently does not happen, even for lice, which are very greatly constrained by their restriction to the host. The European , for example, despite the close association of its young with foster-parents of a number of different species, has not acquired the lice of any of these species. Ash (1960) reports that Philonterus citrinellae. a feather- feeding louse of the chaffinch, cannot survive on feathers of the blackbird, even though 3?. turdi (a normal parasite of the blackbird) feeds exclusively on these. Wilson (1934) found that chicken lice died after being fed on heron feathers, and Eichler (1936) obtained similar results with pigeon lice fed on heron feathers. Ewing (1933) fed specimens of Pediculus (from a spider monkey), (from an Old-World monkey) and Lino gnat hus (from the dog) op man, and all died; Krynski et (1952) demonstrate

that the blood of guinea-pigs is fatal to Pediculus hum anus from man. Murray (1957b, c, d) demonstrates that for Bovicola ovis and Werneckiella equi the diameter of the host hair is important for. successful oviposition 486

and Ludwig (1982) suggests that the spacing and diameter of the host hair is correlated with the development of the legs and claws in the Anoplura. Such observations have led to the supposition that the adaptation of the louse to the environment of the host is a major, if not the only, factor determining host specificity (Lakshminarayana, 1977; Rai & Lakshminarayana, 1980; Ludwig, 1982). As shown in section 4.3.3.2., this assertion is probably not correct, but it is apparent that adaptation to the host does play a part in host specificity. Two aspects of parasite-host coadaptation can be distinguished: the environmental tolerance of the parasite, and the environment for parasites available on or in the host. Overlap between the tolerance of the parasite and the environment of the host must have pre-dated parasitism, otherwise the ancestor of the parasitic group would not have been able to survive upon the first host. Following the initial invasion of the host the parasite tolerance must have been modified as the parasite moved (in evolutionary time) from casual association to facultative parasitism to obligate parasitism. Following the development of obligate parasitism the selective pressure on the parasite to adapt to the host (and on the host to adapt to or resist the parasite) must have, continued, the strength of the pressure to some extent being related to the degree of dependence the parasite necessarily has on the host (see section 4.3«3«2.). The various factors making up the environment offered by the host to lice, and to which lice might become specifically adapted - temperature, humidity, insulation provided by the pelage, hair spacing, hair diameter, hair/feather colour, feather structure, chemical constitution of feathers, blood, skin, sebaceous exudates or other food, odour, host grooming patterns, etc. - have been discussed by a number of authors (e.g. Hopkins, 1949, 1957; Clay, 1949a, 1957; Arora & Chopra, 1957; Ash, 1960; Eichler, 1963; Lakshminarayana, 1977; Rai & Lakshminarayana, 1980; Ledger, 1980; Hennache, 1981; Marshall, 1981; Ludwig, 1982). The effect of adaptations to some of these factors by lice is demonstrated by the examples of specificity given above. However, other tests, particularly on feeding of Pedicuius humanus (Ludwig, 1973), and the records of secondary infestation listed in section 4.3.2.3., indicate that such adaptations 487

do not always result in monoxenia. Any case of secondary infestation must indicate an 'overlap* of parasite tolerance and the environmental conditions provided by the secondary host. There is no reason to suppose that the environment provided by a member of one host taxon should not also be provided by a member of another, or that adaptation by the parasite to the environment provided by one host should preclude the possibility of its tolerating the environment provided by another, even if it is not identical. Furthermore, the adaptation by a parasite to a host (or modification of the parasite's tolerance to fit a new or changing host environment) may change the range of hosts potentially available to the parasite, as some of the features of the host environment may be non-specific, either through patristic similarity or convergence with, other (potential) hosts. The nature of the environmental overlap between different hosts that permits the secondary establishment of a parasite must differ in each case, and cannot with the present level of knowledge be predicted (as can be seen from the list of secondary infestations given in section 4.3.2.3.). It might be expected that the closer the relationship between host taxa (phyletic and phenetic relationship) the more features of the environment offered to parasites would be shared (i.e. the greater the environmental 'overlap'). Colonisation of a 'new' host will therefore more probably be followed by successful establishment of the parasite if the 'new' and 'old' host sire closely related than if they are more distantly related. Natural secondary infestations are more detectable, however, if the hosts are not closely related, as the closer the relationship of the hosts the more probable the distribution of the parasite is explicable in terms of Fahrenholz' Rule. Evidence for successful establishment of lice on closely related hosts following secondary colonisation is consequently difficult to procure. Timm & Price (1980) produce a distribution map of eight species of Geomydoecus parasitising pocket gophers of the Geomys bursarius complex (Fig. 263) • The distributions of lice and pocket gophers seem largely independent (given that a louse record always corresponds to a record), and it appears that the lice have a 488

geographical distribution superimposed upon, and only incidentally coinciding with, the distribution of the pocket gopher taxa. Unless the pocket gophers underwent stasipatric speciation, without concomitant speciation by their lice, it must be inferred that the lice are success- fully colonising and establishing upon different related taxa on a fairly regular basis. Price (pers.comm.) disagrees with this interpretation, and believes that the taxonomy of the gophers is at present weak, and will eventually be modified so that the 'good' taxa produced have distributions that coincide with the distributions of the lice.

In addition to adaptations to the environment provided by the host, two further host-related adaptations must be mentioned, although neither is particularly applicable to lice. Firstly, a parasite that utilises (during some or all of its life-cycle) the local environment of the host must become adapted to that environment as well as to the host. In some cases this adaptation may lead to ecological rather than host specificity as the most apparent causal factor of the parasite's distribution. A prime example of this is the 'nest specificity' of many fleas, where larval and pupal environmental tolerances determine in great part the hosts available to the adult (Smit, 1957), and in which such ecological specificity may have been more important than phylogeny of the host as a major isolating factor in evolution (Hopkins, 1957a). The identity of the host is not entirely irrelevant to fleas exhibiting nest specificity, however, as some polyphagous fleas are more fertile on some of their hosts than on others, a factor that is likely to be related to the composition of the host's blood (Askew, 1971). Whatever the extent of environmental specificity the limitations imposed by this on host specificity do not differ in principle from those imposed by adaptation to the host itself as discussed above. Secondly, a host species, is only 'available' to a parasite if the parasite can detect and recognise it. Marshall (1981: 171) lists the responses to external stimuli (temperature, odour of host, CO^, vibration, air movements, light, gravity, and physical contact) of a number of ectoparasites of birds and mammals. At least some of these stimuli are utilised by the parasites in host location. It is clear that the 489

potential specificity of the stimuli will vary, the response to odour, for example (which is utilised by many parasite groups), being likely to be much more specific than the response to a visual stimulus. Ferhaps significantly, in view of their low level of host specificity, many Hippoboscidae detect their hosts by sight, the importance of this stimulus to the species being related to the vagility of that species as expressed in the development of the wings (Theodor, 1957; Marshall, 1981). The potential specificity of the stimuli used in host location varies within and between parasite taxa, and with it must vary the probability of the parasite detecting and colonising birds and mammals other than the primary host•

4.3.3.4. Availability of suitable hosts

In the last section it was argued that the adaptation of a parasite to the environment provided by one host taxon does not preclude it from tolerating the environment provided by other taxa of the host class. .In addition to the extant host or hosts, therefore, a parasite may have a number of 1 potential hosts* which, although suitable, are not used. Two factors are identified here as possibly instrumental in- preventing an association between a parasite and a potential host: interaction with other ectoparasites, and lack of opportunity to colonise. The first of these factors may be discussed very briefly, although this brevity is a result more of limited information than an accurate assessment of importance. Two interactions are possible between lice and other ectoparasites (including other lice): negative (competition) and positive (•co-operation1). The first of these has not been demon- strated to occur, although there is certainly some niche partitioning among lice that could be explained as the result of direct or indirect competition (Waage, 1979). Ward (1957) demonstrates that competition is not a limiting factor for at least some of the species of lice parasitising tinamous. A more direct negative interaction of lice, predation, has not been observed, but there are several records of lice with fragments of other lice in their crops (Waterston, 1926; Eichler, 1936, 1937) and members of a series of Cebidicola extrarius examined in 490

this study had fragments of an unidentified arthropod present in their alimentary canals. Ward (1957) demonstrates a positive association between pairs of ischnoceran species parasitic on members of the genus Tinamus (Aves). He attributes this in part to a possible co-operative interaction between species, although he accepts that niche partitioning could be the cause. The second factor, lack of opportunity to colonise, can be discussed in rather more detail, and further influences on the availability of the host indicated. The opportunity of the parasite to colonise the potential host must depend to some extent on the geographical distribution of both extant and potential hosts. If the geographical or topographical ranges of the extant and potential hosts do not overlap at any point, clearly the parasite will not be able to colonise the potential host. For example, the range of the original marsupial host of the amblyceran Heterodoxus spirri.ger did not overlap the ranges of any of the present canid and felid hosts until the dingo was introduced into the Australian Region (see section 4.3.2.3«), and thus before that time all of its present hosts other than Wallabia agilis were unavailable, although suitable. In this case a number of other potential hosts may still be unavailable, for Heterodoxus spiniger is not found outside 40°N and 40°S. The more vagile or widely-distributed the host, the greater the number of potential hosts that will become available to the parasite. This relationship provides a further reason for the depressed host specificity of the Hippoboscidae, as many of the family are parasitic on birds, which are frequently more vagile than the mammalian hosts of the other parasites considered in section 4.3.3.2. (including bats). For example, the hippoboscid Ornithomya avicularia has been carried by birds almost all over the world (Richards & Davies, 1977) • Clearly, the more host taxa are used by the parasite, the greater the chances of further potential hosts being colonised. Even if the ranges of a. potential host and the parasite do overlap, colonisation may still not be possible if the extant and potential hosts do not come within a suitable distance of one another. This 'suitable distance' is the distance over which the parasite can locate and colonise a host, and provides a further clue to explain the depressed host

V 491

specificity of the Hippoboscidae. The majority of Hippoboscidae are fairly strong fliers (Marshall, 1981) and are probably capable of travelling much greater distances than species of any of the other parasite families discussed (with the exception of the Oestridae), even as 'reproducing' adults (only scored '1' in Table XVI). This great vagility of the parasite allows colonisation of hosts over much greater distances than would be possible for parasites with a lower vagility. Theodor (1957) shows there to be a relationship in the Hippoboscidae between host specificity and the loss or reduction of the wings. The Philopteridae, which like many Hippoboscidae are parasites of birds, do not exhibit the same depressed host specificity, as their own vagility is very low. The flying ability of the Oestridae has been mentioned, but the level of host specificity does not seem to be depressed as would be expected from comparison with the Hippoboscidae. The reason for this relatively high host specificity may be found in the scarcity of potential hosts. All hosts of Oestridae have a number of features in common: they are terres- trial herd-forming herbivores, at least lm long and with a body weight of at least 20-25 Kg.. * It may be assumed that anything not fulfilling these requirements does not qualify as a potential host. Such hosts are not common, and it is likely that very few suitable species will be present in the range of a fly population, Oestridae being relatively short- lived (Zumpt, 1965).

As already noted, the vagility of lice, and hence the distance over which they can locate and colonise a host, is very low. as lice cannot fly, a host individual can only acquire lice in one of the three following ways: (i) directly from a parasitised animal; (ii) from an environmental feature 'shared' with a parasitised animal; (iii) via an intermediate host (phoresy). These are treated separately:

(i) From a parasitised animal. As already mentioned, lice are restricted to a region close to the skin of the host by a combination of negative phototaxis and positive theimotaxis. When two animals of the host class come into physical contact the temperature and light gradients from the outside of the pelage break down, and the lice are free to move between the animals. Lice may 492

also come to the surface of the fur if one host rubs against another (Mathysse, 1946; Murray, 1963). Such contacts of host-class animals may take place as follows: between parent and offspring; between foster- parent and young (brood parasites); between copulating pairs; during communal roosting, nesting etc.; during casual contact; during territorial interactions; between predator and prey; and during other intra- or inter-species interactions.

(ii) Through an environmental feature. Should a louse leave a host, or become displaced from it, and then come into contact with another individual of the host class, it may manage to colonise successfully. Inter-specific transfer has been demonstrated to occur by way of dust-baths (Hoyle, 1938), and it is possible that intra-specific transfer can occur via the egg-shell in some waders (Rankin, 1982) • It is certainly possible for lice to remain alive after the death of the host or after being separated from it. The collections of the British Museum (Natural History) contain at least two specimens of bird lice collected from leaf-litter. Specimens of Gonoides (Isch- nocera) from a turkey were recently received alive through the post having been removed from the host at least three days before, and Ledger (1980) records a specimen of Laemobothrion (Amblycera) still alive three weeks after the death of its host.

(iii) Phoresy. Ischnocera and Anoplura have been reported on numerous occasions attached to hippoboscid flies (Keirans, 1975a, b), Corbet (1956) reporting 22 lice on a single fly. Keirans (1975a) stresses the frequency of the phoretic association of lice and Hippoboscidae, and, although considering it to be a 'last-ditch' expedient, suggests that phoresy is a survival and dispersal mechanism of lice that is not infrequently exploited. Hippoboscidae may move between different species of host (Marshall, 1981), clearly providing an opportunity for lice to colonise different species of mammal or bird.

Of the methods of colonisation listed above those involving physical contact between two individuals of the host class probably account for 493

most colonisation in the lice. If this is so, there must in most cases be some ecological or environmental 'overlap* of extant and potential hosts for colonisation of the potential host to occur. Such overlap need be only temporary and lice colonise only once for establishment on a potential host to take place, although more repeated colonisations would presumably increase the chance of establishment. Relatively frequent contacts between two species of the host class may lead to the sharing of a species of lice, speciation of the lice being prevented by a sufficient level of gene flow.

4*3»3.5. Summary

The likelihood of a louse (or other .ectoparasite of bird or mammal) colonising a given species of the host 'class1 is linked to the necessary (distributional) dependence of the parasite on the extant host, the geo- graphical location of both extant host and recipient animal, the degree of ecological and environmental overlap between the two animals with reference to the vagility-of the parasite, and the ability of the parasite to recognise the recipient species as being of the host class. The likelihood of the parasite successfully establishing a population on a given species of the host class is dependent on the initial colonisation (which must comprise a pair or fertile female),- the suitability of the species colonised (i.e. overlap of the environment provided and the environmental tolerance of the parasite) and, perhaps, freedom from competition with other, established, parasites. The result of the interaction of all these factors (see Fig. 264) is the observable host specificity. The initial host or hosts is carrying parasites either because it has inherited them (adherence to Fahrenholz* Rule, at least in part), or because the parasites have colonised from another host taxon and have successfully established and speciated. The relative probabilities of these two alternatives are those which govern the likelihood of the parasite colonising and establishing on a 'new' host. 494

4.4. COMPARISON OF TRICHODECTID AND HOST FHYLOGENIES

4.4.1. Introduction

The purpose of the discussion below is to examine the phylogeny of the Trichodectidae (as inferred from the cladogram) in relation to such host phylogenies as are available, and thereby to determine the relationships between the two. As already noted, bifurcations in the cladogram of Trichodectidae are considered to represent bifurcations in the phylogenetic tree of the family, but multiple furcations may represent a number of different topological arrangements of the tree. Comparisons of the trichodectid phylogeny and mammal relationships . are hampered by the methods of analysis and classification generally used by mamma.! systematists. There has been very little cladistic work done on the groups of mammals parasitised by Trichodectidae, relationships generally being determined by palaeontological methods and classification following •evolutionary1 principles (see Simpson, 1945, 1961). The classification of mammals is therefore predominantly •horizontal*, comprising many paraphyletic •grade* groups which cannot be compared meaningfully with the holophyletic groups of lice. The phylogenies of mammals reconstructed from fossil evidence are rarely sufficiently complete for comparison with the trichodectid phylogeny to be valuable. Despite these difficulties some statements can be made, and certain correlations and differences detected. The subfamilies of Trichodectidae will be treated individually below. To determine the identity of the primary host of a holophyletic group of lice parasitising more than one host group the method of maximum parsimony is used, the host taxa being treated as apomorphies of the lice. For example, in the Lorisicola paralaticeps - mungos clade (Fig. 280), •paralaticeps and laticeps parasitise Atilax paludinosus whilst the sister- species of laticeps. L. mungos. is a parasite of Herpestes sanguineus. In this case it is inferred that the association of L. mungos with H. sanguineus is apomorphic for the clade, and therefore marks a speciation event associated with a secondary infestation. Clades of Trichodectidae are identified in the text by the names of the two •extreme* taxa of the clade, reading from the top of the page down (Figs 269, 271, 273, 274, 277, 278, 275, 280, 281). •SUIDAE

•TAYASSUIDAE

•HIPP OPOTAMIDAE

•CAMELIDAE

TRAGULIDAE

-KOSCHIDAE

• H YDR OP OTINAE

•MUNTIACINAE CERVIDAE • ODOCOILEINAE RIMINANTIA PEGORA •CERVINAE

GIRAFFIDAE

•ANTILOCAPRIDAE

•BOVIDAE

Cladogram of the living Artiodactyla. Compiled from Hamilton (197®)» Leinders & Heintz (1980) and Eisenherg (1961). REDUNCINI HIPP OTR AGIN AE HIPP OTR AGINI ] TRAGELAPHINI BttODONTS BOVINI BOVINAE EARLY BOSELAPHINI BOSELAPHINI BOSELAPHINI CEPHALOPHINI •cEPHALCPHINAE _

ANTILOPINI JANTILOPINAE NEOTRAGINI

OVIBOVINI

CAPRINI G APRINI GAPRINAE AEGODONTS

•RUPICAPRINI• J

AEPYOEROTINI ALCELAPHINAE ALCELAPHINI ]

Fig. 267. Phylogeny and classification of the Bovidae. Compiled from Gentry (1978).

vo ON TRAGELAPHINI

'130SELAPHINI BOVINAE

•BOVINI

CEPHALOPHINI

NEOTRAGINI

REDUNCINI

ANTILOPINI

ALCELAPHINI ANTILCPINAE (+ AEPYCEROTINl)

HIPPOTRAGINI

CAPRINI (+ RUPICAPRINI and OVIBOVINI)

Pig. 268. Phylogeny and classification of the Bovidae. After Kingdom (1982).

vo 498

4.4.2. Bovicolinae

All species of Bovicolinae other than Genus n. 3, most Werneckiella and one Bovicola are associated with Bovidae and Cervidae (Artiodactyla: Pecora)• For comparative purposes, possible phylogenies of the Artiodactyla and Bovidae are depicted in Figs 266, 267, 268. The phylogenetic relation- ships of the pecoran families (Moschidae, Cervidae, Antilocapridae, Giraffidae and Bovidae) are much disputed, almost every possible combination of sister- groups having been suggested (e.g. Hamilton, 1978; Leinders & Heintz, 1980; Eisenberg, 1981). For this reason no attempt is made in Fig.266 to resolve the relationships of the Pecora beyond the initial pentafurcation. There are also problems in determining the relationships of the tribes of Bovidae, and two alternative phylogenies for the living members of the family are depicted in Figs 267 and 268;

4.4.2.1. Damalinia

Species of Damalinia are associated with hosts in the Bovidae and Cervidae. Those species of Damaliui a associated with members of the latter family form two small clades in the subgenera D. (Cervicola) and D. (Tricholipeurus), and appear by their positions on the cladogram (Figs

271 9 273) to be the result of independent invasions of Cervidae from Bovidae. The primary host association of Damalinia. therefore, is believed to be with an ancestor of all or part of the Bovidae. Details of the host associations are discussed below by subgenus of Damalinia; the cladograms of the subgenera, with associated host identities, are depicted in Figs 269, 271, 273. Species of D. (Damal inia) are associated with Alcelaphini, Caprini, •Rupicaprini', Antilopini, Neotragini and Reduneini (Fig.269). In no postulated phylogeny of the Bovidae do these tribes foim a holophyletic group, although in Kingdom's (1982) phylogeny the Reduncini are more closely related to the other tribes than is indicated by Gentry (1978). It is notable that Gentry (1980) does raise doubts about the bBodont affinities of the Reduncini. The two species of D. (Damalinia) parasitic on reduncines (D. adenota a^d D. hilli, both associated with Kobus spp.) form an unresolved -theilerl Connochaetes taurinus Bovidae: Alcelaphlni

-neotheileri Connochaetes taurinus Bovidae: Alcelaphlnl

-semitheilerl Connochaetes taurlnus Bovidae: Alcelaphlnl

-crenelatus Damallscus dorcas Bovidae: Alcelaphlnl

-baxl Damallscus lunatus jlmela Bovidae: Alcelaphlni

-harrlsonl Connochaetes gnou Bovidae: Alcelaphlnl

-chorleyl Alcelaphus buselaphus lelwel Bovidae: Alcelaphlnl

-ornata Alcelaphus buselaphus caama . Bovidae: Alcelaphlni

-thompsonl Caprlcornls sumatraensls Bovidae: 'Ruplcaprlni'

•orlentalls Capricornis crispus swlnhoel Bovidae: 1 Ruplcaprlni'

-dimorpha "wild goat from Hangchow, China" Bovidae

-pelea Pelea capreolus Bovidae: Neotraginl

-adenota Kobus kob, K. vardonl Bovidae: Redunclni

-hllll Kobus e. elllpslprymnus, K. e. defassa Bovidae: Redunclnl

-fahrenholz1 Gazella arablca . Bovidae: Antlloplnl

•appendlculata Gazella gazella, G. subgutturosa, G. thomsonl Bovidae: Antiloplnl

Fig. 269- Cladogram and host associations of Damalinia (Damalinia). vo \o 500

trichotomy on the cladogram with JD. pelea, a parasite of the rhebok Pelea capreolus. The position of JD. pelea is tentative, as specimens have not been seen in this study. Pelea is placed by some authors in the Reduncini (Honaki, Kinman & Koeppl, 1982),but it has been placed in the Caprini (Gentry, 1970) and in the Neotragini (Gentry, 1978, Gentry, pers. comm.). If the phylogeny proposed by Kingdom (1982) is correct, Pelea might be a reduncine or neotragine without having to invoke secondary infestation, although if Gentry (1978) is correct some secondary infestation must be invoked to permit Pelea to be anything other than a reduncine. The identities of the remaining hosts of D. (Damalinia). and their relative position on the clado- gram of lice, suggest that some secondary infestation has occurred, and that the association with Reduncini and Pelea is unlikely to be primary. Some secondary absence may be suggested. The two species of Kobus parasitised by D. adenota and I), hilli form a monophyletic group, possibly made paraphyletic by the exclusion of K. leche (Gentry, 1978) • If I), adenota and D. hilli are sister-species, the association may be explained by a single ancestral infestation from another bovid, with possibly secondary absence from K. leche. The sister-group of the D. pelea - hilli clade discussed above is the JD. thompsoni - dimorpha clade. The host of D. dimorpha is not known, although Hopkins (1949) suggests that, as both Naemorhedus goral ('Rupi- caprini1) and Capricornis sumatraensis (•Rupicaprini*) both occur in the neighbourhood of its collection, one of these may be the host. Both D. thompsoni and D. orientalis parasitise Capricomis species, and a single ancestral association with the ancestor of these two hosts seems likely. Naemorhedus and Capricornis are generally placed in the Rupicaprini, but Gentry (1978) would like to transfer its type-genus, Rupicapra. to the Caprini, and Kingdom (1982) includes the whole tribe in the Caprini. Whichever the host of JD. dimorpha. if it is one of the two suggested by Hopkins (1949), a single ancestral association of the D. thompsoni - dimorpha clade with either the ancestor of Capricornis or of Capricomis plus Naemorhedus seems probable. The latter group is almost certainly not sister-group to the hosts of any of the D. pelea - hilli clade, nor of the D. fahrenholzi - appendiculat a clade, which is the sister-group of 501

the D. thompsoni - hilli clade. At least one more secondary association must "be inferred, therefore, either by the ancestor of the D. thompsoni - dimorpha clade or that of the D. f ahrenholzi - appendicul at a clade. The hosts of the I). f ahrenholzi - appendicul at a species pair belong to the genus Gazella. Species of this genus are among the few bovids to span both. Africa and Eurasia in their distribution (Groves, 1969), and this perhaps offers an explanation of the association of the sister-group of the lice with hosts in both Africa and China. The I). (D.) theileri - ornata clade parasitises the tribe Alcelaphini (Bovidae: Alcelaphinae), and comprises the only Trichodectidae known to do so. The sister-group of the Alcelaphini is the Aepycerotini (Vrba, 1979), which comprises only one species, Aepyceros melampus. This species is parasitised by two Trichodectidae, both of them in D. (Tricholipeurus) and not in a sister-group relationship with-the lice of Alcelaphini. These lice are discussed below. The cladogram of the lice of the Alcelaphini (Fig.269) may be compared to the cladogram of the hosts produced by Vrba (1979) (Fig.270a). The two cladograms do not entirely accord, some of the differences arising from the lack of full resolution of the trichodectid cladogram. If this lack of resolution is corrected to a series of dichotomies by use of host relationships, a cladogram rather more similar to that of the hosts is produced (Fig.270b). The only anomaly in this cladogram is in the position of the species parasitising Damaliscus, which is clearly not in accordance with any conception of Connochaetes as a holophyletic genus. To achieve full accord between host and parasite cladograms a secondary infestation of Damaliscus from Connochaetes is postulated. Y/ith this hypothesis, the association of the D. theileri - ornata clade with the Alcelaphini, with the exception of a single secondary infestation, can be adequately described by reference to Fahrenholz* Rule. It is predicted that lice of the clade will be found on Sigmoceros lichtensteini. but not necessarily on Damaliscus hunteri.

Species of D. (Cervicola) are associated with Tragelaphini, Reduncini, Cephalophini and Cervidae (Fig.271). The association with Cervidae (of the D. maai - muntiacus clade) has been noted above as secondary. The four described and one undescribed 502

CO ra •H 3 s—X ft ft CO s ra U 03 ra ra ft ft •H PJ'-N m 3 ra 3 rH 2 ca •H ca G -P CQ 3 •H G rH 53 •H G rH >jH O ca ca -p U -H •H »H ra -h ca O (B <]} m 3 rH o ca +> ca rH 01 •H 0 <0 Jh 5 U A & SS ca 3 U 3 g o •H B •h ca -p ca •h ca G ft ra tJ ra -p 0 -p S -P U U ft ra ra 0 0 • ft • o ft • •p • G • • ra • ' • x^ . p p P p. ^^ p p P

Fig. 270. Cladograms of lice and their hosts, (a) Cladogram of Alcelaphinae, after Vrba (1979)• Extinct (fossil) species are marked "by short "branches. (b) Cladogram of the Damalinia theileri - ornata clade if unresolved trichotomies resolved using host identity as an apomorphy (c.f. cladogram in Fig. 269«)« •hendrickxl Cephalophus nigrlfrons Bovldae: Cephalophlnl

-undescrlbed Kobus megaceros Bovldae: Redunclnl

-martlnaglla Kobus leche Bovldae: Redunclnl

-maal Cervus nlppon Cervldae: Cervlnae

-forficula Axis axis, A. porclnus Cervldae: Cervlnae

-? undescrlbed Hydropotes lnerals Cervldae: Hydropotlnae

-meyeri Capreolus capreolus Cervldae: Odocollelnae

-muntlacus Muntlacus muntjac Cervldae: Muntlaclnae

-reduncae Redunca arundlnum Bovldae: Redunclnl

-ugandae Redunca redunca Bovldae: Redunclnl

-trabeculae Redunca fulvorufula Bovldae: Redunclnl

-lerouxl Sylvlcapra grlramla Bovldae: Cephalophlnl

-annectens Tragelaphus s. scrlptus, T. spekel Bovldae: Tragelaphlnl

-hopklnsi Tragelaphus angasl, T. strepslceros, Taurotragus oryx Bovldae: Tragelaphlnl

-natalensls Tragelaphus scrlptus sylvatleus Bovldae: Tragelaphlnl

Fig. 271. Cladogram and host associations of Damalinia (Cervicola). vji o 504

species in the clade parasitise members of the Cervinae (Cervus. Axis), Odocoileinae: Capreolini (Caproeolus) , Muntiacinae (Muntiacus) and Hydro - potinae (Hydroootes) • Because the D. maai - muntiacus clade has not been fully resolved, the hypothesis of a primary association of the clade with the ancestor of the Cervidae, followed by co-evolution, is not refuted directly by the cladogram. Militating against this is (i) the absence of members of the clade from most Cervidae, including species very closely related to some of those parasitised, (ii) the association of the sister- group of the clade with hosts that evolved much more recently than the Cervidae, and (iii) the possible identity of the specimens from Hydropotes with D. meyeri from capreolus. The hypothesis of several secondary infestations between cervids would also account for the present host associations, and is regarded as more likely, especially as all of the deer concerned other than Axis spp. are sympatric in China and the Far East, and Axis and Muntiacus are sympatric in India (Corbet, 1978; Corbet & Hill, 1980). The D. maai - muntiacus clade is sister-group to the JD. hendrickxi - martinaglia clade, species of which are parasitic on Cephalophini and Reduncini. Species of the D. re dune ae - lerouxi clade, which is sister- group to the D. hendrickxi-muntiacus. clade are, surprisingly, also parasitic on members of the same two tribes. There are two possible explanations for this anomalous host distribution: (i) The ancestor of the clade was associated with (at least) two host species, a primitive reduncine and a primitive cephalophine. This association may have been from a common ancestor, or it may have been secondary on one of the two. Both host species and the louse species underwent speciation at roughly the same time, perhaps as the result of some geological or climatic event. The two louse species resulting were both still associated with two hosts, although further speciation may have followed fairly rapidly in the ancestor of the re dune ae - lerouxi clade. In the other branch, the ancestor of the maai - muntiacus clade may have moved from either the caphalophine or the reduncine to a cervid. The course of these postulated evolutionary events is depicted in Fig.272a. (ii) The ancestor of the clade was parasitic on either a primitive 505

Fig. 272. Two alternative possible phylogenetic associations between lice in Damalinia (Cervicola) and their hosts in the Reduncini, Cephalophini and Cervidae (R, Cep and Cerv respectively). 506

reduncine or a primitive cephalophine, and speciated with it; secondary infestations of the other bovid tribe and of the Cervidae followed to give the pattern of associations seen today (Fig. 272b). The Cephalophinae have very little ascertainable fossil history, and it is not known whether the group was ever distributed outside Africa. Fossils that probably represent primitive reduncines have been found in Asia (Gentry, 1978), and perhaps account for the transfer of lice to the Cervidae. The remaining two clades (three species) of D. (Cervicola) are associated with the Tragelaphini, although a single primary infestation of this tribe is not certain. • The association of D. hopkinsi with species of both Tarurotragus and Tragelaohus may be the result of secondary infestation of the former genus from the latter, or may be the result of a single primary association with the- common ancestor of the two genera (which may, if Tragelaphus is paraphyletic, lie within this genus). The association of D.annectens and D. natalensis with different subspecies of Tragelaphus scriptus is not compatible with the cladogram, and neither is the association of D. annectens and D. hopkinsi with JT. spekei and JP. angasi respectively, as these hosts are probably sister-species (Gentry, 1978). Species of I). (Tricholipeurus) are associated with Aepycerotini, Antilopini, Neotragini, Cephalophini and Cervidae (Fig.273). The association with Cervidae (of the D. albimarginata - indica clade) has been noted above as secondary. Lice of this clade are associated with three genera of the New World Odocoileini (the D. albimarginata - parallela clade) and the Old 'World Muntiacinae (D. indica parasitising Muntiacus muntjac). The cladistic relationships of these two host groups (Fig.266), coupled with the absence of lice of the clade from other Cervidae examined, suggests that the hypothesis of a single ancestral infestation of the Muntiacinae - Cervinae clade is unlikely. The hypo- thesis of a single ancestral infestation of the Odocoileini before their migration to America is more likely, and may be tested by examination of more deer of this tribe for lice. The association may be derived by secondary infestation from Muntiacinae or vice versa, or of both from an unknown third host. The initial infestation of the Cervidae must have •aepycerus Aepyceros melampus Bovidae: Aepycerotlni

-antidorcus Antldorcas marsupialIs Bovidae: Antlloplnl

• c. cornutus Ant Hope cervlcapra Bovidae: Antlloplnl

-c. ourebiae Ourebla ourebi Bovidae: Neotraglnl

.parkerl Gazella thomsonl Bovidae: Antlloplnl

-longlceps Gazella arablca Bovidae: Antlloplnl

-splnlfer Gazella grantl Bovidae: Antlloplnl

•albimarglnata Mazama americana, M. gouazoubira Cervidae: Odocollelnae

.dorcephall Ozotoceros bezoartlcus Cervidae: Odocollelnae

•llpeuroides Odocoileus hemlonus, O. vlrglnlanus Cervidae: Odocollelnae

•parallela OdocoIleus hemlonus, 0. vlrglnlanus Ceividae: Odocollelnae

-lndica Muntlacus muntjac Cervidae: Huntiaclnae

•llneata Raphlcerus campestris, R. sharpei Bovidae: Neotraglnl

-vlctorlae Madoqua gumtheri, M. klrki Bovidae: Neotraglni

•conectens Oreotragus oreotragus Bovidae: Neotraglnl

•pakenhaml Cephalophus adersl, C. montlcola, sylvlcapra grinmla Bovidae: Cephalophinl

•bedfordi Cephalophus montlcola Bovidae: Gephalophlnl

clayl Neotragus pygmaeus Bovidae: Neotraglnl

.moschatus Neotragus moschatus Bovidae: Neotraglni

•elongata Aepyceros melaropus Bovidae: Aepycerotlni VJl Fig. 273- Cladogram and host associations of Damalinia (Tricholipeurus). o 508

been from a bovid, and it is interesting in this context to note the suggestion of Gentry (1978) that the decline of the 1 african* bovids in India may have been linked to the burgeoning of the deer. The association of both D. lip euro ides and I), parallels with Odocoileus hemoinus and _0. virginianus indicates that speciation of the lice parasitising the Odocoileini has not entirely followed the predictions of Fahrenholz* Rule. The remaining hosts of the subgenus belong to the Aepycerotini, Antilopini, Neotragini and Cephalophini. Gentry (1978) regards the first three of these as having aegodont affinities and the last as bflodont (Fig.267), in which case the two lice parasitic on Cephalophini (D« paken- hami and I), bedfordi) are almost certainly associated with their hosts as a result of an ancestral secondary infestation from a neotragine. Kingdom (1982), however, figures the Cephalophini and Neotragini as sister-groups (Fig. 268), although referring in the text to a •neotragine* ancestry of many of the bovines. If these two tribes are sister-groups, secondary infestation of Cephalophini or Neotragini must still be invoked to explain the distribution of host associations on the cladogram (Fig. 273). If there have been no secondary infestations, Neotragus must be paraphyletic with respect to the Cephalophini and some of the other neotragine genera, given the position on the cladogram of D. clayi and D. moschatus (both parasites of Neotragus). Lice have been collected from only eight of the fourteen species of living Neotragini, and further collecting may elucidate the situation. Whilst both D. pakehhami and D. bedfordi parasitise species of Cephalophus, one of them (D. pakenhami) is also associated with Sylvicapra grimmia. If the genus Cephalophus is holophyletic, this association must be identified as secondary (see section 4«3«2.3. above). There is very little fossil record of the Cephalophini (Gentry, 1978), and the acceptance by mammal systematists of paraphyletic groups makes it possible that Sylvicapra has an ancestry in Cephalophus. If this is so, the association with D. pakenhami may be primary. Cephalophus and Sylvicapra live in slightly different environments (Gentry, 1978). The sister-group of the D. albimarginata - moschatus clade discussed above comprises a single species, D. elongata. parasitic on Aepyceros melampus (Alcelaphinae; Aepycerotini). The same host is parasitised by 509

D. (jT •) aepycerus, a member of the single remaining clade in D. (Tricho- lipeurus) , the D. aepycerus - spinifer clade. Species of this clade are also associated with members of the Antilopini and Neotragini but, because of the poor resolution of the cladogram, no suggestion can be made of primary host association on internal evidence alone. A possible hypothesis of the primary host of the D. (JC.) aepycerus - spinifer clade (and of the subgenus Tricholipeurus) is suggested by the sister-group relationship of Aepyceros and the Alcelaphini, which are parasitised by the I). (D.) theileri - ornata clade as described above. This hypothesis may be broken down into the following elements: (i) The ancestor of D. (Tricholipeurus)was associated with Aepyceros melampus. a primary infestation derived from its common ancestor with the D. (I).) theileri - ornata clade, which itself parasitised the ancestor of the Alcelaphinae. (ii) The first speciation event in the history of D. (Tricholipeurus) to be discernible by study of living species gave rise to two species on Aepyceros melampus (I), elongata and D. aepycerus) • (iii) All species of the D. (jC.) aepycerus -spinifer clade other than I). aepycerus became associated with their present hosts following secondary infestation from Aepyceros to one or more members of the Antilopinae. (iv) All species of the I). (T^.) albimarginata - moschatus clade became associated with their present hosts following secondary infestation of an unknown host (of the Neotragini or Cervidae) from Aepyceros. (v) All species of the D. (D.) thompsoni - aop endiculat a clade became associated with their present hosts following secondary infestation of an unknown host from an ancestor of the Alcelaphini. An alternative hypothesis is available, which explains the predominance of the association of the D. (T.) aepycerus - spinifer clade with Antilopini and that of the D. (JT.) lineata - moschatus clade with Neotragini. The hypothesis depends on a sister-group relationship between these two tribes, as may be inferred from the phylogeny in Fig.267, but which is denied by that in Fig. 268. The hypothesis postulates that the ancestor of D. (Tricholipeurus) was parasitic on the ancestor of the Antilopini plus Neotragini (Antilopinae sensu Gentry, 1978). The ancestor of the D. aepycerus - spinifer clade was associated with the ancestor of the 510

Antilopini, and secondary infestations of Aepyceros and Ourebia took place from antilopines. The ancestor of the D. albimarginata - elongata clade was associated with the ancestor of the Neotragini, and secondary infest- ations of the Cervidae, Cephalophini and Aepyceros took place from members of this tribe. Speciation on the common ancestor of the Neotragini and Antilopini might account for a primary association of D. (Damalinia) with this subfamily (Antilopinae sensu Gentry), and again division with the divergence into Neotragini and Antilopini took place (today represented by the lice on Pelea and Gazella). Secondary infest- ations of the Alee lap hini, Reduncini and 'Rupicaprini* took place. The hypotheses of Alcelaphinae and Antilopinae (sensu Gentry) as ancestral hosts each require eight secondary infestations to be postulated, so no choice can be made between them on the grounds of parsimony. The problem will probably not be resolved without further phylogenetic research into relationships within the Bovidae. It is notable that, if Gentry's (1978) divsion of the Bovidae into aegodont and bOodont tribes (Pig. 267) is accepted, D. (Damalinia) and D. (Tricholipeurus) are primarily associated with aegodonts and D. (Cervicola) is, as regards Bovidae, entirely associated with bOodonts. If Kingdom's (1982) division of the Bovidae into subfamilies (Pig. 268) is accepted, all members of Damalinia that sire aissociated with Bovidae, except three species of D. (Cervicola). are parasitic on members of the Antilopinae. The cladogram of Damalinia does not support one phylogeny of the Bovidae rather than the other.

4.4.2.2. Bovicola

Species of Bovicola are associated with Bovidae, Cervidae and Camelidae (Pig.274). Only one species, B. breviceps. is parasitic on a camelid, this being the only trichodectid associated with the family.

This fact, together with the position on the cladogram of

-2 s. sedecimdecembrii Bison bonasus Bovinae: Bovlni -3 tragull Tragulus javanicus, T. napu Tragulldae •W. aspllopyga ' Dquus (E.) burchelll boehml Equidae Equus (E.) caballus, E. (E.) przewalskli, -W. equl EJquldae E. (A.) hemlonus kulan Equus (A.) aslnus, E. (E.) b. burchelll, -W. ocellata Bquidae E. (E.) caballus -W. fulva Ammo tragus lervla Caprlnae: Caprlnl •W. neglecta Ammotragus lervla Caprinae: caprlnl -w. zebrae Bquus (E.) zebra hartmannae Equidae •w. zuluensis Equus (E.) burchelll antlquorum Equidae v_n

Fig. 274. am and. host associations of Bovicolinae other than Darnalinia. 512

Bovidae. The deer concerned belong to the Cervinae and the Rangiferini (Odocoileinae), and do not form a holophyletic group (Fig. 266). The absence of lice of this clade from other species of Cervinae, particularly Cervus, suggests at least one and probably two secondary infestations within the clade to account for the present distribution of host associations. It is notable that the deer parasitised are all distributed in the Western Palaearctic and Nearctic. The secondary infestations of Elaphurus davidianus and 0docoileus columbianus have been discussed in section 4.3.2.3.. The remainder of the species of Bovicola are associated with Caprini, Rupicaprini (although Gentry,1978, suggests that Rupicapra itself might be better placed in Caprini) and Bovini. The last-named tribe is not closely related to the other two (Figs 267 , 268), and the position on the cladogram of the single species parasitising a bovine (13. bovis, on domestic cattle), suggests the association to have been derived by secondary infestation from a caprine. The poor resolution of the cladogram prohibits much further speculatioi but it is notable that a number of the species are associated with the domestic goat. It may be that some of these are secondary infestations from wild species of Capra. but further collecting will be needed to confirm this. The ancestor of Bovicola was almost certainly parasitic on an ancestor of some or all of the Caprinae.

4.4.2.3. Werneckiella

Members of this genus parasitise species of Equidae (Perissodactyla) and Bovidae (Artiodactyla), five of them being associated with the genus Equus and the other two with Ammo tragus (Bovidae: Caprini) (Fig. 274 ). Cladograms of Eouus are provided by Bennett (1980) and Eisenmann (1980) (Figs 275a, b). Although the poor resolution of the cladogram of Werneckiella makes detailed comment impossible, the postulated sister- species relationship of W. eaui and W. ocellata is not compatible with either of the host cladograms. The primary hosts of these two species have been established by Moreby (1978). He notes that W. equi, commonly found on the domestic horse E. (E.) cab alius, has been taken from captive -ONAGER -QUAGGA

•HEMIONUS -BURCHELLI*

-ASINUS* tGREVYI

-QUAGGA -ZEBRA*

-CABALLUS* -ASINUS*

•ZEBRA* -HEMIONUS

•BURCHELLI* -GABALLUS*

•GRSVYI

Pig. 275. Cladograms of living (and one recently extinct) species of Equus. (a) after Bennett (1980); (b) after Eisenmann (1980). Species of Equus known to be parasitised by Vferneckiella are indicated by an asterisk (*). 514

specimens of El. (A.) hemionus kulan but suggests that, as these may have come into contact with domestic horses, the infestation is. probably secondary and EI. caballus the primary host (subgeneric placements of Bennett, 1980). Moreby is uncertain of the identity of the primary host of W. ocellata, but notes that numerous samples have been taken from the domestic ass E. (A.) asinus and only one from the zebra Ej. (E.) burchelli (the type host). The most probable primary host of the louse is therefore the ass. Compatibility with both cladograms can be obtained by postulating a secondary infestation between the domestic horse arid domestic ass, followed by speciation. If Bennett's cladogram is correct, the most likely direction of the transfer would be from El. caballus to E». asinus, leaving Werneckiella primarily associated with one branch of the cladogram (comprising species of the nominate subgenus). Discovery of species of Y/erneckiella on E. hemionus, kiang and onager would not necessarily be expected. If Eisenmann*s cladogram is correct, there is equal likelihood that the transfer took place in either direction, and associations between species of Werneckiella and El. hemionus, kiang and onager would be expected. An alternative hypothesis is also possible if this cladogram is correct. This is that the ancestor of the El. zebra - caballus clade was parasitised by an ancestor of W. equi, ocellata and zebrae, but speciation did not occur at the primary dichotomy of the equid cladogram. Speciation of the louse did occur when the ancestor of E. zebrae diverged from the ancestor of the El. asinus - hemionus clade, and when E. asinus and E. hemionus diverged. If this- is the case, E. hemionus will be found to be naturally infested with Y7. equi or a sister-species to this or YV. ocellata, and W. zebrae will be found to be the sister-species to these two or three species.

In the absence of a large number of secondary infestations of horses from other hosts, the association of Y/erneckiella with Bquus may date to the Hemphillian (early Pliocene) period in North America, as Bennett (1980) suggests this origin for the genus. The remaining two species of V/erneckiella both parasitise Ammo tragus lervia (Capra lervia of Gentry, 1978). Although the two lice are not indicated as sister-species on the cladogram, they are very close 515

morphologically, the females apparently being indistinguishable. The association of Werneckiella with Ammo tragus may be primary and that with

Squus secondary, or vice-versaf Ammotragus being infested once or twice from horses. Gentry (1978) refers to late Pleistocene fossils of Ammotragus lervia in North Africa, and Newbiggin (1936) suggests a Central Asian origin for the species. A "North American origin for Ammotragus seems unlikely, therefore, so the hypothesis of an infestation of horses from this host is rejected. Far more likely is that Ammotragus acquired Werneckiella once or twice from a more recent equid in North Africa where it is, or has been, sympatric with several species (Ansell, 1971).

4.4.2.4. Bovicolinae

The host associations of two of the five genera of Bovicolinae have not so far been, discussed. Genus n. 2 comprises one species (with two subspecies) parasitic on the bison and buffalo (Bovidae: Bovini), and Genus n. 3 comprises a single species parasitic on the mouse deer (Tragulidae). Both of these associations must be considered primary for the modern host species, but the poor resolution of the cladogram does not allow any statement regarding the acquisition of the parasites by the ancestors of the modern hosts. • The primary associations of the major clades of the subfamily (including each subgenus of .Damalinia) thus comprise the Bovidae, Tragulidae and Equidae. The identity of the host of the ancestor of the subfamily is not clarified by the cladogram because of poor resolution. It is clear that there is an early association with the Bovidae, although it is likely that some secondary infestation within the family (other than that already discussed above) has taken place. If the association of Genus n. 3 with Tragulidae is primary, the ancestor of this genus and the rest of the Bovicolinae (with the possible exception of Werneckiella) may have been with the ancestor of the Ruminantia (Fig.266). If this is the case, lice must be secondarily absent from the Cervidae, Moschidae, Giraffidae and Antilocapridae. If the association between Werneckiella and Equidae is primary, an infestation of the common ancestor of the

Perissodactyla and Artiodactyla might be postulated, in which case lice

V 516

are secondarily absent from other Perissodactyla and the artiodactyls Suidae, Tayassuidae, Hippopotamidae and Camelidae as well as those mentioned above. The sister-group of the Perissodactyla has been held to be the Hyracoidea (McKenna, 1975; Eisenberg, 1981) and, if the above hypothesis is correct, the association of Trichodectidae with the hyraxes must be secondary. This problem is discussed further below, but no conclusion is reached as to the identity of the ancestral host of the Bovicolinae.

4.4.3« Eut ri cho philinae

The single genus is parasitic on new-world procupines (Erethizontidae: Rodentia) • The origin of this association will be discussed further below. No attempt has been made in this study to resolve the relationships of the lice in this subfamily, and therefore little comment can be made. The presence of the same three species of Butrichophilus on Coendou prehensilis and soinosus is notable.

4.4.4. Dasyonyginae

The initial dichotomy of the clade divides Cebidicola. parasitic on Cebidae (New World Primates) from the hyrax lice of the Old World. No phylogenetic implications for the hosts can be drawn from this, and it must be that one or both infestations are secondary. This will -be discussed below. The uncertainties attached to the taxonomy and systematics of Procaviidae (Hyracoidea) and the dubious nature of some of the host identifications for the lice preclude any discussion of host - parasite coevolution in this group, although with one exception secondary infestations, if there have been any, have been confined to the one host group. The single exception is the presence of JP. (Meganarion- oides) colobi on a monkey (see discussion in section 3.3.4.3.) • This association must be secondary, as all other members of the clade (Procavicola - Eurytrichodectes clade) are parasites of hyraxes. viverrinae

hemigalinae viverridae ttrr galidiinae herpestidae cryptoproctidae protelinae hyaenidae hyaeninae ] felidae caninae 1 ganidae otocyoninae ursidae otariidae ailuridae ailuropodidae procyonidae mustelinae mellivorinae

melinae mustelidae mephitinae lutrinae phocidae

Pig. 276. Phylogeny of the living Carnivora. Compiled, from Tedford (1976), Eisenberg (1981) and Plynn & Galiano (1982). VJ1 —x -J 518

4.4.5* Trichodectinae

Almost all Trichodectinae parasitise Carnivora so, for comparative purposes, a cladogram of the living carnivoran families is presented in

Fig. 276.

4.4.5.1. Trichodectes

Species of this genus parasitise the canoid families Ursidae, Canidae, Procyonidae and Mustelidae, the predominant association being with the latter (Figs 277, 278). Although all these families are members of a monophyletic (paraphyletic) group and thus the association might be assumed to have resulted from a single primary infestation, 'the distribution of the host taxa on the cladogram of the lice indicates that some secondary infestation must have taken place. The three species parasitising Procyonidae comprise, a single clade, sister-group to some of the musteline lice. The Procyonidae are not the sister-group of the Mustelinae, but of the Mustelidae plus Phocidae (Tedford, 1976). The distribution on the cladogram of other lice parasitising Mustelidae suggests that the association with Procyonidae is secondary. The sister-species relationship between the two species parasitising Proc yon is suggestive of a primary infestation, although the relationship between the two raccoons is not known. Further collecting from other species of Procyon, coupled with a cladistic analysis of the genus, would ena.ble the degree of phylogenetic association to be determined. Without such collecting and analysis the status of the association with Potus is difficult to assess. Simpson (1945) suggests that this genus is an early offshoot from the line which gave rise to other Procyonidae but, if this were so and the infestation with Trichodectes were primary, the other genera of Procyonidae might be expected to be parasitised by the same genus also. A resolution of this problem must await further collecting and a cladistic analysis of procyonid relationships. The remainder of Trichodectes (Stachiella) is associated with mustelines of the closely-related genera Martes and Mustela. If Martes is holophyletic, however, a secondary infestation of Must el a from this genus must be inferred. •T. (T.) p. plnguis Ursus arctos Ursldae

•T. (T.) p. euarctldos Ursus amerlcanus Ursldae Canls 'famlllarls1 etc. •T. (T.) canls Canldae, Vlverrldae (full list ln Table XIV)

•T. (T.) mells Meles oeles Mustelidae: Mellnae

• T. (T.) vosseleri Mellivora capensls Mustelldae: Melllvorlnae

•T. (T.) emersonl Melogale personata orientalls Mustelidae: Mellnae

•T. (T.) kuntzi Melogale moschata Mustelldae: Mellnae

T. (T.) gallctldis Gallctls vlttata, G. cuja Mustelidae: Mustellnae

T. (T.) undescrlbed Grlsonella furax Mustelldae: Mustellnae

T. (5 ) ovalls Ictonyx striatus Mustelldae: Mustellnae

T. (5 ) ugandensls Poecllogale alb1nucha Mustelldae: Mustellnae

T. (5 ) zorlllae Poeclllctls lyblca Mustelldae: Mustellnae

T. (Stachlella) - full details on Fig. 278. Mustelldae, Procyonldae

•W. ferrlsi Tremarctos ornatus Ursldae

k lutrae Pteronura braslllensls Mustelldae: Lutrlnae

L. exilis Lutra lutra Mustelldae: Lutrlnae

L. matschlel Lutra macullcollls, Aonyx conglca Mustelldae: Lutrlnae

P. lntennedius Proteles crlstatus Hyaenldae

P. hyaenae Hyaena brunnea Hyaenldae

Fig. 277. Cladogram and host associations of Trichodectini• VJl• vo T. (s. emeryl Martes flavlgula Mustelldae: Mustellnae

T. (S. r. martIs Martes americana Mustelidae: Mustellnae

T. (S. r. salfll Martes martes Mustelldae: Mustellnae

T. (S. r# retusus Mustela foina Mustelldae: Mustellnae

T. (S. jacobl Martes putorlus Mustelldae: Mustellnae

T. (S. ermlnlae Mustela ermlnea Mustelldae: Mustellnae

T. (S. klngl Mustela nivalis Mustelldae: Mustellnae

T. (S. larsenl Mustela vison Mustelldae: Mustelinae

T. (S. mustelae Mustela nivalis, M, slblrlca Mustelldae: Mustellnae

T. (S. fallax Procyon cancrlvorus Procyonldae

T. (S. octomaculatus Procyon lotor Procyonldae

T. (S. potus Potos flavus Procyonldae

Cladogram and host associations of Trichodectes (Stachi'ella). 521

The three species of jF. (n.5) a*"6 associated with three closely- related African Mustelinae, and may have co-evolved with them.

The host associations of the species of T>. (Trichodectes) are rather more confusing. One species, jr. canis, is parasitic on a number of Canidae and a single species of Viverridae; there is no doubt that the latter infestation is secondary. All but one other species of jP. (Trichodectes) are parasitic on Mustelidae, and it is therefore probable that the association with Canidae is secondary. The same argument holds for the species parasitic on Ursidae, although the existence of two host- related (and morphologically distinct) subspecies suggests that infestation was initially with the common ancestor of the bears concerned. The possible primary association of the jT. pinguis - vosseleri clade lies therefore with the Mustelidae, either with the Melinae or the Mellivorinae. The other two clades of jF. (Trichodectes) are associated with Melinae and Galictinae (the latter sometimes incorporated into Mustelinae) respectively, but the poor resolution of the cladogram at this level makes conjecture difficult. The association with the South American Galictis and Grisonella is anomalous, as most of the lice in the subgenus (excepting only jP. j). euarctidos) are Old World, and Galictinae may be more closely- related to the hosts of jP. (n. 5) than other Mustelinae. The relationship of Melinae and the monobasic Mellivorinae is not known.

The host of the ancestor of Trichodectes was therefore almost certainly a mustelid, although possibly not a member of one of the modern subfamilies.

4.4.5.2. Trichodectini (Figs 277, 278)

Members of this tribe are associated with Mustelidae, Ursidae and Hyaenidae (and secondarily, as demonstrated above, with Canidae, Procyonidae and other Ursidae). The position on the cladogram of Werneckodectes, the single species of which is parasitic on a South American bear, is anomalous, and the association with an ursid must be assumed to be secondary. The primary host association of the Trichodectes - Lutridia clade is therefore with the Mustelidae. The cladistic position of Lutridia and genus n. 4 is confusing, and suggests that either the 522

Lutrinae is paraphyletic with respect to the rest of the Mustelidae, or that a secondary infestation of Lutra or Pteronura has taken place. The association of Lutridia matschiei with both Lutra and Aonyx suggests a secondary infestation of the latter genus from the former, as the two otter species concerned are at least partially sympatric, and the sister- species of matschiei is also associated with Lutra. Collection from other species of otter may clarify the situation. The Trichodectes - Lutridia clade, therefore, is believed here to have a primary association with Mustelidae, and secondary associations with Ursidae in the Neotropical and Holarctic Regions, Procyonidae in the Neotropical Region, and Canidae at an unknown site. The association with Mustelidae has proceeded partially by co-evolution, but some secondary infestation must be invoked to explain discrepancies in the association. The cladistic relationships of some species of (Trichodectes) are difficult to reconcile with the relationships and distributions of the hosts. The single genus of the Trichodectini not yet considered, Protelicola, has been recorded from two of the four living species of Hyaenidae.. The discovery of further louse species on the other hyaenids and an analysis of their relationships is necessary before any comment on co-evolution can be made. The significance of the host association of this clade. with hyaenids will be discussed below.

4-4.5.3. Felicola

Species of this genus parasitise Herpestidae, Viverridae, Felidae and Canidae, most species being associated with members of the first family listed (Fig. 279). The parasites of Canidae form a single clade in F. (Suricatoecus). and the most parsimonious explanation of the association is of a single infestation of a canid from a herpestid. The phylogenetic relationships of the species of canid concerned are difficult to determine. Otocyon is generally considered distinct from other Canidae (Van Gelder, 1978; Eisenberg, 1981), but the other canids infested may all be placed in a single genus (Van Gelder, 1978). The typological and phenetic basis of •F. (F. minimus Atllax paludlnosus Herpestldae •F. (F. rahml At 1lax paludlnosus Herpestldae •F. (F. calogaleus Herpestes sanguineus, pulverulentus Herpestldae •F. (F. zeylonicus Herpestes vlttlcollls Herpestldae •F. (F. rohani Herpestes auropunctatus, edwardsi, Javanlcus, urva Herpestidae •F. (F. lnaequalls Herpestes Ichneumon Herpestldae •F. (F. vlverriculae Viverrlcula Indica Vlverrldae •F. (F. cynictls Cynictls penlclllata Herpestldae -F. (F. setosus Paracynlctls selousl Herpestldae •F. (F. liberlae Llberllctls kuhnl Herpestldae •F. (F. robertsl Rhynchogale mrlleri Herpestldae •F. (F. hopkinsi Nandlnla blnotata Vlverrldae Fells catus, Ichneumla alblcauda etc. Herpestidae, F. (F. subrostratus (full list in Table XIV) •F. (F. congoensls Crossarchus alexandrl Herpestldae F. (F. helogaloldls Helogale p. undulata Herpestidae •F. (F. helogale Helogale p. parvula Herpestldae F. (F. occidentalis Crossarchus obscurus Herpestldae •F. (S. becifordi Bdeogale crass1cauda, Jacksonl, nigrlpes Herpestldae F. (S. acutlrostrls Atllax paludlnosus Herpestldae F. (s. macrurus Atllax paludlnosus Herpestldae F. (S. pygldialls Atllax paludinosus Herpestidae •F. (S. cooleyl Surlcata surlcatta Herpestldae F. (s. dec!piens Mangos mungo Herpestldae F. (s. fahrenholzi Dusicyon grlseus, thous Canldae F. (S. guinlel Orocyon megalotls Canldae F. (S. vulpls Vulpes vulpes Canldae VJ1 •F. (s. quadratlceps Vulpes velox, Urocyon clnerorgentatus Canldae oroj Fig. 279* Cladogram and host associations of Felicola. 524

canid systematics prohibits useful discussion at this level. The positions on the cladogram of species parasitising Viverridae (P. viverriculae, ]?. hopkinsi and P. subrostratus) suggest that these associations are secondary, resulting from two or three secondary infestations of viverrids from herpestids. A further species of Pelicola, an undescribed member of the P. zeylonicus - viverriculae clade, parasitises the viverrid Paradoxurus hermaphro dytus. P arado xuru s.- and Viverricula (the host of P. viverriculae) are not closely related and, even if _F. viverriculae and the undescribed species are found to be sister-species, a single primary infestation of Viverridae by the common ancestor of these lice seems unlikely. The status of the associations with Viverfidae of P. hopkinsi and its sister-species, P. subrostratus, are discussed below. Pelicola subrostratus is the only species of the genus to parasitise Pelidae, and its position on the cladogram indicates that this association is secondary. The host range of this louse includes members of the Pelidae, Viverridae, and Herpestidae (Pig. 279), however, and the identity of the primary host is not certain. The sister-species of P. subro stratus parasitises the viverrid Nandinia binotata. This species is not closely related to either of the two viverrid hosts of P. subro stratus, and these two are not closely related to each other (Gregory & Hellman, 1939). Hopkins (1949) identifies the association of P. subrostratus with one of these, Civettictis civetta, as secondary on personal observation. Unless secondary absence from a large number of other Viverridae is postulated, a primary association by the common ancestor of P. subrostratus and ]?. hopkinsi with the ancestor of Nandinia and Eupleres (the remaining viverrid host of ]?. subrostratus) seems unlikely. The association of each of these lice is therefore believed to result from independent secondary infestations of the present viverrid hosts. The two remaining hosts of 3?. subro stratus are the herpestids Ichneumia albicauda and Salanoia concolor, either of which may be identified as the primary host of the louse. Salanoia is placed in the Madagascar subfamily Galidictinae (= Galidiinae) (Petter, 1962), from which no other lice are known, whereas Ichneumia is African and more closely related to the other genera 525

parasitised by Felicola. Further, Ichneumia is frequently domesticated whereas Salanoia is not (Walker, 1964) t therefore the chances of its lice coming into contact with potential hosts, especially other domestic animals such as the cat, are enhanced. For these reasons Ichneumia is identified as the primary host of I?, subrostratus. The remaining species of Felicola are associated with Herpestidae, and most genera of the family have been recorded as hosts to lice of this genus. There is a clear inference that the association of Felicola and Herpestidae is primary. Despite that, there are indications of secondary infestation. The F. congoensis - occidentalis clade is associated with Crossarchus and Helogale but, if both of these genera are considered to be holophyletic, at least one secondary infestation must be postulated (Fig. 279). The most parsimonious hypothesis is that the association of F. congo ensis and Cross archus is secondary, although the existence of an underscribed sister-species to either-F. congoensis or j?. helogale and parasitic on Crossarchus obscurus poses problems. Two of the jF. rahmi - viverriculae clade have been identified above as having a secondary association with Viverridae. Apart from these, one species is parasitic on Atilax paludinosus and the others on Herpestes species. The distribution of species of this clade on the cladogram suggests a primary association with Herpestes. and therefore a secondary infestation of Atilax might be suspected. However, Herpestes may be paraphyletic with respect to Atilax (Gregory & Hellman, 1939), and the association therefore primary. Additional evidence suggesting a secondary infestation of Atilax paludinosus from Herpestes sanguineus is discussed below under Lorisicola. At least two secondary infestations of Atilax paludinosus with Felicola must be postulated, as lice associated with this species appear at three points on the cladogram (Fig. 279). The lack of a cladistic treatment of the Herpestidae precludes further comment.

4.4.5.4. Lorisicola

Species of this genus are associated with Viverridae, Felidae, Herpestidae and Primates (Fig.280). The distribution of host associations on the cladogram suggests clearly that the last two families named are L. (P. bengalensis Paradoxurus henna phrodyt us Viverrldae L. (P. phlllpplnensls Paradoxurus phllipplnensls Vlverrldae L. (P. Juccii Paguma larvata Vlverridae L. (P. aspldorhynchus Prionodon llnsang Vlverrldae L. (P. sumatrensls Prlonodon llnsang Vlverrldae L. (P. lenlcomls Genetta tlgrlna, G. vlctorlae Vlverrldae L. (P. werneckl Genetta tlgrlna, G. thlerryi Viverrldae L. (P. acutlceps Genetta tlgrlna, 0. genetta, G, abysslnica Vlverrldae L. (P. afrlcanus Genetta genetta Vlverrldae L. (P. neoafrlcanus Genetta tlgrina Vlverrldae L. (P. paralatlceps At1lax paludlnosus Herpestldae L. (P. latlceps At1lax paludlnosus Herpestldae L. (P. mangos Herpestes sanguineus Herpestldae L. (L. malayslanus cynogale bennettl Vlverrldae L. (L. mjobergl Nyctlcebus coucang Lorlsldae L. (L. amerlcanus Fells rufa Felldae L. (L. brazl11ensis Fells colocola Felldae L. (L. caffra Fells lyblca Felldae L. (L. fells Fells pardalls, ? F. concolor Felldae L. (L. neofells Fells georfroyl Felldae L. (L. slmllls Fells yagouaroundl Felldae L. (U spencerl Fells canadensis Felidae L. (L. sudamerlcanus Fells tlgrlna pardlnoldes Felldae L. (L. hercynlanus Fells sylvestris Felldae L. (L. slamensls Fells bengalensis Felldae

vji Fig. 280. Cladogram and host associations of Lorisicqla. ro ON 527

associated with Lorisicola following secondary infestations. The association with a probably took place in South East Asia, as both the louse concerned and its sister-species are distributed there, and that with Herpestidae in Africa as, again, both the clade concerned and its sister-group are sympatric in that continent. Further comment must be made on the three species parasitic on the Herpestidae, the paralaticeps - mungos clade. Two of them are parasitic on Atilax paludinosus and the third on Herpestes sanguineus, their postulated cladistic relationship suggesting an ancestral association with the former host. Herpestes may be paraphyletic with respect to Atilax. as noted above, but this relationship is not compatible with the cladistic relationships of the species in the L. paralaticeps - mungos clade. A secondary infestation of Herpestes sanguineus from Atilax paludinosus is the most parsimonious hypothesis to explain the distribution of lice on the hosts. This being the case, the associations of the Felicola rahmi - viverriculae clade may be re-examined. In this clade also there is a sister-species relationship between lice parasitic on Atilax paludinosus and Herpestes sanguineus. Although the possibility of the association of F. rahmi and Atilax being primary due to the paraphyly of Herpestes cannot be dismissed, the identification of a transfer of lice from Atilax paludinosus to Herpestes sanguineus in Lorisicola must increase the likelihood attached to the proposition of a similar transfer, in the reverse direction, taking place in Felicola.

The association of the bengalensis - nuccii clade with their hosts may have involved co-evolution, as Paguma and Paradoxurus are sometimes placed together in the subfamily Paradoxurinae (Gregory & Hellman, 1939; Honacki, Kinman & Koeppl, 1982). Nandinia, parasitised by Felicola hopkinsi, is also placed in the same subfamily, as are several genera not known to be parasitised at all. In these cases either secondary absence or failure in collection may be suggested to explain the apparent absence of Lorisicola. However, an initial secondary infestation of the common ancestor of the two genera parasitised (which may, with Arctitis. form a holo- phyletic group), is also possible. Notably, Paradoxurus is also 528

parasitised by an undescribed Felicola (F. inaeaualis - zeylonicus clade), so multiple infestation is quite possible in the Viverridae. The identity of the primary host of Lorisicola is not certain, even though the associations with Primates and Herpestidae have been identified as secondary. A single species of viverrid and eleven species of felid are known to be parasitised by species of L. (Lorisicola), the lice associated with the two families being (if L. m.jobergi on the loris is ignored) sister-groups. If the Viverridae are assumed to be holophyletic, or at least not paraphyletic with respect to the Felidae, a secondary infestation of either Viverridae or Felidae must be postulated if the cladistic relationships of Lorisicola species are accepted as postulated. The cladistic relationship between Felidae and Viverridae is not certain (Fig.276), but the association of Lorisicola with the two families could be primary (with one secondary infestation of Viverridae); even if the two families form a paraphyletic group, secondary absence from other families may be invoked to preserve the hypothesis of primary infestation. The relatively large number of Felidae parasitised, and their wide

distributions, support the hypothesis of an early association of lice with members of this family, although all of the species conerned are in the genus Felis; further species of lice collected from the other genera of Felidae will strengthen the proposition. The cladistic position of the viverrid Cynogale (the louse of which is sister-group to the lice of Felis) is not known, but Gregory & Hellman (1939) suggest that it lies in the 'Hemigalida*, which they put in a sister-group relationship with the 'Paradoxurida* (including Paradoxurinae but not Genetta or Prionodon); they also imply a closer phylogenetic relationship between Genetta and Prionodon than between either of these and the Hemigalida and Paradoxurida. If the relationships suggested by Gregory & Hellman are correct, the association of Lorisicola with Genetta and Prionodon can be considered primary, dating to their common ancestor, and two alternative hypotheses proposed to explain the host distribution of Lorisicola: i) The association of Lorisicola with the Hemigalida plus Paradoxurida dates to their common ancestor, the lice being derived either from their common ancestor with other Viverridae or by secondary infestation from 529

the Herpestidae. If the lice are primary for the Viverridae, they must be secondarily absent from groups other than the two named. The common ancestor of Genetta and Prionodon was colonised by lice from the Para- doxurida and the ancestor of Felis by lice from the Hemigalida. Lice of the genus Lorisicola will be found on other Viverridae and possibly on other Felidae; if the association of Felicolini with Feloidea is primary, Herpestidae and Viverridae may be found to be cladistically closer to each other than either is to Felidae. ii) The association of Lorisicola with the common ancestor of Viverridae and Felidae is primary, but the lice primarily present on Cynogale have not yet been discovered, L. malaysianus being a secondary acquisition from the Felidae. Lice will be found on other Viverridae and Felidae; Felidae will be found to be cladistically closer to Viverridae than either of these families is to Herpestidae. The resolution of this problem awaits further collecting and a cladistic analysis of the Feloidea.

4«4.5«5. Trichodectinae

The association of the ancestor of the Felicolini with the common ancestor of the Herpestidae and the Viverridae, or possibly with that of the Herpestidae, Viverridae and Felidae, seems likely. Further collecting is needed to determine the presence of Felicolini on Cryptoproctidae. The primary host of the Trichodectes - Lutridia clade was an ancestor of some or all of the Mustelidae (Canoidea), but its sister-group (Protelicola) is parasitic on the Hyaenidae (Feloidea). Allowing that any primary association of Trichodectidae with Otariidae and Phocidae would have been lost on the assumption by these mammals of an aquatic life, no family of terrestrial Canoidea other than the Mustelidae (i.e. the Ursidae, Procyonidae, Ailuridae, Ailuropodidae and Canidae) is known to have a primary association with Trichodectidae. If the hypothesis is made of a primary association between the ancestor of the Trichodectinae and that of the Caraivora, the association between Hyaenidae and Protelicola must be deemed to have resulted from a secondary infestation, and lice must be secondarily absent from the families named above. The alternative hypothesis is that the Trichodectinae are primarily associated with the 530

Feloidea, and became associated with the ancestor of the Mustelidae by secondary infestation from the ancestor of the Hyaenidae; no further secondary absences need be postulated. The latter hypothesis is favoured on the grounds of parsimony.

4.4.6. Subfamily n.

4.4.6.1. Geomydoecus

Geomydoecus is restricted to rodents of the family Geomyidae. As stated above, the cladistic relationships between most members of the genus were not analysed in this study, and those between the hosts are largely (or completely) unknown. The distributional information available from publications on Geomydoecus by Price and his associates (e.g. Fig. 263) • indicates that there has been repeated secondary infestation between members of the Geomyidae, and the same authors note several instances of mustelid predators being found with straggling specimens of Geomydoecus.

The only other rodents parasitised by Trichodectidae are Erethizontidae, which are associated with the genus Eutrichophilus. The Geomyidae and Erethizontidae are not closely related (Simpson, 1945; Eisenberg, 1981), but Hopkins (1949, 1957) suggests that their association with Trichodectidae is primary, and that these lice are secondarily absent from other Rodent families. The position of Geomydoecus and Eutrichophilus on the cladogram does not support this hypothesis, and parsimony suggests that the two associations result from separate secondary infestations of rodents from other mammals.

4.4.6.2. Neotrichodectes

Species of Neotrichodectes parasitise Mustelidae and Procyonidae (Camivora) and Bradypodidae (Edentata) in the New World (Fig.281). The two species parasitic on sloths (subgenus Lakshminarayanella) are the only lice known from Edentata, and their position on the cladogram makes it virtually certain that lice have been secondarily acquired by sloths from a member of the Camivora. Speciat ion within N. (Lak shminarayanell a) may have proceeded according to Fahrehholz1 Rule, although if the Rule has been fully adhered to, further species of the subgenus must be Geomydoecus Geomyldae Rodentla: Geomyldae

N. (N.) thoraclcus Bassarlscus astutus Procyonldae

N. (N.) mlnutus Mustela frenata, M. nigrlpes Mustelldae: Mustellnae

N. (N.) osbornl Spilogale putorlus Mustelldae: Mephltlnae

N. (N.) roephltldls Mephitis macroura, M. mephitis Mustelidae: Mephltlnae

N. (N.) wolffhuegell Conepatus chlnga Mustelldae: Mephltlnae

N. (T.) barbarae Elra barbara Mustelldae: Mustellnae

N. (6 ) pallldus Nasua narlca Procyonldae

N.

N. (L.) cummlngsi Bradypus trldactylus Edentata: Bradypodldae

N. (7 ) 1nterruptofasc1atus Taxidea taxus Mustelldae: Mellnae

N. (7 ) chilensls conepatus chlnga, C. humboldti Mustelidae: Mephitlnae

N. (7 ) semistrlacus Conepatus semlstriatus Mustelldae: Mephltlnae

N. (7 ) arizonae Conepatus leuconotus, C. mesoleucus Mustelldae: Mephitlnae

Cladogram and host associations of Subfamily n. 532

expected from the other members of Bradypus and Choloepus. The supposed similarity of the sloth lice to hyrax lice, particularly Procavicola (Hopkins, 1949; Vanzolini & Guimaraes, 1955; Eichler, 1963) does not reflect any phylogenetic relationship, and the suggestion that the association of sloths and Trichodectidae is "primary and ancient" (Vanzolini & Guimaraes, 1955) is unfounded. The remaining eleven species of Neotrichodectes are associated with members of the Procyonidae and Mustelidae, and within the latter family to members of the Mustelinae, Melinae and Mephitinae. Both families are also parasitised by Trichodectes, although the association of this genus with Procyonidae has been shown to be secondary. The positions on the cladogram of the two species parasitic on Procyonidae suggests these associations to be the result of independent secondary infestations from Mustelidae. This suggestion is supported by the lack of any close relationship between the procyonids (Simpson, 1945). The position on the cladogram of the single species parasitising a badger (Melinae), which is sister-species to a parasite of the mephitine genus Conepatus, suggests a secondary infestation from Conepatus to Taxidea to account for the association (Fig.281). The two remaining subfamilies parasitised, Mephitinae and Mustelinae, both appear on the two branches of the primary dichotomy of the Neotricho- dectes cladogram, so some secondary infestation is almost certain. The number of secondary infestations is minimised by postulating that the lice were primarily parasitic on members of the Mephitinae, transferring to Mustelinae twice, and possibly once between different members of the Mephitinae (to account for the host association of N. wolffhuegeli). If Conepatus is paraphyletic with respect to a holophyletic group comprising the other two skunk genera, no secondary transfer between members of the Mephitinae need be postulated at all. If Neotrichodectes is primarily parasitic on Mustelinae, four secondary infestations are necessary to explain the distribution of host associations on the cladogram. The two mustelines parasitised are not closely related, Eira being more closely related to some African genera than it is to Mustela (Ray, Anderson & Y/ebb, 1981), but the mephitine genera probably fonn a holophyletic group, restricted to the New Y/orld. The identification of the Mephitinae • EDENTATA* 'LEPICTIMORPHA •LAGOMORPHA RODENTIA* G/iRNIVORA* • IIISECTIVORA •SCANBENTIA •DERMOPTERA CHIROPTERA PRIMATES* TUEULIDENTATA ARTIODACTYLA* CETACEA PERISSODACTYLA* HYRACOIDEA* PROBOSCIDEA SIRENIA

Fig. 282. Cladogram of the Eutheria (after McKenna, 1975 and Esenherg, 1981)• Groups with members parasitised by Trichodectidae are indicated by an asterisk (*). 534

as the probable ancestral hosts of Neotrichodectes is consistent with the geographical distribution and probable origin of the Geomydoecus - Neotrichodectes clade.

4.4.7. Trichodectidae

A cladogram of the Eutheria is depicted in Fig.282 for comparative purposes. The primary associations postulated above for each major clade of the Trichodectidae are depicted in Fig. 283 and, as can be seen, are not immediately compatible with the host cladogram. The associations of Trichodectidae with Primates have, apart from the three species of Cebidicola, been identified as secondary. If this remaining association is an indication of a primary association with the Order, secondary absence of the lice on a large scale must be postulated. For this reason, the association is believed to be primary only for the family Cebidae, although the sister-group relationship between the Central American cebid lice and the African hyrax lice is very difficult to account for and the primary host association of the Dasyonyginae is not identified. The two families of rodents parasitised by Trichodectidae are in different suborders and thus, if these are believed to be relict associations of a primary association of the Rodentia, a large number of secondary absences must be postulated. The hypothesis of independent secondary infestations of Geomyidae and Erethizontidae (or their respective ancestors) is considered more parsimonious. The primary host association of the Geomydoecus - Neotrichodectes. clade may therefore have been a geomyid or a mephitine. The primary host association of the sister-group to the clade, the Trichodectinae, has been identified as either the ancestor of the Feloidea or the Carnivora. If the primary host association of the Subfamily n. is postulated to be with Mephitinae or even Mustelidae, and is believed to be derived with that of the Trichodectinae from a common ancestral host, this ancestral host must have been a member of the Canoidea. As the primary host association of the Trichodectes - Lutridia clade has been identified as a Mustelid, the most parsimonious identification Geomydoecus Rodentia: Geornyidae

ITeotrichodeotes Carnivora: Mephitinae

Trichodectini Carnivora: Feloidea/ Mustelidae

Felicolini Carnivora: Feloidea

Eurytrichodectes — Hyracoidea Procavicola clade

Cebidicola Primates: Cebidae

Eutrichophilinae Rodentia: Erethizontidae

Bovicola Artiodactyla: Bovidae (part)

Genus n. 2 Artiodactyla: Bovidae (part)

Damalinia Artiodactyla: Bovidae (part)

Genus n. 3 Artiodactyla: Tragulidae

Uerneckiella Perissodactyla: Equidae

Fig. 283. Primary host associations of the major clades of Trichodectidae,

VJl UJ KJ\ 536

of the common ancestral host must be as a Mustelid. This hypothesis demands that infestation of the Feloidea took place twice from a Mustelid, once to give rise to the Felicolini and once to give rise to Protelicola. Considering that the primary host association of the Felicolini is with a postulated ancestor of most of the Feloidea, and that Feloidea is the sister-group of Canoidea, the additional hypothesis must be made that the Mustelidae are the sister-group of most of the rest of the Canoidea. This is necessary to allow the Mustelidae to be distinct from the other Canoidea at an early enough period to provide a source of lice to the ancestor of the Herpestidae, Viverridae and Felidae. This hypothesis is not supported by the cladogram of Camivora (Fig.276). The alternative hypothesis is that the two clades of lice parasitising Mustelidae became associated with their hosts independently, once as described from the Feloidea, the other from an unknown host. It is notable that the plesio- morphic form of the female external genitalia in ITeotrichodectes and Geomydoecus is found in most species of Geomydoecus but very few of Neotrichodectes. If the-form of the female genitalia is determined at least partially by the form of the host hair on which the egg must be

cemented, as seems possible, the assumption can be made that the form of the female genitalia found in numerous species of geomyid lice is adapted to oviposition on Geomyidae, and that selection pressure on lice of other hosts (of Neotrichodectes) has caused modification to this (plesiomorphic) form. If the plesiomorphic form of the female genitalia is an adaptation to oviposition on geomyid rodents, it follows that the primary host of the Subfamily n. was a geomyid or an ancestor of that family, from which the ancestor of Neotrichodectes was acquired by an ancestor of the Mephitinae. The primary host association of the Bovicolinae is not readily apparent, but may lie with the Bovidae, Ruminantia or a common ancestor of the Perissodactyla and the Artiodactyla. If the latter hypothesis is correct, the association with,Hyracoidea must be secondary for the cladogram of lice to be compatible with the cladogram of hosts, since the Hyracoidea is held to be the sister-group of the Perissodactyla. If it is not correct, and the primary association of Bovicolinae lies within the ancestry of the Artiodactyla, the association of lice with Hyracoidea can 537

only be primary if that with Perissodactyla was secondarily lost (and replaced by a secondary infestation with Werneckiella from an artiodactyl). The apparent restriction of Trichodectidae to Equus does not suggest a primary association with Perissodactyla, although the environment provided by Rhinocerotidae may not be suitable for Trichodectidae. Specimens of the other family of extant perissodactyl, the Tapiridae, have been examined for lice without success (Hopkins, 1949), but further examination of these hosts must be made to provide more evidence of the status of the association between Trichodectidae and Perissodactyla. The poor resolution of the cladogram of Bovicolinae prohibits further hypotheses, but it is clear that either secondary absence or secondary infestation must be invoked to explain the present associations of Trichodectidae with Artiodactyla, Perissodactyla and Hyracoidea, or the structure of the host or parasite cladograms must be changed. No clear hypothesis can yet be proposed regarding the ancestral host association of the Trichodectidae, other than that it was with an eutherian and probably not an insectivore, primate, dermopteran, chiropteran or tupaiid. The primary host association of Subfam. n. was probably with a rodent in America, from which a predatory ancestor of the Mephitinae acquired the ancestor of Neotrichodectes. The primary host association of the Trichodectinae was with the ancestor of the Feloidea; either the Canoidea were also parasitised but lost their Trichodectidae, or the Feloidea acquired their lice, as did the Mephitinae, through predation on another animal. If the latter hypothesis is correct, this prey animal may be postulated also to be the rodent ancestor of the Geomyidae. The hypothesis of independent secondary infestation of this geomyid ancestor and of the Erithezontidae is maintained, but the source of these infestations cannot be identified. The large number of extinct Hyracoidea may have included these hosts, but it is difficult to see how this hypothesis can be tested. The Cebidae (or ancestor of the Platyrrhin monkeys) may have acquired their lice from a hyrax just as has the colobus more recently, and must have done this in Africa, as no fossil record of the hyraxes exists outside this continent. The primary host association of the Bovicolinae is still more confusing. Although the strongest 538

evidence is perhaps for a primary association with an ancestor of the Bovidae, the earliest fossil occurrence of this family is in the Miocene (Patterson, 1957), later than the most recent common ancestor of the Peloidea (Plynn & G-aliano, 1982) and thus after the initial dichotomy of the cladogram of the Trichodectidae. An earlier host of the Bovicolinae, and thus some secondary absence, must be postulated, although the identity of this host is not known. Future work to resolve these problems must involve: (i) Further collection of lice, with particular emphasis on Artiodactyla, Perissodactyla and Carnivora. (ii) Further investigation of the cladistic relationships of the Tricho- dectidae, particularly at the primary furcations of the cladogram proposed in this study > where apomorphies have been very difficult to discern, and in the Bovicolinae as here constituted. (iii) Production of a more fully resolved cladogram of the Eutheria, and the identification of holophyletic mammalian groups at all levels. section 5

summary 540

SUMMARY

The systematic (phylogenetic) position of the Phthiraptera with respect to other insects is investigated by examining the validity of autapomorphies proposed in the literature for Acercaria, Psocodea, Psocoptera and Phthiraptera. The Permopsocida are treated as an assemblage of species belonging to the stem-groups of Acercaria, Psocodea, Psocoptera and perhaps other acercarian Orders.

The distribution of available apomorphies indicates the holophyly of the Acercaria ( Fsocodea plus Thysanoptera and Hemiptera)• An alternative hypothesis is suggested by proposed homologies of the male genitalia of Psocodea and Thysanoptera (section 2.2.4.4.). This is that the Condylognatha (Thysanoptera plus Hemiptera) are the sister-group of the Holometabola, united by the common possession of an aedeagus (fused mesomeres), and the Psocodea are the sister-group of the Condylognatha plus Holometabola, linked by the common possession of an endophallus. The evidence may also be explained by convergence between Condylognatha and Holometabola, or character-state reversal in the Psocodea. The present data do not permit the selection of any one hypothesis as most parsimonious, and the traditional classification is retained.

The distribution of available apomorphies suggests the holophyly of the Psocodea (Psocoptera plus Phthiraptera), although the supposedly autapomorphic states of the ovarioles, cibarium, lacinia and male genitalia are rejected.

The distribution of available apomorphies suggests the holophyly of the Phthiraptera, although two supposed apomorphies (method of production of egg-cement and adoption of ectoparasitism) are rejected, and three (loss of wings, loss of ocelli and reduction of compound eyes) are believed to be synapomorphies of the Phthiraptera and Liposcelidae (Psocoptera)•

Sixteen autapomorphies proposed in the literature for the Psocoptera are examined and ten rejected. The remaining six are aspects of two structural novelties in the psocopteran egg. 541

6. Eleven synapomorphies of Liposcelidae and Phthiraptera are identified, and the apomorphies proposed by Smithers (1972) for Liposcelidae and associated Psocoptera briefly examined. Phthiraptera and Liposcelidae are identified.as sister-groups, but the placement of the Liposcelidae with respect to other Psocoptera is not challenged. The Psocoptera is therefore considered paraphyletic with respect to Phthiraptera.

7. Thirty-one supposed apomorphies within the Phthiraptera are examined and nine of these rejected as not indicating holophyletic groups. Anoplura, Rhyncophthirina, Amblycera and Trichodectidae are holo- phyletic groups; Anoplura and Rhyncophthirina are sister-groups, and Amblycera is the sister-group to all other lice. The Ischnocera, ' with or without the Trichodectidae, is not demonstrably holophyletic.

8. The ordinal name 'Mallophaga1 is rejected and all lice considered to be a single Order comprising four suborders, including the possibly paraphyletic Ischnocera.

9. Published and original observations on the morphology of Phthiraptera are examined v/ith particular reference to Trichodectidae. Problems associated with the major structural features are briefly reviewed and solutions suggested where possible; where appropriate the morphological terminology is clarified.

10. The antennae of many male Ischnocera and Anoplura are modified to clasp the female during copulation by expansion of the intrinsic and extrinsic scape musculature and development of various projections. In Trichodectidae tv/o apical flagellar setae are found to be greatly modified into tooth-like structures for this purpose. It is suggested that the reduction of the flagellum in male Trichodectidae to a single flagellomere distal to the pedicel increases the mechanical efficiency of the clasping system. The apex of the pedicel is elongate, and this is recognised as part of a mechanism for controlling the direction of movement of the antenna during clasping.

11. The mandibles of Trichodectidae have two functions: feeding and anchoring the louse to a hair. Trichodectidae at rest lock the 542

mandibles around a hair and release the hold of their legs. Ridges on the mandibles are not positioned in such a way that they can act in opposition to prepare food, but probably function as a frictional surface to increase the strength of the hold on a hair.

12. Various hypotheses of the function of the lingual sclerites, sitophore sclerite and cibarial pump in Phthiraptera are briefly reviewed. Rudolph (1982b) demonstrates that homologous structures in Psocoptera function in the uptake of atmospheric water vapour, and it is suggested that this is the function of the structures in Phthiraptera. This proposition has been confirmed since writing by Rudolph (1983).

13. The heavily-sclerotised internal bar on the second sternum of some Trichodectidae parasitic on hyraxes is identified as, in some cases, an apophysis developed from the sternite, and in others an enlargement of the metathoracic postcoxale. The failure on the part of previous workers to distinguish these two structures had led to the recognition of a polyphyletic assemblage of lice as a genus (Procavicola s. lat.).

14. The subgenital plate of female Trichodectidae is formed from sternite VIII or sometimes sternites VII and VIII. The lobulate projection on its posterior margin in some Trichodectidae is discussed and termed a 'subgenital lobe'. The presence and form- of this structure is of considerable value in systematic and taxonomic studies, although not greatly utilised in past works.

15. The homologies of the reduced elements of the phthirapteran ovipositor are established. The ' copulatory valve' or •gonapophysis' of Trichodectidae is recognised as gonapophysis VIII, and homologous structures located in Rhyncophthirina, Anoplura, and some members of the Amblycera and Ischnocera. The • gonapophysis' of Boopiidae (Amblycera) is identified as a modified paraproct, and homologous structures located in other Amblycera and some Ischnocera and Anoplura. Homologues of the gonoplac and gonapophysis IX cannot be identified with certainty in the Phthiraptera, but the 'genital lobe' of many Ischnocera, Amblycera and Anoplura may be homologous with one or both of these. 543

16. The post-genital sternal sclerites in female Boopiidae are homologised with the gonangulum and sternite IX+X of Psocoptera, but a similar homology cannot be made for the single 'post-vulval sclerite' of the Trichodectidae.

17. The common oviduct of Trichodectidae is continuous posteriorly with the genital chamber and branches anteriorly into the paired oviducts. The paired oviducts lie dorsal to the common oviduct which in turn lies dorsal to the genital chamber. The ventral wall of the genital chamber is frequently indistinct, and some workers have misidentified the dorsal wall as the ventral wall and the common oviduct as the dorsal wall ('genital sac* auctt.). A spermatheca is absent from Trichodectidae, except perhaps for Dasyonyx.

18. The opening of the male genital chamber in most Trichodectidae has migrated dorsally in response to a shortening and broadening of the abdomen. The stresses caused by curvature of the abdomen during copulation are considered and the optimum form of the tergal sclerites determined; this foim is found in some Trichodectidae.

19. The structure and component sclerites of the subgenital plate of male Trichodectidae are described for the first time. The plate comprises the sternites of segments' VII, VIII and IX linked by a pair of lateral rods, the two pleurites of segment IX, and the 'post-genital sternite', which is of uncertain homology. Any or all of these elements may be absent in individual species.

20. The two posterior extensions of the abdomen of some male Trichodectidae sometimes known as • styli' are not homologous v/ith true styli and a new term, 'pseudostyli', is coined.

21. The structures named 'abdominal lateral flecks' by Moreby (1978) in Werneckiella are found to be present in males of all species of Trichodectidae, but restricted to this family.

22. The structure of the male genitalia in Psocodea is discussed at some length and the homologies of the parts established for the first time. The genitalia comprise a basal apodeme, a pair of parameres,a pair of 544

mesomeres and an eversible endophallus. Despite statements to the contrary (Matsuda, 1976) a phallotheca is not present. An aedeagus (sensu Snodgrass) is also absent. The aedeagus is recognised as an apomorphic development of the Holometabola and, perhaps convergently, the Condylognatha. The homologies of the parts of the thysanopteran male genitalia are discussed and the structure identified as a phallotheca by Heming (1970) and Matsuda (1976) is re-identified as a true aedeagus.

23« The copulation of Trichodectes canis is described for the first time. The lice assumed a ' back-to-back* position, unexpected as other lice ' have been observed to copulate - with the male underneath the female. The parameres fonn a cone-shaped structure when the endophallus is not everted, and it is suggested that this protects the delicate endophallus in the initial stages of copulation from damage by the spines surrounding the vulva.

24. The posterior tracheal commisure, previously reported only in some Trichodectidae, is found in all members of the family examined in this study (93-5% of all species and subspecies).

25. The states of 279 characters are recorded for most species of Trichodectidae. One hundred and eighty-seven of these characters are used in a cladistic analysis of .the family at species level, many being employed for the first time in studies of lice. This is the first cladistic study of Phthiraptera below the subordinal level.

26. The 351 species and subspecies of Trichodectidae are re-classified according to the results of the cladistic analysis. Five subfamilies are used to partition the twenty genera employed, ten of the latter being sub-divided into twenty-seven subgenera. This necessitates the description of one new subfamily, three new genera and four new subgenera. Three genera are placed in synonymy, eight genera and subgenera are raised from synonymy, and four genera are reduced to subgenera. The generic placements of 106 species and subspecies sire changed. All genera and subgenera sire described. 545

27. A key is provided to subfamilies of Trichodectidae.

28. A key is provided to genera and subgenera of Trichodectidae, the first such since 1938, and the first to include all genera of Trichodectidae known at the time of writing.

29. Fahrenholz* Rule, predicting the topological identity of host and parasite phylogenies, is tested rigorously for the first time. Testing is by the identification of non-simultaneous speciation events of host and parasite, identification of secondary infestations, and by direct comparison of host and parasite phylogenies.

30. Fahrenholz' Rule necessitates the prediction that co-speciation of host and parasite must occur. Evidence suggesting unlinked host and parasite speciation is presented for a number of louse-host associations, falsifying this prediction of the Rule and hence the Rule itself.

31. Fahrenholz' Rule also necessitates the prediction that lice do not successfully colonise •new' host species, but are associated with a host species only because they evolved upon it. Examples of such •secondary infestation' by Trichodectidae and other groups of lice are cited, falsifying this prediction of the Rule and hence the Rule itself.

32. The comparison of the trichodectid cladogram with such host relation- ships as are known is the first such comparison in the lice and the most extensive for any parasite group. The comparison reveals mismatches that require the postulation of a minimum of 41 secondary infestations and associated speciation events. This is 20.7% of all speciation events postulated in the history of the Trichodectidae (Geomydoecus excluded), and is much higher than would have been predicted by any worker on lice. Such a high degree of secondary infestation necessitates the rejection of Fahrenholz' Rule in any but the most general sense.

33. Fahrenholz1 Rule is falsified and, whilst probably describing a proportion of host-parasite evolutionary associations, is considered 546

to have no value as a precise tool in phylogenetic reconstruction. Methods involving use of a number of parasites to determine the phylogenetic relationships of the hosts (Hopkins' principle) are considered justifiable only if the phylogenies of all the parasites have been established using cladistic principles, and if any non- phylogenetic factors which may explain the host distributions of some of the parasites do not explain the majority of them.

34. Trichodectidae parasitise eutherian mammals in the Carnivora, Rodentia, Hyracoidea, Artiodactyla, Perissodactyla, Primates and Edentata. The supposed primacy of some of these relationships has been used in zoo geographical and phylogenetic studies.

35. The three associations with Primates are shown to be the results of secondary infestations from different hosts.

36. The associations of Geomydoecus and Eutrichophilus with rodents are believed to be the result of two independent secondary infestations of the ancestors of the Geomyidae and Erethizontidae respectively, from unknown hosts.

37. The primary host association of Neotrichodectes is with the ancestor of the Mephitinae, and secondary infestations of Mustelinae, Procyonidae and Edentata have taken place. The supposed relationship of the species parasitising Edentata with hyrax lice is rejected. The association with Mephitinae is probably derived from an initial secondary infestation from a geomyid rodent.

38. The primary host association of Lorisicola is with an ancestor of the Viverridae and/or Felida^ the association of one species with a Primate being the result of secondary infestation. The primary host association of Felicola is with an ancestor of the Herpestidae, secondary infestation of Viverridae, Felidae and Canidae having taken place. The primary host association of the Felicolini (Lorisicola plus Felicola) is therefore probably, with an ancestor of one or all of the Herpestidae, Viverridae and Felidae.

39. The primary host association of the Trichodectes - Lutridia clade is with an ancestor of all or some of the Mustelidae, secondary 547

infestations of Ursidae, Canidae and Procyonidae having taken place. The primary host association of the sister-group of this clade is with an ancestor of the Hyaenidae (Feloidea). The primary host association of the Trichodectinae (Felicolini plus Trichodectini) is postulated to have been with an ancestor of the Feloidea, with secondary infestation of a mustelid from a hyaenid taking place.

40. The primary host association of the Bovicolinae is uncertain but may have been with an ancestor of the Bovidae, the Pecora, the Ruminantia or a common ancestor of the Peris so dactyl a and Artiodactyla. The associations of Bovicolinae with Cervidae and Camelidae result from secondary infestations from Bovidae.

41. The primary host association of the Burytrichodectes - Procavicola clade is with an ancestor of the Procaviidae (Hyracoidea) and that of its sister-group, Cebidicola, with an ancestor of the Cebidae (Primates). The host of the ancestor of the whole clade (the Dasyonyginae) is not known.

42. The factors governing host specificity in parasites of vertebrat.es, particularly lice, are discussed. A direct relationship is shown between the extent to which the location of the parasite is governed by the location of the host and the- level of monoxenia exhibited in a parasite family. The adaptation of a parasite to a given host taxon does not prohibit its survival on a different host taxon,- and parasites may be considered to have, in addition to their observed hosts, a number of potential hosts which could support the parasites if they were infested. The observable host specificity of a parasite species is the result of the interaction between its distributional dependence on the host, the availability of potential hosts, and the chances of it colonising these hosts. The inter- relationships between these and several other factors are indicated. section 6

bibliography 549

6. BIBLIOGRAPHY

References not seen are indicated lay an asterisk (*)•

ABERCRCMBIE, M.f C. J. HICKMAN & M. L. JOHNSON, 1951. A dictionary of "biology. Penguin Books, Harmondsworth. pp. 284. ANSELL, W. F. H., 1971. Order Artiodactyla. In: Meester, J. & H. VT. Setzer (eds) The Mammals of Africa. An identification manual. Smithsonian Institution Press, Vfashington, D. G.. Part 15, PP. 1-84. ARORA, G. L. & N. P. CHOPRA, 1959. Observations on the life-history of Lipeurus tropicalis Peters (Mallophaga: Ischnocera). Research Bulletin (N.S.) of the Pan.jab University. JO: 179-187-

ASH, J. S., 1960. A study of the Mallophaga of birds with particular reference to their ecology. Ibis. 102: 93-110. ASELOCK, P. H. & D. J. BROTHERS, 1979*- Systematization and higher classification in evolutionary systematics through cladistic and anagenetic analysis. Manuscript quoted by Wiley, 1981. ASKEH, R. R., 1971. Parasitic Insects. Heinemann, London, pp. xvii + 316* BADONNEL, A., 1934- Recherches sur l'anatomie des Psoques. Supplements au Bulletin Biologique de la France et de la Belgicrue. 18. pp. xviii + 241. BEDFORD, G. A. H., 1929. Anoplura (Siphunculata and Mallophaga) from South African hosts. Annual Report of the Director of Veterinary Services, Union of South Africa. 15: 501-549-

BEDFORD, G. A. H., 1932a. Trichodectidae (Mallophaga) found on African Carnivora. Parasitology. 24: 350-364. BEDFORD, G. A. H., 1932b. Trichodectidae (Mallophaga) parasitic on

Procaviidae. Proceedings of the Zoological Society of London. 3:

709-730. BEDFORD, G. A. H., 1936. Notes on species of Trichodectidae with descriptions of new genera and species. Onderstepoort Journal of Veterinary Science and Animal Industry. J.1 33-58. BEDFORD, G. A. H., 1939. Further notes on Trichodectidae with descriptions of new species. Cnderstepoort Journal of Veterinary Science and Animal Industry. J2: 103-119- BENNETT, D. K., 1980. Stripes do not a zebra make, part I: A cladistic analysis of Squus. Systematic Zoology. 29? 272-287. BEQUAERT, J. C., 1956. The Hippoboscidae or louse-flies (Diptera) of mammals and birds. Part II. Taxonomy, evolution and revision of American genera and species. Entomologica Americana. 36: 417-611. 550

BERTRAM, D. S., E. VI. ROBERTS & M. EDWARDS, 1952. A rapid method of sampling for Damalinia (Bovicola) bovis L. on cattle, and its use in a comparison of the effectiveness of 'Gammexane* and D.D.T. against this parasite. Annals of Tropical Medicine and Parasitology. 46: 7-24. BLAGOVESETCHENSKI, D. I., 1956. Structure and taxonomic significance of the genitalia of Mallophaga. Paraz.itolog.jcheskiy Sbornik Zoologicheskogo Instituta Akademi.ya Nauk. S.S.S.R. 16: 1-88. (in Russian) • BOCK, VI., 1969• Discussion: The concept of homology. Annals of the New York Academy of Sciences. 167s 71—73* BOCK, VI., 1974* Philosophical foundations of classical evolutionary classification. Systematic Zoology. 22: 375-392. (1973) BOCK, VI., 1977* Foundations and methods of evolutionary classification. In: Hecht, M. K., P. C. Goody & B. M. Hecht (eds) Major Patterns in Vertebrate Evolution. N.A.T.O. Advanced Study Institutes Series: Series A, Life Sciences. Volume 14* PP* 851—895• BODEMER, C. VI., 1968. Modern Embryology.' Holt, Rinehart & Vrinton, New York. pp. xviii + 475* BURNER, C., 1904« Zur Systematik der Hexapoden. Zoologischer Anzeiger. 21: 511-533. BOUDREAUX, H. B., 1979* Arthropod phylogeny with special reference to Insects. John Vliley fr Sons, New York. pp. 320. BOUVIER, G., 1945- De l'h&nophagie de quelques Mallophages des animaux domestiques. Schweizer Archiv f\lr Tierheilkunde. 87: 429-434* BOYD, E. M., 1951* A survey of parasitism of the starling Sturnus vulgaris L. in North America* Journal of Parasitology. 37: 56-84. BRINCK, P., 1956.* The reproductive system and mating in Plecoptera. I. Opuscula Entomologica. 21_ (1): 57-127. BROOKS, D. R., 1979. Testing the context and extent of host-parasite coevolution. Systematic Zoology. 28: 299-307. BROOKS, D. R., 1981. Hennig's parasitological method: A proposed solution. Systematic Zoology. 30: 229-249. ERUNDIN, L., 1966. Transantartic relationships and their significance, as evidenced by chironomid midges. Kungliga Svenska Vetenkapsakademiens Handlingar. ser. 4, vol JM, no. 1: 1-472. BUCHNER, P., 1965* Endosymbiosis of Animals with Plant Microorganisms. Revised English version. Miley-Interscience. New York, London, Sydney, pp. xvii + 9°9* 551

BUCKUP, L., 1959. Der Kopf von Myrsidea cornicis (De Geer) (Mallophaga- Amblycera). Zoologische Jahrbticher. (Anatomie). 77:241-288. BURMEISTER, H., 1838. Handbuch der Entomologie. Band 2, 2. Gymnognatha. pp. 418-443. BUSVINE, J. R., 1948. The 'head1 and 'body1 races of Pediculus humanus L.. Parasitology. 39: 1-16. BUSVINE, J. R., 1978. Evidence from double infestations for the specific status of human head lice and body lice (Anoplura). Systematic Entomology. _3: 1-8. CAIN, A. J. & G. A. HARRISON, 1958. An analysis of the taxonomist's judgement of affinity. Froceedings of the Zoological Society of London. 131: 85-98. CHAPMAN, R. F., 1982. The Insects. Structure and Function. Hodder & Stoughton, London, Sydney, Auckland, Toronto, pp.xiv + 919* CHIN, T.-H., 1980. Studies on Cinese Anoplura IV. The description of two new species and proposal of new families and new suborder for the lice of Typhiomys cinereus Milne—Edwards. Acta Academiae Medicinae Kweiyang. ^(2): 91-100. " CLAY, T., 1946. Mallophaga Miscellany. - No. 3. The Trabecula. Annals and Magazine of Natural History. (11) _13: 355-359* CLAY, T., 1949* Piercing mouthparts in the biting lice (Mallophaga). Nature, London. 164: 617. CLAY, T., 1949a. Some problems in the evolution of a group of ectoparasites. Evolution. 3:279-299. CLAY, T., 1951* An introduction to a classification of the Avian Ischnocera (Mallophaga): Part I. Transactions of the Royal Entom- ological Society of London. 102 (2): 171—194- CLAY, T., 1953. Revisions of the genera of Mallophaga. I. The Rallicola complex. Froceedings of the Zoological Society of London. 123: 563-587. CLAY, T., 1954* The post-spiracular seta and sensillus in the Mallophaga (insecta). Annals and Magazine of Natural History. ^2) 2: 716—718. CLAY, T., 1957* The Mallophaga of Birds. In: First symposium on host specificity among parasites of Vertebrates, Neuchatel. pp. 120-155* CLAY, T., 1958. Revisions of Mallophaga genera. Degeeriella from the Falconiformes. Bulletin of the British Museum (Natural History). (Entomology). 2: 121-207. CLAY, T., 1962. A key to the species of Ferris with notes and descriptions of new species. Bulletin of the British Museum (Natural History). (Entomology). _1_1: 191-244* 552

CLAY, T., 1963. A new species of Haematomyzus Piaget (Phthiraptera: Insecta)• Proceedings of the Zoological Society of London. 141? 153-161. CLAY, T., 1964. Geographical distribution of the Mallophaga (insecta). Bulletin of the British Ornithologist's Club. 84: 14-16. CLAY, T., 1966. The species of Strigiphilus (Mallophaga) parasitic on the barn (Tyto)• Journal of the Entomological Society of Queensland. 10-17. CLAY, T., 1969* A key to the genera of Menoponidae (Amblycera: Kallophaga: Insecta). Bulletin of the British Museum (Natural History). (Entomology). 24: 1 - 76. CLAY, T., 1970. The Amblycera (Phthiraptera: Insecta). Bulletin of the British Museum ( Natural History). (Entomology). 2^: 75-98. CLAY, T., 1971* A new genus and two new species of Boopidae (Phthiraptera: Amblycera). Pacific Insects. 13: 5529• CLAY, T., 1973. Fhthiraptera (Lice). Iru Smith, K. G. V. (ed.) Insects and Other Arthropods of Medical Importance. British Museum (Natural History), London, pp. 395-397. CLAY, T., 1976. Geographical distribution of the avian lice (Phthiraptera): a review. Journal of the Bombay Natural History Society. JVt 536-547. CONCI, C., 1942. Diagnosi preliminari di tre nuovi generi e di una

nuova specie di Trichodectinae. Bolletino della Societa Entomologica Italiana. 140-142. CONCI, C., 1946. Un nuovo genere di Trichodectidae Sudamericano. Bolletino della Society Entomologca Italiano. 76: 59* CCOK, E. P., 1954* A modification to Hopkins' technique for collecting ectoparasites from mammalian skins. Entomological News. 65:35-37. CCOK, E. F.

CORBET, G. B., 1978- The mammals of the Palaearctic Region: a taxonomic review. British Museum (Natural History), Cornell University Press, London and Ithaca, pp.314* CORBET, G. B. Sr. J. E. HILL, 198O. Ato or Id list of mammalian species. British Museum (Natural History), Comstock Publishing Associates, London and Ithaca, pp.viii + 226. CRACRAFT, J,, 1974- Phylogenetic models and classification. Systematic Zoology. 23: 71-50. CUMMINGS, B., 1913- On some points of the anatomy of the mouth-parts of the Mallophaga. Proceedings of the Zoological Society of London. 1913 PP. 128-141. de BEER, G. R., 1951** -Embryos and Ancestors. Clarendon Press, Oxford, pp. xii + 159. de BEER, G. R., 1958.* Embryos and Ancestors. Clarendon Press, Oxford, pp. xii + 197. EICHLER, Y!., 1936. Die Biologie der Federlinge. Journal ftir Ornithologie. 84: 471-505* EICHLER, YJ., 1937* Nutrition and position of eggs among Mallophaga. Sitzungsberichte der Gesellschaft Naturforschender Freunde zu Berlin. 1937* pp.80-111. (in German).

EICHLER, Y1.t 1940. Notulae Mallophagologicae. I. Neue Gattungen und Subfamilien von Haarlingen. Zoologischer Anzeiger. 129: 158-162. EICHLER, H., 1940a. Uirtsspezifit&t und stammegesgeschtliche Gleichl&ufigkeit (Fahrenholzsche Regel) bei Parasiten im allgemeinen und bei Mallophagen im besonderen. Zoologischer Anzeiger. 132: 254-262. EICELER, YI., 1941. Zur Klassifikation der Lauskerfe (Phthiraptera Haeckel: Rhynchophthirina, Mallophaga und Anoplura). Archiv fVlr Naturgeschichte. N.F. JO: 345-398. EICHLER, V.7., 1948. Some rules in ectoparasitism. Annals and Magazine of Natural History. (12) J,: 588-598. EICHLER, YT., 1954. Ectopias it en von Zoo-tieren. III. Ein haarling von trampeltier. Monatsheften fttr yeterin&rmedizin. 330-331.

EICHLER, YI.f 1963. Phthiraptera. 1: Mallophaga. Bronns, Klassen und Ordnungen des Tierreichs. 5 Band, III Abt., 7 Buch. pp. vii + 291. EICKLER, YJ., 1966. Evolutionary terms for parasite species. Systematic Zoology. £5: 216-218. EICHLER, VI., 1582. Notulae Mallophagologicae. XIII. Goliathipon nov. gen. und weitere neue Taxa der Gattungsstufe. Deutsche Entomologische Zeitschrift. 281-88. 554

EISENBERG, J. F., 1981 • The Mammalian Radiations. Athlone Press, London, pp. xx + 610. EISENMANN, V., 1980. Les chevaux (Equus sensu lato) fossiles et actuels: cranes et dents .jugales supgrieures. Cahiers de Paldontologie (Editions du C.N.R.S.). pp.186. EMERSCN, K. C., 1965. A new species of Mallophaga from the Philippine Islands. Journal of the Kansas Entomological Society. £8: 68-69. EMERSON, K. C. & R. D. PRICE, 1979- Two new species of Bovicola (Mallophaga: Trichodectidae). Journal of the Kansas Entomological Society. £2: 747-750. EMERSON, K. C. & R. D. PRICE, 198O. A new species of Suricatoecus (Mallophaga: Trichodectidae) from the western cusimanse, Crossarchus obscurus (Carnivora: Viverridae). Florida Entomologist. 63? 505~508. EMERSON, K. C. & R. D. PRICE, 1981. A host-parasite list of the mallophaga on mammals. Miscellaneous Publications of the Entomol- ogical Society of America. 12: 1-72. EI'IERSON, K. C. & R. D. PRICE, 1982. A hew species of Bovicola (Mallophaga: Trichodectidae) from the Formosan Serow, Capricornis crispus swinhoei (Artiodactyla: Bovidae). Pacific Insects. 24: 186-188. EMERSON, K. C. I- R. D. PRICE, 1983. A review of the Felicola felis complex (Mallophaga:'Trichodectidae) found on New Nor Id cats (Carnivora: Felidae). Proceedings of the Entomological Society of Vlashington. 85: 1-9. ESTABR00K, G. F., 1977- Does common equal primitive? Systematic Botany- 2: 36-42. ESTABRCOK, G. F., 1979- Some concepts for the estimation of evolutionary relationships in systematic botany. Systematic Botany. .3(2) (1978) 146-158. ESTAERCCK, G. F., J. G. STRAUCH, Jr. & K. L. FIALA, 1977- An application of compatibility analysis to the Blackiths1 data on orthopteroid insects. Systematic Zoology. 26: 269-276. EVELEIGH, E. S. & W. THRELFALL, 1976. Population dynamics of lice (Mallophaga) on auks (Alcidae) from Newfoundland. Canadian Journal of Zoology. £4: 1694-1711. EYING, E. E., 1924- On the taxonomy, biology, and distribution of the biting lice of the family Gyropidae. Proceedings of th£ United States National Museum. 63 (20): 1 - 42. ETING, H. E., 1929. A Manual of External Parasites. Charles C. Thomas, Springfield and Baltimore, pp. xiv + 225. 555

EKING, H. E., 1933. Some peculiar relationships between ectoparasites and their hosts. American Naturalist. 67: 365-373. EJING, H. E., 1936. The taxonomy of the Mallophagan family Trichodectidae, with special reference to the New World fauna. Journal of Parasitology. 22: 233-246. FAHRENHOLZ, H., 1936. Zur Systematik der Anopluren. Zeitschrift fVlr Parasitenkunde. ° (l): 5°-57- FARRIS, J. S., 1980. The information content of the phylogenetic system. Systematic Zoology. 28: 483-519* FENST'ERMACHER, R. Z W. L. JELLISON, 1932. Porcupine louse infesting the monkey. Journal of Parasitology. 18: 294. FERRIS, G. F., 1929. Review: A Manual of External Parasites, by Henry Ellsworth Ewing. Entomologists' News. 40 (10): 337-341. FERRIS, G. F., 1931. The louse of Elephants. Haematomyzus elephantis Piaget (Mallophaga: Haematomyzidae). Parasitology. 23 (l): 112-127. FERRIS, G. F., 1933. The Kallophagan genus Trichophi1opterus. Parasitology. 2^ (4): 468-471. FERRIS, G. F., 1935* Contributions towards a monograph of the-sucking lice. viii. Stanford University Publications, University Series, Biological Sciences. II (8): 529-634. FERRIS, G. F., 1951. The Sucking Lice. Memoirs of the Pacific Coast Entomological Society. 1 pp. 320. FLORENCE, L., 1921. The hog louse, Haematopinus suis Linng: Its biology, anatomy and histology. Memoirs. Cornell University Agricultural Experimental Station. 51: 635-743. FLORKIN, M., 1962.* Isologie, homologie, analogie et convergence en biochimie compar^e. Bulletin de I'Academie Royale de Belgicrue. Classe des Sciences. (5) 819-824.

FLYNN, J. J. <3: H. GALIANO, 1982. Phylogeny of Early Tertiary Carnivora, with a description of a nev; species of Frotictis from the Middle Eocene of Northwestern Wyoming. American Museum Novitates. 2725: 1 - 64. FOERSTER, 1868. Synopsis der Familien und Gattungen der Ichneumonen. Verhandlungen des Naturhistorischen Vereins der Preussischen Rheinlande und Westfalens. 2£: 135-221.

GANS, C., 1969. Discussion: Some questions and problems in morphological comparison. Annals of the New York academy of Sciences. 167: 506- 513. 556

GENTRY, A. VI., 1970.* Revised classification for Makapania broomi Wells and Cooke (Bovidae, Mammalia). Palaeontologia Africana. J3: 63-67. GENTRY, A. VI., 1978- Bovidae. In: Maglio, V. J. & E. B. S. Cooke (eds) Evolution of African Mammals. Harvard University Press, London. PP. 540-572. GENTRY, A. W., 1980. Fossil Bovidae (Mammalia) from Langebaanweg, South Africa. Annals of the South African Museum. ££ (8): 213-337. GCSS, R. J., 1953* The advanced embryology of the book louse, Liposcelis divergens Badonnel (Psocoptera: Liposcelidae)• Journal of Morphology. £2: 157-206. GRAY, J. E., 1825. An Outline of an Attempt at the Disposition of Mammalia into Tribes and Families, with a List of the Genera apparently appertaining to each Tribe. Annals of Philosophy. 26: 337-344. GRAY, J. E., 1843. List of the specimens of Mammalia in the collection of the British Museum. George VIoodfall and Son, London, pp. xxviii + 216. GREGORY, VI. K. & M. HELL MAN, 1939- The evolution and major classification of the civets (Viverridae) and allied fossil and Recent Carnivora? a phylogenetic study of the skull and dentition. Proceedings of the American Philosophical Society. 81 (3): 309-392. GRIFFITHS, G. C. D., 1972. The phylogenetic classification of Diptera Cyclorrhapha. Br W. Junk, N. V., The Hague, pp. 340. GROVES, C. P., 1969.* On the smaller gazelles of the genus Gazella de Blainville, 1816. Zeischrift fttr SHugetierkunde. 34? 38-60. GATHER, K. K., 1974. StaublHuse, Psocoptera. Die Tierwelt Deutschlands. 61: 1-314. HAMILTON, VI. R., 1978. Cervidae and Palaeomerycidae. In: Maglio, V. J. & H. B. S. Cooke (eds) Evolution of African Mammals. Harvard University Press, London, pp. 496-508. HARRISON, L., 1915* The respiratory system of Mallophaga. Parasitology. 8: 101-127. HARRISON, L., 1916. The relation of the phylogeny of the parasite to that of the host. Report of the British Association for the Advancement of Science. 85: 476-477. HARRISON, L., 1916a. The genera and species of Mallophaga. Parasitology. 1 - 156. 557

FAUB, F., 1967* Der Kopf von Pseudomenopon pilosum (Scopoli) (Mallophaga - Amblycera). Zoologische Jahrbttcher. (Anatomie). 84: 493-558. EAUB, F., 1971* Der Kopf von Ornithobius cygni (Denny) (Mallophaga - Ischnocera). Zoologische Jahrbticher. (Anatomie). 88: 450-504*

KAUB, F., 1972. The sitophore (Cibarialsklerit) of the Mallophaga - Amblycera and the Mallophaga - Ischnocera (Kellogg) (insecta). Zeitschrift ftir Morphologie und Okologie der Tiere. 73: 249-261. (In German). FAUB, F., 1973. The cibarium of Mallophaga, investigations about the Morphological differentiation. Zoologische Jahrbttcher. (Anatomie). _20: 483-525. (In German). HAUB, F., 1977* The sitophore of species from the family Colpocephalidae (Burmeister, 1838) Mallophaga/ Amblycera. Investigations about Dimension- and Shape-Stability of Function-connected Structures in Psocodea. Zoologische Jahrbttcher. (Anatomie). 97: 294-322. FAUB, F., 1980. Letter to the Editors: Concerning "Phylogenetic relationships of parasitic Psocodea and Taxonomic Position of the Anoplura" by K. C. Kim and H. Li. Ludwig. Annals of the Entomological Society of America. 73: 3-6. HAUG, G., 1952. Morphologische und histophysiologische Untersuchungen an den Verdauungsorganen der Kallophagen und Anopluren. Zoologische Jahrbllcher. (Anatomie). 22: 289-344. HEBERDEY, R. F., 1931. Zur Entwicklungsgeschichte, vergleichenden Anatomie und Physiologie der weiblichen Geschlechtsausftihrwege der Insekten. Zeitschrift fVlr Morphologie und Okologie der Tiere. 22: 416-586. EECHT, M. K. & J. L. ED.iARDS, 1976. The determination of parallel or monophyletic relationships: the proteid salamanders — a test case. American Naturalist. 110 (974): 653-677* FECHT, M. K. &: J. L. ED/TARDS, 1977. The methodology of phylogenetic. inference above the species level. In Hecht, M. K., P. C. Goody & B. M. Hecht (eds) Major Patterns in Vertebrate Evolution. N.A.T.O. Advanced Study Institutes Series: Series A, Life Sciences. Volume 14* pp*3-51* FELLENTHAL, R. A. i R. D. FRICE, 1976. Louse-host associations of Geomydoecus (Mallophaga: Trichodectidae) with the yellow-faced pocket gopher, Pappageomys castanops (Rodentia: Geomyidae). Journal of Medical Entomology. 13: 331-336. 558

EELLSNTHAL, R. A.

HENNIG, VI., 1965. Phylogenetic Systematics. Annual Review of Entomology. JO: 97-116. HENNIG, VT., 1966. Ph7/logenetic Systematics. Translated by Davis, D. D. and R. Zangerl. University of Illinois Press, Urbana, Chicago and London, pp. 263. EENNIG, W., 1981. Insect Phylogeny. Translated and edited by Pont, A. C. John Wiley and Sons, Chichester, pp. xxii + 514- HCNACKI, J. H., K. E. KINMAN

HOPKINS , G. F. E.t 1960. Notes on some Mallophaga from Mammals. Bulletin of the British Museum (Natural History). (Entomology)• JO: 75-55. HOPKINS, G. H. E. & T. CLAY, 1552. A check list of the genera and species of Mallophaga. British Museum (Natural History), London, pp. 362. HOPKINS, G. H. E. &. T. CLAY, 1553. Additions and corrections to the check list of Mallophaga. Annals and Magazine of Natural History. (12) 6: 434-448. HOPKINS, G. H. E & T. CLAY, 1555. Additions and corrections to the check list of Mallophaga - II. Annals and Magazine of Natural History. (12) 8: 177-150. HOPKINS, G. H. E. & M. ROTHSCHILD, 1962. An illustrated catalogue of the Rothschild collection o"f fleas (Siphonaptera) in the British Museum (Natural History). Hystrichopsyllidae, Vol. 3. British Museum (Natural History), London, pp. viii + 56O. HOPKINS, G. H. E. & M. ROTHSCHILD, 1966. An illustrated catalogue of the Rothschild collection of fleas (Siphonaptera) in the British Museum (Natural History). Hystrichops.yllidae, Vol. 4. British Museum (Natural History), London, pp. viii + 545* HOYLE, N. L., 1938. Transmission of poultry parasites by birds with special reference to the "English" or house sparrow and chickens. Transactions of the Kansas Academy of Science. 41: 375-384- HULL, D. L., 1968. The operational imperative: sense and nonsense in operationism. Systematic Zoology. 17: 438-457. IMMS, A. D., 1925. A general textbook of Entomology. Methuen & Co. Ltd., London, pp. xii + 698. IMMS, A. D., 1539. Cn the antennal musculature in insects and other arthropods. Quarterly Journal of Microscopical Science. 8j: 273- 320. 560

INGER, R. F., 1967. The development of a phylogeny of frogs. Evolution. 2_U 369-384. JARDINE, N., 1967- The concept of homology in "biology. British Journal for the Philosophy of Science. j8: 125—139• JARDINE, N., 1969- The observational and theoretical components of homology: a study based on the morphology of the dermal skull- roofs of rhipidistian fishes. Biological Journal of the Linnean Society. U 327-361. JENSEN, R. J., 1981. VTagner networks and ITagner trees: a presentation of methods for estimating most parsimonious solutions. Taxon. 30 (3): 576-590.

JENTSCH, S., 1940. Zur Morphologie des Gehirns und der Lichtsinnes- organe der Psocopteren. Zoologische JahrbMcher. (Anatomie). 66 (3): 403-436. JOBLING, B., 1949- Host-parasite relationship between the American Streblidae and bats, with a new key to the American genera and a record of the Streblidae from Trinidad, British Nest Indies (Diptera). Parasitology. 3°: 315-329. JOHNSON, L. A. S., 1970. Rainbow's end: the quest for an optimal

taxonomy. Systematic Zoology. 19s 203-239* JOHNSON, P. T., 196O. The Anoplura of African Rodents and Insectivores. United States Department of Agriculture Technical Bulletin. 1211: 1 - 116. JOHNSTON, T. H. £ L. HARRISON, 1911. Notes on some Mallophagan generic names. Proceedings of the Linnean Society of New South Hales. 36 (2): 321-328. KEIRANS, J. E., 1567. The Mallophaga of New England birds. Bulletin of the New Hampshire Agricultural Experimental Station. 492: 1-179* KEIRANS, J. E., 1975a. A review of the phoretic relationship between Mallophaga (Phthiraptera: Insecta) and Hippoboscidae (Diptera: Insecta)• Journal of Medical Entomology. 12: 71-76. KEIRANS, J. E., 1975D. Records of phoretic attachment of Mallophaga (Insecta: Phthiraptera) on insects other than Hippoboscidae. Journal of Medidal Entomology. 12: 476. KELER, S. von, 1934* Mallophaga von Folen. Die Familie Trichodectidae. Bulletin de l'Academie Polonaise des Sciences et des Lettres. (B) 2: 259-267.

KELER, S. von, 1938. Baustoffe zu einer Monographie der Mallophagen. I Teil: Uberfamilie der Trichodectoidea. Nova Acta Leopoldina. Nr. 32: 395-467. 561

K3&LER, S. von, 1938a. frber einege Mallophagen aus Paraguay und Kamerun• Arbeiten Tiber Morphologische u. Taxonomische Entomologi• Ber1in-Dahlem• 228-241. KILLER, S. von, 1939* Baustoffe zu einer Nonographie der Mallophagen. II Teil: frberfam. Nirmoidea (1). Nova Acta Leopoldina. (n.f.). 8: 1-254. KELER, S. von, 1944* Bestimmungstabelle der Uberfamilie Trichodectiodea (Mallophaga). Stettiner Entomologischer Zeitung. 105:167—191• * 11 KELER, S. von., 1952. Uber den feineren Bau der Tarsen bei Pseudomenopon rowanae Kdler. (Mallophaga). BeitrBge zur Entomologie. 2 (6): 573- 582. » KELER, S. von, 1955- Einige Bemerkungen Tiber den Bau der Tarsen von Gyropus und Gliricola. (Mallophaga). BeitrTjge zur Entomologie. 1 (3/4): 293-308. KELER, S. von, 1957* Der Haarling der Tiildkatze (Felicola hercynianus n. sp.). Mallophaga: Trichodectidae. Deutsche Entomologische Zeitschrift. 172-178. KELER, S. von, 1957a. tiber die Deszendenz und die Differenzierung der Mallophagen. Zeitschrift fUr Parasitenkunde. 18: 55-160. » KELER, S. von, 1960a. Tiber die Nymphe von Nesiotinus demersus (Kellogg) (Mallophaga) • Deutsche Entomologische Zeirtschrift.

It 177-184.

KELER, S. von, 1960b. Bibliographie der Mallophagen. Mittelungen aus dem Zoologischen Museum in Berlin. 36 (2): 145-403. KELER, S. von, 1966. Zur mechanik der nahrunsaufname bei Corrodentien. Zeitschrift fTTr Parasitenkunde. 2J: 64-79. KELER, S. von, 1969. Mallophaga (Federlinge und Haarlinge). Handbuch der Zoologie. Berlin. 4 (2) 11/17: 1-72. KELER, S. von, 1971- A revision of the Australasian Boopiidae (insecta: Phthiraptera) with notes on the Trimenoponidae. Australian Journal of Zoology. Suppliment 6: 1-126. KELLOGG, V. L., 1896. New Mallophaga, I, - with special reference to a collection made from maritime birds of the bay of Monterey, California. Proceedings of the California Academy of (Natural) Sciences. (2) 6: 31-168. KELLOGG, V. L., 1908. Mallophaga. In TPytsman, P. Genera Insectorum. 66: 1-87.

KELLOGG, V. L., 1910 (1908). Corrodentia: Mallophaga. V.'issenschaftliche

Srgebnisse der Scbwedischen Zoologischen Expedition dem Kilimand.jaro

dem Meru und dem Umgebenden Massaisteppen Deutsch-Ostafrikas 1905 -

1906. J£ (4): 43-56. 562

KELLOGG, V. L., 1913* Distribution and species-forming of ecto- parasites. American Naturalist. 47: 129-158. KELLOGG, V. L., 1914* Ectoparasites of mammals. American Naturalist. 48: 257-279. KENNEDY, C. R., 1975. Ecological Animal Parasitology. Blackwell Scientific Publications, Oxford, pp. ix + 163. KETTLE, P. R., 1977. A study on Phthiraptera (chiefly Amblycera and Ischnocera) with particular reference to the evolution and host- parasite relationships of the Order. Ph. D. thesis, University of London, pp. xiii +510. KEY, K. H. L., 1967* Operational homology. Systematic Zoology. 16: 275-276. KIM, K. C., 1965. A review of the Hoplopleura hesperomydis complex. Journal of Parasitology. 871-887. KB!, K. C. & H. VI. LUDUIG, 1978a. The family classification of the Anoplura- Systematic Entomology. _3: 249-284. KIM, K. C. <5: H. K. LU3X7IG, 1978b. Phylogenetic relationships of parasitic Psocodea and taxonomic position of the Anoplura. Annals of the Entomological Society of America. 71: 910-922. KIM, K. C. & H. Yf. LUDWIG, 1982. Parallel evolution, cladistics, and classification of parasitic Psocodea. Annals of the Entomological Society of America. 75: 537-548. KINGDOM, J., 1982. East African Mammals. An Atlas of Evolution in Africa. Vol. Ill, part C. (Bovids). Academic Press. London, New York and San Fransisco. pp. 393. KLIER, E., 1956. Zur Konstruktionsmorphologie des mSLnnlichen Geslechts- apparates der Fsocopteren. Zoologische Jahrbucher. (Anatomie). 75: 207-286. KLOCKENHOFF, H., 1972. Zur Taxonomie der auf der Salvadorikrdhe Corvus orru lebenden Mallophaga-Gattung Myrsidea Yfaterston, 1915* Bonner Zoologische Beitr£ge. 23: 253-266. KLOET, G. S. & Yi. D. HBTKS, 1964. A check list of British Insects Part 1: Small orders and Hemiptera. Handbooks for the Identification of British Insects. 11( 1) pp. xv + 119. KLUGE, A. G. & J. S. FAHRIS, 1969. Quantitative phyletics and the evolution of anurans. Systematic Zoology/. 18: 1-32. m KOUIGSHANN, E., 1960. Zur Fhylogenie der Parametabola unter besonderer

Berllcksichtigung der Phthiraptera. Beitrhge zur Entomologie. 10:

705-744. 563

KRISTEMSEN, N. P., 1975- The phylogeny of hexapod "orders". A critical review of recent accounts. Zeitschrift fUr Zoologische Systematik und Evolutionsforschung. 13: 1-44* KRISTENSEN, N. P., 1981. Phylogeny of Insect Orders. Annual Review of Entomology. 26: 135-157. KRYNSKl, S., A. KUCHTA & E. BECLA, 1952;* Research on the noxious action of guinea-pig "blood on the body-louse. Bulletin of the Institute of Marine Medicine in Gdansk. 104-107* (in Polish). KUHN, H.-J. & H. VI. LUB-JIG, 1964. Mallophaga "on Catarrhine monkeys: Colobus guereza, a natural host of Procavicola colobi. Nature. 203 (4S43): 424-425* LAKSHMINARAYANA, K. V., 1976. Nomenclatural changes in Phthiraptera - some suggestions. Angewandte Parasitologic. 17 (3): 160-167. LAKSHMINARAYANA, K. V., 1977* Factors involved in the host-specificity in Mallophaga sens, lat., (Phthiraptera: Insecta) infesting birds. In: Ananthakrishnan, T. N. (ed.) Insects and Host-Specificity. The Macmillan Company of India Ltd. pp. 101-109. LANE, R. P. & J. E. MARSHALL, 1981. Geographical Variation, races and subspecies. In: Forey, P. L. (ed.) The Evolving Biosphere. British Museum (Natural History), Cambridge University Press. PP. 9-19. LEDGER, J. A., 1980. The Arthropod Parasites of Vertebrates in Africa South of the Sahara. Volume IV. Phthiraptera (Insecta). Publications of the South African Institute for Medical Research No. 56. pp. 327. LEINDERS, J. J. M. & E. HEINTZ, 1980. The configuration of the lachrimal orifices in Pecorans and Tragulids (Artiodactyla, Mammalia) and its significance for the distinction between Bovidae and Cervidae. Beaufortia. 30: 155—162. L0VTRUP, S., 1977. Phylogeny of the Vertebrata. John YTiley and Sons. London, pp. xii + 330. LUIHIG, H. VI., 1573. Viability of lice feeding on blood of nonspecific hosts. Proceedings of the International Symposium on the Control of Lice and Louse-borne Diseases, Washington, 1972. pp. 159-162. LUIWIG, H. 1982. Host specificity in Anoplura and coevolution of Anoplura and Mammalia. H&noires. Museum National d'Histoire Naturelle. (n.s.) A. _123: 145-151. LYAL, C. H. C., 1982. Lice. In: Manson-Bahr, P. E. C..& F. I. C. Apted (eds) Manson's Tropical Diseases, 18th Edition. Bailli&re Tindall, London, pp. 805-808. 564

MAA, T. C., 1963. Genera and species of Hippoboscidae (Diptera): "types, synonymy, habitats and natural groupings. Pacific Insects Monographs. (5. pp. 186. MAA, T. C., 1964. A review of the Old-Uorld Polyctenidae (Hemiptera: Cimicoidea). Pacific Insects. 6 (3): 494-516. MC KENT!A, M. G., 1975* Towards a phylogenetic classification of the Mammalia. In: Luckett, VI. P. & F. S. Szaley (eds) Phylogeny of the Primates: a multidisciplinar,y approach. Plenum, New York and London, pp. 21-46- MANTER, H. W., 1955• The zoogeography of trematodes of marine fishes. Experimental Parasitology. 4: 62-86. MANTER, H. Vf., 1966. Parasites of fishes as biological indicators of recent and ancient conditions. Proceedings of the 26th Annual Biological Colloquium, Oregon State University, pp. 55-71* MARSHALL, A. G., 1980. The comparative ecology of insects ectoparasitic upon bats in Viest. Malaysia. Proceedings of the fifth international bat research conference, pp. 135-142. MARSHALL, A. G., 1981. The ecology of ectoparasitic insects. Academic Press, London, pp. xvi + 459* MARX, H. & G. B. RABB, 1972. Phyletic analysis of fifty characters of advanced snakes. Fieldiana: Zoology. 63: 1-321. MATTHYSSE, J. G., 1946. Cattle lice, their biology and control. Bulletin of the New York State Agricultural Experiment Station. 832. MATSUDA, R., 1958* On the origin of the external genitalia of insects. Annals of the Entomological Society of America. 51:84-94. MATSUDA, R., 1965* Morphology and Evolution of the Insect Head. Memoirs. American Entomological Institute. 1-334* MATSUDA, R., 1970. Morphology and Evolution of the Insect Thorax. Memoirs of the Entomological Society of Canada. 76: 1—431• MATSUDA, R., 1976. Morphology and Evolution of the Insect Abdomen - with special reference to developmental patterns and their bearings upon s.ystematics. International series in pure and applied biology, 56. Pergamon Press, Oxford and New York. pp. viii + 532. MAYER, C., 1954* Vergleich Untersuchungen am Skeletl-Huskelsystem des Thorax der Mallophagen unter Berttcksichtigung des Nervensystems. Zoologische Jahrbttcher. (Anatomie). 74 (l): 1-188. MAYR, E., 1942. S.ystematics and the Origin of Species. Columbia University Press, New York. pp. 334* 565

MAYR, E., 1945•* Birds of the Southwest Pacific. Macmillan Co., New York. pp. xix +316. MAYR, E., 1963. Animal species and evolution. Harvard University Press, Cambridge, Massachusetts, pp. xiv + 797*

MAYR, E., 1969. Principles of systematic zoology. TATA McGraw Hill, Bombay, pp. x + 428. MAYR, E., 1970. Populat ions, species, and evolution; An abridgement of Animal species and evolution. Harvard University Press, Cambridge, Massachusetts, pp. xv + 453.

MAYR, E., 1574- Cladistic analysis or cladistic classification? Zeitschrift fUr zoologische Systematik und Evolutionsforschung. J2: 94-128. MAYR, E., E. G. LINDSLEY & R. L. USINGER, 1953. Methods and Principles of Systematic Zoology. McGraw - Hill, New York. pp. x + 328. MEACHAM, C. A., 1980. Phylogeny of the Berberidaceae with an evaluation of classifications. Systematic Botany. £ (2): 149-172. MEHTANDER, M., 0. HALKKA &. V. S&DERLUND, 1974. Chromosomal evolution in the Psocoptera. Notulae Entomologicae. 54: 81-84. MILLER, P. H., JR'., 1970a. Scanning electron microscopy of antennal structures of Pol.yplax serrata (Burmeister) (Anoplura: Hoplopleuridae). Journal of the Hew York Entomological Society. j8: 33-37. MILLER, P. H., JR., 1970b. Scanning electron microscopy of . Journal of the Hew York Entomological Society. j8: 139-145. MILLER, P. H., JR., 1971a. Scanning electron microscopy of Echinophthirius horridus (von Clfers), Antarctophthirus callorhini (Osborn), and Proechinophthirius fluctus (Ferris) v/ith emphasis on the antennal structures (Anoplura: Echinophthiriidae). Journal of Parasitology.

668-674. KILLER, F. H., JR., 1971b. Scanning electron microscopy of five Haematopinus species v/ith special attention to their antennal structures. Journal of the Hew York Entomological Society. 79: 19-26. MILLER, F. H., JR., 1971c. Antennal sensilla coeloconica of Fedicihus obtusus obtusus (Anoplura: Foplopleuridae). Annals of the Entomological Society of America. 64: 973-975• f« KJOBERG, E., 1910. Studien ttber Mallophagen und Anopluren. Arkiv fbr Zoologi. 6 (13): 1-256, 5 plates. MCREBY, C., 1978. . The biting louse genus Werneckiella (Phthiraptera: Trichodectidae) ectoparasitic on the horse family Equidae (Mammalia: Perissodactyla). Journal of Natural History. 12: 395-412. 566

FUKERJI, D. & P. SEN-SARMA, 1555. Anatomy and affinity of the elephant louse Faematomyzus elephantis Piaget (insecta: Rhyncophth- iraptera). Parasitology. 45: 5-30.

MURRAY, M. D., 1957a* The distribution of the eggs of Mammalian lice on their hosts. I. Description of the oviposition behaviour. Australian Journal of Zoology. £ (l): 13—18• MURRAY, M. D., 1957b. The distribution of the eggs of Mammalian lice on their hosts. II. Analysis of the oviposition behaviour of Damalinia ovis (L.). Australian Journal of Zoology. £ 0): 19-29*

MURRAY, M. D., 1957c. The distribution of the eggs of Mammalian lice on their hosts. III. The distribution of the eggs of Damalinia ovis (L.) on the sheep. Australian Journal of Zoology. £ (2): 173-182. MURRAY, M. D., 1957d. The distribution of the eggs of Mammalian lice on their hosts. IV. The distribution of the eggs of Damalinia ecrui (Denny) and Haematopinus asini (L.) on the horse. Australian Journal of Zoology. 5 (2): 183-187. MURRAY, M. D., 1963. The ecology of lice on sheep. IV. The establishment and maintenance of populations of Linognathus ovillus (Neumann). Australian Journal of Zoology. 11: 157-172. MURRAY, M. D. & J. H. CALABY," 1971. The host relations of the Boopiidae. In: K61er, S. von, 1971 (q.v.) Appendix II pp. 81-82. NELSON, B. C., 1972. A revision of the New Nor Id species of Ricinus (Mallophaga) occurring on Passeriformes (Aves). University of California Publications in Entomology. 68: 1-175* NELSON, G., 1972. Phylogenetic relationship and classification. Systematic Zoology* .21: 227-231. NELSON, G., 1974a. Classification as an expression of phylogenetic relationship. Systematic Zoology. 22: 344-359 (1973)• NELSON, G., 1974b. Darwin-Hennig Classification: A reply to Ernst Mayr. Systematic Zoology. 23: 452-458. NEN, T. R., 1974. Structural variation in Psocopteran wing-coupling mechanisms. International Journal of Insect Morphology and Embryology. 3 (2): 193-201. NEN3IGGIN, M. J., 1936*. Plant and Animal Geography. Kethuen, London, pp. 258. NITZSCH, C. L., 1818. Die Familien und Gattungen der Thierins'ecten (insecta epizoica)* Magazin der Entomologie (Germar). 3: 261—316'. 567

NUSSBAUM, J., 1882. Zur Entwicklungsgeschichte der AusfVlhrgange des Sexualdrttsen bei den Insekten. Zoologischer Anzeiger. 2} 637-643. OLDROYD, H., 1964. The Natural History of Flies. Weidenfield and Nicolson, London, pp. xiv + 324* OLSON, E. C., 1971. Vertebrate Paleozoology. 7£iley-Interscience, New York. pp. xv + 839. PALMA, R. L., 1978. Slide-mounting of lice: a detailed description of the Canada balsam technique. New Zealand Entomologist. 6: 432-436. PATTERSON, E., 1957- Mammalian phylogeny. In: First symposium on host specificity among parasites of vertebrates, Neuchatel. pp. 15-48. PATTERSON, C., 1981. Significance of fossils in determining evolutionary relationships. Annual Review of Ecology and Svs'tematics. 12; 195-223.

FETTER, G., 1962. Le peuplement de Carnivores de Madagascar. Collogues Internationaux du Centre National de la Recherche Scientifiaue. 104: 331-342. PICTRGJSKI, F., 1961. On the morphogenesis of the genital efferent system in Pedicuius humanus L. (Anoplura). Acta Zoologica Cracoviensia. 5 (13): 583-659. PLATNICK, N. I., 1977. Cladograms, phylogenetic trees, and hypothesis testing. Systematic Zoology. 26: 438-442. PLCMLEY, N. J. B. & G. B. THOMPSON, 1937. Distribution of the biting louse. -Nature, London. 140: 199*

PRICE, P. W.t 1977* General concepts on the evolutionary biology of parasites. Evolution. 31: 405-420. PRICE, P. W., 1980. Evolutionary Biology of Parasites. Monographs in population biology 15• Princeton University Press, Princeton, pp. xi + 237. PRICE, R. D., 1974. Two new species of Geomydoecus from Costa Rica pocket gophers (Mallophaga: Trichodectidae). Proceedings of the Entomological Society of Washington. 76: 41-44. PRICE, R. D., 1975. The Geom.ydoecus (Mallophaga: Trichodectidae) of the Southeastern U.S.A. (Rodentia: Geomyidae). Proceedings of the

Entomological Society of VFashington. 77: 61-65. PRICE, R. D., 1975a. The Kenacanthus eurysternus Complex (Mallophaga: Menoponidae) of the Passeriformes and Piciformes (Aves). Annals of the Entomological Society of America. 68: 617-622. PRICE, R. D. <3- J. R. BEER, 1963. Species of Colpocephalum (Mallophaga:

Menoponidae) parasitic upon the Falconiformes. Canadian Entomologist.

25s 731-763. 568

FRICE, R. D. & K. C. EMERSON, 1971. A revision of the genus Geomydoecus (Mallophaga: Trichodectidae) of the New VTorld Pocket Gophers (Rodentia: Geomyidae)• Journal of Medical Entomology, Honolulu. 8: 228-257. PRICE, R. D. & K. C. EMERSON, 1972. A new subgenus and three new species of Geomydoecus (Mallophaga: Trichodectidae) from Thomomys (Rodentia: Geomyidae)• Journal of Medical Entomology, Honolulu.

o: 463-467. PRICE, R. D. & R. A. HELLENTEAL, 1975a. A reconsideration of Geomydoecus expansus (Duges) (Mallophaga: Trichodectidae) from the yellow- faced pocket gopher (Rodentia: Geomyidae). Journal of the Kansas Entomological Society. 48? 33-42. PRICE, R. D. & R. A. HELLENTHAL, 1975"b. A review of the Geomydoecus texanus complex (Mallophaga: Trichodectidae) from Geomys and Pappageomys (Rodentia: Geomyidae). Journal of Medical Entomology, Honolulu. J2: 401-408.

PRICE, R. D. & R. A. KELLENTHAL, 1976. The Geomydoecus (Mallophaga: Trichodectidae) from the hispid pocket gopher (Rodentia: Geomyidae).

Journal of Medical Entomology9 Honolulu. _12: 695-700- PRICE, R. D. & R. A. KELLENTHAL, 1979- A review of the Geomydoecus tolucae complex (Mallophaga: Trichodectidae) from Thomonws (Rodentia: Geomyidae) based on qualitative and quantitative characters. Journal of Medical Entomology, Honolulu. j6: 265-274- PRICE, R. D. & R. A. HELLENTHAL, 1980a. The Geomydoecus oregonus complex (Mallophaga: Trichodectidae) of the Western United States pocket gophers (Rodentia: Geomyidae). Proceedings of the Entomological Society of Washington. 82: 25-38. PRICE, R. D. & R. A. HELLENTHAL, 1980b. The Geom:/-doecus neocopei complex (Mallophaga: Trichodectidae) of the Thornonxys umhrinus pocket gophers (Rodentia: Geomyidae) of Mexico. Journal of the Kansas Entomological Society. 53: 567~580. PRICE, R. D. R. A. HELLENTHAL, 1980c. A review of the Geomydoecus minor complex (Mallophaga: Trichodectidae) from Thomomys (Rodentia: Geomyidae). Journal of Medical Entomology, Honolulu. 298-313. PRICE, R. D. 2 R. A. KELLEWTHAL, 1S8la. Taxonomy of the Geomydoecus umbrini complex (Mallophaga: Trichodectidae) from Thomomys umbrinus (Rodentia: Geomyidae) in Mexico. Annals of the Entomological Society of America. 74: 37-47- PRICE, R. D. & R. A. HELLENTFAL, 1981b. A review of the Geomydoecus californicus complex (Mallophaga: Trichodectidae) from Thomomys (Rodentia: Geomyidae). Journal of Medical Entomology, Honolulu. _l8: 1-23.

PRICE, R. D. & R. M. TIMM, 1979. Description of the male of Geomydoecus scleritus (Mallophaga: Trichodectidae) from the Southeastern pocket gopher. Journal of the Georgia Entomological Society. 14? 162-165. RAI, R. K. & K. V. LAKSH-'INAR AYANA, 19 80. A note on the In Vitro studies of the chewing-lice (Phthiraptera). Proc. TTorkshop Tech. Parasitol. zool. Surv. Indian pp. 55-59* RAMCKE, J., 1964* Der Kopf der Schweinelause (Kaematopinus suis L., Anoplura). Zoologische JahrbVlcher. (Anatomie). 82 (1965): 547-663* RANKIN, G. D., 1982. Mallophaga on the eggs of wading birds. Ibis. •124: 183-187. RAY, C. E., E. A. ANDERSON & S. D. VOTE, 1981. The Blancan Carnivore Trigonictis (Mammalia: Mustelidae) in the 'Eastern United States. Brimleyana. 1-36. RENDEL, J. M., 1959* Canalisation of the scute phenotype of Drosophila. Evolution. J_3: 425-439. RENDEL, J. M., 1967* Genetic control of a developmental process. In: Lewontin, R. C. (ed.) Population biology and Evolution. Syracuse University Press, New York. pp. 47-66. RICHARDS, 0. Li. & R. G. DAV1ES, 1957. A. D. Imms, A General Textbook of Entomology. Jth Edition, Methuen, London, pp. 886. RICHARDS, C. Li. & R. G. DAVIES, 1977* Irons' General Textbook of

Entomology. 10th Edition, Chapman and Hall, Ltd., London. 2 volumes,

pp. vii + 1354* 11 RIES, E., 1930. Uber die Symbioten der LHuse und Federlinge. Zentralblatt fUr Bakteriologie, Parasitenkunde und Infektionskrankheiten. Abt. 1, 117: 286-298. RIES, E., 1931. Die Symbiose der LHuse und Federlinge. Zeitschrift ftir ti Morphologie und Okologie der Tiere. 20: 233-367. RIES, E., 1932. Die Prozesse der Eibildung und der Eiwachsturns bei Fediculiden und Mallophagen. Zeitschrift fUr Zellforschung und Mikroskopische Anatomie. 16: 314-338. RIGLER, P., 1951. Der Kopf von Bovicola caprae (Gurlt.) (Kallophaga)• Zoologische Jahrbttcher. (Anatomie). 71: 289—426. RISLER, von H. &• K. GIZSINGER, 1965. Die Mundwerkzeuge von Gliricola gracilis N. (Mallophaga - Amblycera), ein Beitrag zur Kopfmorphologie der Tierl&use (Phthiraptera). Zoologische JahrbUoher. (Anatomie). 82: 536-546. 570

ROTHSCHILD, JI. & T. CLAY, 1952. Fleas, Flukes and . Collins (New Naturalist Series) London, pp. xiv + 304* RUDOLPH, D., 1982a. Occurrence, properties and biological implications

of the active uptake of water vapour from the atmosphere in Psocoptera. Journal of Insect Physiology. 28: 111-121. RUDOLPH, D., 1982b. Site, process and mechanism of active uptake of water vapour from the atmosphere in the Psocoptera. Journal of Insect Physiology. 28: 205-212. RUDOLPH, D., 1983. The water-vapour uptake system of the Phthiraptera.

Journal of Insect Physiology. 2£: 15-25. SCH8FER, L., 1964* Sur la phylogdnie des mallophages. Bulletin de la

SociStd d'Histoire Naturelle de Toulouse. 99: 312-318. SCHAEFER, C. VI., 1578. Ecological separation of the human head lice and body lice (Anoplura: Pediculidae). Transactions of the Royal Society of Tropical Medicine and Hygiene. 72: 669-670. SCHLEIN, Y., 1970. A comparative study of the thoracic skeleton and

musculature of the Pupipara and the Glossinidae (Diptera) • Parasitology.

60: 327-373. SCFMUTZ, YJ., 1955» Zur Konstruktionsmorphologie des m&nnlichen Geschlechtsapparates der Mallophagen. Zoologische Jahrbllcher. (Anatomie). 189-338. SCUDDER, G. G. E., 1961. The comparative morphology of the insect ovipositor. Transactions of the Royal Entomological Society of London. 113: 25-40. SCUDDER, G. G. E., 1964- Further problems in the interpretation and homology of the insect ovipositor. Canadian Entomologist. 96:

405-417. SCUDDER, G. G. E., 1971. Comparative morphology of insect genitalia. Annual Review of Entomology. 16: 379-406. SEEGER, W., 1975* Funktionsmorphologie an Spezialbildungen der FVlhlergeissel von Psocoptera und anderen Paraneoptera (insecta); Psocodea als monophyletische Gruppe. Zeitschrift fBr Morphologie und Ckologie der Tiere. 81: 137-159* SEEGER, Vi., 1979• Spezialmerkmale an EiMlllen und Embryonen von Fsocoptera im Vergleich zu anderen Paraneoptera (insecta); Psocoptera als monophyletische Gruppe. Stuttgarter Beitr^ge zur Naturkunde aus dem Staatlichen Museum fUr Haturkunde in Stuttgart.

329: 1-57* 571

SIKORA, H. & VI. EICHLER, 1941 • frber Kopulationseigentttmlichkeiten der Mallophagen. Zeitschrift flir Morphologie und 3kologie der Tiere. 38: 80-84. SIMPSON, G. G., 1945* The principles of classification and a classification of mammals. Bulletin of the American Museum of Natural History. 8$: 1-350. SIMPSON, G. G., 1961. Principles of animal taxononw. Columbia University- Press, New York. pp. 247. SMIT, F. G. A. M., 1957. Siphonaptera. Handbooks for the Identification of British Insects. J (16) pp. 1-94- SMITHERS, C. N., 1972. The classification and phylogeny of the Psocoptera. Memoirs of the Australian Museum. 14? 1-349* STTEATH, P. H. A. <5: R. R. SCKAL, 1973. Numerical Taxonomy - The Principles and Practice of Numerical Classification. Freeman, San Francisco, pp. xv + 573. SNODGRASS, R. E., 1896. In: Kellogg, V. L., 1896a. New Mallophaga II, - From Land Birds; together with an account of the Mallophagous mouth-parts. Proceedings of the California Academy of (Natural) Sciences. (2), 6: 431-548. SNODGRASS, R. E., 1899. The anatomy of the Mallophaga. Occasional Papers of the California Academy of Sciences. 6: 145-229.

SNODGRASS, R. E., 1905. A revision of the mouth-parts of the Corrodentia and the Mallophaga. Transactions of the American Entomological Society. 31: 257-307. SNODGRASS, R. E., 1935. Principles of Insect Morphology. Mc Graw - Hill, New York. pp. 667. SNODGRASS, R. E., 1944. The feeding apparatus of biting and sucking insects affecting man and animals. Smithsonian Miscellaneous Collections. 104: 1-113. SNCDGRASS, R. E., 1957. A revised interpretation of the external reproductive organs of male insects. Smithsonian Miscellaneous Collections. 135: 1-60. SMOBGF.ASS, R. E., 196O. Facts and theories concerning the insect head. Smithsonian Miscellaneous Collections. 142: 1-61* STEilRAM, H., 1956. The ecology of Oolumbicola columbae L. (Ma'llophaga). Opuscula Entomologica. 2_1: 170-150. STENRAM, H., 1964. The evolution of the Mallophaga and the phylogeny of their hosts. Zoologiska Revy. 26: 23-32. (in Swedish). 572

STOBBE, R., 1913a. Mallophagen. 1. Beitrag: Neue Formen von S&ugetieren (Trichophilopterus und Eurytrichodectes nn. gg.). Entomologische Rundschau. 30 (18): 105-106, 111-112. STOBBE, R., 1913b. Mallophagen. 3. Beitrag: Die Trichodectiden des Berliner Museums fttr Naturkunde• Sitzungsberichten der Gesellschaft naturforschender Freunde, Berlin. 8: 365-383.

STOJANOVICH, C. J., JR., 1945. The head and mouthparts of the sucking lice (insecta: Anoplura). Microentomology. 10: 1-46. ST&IE, E., 1943* Der Kopf von Trimenopon jenningsi Kellogg and Paine (Mallophaga). Zoologische Jahrbllcher. (Anatomie). 68: 177-226. STRINDBERG, H., 1916. Zur Entwicklungsgeschichte und Anatomie der Mallophagen. Zeitschrift flir Nisenschaftliche Zoologie. 115: 383-459* STURTEVA1TT, A. H., 1929. The genetics of Drosophila simulans. Publications. Carnegie Institution of Washington. 399: 1-62. SYMMCNS, S., 1951* Comparative anatomy of the Mallophagan head. Transactions of the Zoological Society of London. 27: 345-436. TARRY, D. 7f., 19^7* The occurrence of Linognathus setosus (Anoplura, Siphunculata) on poultry. The Veterinary Record. 81: 641. TELFORD, R. H., 1976. Relationship of pinnipeds to other carnivores (Mammalia). Systematic Zoology. 25: 363-374*

THEODOR, 0., 1957* Parasitic adaptation and host-parasite specificity in the pupiparous Diptera. In: First symposium on host specificity among parasites of Vertebrates, Neuchatel. pp. 50-62. TEEODCR, 0. & M. COSTA, 1967. A survey of the parasites of wild mammals and birds in Israel. Part I. Ectoparasites. Israel Academy of

Sciences and Humanities, Jerusalem, pp. 117. THOMPSON, G. B., 1940. The distribution of Feterodoxus spiniger (Enderlein). Papers and Proceedings of the Royal Society of Tasmania. (1939) pp. 27-31.

TIMM, R. M. &: R. D. PRICE, 1979- A new species of Geomydoecus (Mallophaga: Trichodectidae) from the Texas pocket gopher, Geomys personatus

(Rodentia: Georr.yidae). Journal of the Kansas Entomological Society. £2: 264-268.

Tin:, R. M. £ R. D. PRICE, 1580. The taxonomy of Geomydoecus (Mallophaga: Trichodectidae) from the Geomys bursarius complex (Rodentia: Geomyidae). Journal of Medical Entomology, Honolulu. 17: 126-145. 573

TBETERMANN, G. von, 1957- Stellung und Gliederung der Regenfeifervttgel (Ordnung Charadriiformes) nach massgabe des Mallophagologischen befundes. In: First symposium on host specificity among parasites of Vertebrates, Neuchatel. pp. 159-168. TCULECHKOFF, K., 1955. The Trichodectidae (Trichodectoidea, Mallophaga) ectoparasitic on mammals of Bulgaria. Izvestiya na Zoologicheskiya Institut. Bfilgarska Akademiya na Naukite. Sofiya. Bd. 4-5: 423-434* (in Russian). TRAUB, R., 1980. The zoogeography and evolution of some fleas, lice and mammals. In: Traub, R. & H. Starke (eds) Fleas. Proceedings of the International Conference on Fleas, Ashton IT old. A. A. Balkema, Rotterdam, pp. 93-172. TUXETT, S. L., (ed), 1970. Taxonomist's glossary of genitalia in insects. Munksgaard, Copenhagen, pp. 359• UESEBTA, TT., 1972. New World Polyctenidae (Hemiptera), with special reference to Venezuelan species. Brigham Young University Science Bulletin, Biological Series. 17: 13-21. USINGER, R. L., 1966. Monograph of Cimicidae (Hemiptera, Heteroptera). Entomological Society of America (Thomas Say Foundation). Volume 7. PP. 585. VAN GELDER, R. G., 1978. A review of canid classification. American

Museum Novitates. 2646: 1-10. VANZOLBTI, P. E. & L. R. GUBIARAES, 1955. Lice and the history of South American land mammals. Revista Brasileira de Entomologia. .3: 13-46. VRBA, E. S., 1979* Phylogenetic analysis and classification of fossil and recent Alcelaphini. Mammalia: Bovidae. Biological Journal of the Linnean Society. JJ: 207-228. 7JAAGE, J. K., 1979. The evolution of insect/vertebrate associations. Biological Journal of the Linnean Society. 12: 187-224. WAGNER, VI. J., JR., 1961.* Froblems in the classification of ferns. In:

Recent Advances in Botany. (From lectures ard Symposia presented to the IX International Botanical Congress, Montreal, 1959)- University of Toronto Press, Montreal, pp. 841-8A4. WALKER, E. P., 1964. Mammals of the World. Johns Hopkins Press, Baltimore. Volume II, pp. 647-1500. WARD, R. A., 1957. A study of the host distribution and some relationships of biting lice (Mallophaga) parasitic on birds of the order Tinamiformes. Part II. Annals of the Entomological Society of America. j)0: 452-459- 574

VIARD, R. A.| 1958- Preliminary observations on the origin of some Nearctic Bird Lice (Kallophaga)• Proceedings of the 10th International Congress of Entomology, Montreal. 1956, 1. pp. 745-749* WATERSTCN, J., 1926. On the crop contents of certain Mallophaga. Proceedings of the Zoological Society of London. 4: 1017-1020. WATSON, G. E. i A. 3. AMERSON, 1967. Instructions for collecting bird Parasites. Smithsonian Institution Information Leaflet 477. pp. 12. WEBB, J. E., 1946. Spiracle structure as a guide to the phylogenetic relationships of the Anoplura (biting and sucking lice), ltfith notes on the affinities of the mammalian hosts. Proceedings of the Zoological Society of London. • 116: 49-119* WEBB, J. E., 1948. Eyes in the Siphuriculata. Proceedings of the Zoological Society of London. 118: 575-577. WEBB, J. E., 1949* The evolution and host-relationships of the sucking lice of the Ferrungulata. Proceedings of the Zoological Society of London. 119: 133-188. WEBER, H., 1936. Copeognatha. In: Schulz, P. (ed.) Biologie der Tiere Deutschlands. 3£: 1-50. WEBER, H., 1938. Beitr&ge zur Kenntnis der t5berordnung Psocoidea. 1. Die LabialdrTlsen der Copeognathen. Zoologische Jahrbttcher. (Anatomie). 64: 243-286. WENZEL, R. L., 1976. The streblid batflies of Venezuela (Diptera: Streblidae)• Brigham Young University Science Bulletin, Biological Series. 20: 1-177* WENZEL, R. L., V. J. TIPTON & A. KIEWLICZ, 1966. The streblid batflies of Panama (Diptera Calypterae: Streblidae). Ln: Wenzel, R. L. & V. J. Tipton (eds) Ectoparasites of Panama. Field Museum of Natural History, Chicago, pp. 405-675* WERNECK, F. L., 1936. Contribuicg.0 ao conhecimento dos Mallophagos

encontrados nos mamrniferos sul-americanos. Memorias do Instituto Oswaldo Cruz. 31: 391-589* WERNECK, F. L«, 1941* Os Maldfagos dos Frocaviiaeos. Memorias do Instituto Oswaldo Cruz. 36: 445-576. WERNECK| F. L., 1946. Sohre alguns mal6fagos de procaviideos. Revista Brasileira de Biclogia. 6: 85-97. WERNECK, F. L., 1943. Os Maldfagos de Mamlferos. Parte I: Amblycera e Ischnocera (Philopteridae e Parte de Trichodectidae),. Edigao da Revista Brasileira de Biologia, Rio de Janeiro, pp. 243. 575

WERNECE, F. L., 1950. 0s MaI6fagos de Mamfferos, Parte II: Ischnocera (continuagao de Trichodectidae) e Rhyncophthirina. Edicao de Memorias do Instituto Oswaldo Cruz. pp. 207• WESTROM, D. E., B. C. NELSON & G. E. CONNOLLY, 1976. Transfer of Bovicola tibialis (Piaget) (Mallophaga: Trichodectidae) from the introduced fallow deer to the Columbian black-tailed deer in California. Journal of Medical Entomology, Honolulu. _13: 169-176. WHITE, I. M., 1980. Nymphal taxonomy and systematics of the Psylloidea (insecta: Fomoptera). Ph.D. Thesis, Liverpool Polytechnic, pp. 340. WHITE, M. J. D., 1957* Genetics and systematic entomology. Annual Review cf Entomology. 2: 71-90. WI0GLE5WCRTH, V. 3., 1941. The sensory physiology of the human louse Pedi cuius hum anus corporis De Geer (Anoplura). Parasitology. 33? 67-109. WILEY, E. 0., 1979a* Cladograms and phylogenetic trees. Systematic Zoology. 28: 88-92. WILEY, E. 0., 1979b. An annotated Linnaean hierarchy, with comments on natural taxa and competing systems. Systematic Zoology. 28: 308-337* WILEY, E. 0., 1981. Phylogenetics. The theory and practice of phylogenetic systematics. Miley-Interscience. John Wiley and Sons, New York, pp. xv + 439- WILLIAMS, R. T., 1971* In vitro studies on the environmental biology of Goniodes colchici (Denny) (Mallophaga: Ischnocera). III. The effect of temperature and humidity on the uptake of water vapour. Journal of Experimental Biology. 55: 553-568. WILSON, E. 0., 1965. A consistency test for phylogenies based on contemperaneous species. Systematic Zoology. 14: 214-220. WILSON, F. H., 1934. The life-cycle and bionomics of Lipeurus heterographus Nitzsch. Journal of Parasitology. 304-311. WILSON, F. H., 1936. The segmentation of the abdomen of Lipeurus heterographus Nitzsch. Journal of Morphology. 60: 211-219. WILSON, F. H., 1939* The life-cycle and bionomics of Lipeurus caponis (Linn.). Annals of the Entomological Society of America. 32: 318-320. WONG, S. K. & I. W. B. THORNTON, 1968. The internal morphology of the reproductive systems of some psocid species. Proceedings of the Royal Entomological Society of London, A. 43: 1-12. WUNDRIG, G., 1936. Die Sehorgane de Mallophagen, nebst vergleichenden

Untersuchungen an Liposcleiden und Anopluren. Zoologische Jahrbttcher.

(Anatomie)i 62: 1-172. 576

YOUNG, J. E., 1953. Embryology of the mouthparts of Anoplura. Microentomology. 18: 85—133•

ZL0T0R2YCKA, J., 1964. Mallophaga parasitising Passeriformes and Pici.

1. Subfamilies Dennyinae, Machaerilaeminae, Colpocephalinae. Acta Parasitologica Polonica 12: 165—192- 2L0TCRSYCKA, J., 1972. Klucze do Oznaczaniz Cwad6w Polski, XV, Mallophaga, 3. Goniodidea i Trichodectidea. Polskie Towarzystwo Entomologiczne. 24: 1-48. ZUMPT, P., 1565. Myiasis in man and animals in the Old World. A textbook

for physicians, veterinarians, and zoologists. Butterworths, London, pp. xv + 267. SECTION 7

APFENDICES 578

7.1. APPENDIX A - DATA MATRIX

The matrix displays the states of the 279 characters listed in section 2.3. (columns) for 327 species of Trichodectidae (rows). Most of the species of G. (Geomydoecus) and of G. (Thomomydoecus) exhibit the same states for the characters surveyed, and therefore instead of listing each species just the two subgenera are included in the list, reducing the number of rows necessary from 327 to 233* The coding applied-to the characters is discussed in section 1.4.3..

The columns are arranged in 27 blocks of ten and one of nine characters although, as the coding required for some characters occupies more than one column, the blocks are of varying widths. The first character (one or more columns) in each block is identified by number (from section 2.3.), as are characters requiring more than one column. In the latter case the columns are linked by parentheses, the lower Parenthesis uniting the columns coding for LTagner tree analysis, the upper uniting these with the coding for phenetic analysis; the coding for phenetic analysis always preceeds that for Magner tree analysis. For example, the full heading of the first'column would be:

o 7"3 ®e g® g IB

0000000000000

The omission of most of the column headings leads to a considerable saving of space, without hinderance to the use of the matrix. To display the full matrix 18 pages are required, the characters extending across three pages, and each page listing 40 species. The first three pages below list the states of characters 1 - 90» 91 — 200 and 201 - 279 respectively for the first 40 species, the following

nd three pages the same for the next 40 species, a so on. On the first of each set of three pages the species names are given, with a number (1 - 233); on the following two pages only the number is given as a reference. 579

— — — O o — OOOOOO oooooooo oooooooo

o\ O 8\ ffi0} g OC n ff\ O O O O O O O o

0o) 01 AS 01 p3 l-l K Oft) bO 0) si BJ a) d *H —o4 p01 P 01 <0 e w 3 — o E c to c3r hi g OG u 3 U 0o] C r-l O 3 I-l H m m a o .C >» G p •O XI X3 o. C o a> >> o o > hi M o 01) pH hi 031 jG . s • . • , c J J O. O, b. Cbb. Cr .C b Cb u. Cx.C b Cb Cb Cb CM m vo r- coc n o — cm f^ -3- m vo js»co cn KV Drixvxvxixvxvxvxvxvxvx ro ro rv> ro rroo to to ro 10 — — — — — — — — — — voor>-Jovvn.e-v>iio —

Soobco8SS^oSoooo§§oo o 000 — — — — — — — -*— — — — £ 0000000000000000000 — 0000 — — 00000 0 OOOOOOOOOOOOOCOC-OOOCOOOOOQOOOOOOOOO OQQOOOOOOOOOOOOO •-JOOOOOOOCOOOOOOCOO OOOOOOOOOOOOOOOOCOOCOOOOOOOOOOO 0000.----OOOOOOCOOOOOO^0 w - - w 0^ 0^ — —— wv^w^^-w0000000w — 000 — — — — — -*-»-»o — — —• — — — — — — -1 PD o 000000000 — o 00000000 o o o 0000000000000000 s s § vo voVOVOVOVDVOVDVDVOV D 0000 _ o00 o 0o o— o o ro o o o ru ro1 8 10 ror o o 8 VD VD VQ _ 0088 8888222 o o o o o o 0008 8 80 8 O VO VO vC VO VD O— —O VDO VOO O vo VD o O o — o 0000 — 0000 — — — — 00 — — o o — — — o o o o 88 0000 OOOOO — — 00— — OOOOOOO o o 000 o— o o o O — — O o o o 0000000000000000000000 2 <63 o o o o o —o — o O — — — — o o o o o o O00000 O O Q O 0O 0000000000000 o o o o OOOOOOO 0000000000000 o o o o OOOOOOO 0000000000000 o o OOOOOOO 0000000000000 o o o8 8o 8 8 OOOOOOO 0000000000000 o o o o 0000 — — — — — — — — — 8 8 o o o o 8008888 OOOOOQOOOOOOO — — — OOOO — — 00O 0 o o 000 o o 8000 0000 0000000000000 o o 0000 0000 0000000000000 o o o— p co —o 0000 ooooooooopooo o o 000- 8 o 8 8 S o — 000 § — — — — — - - - - 0888668888—o VDVOVDVOVOVDVOVDVOVDVDV8 8O8 o o 8000 VO vVoD — o v ovo v o vo8800000 0 O VO VD VO VD VD VO VO VO VQ VD VD o— — — — IO fw 8 8 8 8 8wiu — — MM- — MwioioMMwio-' — — — — 00 — — — — — ro o 00 — — ogrooo — — — — — — — — w M ro ro ro -» — p p ro to ro - o 00000 — — — — — 08 000 — o o Oo VvDo — — — —8S8838S83 o o o OOO_ _ _ O„ O o O O O O O 8 O O O P 000000000000— — — — VQVOVOVDOVOVGVOV0D 8 8 8 8 o S o o — — — o vo 00888 -88888388888 VO vo VO VO vo o vo o00000 0 VOVOVOVOVOVOVDVDVOOO o 000000 — — — VO VO VO vo 8OOOOO — — OOOO o "VO VOVOCDVOVD — — OOOOVDVDVO — —OOOO VO VO VD VO VD VO OOOOO O VO vo vo OVDVOVOVOVOVDVO VOVOVDVOVDVOVOVD'OvDVOVVD VD VO VO VO VO — — VO voO V OVO V D VD VO VOVOlOVOVDVDVDVOVD OVDVDVD VO VO VO VOvVoD vVoD vVoD""VOVDVDVDVDVOVOVOvDVOV D vVoO S 8888888888 0888 8VO VO VO VO o VO VOVOVOVOVOVOVOVOVOVOMOVOVOVO — — 00 o o 000S8888888 qS'8 8 vo VO VOVOVDVOVDVOVDVDVDVO o VO VD VD O— —o V—D —V*D —VO —V O— — — VVDVDVDVOVDVDVOVO'DVOVOVO VO VD VO VD VO VO VO O VOO V vOo VO o 00000000000000 o o p o o o OOOOOOO — — — — vo vo VO 'O 88 888888888 8 8 8 8 8 8 88388388888 VoD VDVOVDVDVOVDVDVOVDVOVDVDVOV000 C- 000000000O0 VD VD VO VO VO VD VO VD VO VO VO VO VO VO VO VD VD VO VO VO VO VQ O OOOOOOOOOOO--- W W W —0000000000W 1—W —W —W W W W W W wW/w0 W w pooooogpoW u K —t — — P Oo P Oo — — .Q — — — — — — — — — —OOO pOOPO P — — OOOOO 8 88888888888880 — — — — — — — — — uuuuu00000u0 VD— — u u u y t) VO VOVOVOVOVOVOVSVOVOVOVOVOVO o OOOOOOOOOVOVDVOVOVDVOOPVOVDVOVOVooopoppooovo VD VD VO . — -* C5 VDO C o C vc o opoooooooovooo — — — — — — — — — — oooooopoooooo 8 8 8 88 8888888 — — o 00 — — — — — — — 8888 — —8888838 VD — — vo vo — vo —.— VDVDVOVOVDVDVOVOVDVDVOVDVD VD VO VO VO VO 'O VD VO VO VVO DV OVD VQ — — VDVPVDvOVOVDVnvDvOVDVDVDvO VD VDVDVDVDVOVOVDVDVDVD VO VO VD — — VDVDVOVDVOVDVOVOVOPVDVDVD VO VvOo VOD v'Oo \VDo v'o VvOo VvDo VvOp VvDVoD v v3VoD VO OVDVOVOVDVOVOVDVDVOvOVDVD VD VD VC VOVOVOVOVOVOVOVOVO VvOo vVOo VvoD O OVDVOVOVOVDVOVOVOVDVOVDVOVD lis Ooooopopppopopoooopoopooooooopoooooooooo 0000000000 OOOO 0OVOVDVOVDVOVOVOVOVOVOVOVOV „ „ „ — _ — _ _o _O o2222222222222222222222222222222222°°°° ooocoooooo OOOO O °

2— — SKK^K^iS^GK^OpOOO-WMOOOOOOOMMrO-NNNIOHMWWpo — OOOOOOO — — — — — — po — — — — — — — OOO — OOOOOOOOOTN o OCOOOOOOO— — OOOOOOO OOOOOOOOOOOOOO

o o o o 00000p o o o 0o o o BOOOOO

08^ 581

w i— cj ni -3"u unv) or ^r«-c o o oj ffv -a- m vo r^- co cr> o»- cy i^^r in vo nco o c\j M^-a-mvo p-co t- ^-T- CMCMCMCMCMCVJCXJCMCVJCVJ K^n ^ jrjTi n ^ ^ 1 4 18 19 20 61 11 14 I II—I—I 21 31 50 51 56 65 71 73 78 81 63 I 1 r-r 41 F. congoensis 0000000010000 10000911201000000 000009010000 000100000000000 19201000000000 010000000190 000000000000 0000001112101 000000000010 42 F. helogaloidls 0010000110000 10000911201000000 000009000000 000100000000000 19201000000000 010000000190 000000000000 0000001112101 000000000010 43 F. helogale 0003001110000 10000910201000000 000009000010 000000000000000 19201000000000 000000000190 000000000100 0000010112101. 000000000000 44 f> occidental is 0000000100000 00000910201000000 210009000010 000000000000000 19201000000000 000000000190 000000000000 0000000912101 000000000000 45 F. minimus 0019999000900 00000991000999000 210009000010 000000000000001 99999999599999 999999990999 000000000000 0000000910001 000000000000 46 F. fahrenholzl 0000000100900 00000991201999000 210009000010 000000000000001 99999999999999 999999990999 000000000000 0000000910001 000000000000 47 F. guinlei 0000000000000 00000911201000000 000009000010 000000000000000 19201000000000 000000000190 000000000100 0000000910001 000000000000 48 F. cooleyi 0000000000000 00000910110000000 000009000000 001000000000000 19201000000000 000000000190 000000000100 0000000910001 000000000000 49 F. decipiens 0000000000000 00000910110000000 000009000000 001000000000000 19201OOGOOOOOO 000000000190 000000000100 0000010110000 010000000000 50 F. quadratlceps 0000000000000 00000910201000000 000009000010 000002010000000 19201000000000 000000000190 000000000100 0000000910000 010000000000 51 F. vulpis 0000000000000 00000910201000000 000009000010 000002010000000 19201000000000 000000000190 000000000100 0000000910000 010000000000 52 F. acutlrostris 0000000000000 01000911201000000 000009000010 000000000000000 19000100000000 000000000190 000000000000 0000000910001 000000000000 53 F. macrurus 0000000000000 01000911201000000 000009000010 000000000000000 19000100000000 000000000190 000000000000 0000000910001 000000000000 54 F. pygidielis 0000000000000 01000911201000000 000009000010 000000000000000 19000100000000 000000000190 000000000000 0000000910001 000000000000 55 F. bedford1 0119999000900 00000591201999000 210109000000 000000000000001 99999999999999 999999990999 000000000000 0000000910001 000000000000 56 L. bengalensis 0011100200000 00000910110000000 301009000000 000000000000000 19201000010000 000000000000 000000000000 0000000910001 000000000000 57 L. philippinersls 0011100200000 00000910110000000 301009000000 000000000000000 19201000010000 000000000000 000000000000 0000000910001 000000000000 53 U juccil 0001100200000 00000910000000000 301009000000 000000000000000 19201000000000 000000000000 000000000000 0000000910001 000000000000 59 L. aspidoirhynchu3 0001100200000 00000910000000000 000009000000 oooocoooooooooo 19201000000000 000000000010 000011100000 0000001100000 000000000000 60 L. sumatrensis 0001100200000 0000091cooooooooo 000009000000 000000000000000 19201000000000 000000000010 000011100000 0000001100000 000000000000 61 L. lenlcornls 0019999000000 00010110000110000 000009000000 000000000000000 10201001000000 000000000000 000011100000 0000000910001 000000000000 62 L. wemecki 0019999000000 00010110000110000 000009000000 000000000000000 10201001000000 000000000000 000011100000 0000000910001 000000000000 63 L. acuticeps 0019999000000 00010110000110000 0000090G0000 000000000000000 10201001000000 000000000000 000011100000 0000000910001 000000000000 64 L. afrlcanus 0019999000000 00010110000110000 000009000000 000000000000000 10201001000000 000000000000 000011100000 0000000910001 000000000000 65 L. neoafrlcanus 0019999000000 00010110000110000 000009000000 000000000000000 10201001000000 000000000000 000011100000 0000000910001 000000000000 66 L. laticeps 0000000000000 00000910000000000 000009000000 000000000100000 19201000001000 000000000190 000011000000 0000C10110000 010000000010 67 L. mungos 0000000000000 00000910000000000 000009000000 000000000100000 19201000001000 000000000190 000011000000 0000010110000 010000000000 68 L. parolatlceps 0000000000000 0000091cooooooooo 000009000000 000000000100000 19201000001000 000000000190 000011000000 0000010110000 010000000010 70 L. fells 0000000200000 00000910000000000 110009000000 000000000000000 19201000000000 000000000000 000000000000 0000010112101 000000000000 0000000200000 71 L. spencerl 0000091cooooooooo 110009000000 000000000000000 19201000000000 000000000000 000000000000 0000010112101 000000000000 0000000200000 72 L. hercynianus 00000910000000000 110009000000 000000000000000 19201000000000 000000000000 000000000000 OOGOOIOt12101 000000000000 0000000200000 73 L. 3 i aniens 13 00000910000000000 110009000000 000000000000000 19201000000000 000000000000 000000000000 0000010112101 000000000000 0019999000000 74 L. malaysionus 00000901201000000 110009000000 000000000000000 09110000000000 000000000000 000000000000 0000001100000 000000100000 0019999210000 75 L. mjobergl 00000901201000000 000009000000 000000000000000 19000000000000 000002110000 000000001000 0000010100000 000000100010 0019999010100 76 G.(Geomydoecus) 00000901000000000 210010000000 000000000000000 19000000000000 000000000000 000000000000 0000011101000 000000000900 0019999010100 G. mexlcanus 00000901000000000 210110000000 000000000000000 19000000000000 000000000190 000000000000 0000011101000 000000000900 77 0019999010100 78 G. perotensis 210110000000 000000000000000 19000000000000 000000000190 000000000000 0000011101000 000000000900 0019999010100 00000901000000000 0. coronadoi 210010000000 19000000000000 000000000000 0000011101000 000000000900 79 0019999010100 00000901000000000 000000000000000 000000000000 80 0. fulvescens 00000901000000000 210010000000 000000000000000 19000000000000 000000000000 000000000000 0000011101000 000000000900

00 ro 583

O 0 O O — — — « Lo O 0 O O *— 888888822888888 8 L CM CM CM CM CM O O O O CMCMCMCMCMCMCMOOCMCMCMCMCMCMCM T— •— <0 L— *"L 0O O O O O O O O O O O O o o o o O OSOQQQQOOQ O O O O O O O 8 OOOOO ^ O OOOOOOOOOO CN ON CNO N ON CN CN CN GN CT\ ON GN 0008 =r QN S CN ON ^ ON CN CN O CN ON CN CN CN ON CN CNC N CN 8 ON CN CN CN ON ON CN O ON O ON CN _ 'JN CN CN CN CN ON ON ON ON CN CN (TV O O O O O OCN OO CONN C NO NO O OW CN O^N OCN CON OONN CONN OCNN O8N ON m ON ON GN ON CN »— ^ CN CN— *—»—*—»— — *—*— GNQGNONCNCN'— CNCNCNCNON CN CN CN CN OCN CN CON OCN CONN CN — — — — — — ON- GNGNOOGNCNCN — CN CN CN ON CN to CN 8 ON CNO N CN CN CN CN CN ON 88 CN — — — t— r- -w— r- »— CN CN CN ON ON 0 — 2,— 22222oooooo2ooooo OOOOOOCNOCNGNOOOCNCNOCN CN CN CN CN 222222828822888288888 --—--«--*--.— .- — T- — *-,— OOOOO OOOOOOOOOOOOOOOOOOOOO CN CN CN CN ON CN CN ON CN ON CN ON CN CN CN CN CN CN ON ON CN 88888888888 88888 88888 00000000000 00000 00000 O ON CN CN r— CN CN CN ON CN CNOOOOOOOOCNO O ON ON ON — CN ON CN CN CN moooooooogo 2 8882 8 8 8 8 8 ON QN CN CN O cnonqnoncncncncnoncno CN O O CN O _ CNCNCNCNCNCNCNCNCNO O ON «— .— CN *— OOO 000 O O O O «— 0000 CN CN CNO N CN CNC N ON ON ON CN Cncncncncncncncncnt- — - ss 888 88888 ooooooooor-.- 2228000000 o o OOO 00000 0000000S522 OQ080OOOOO CN CN CN CNC N CN CN on — o o *— O O CN *— r— T— *— T— T— 82888 888 00000000000 — — O CM CMC M CM CMg N O— Or- — T-*-T-—*-0000*- 2oo22 2 2222 2 8 8 8 2 O25585oooqo555O22222222228882222282 OOOOo O O Q Oo ~ ~ O O CN o O ON ON CN OO N CN CN CN CN 2228888888§ O O ON — O ON CN CN ON CN CN JN CN o O CN o CN CN CN CN "TN 2 8882 O O O O Q §8888§§§8§5 OOOOO OOOOO 00000000000 OOOOO 00000 ooooogogooo OOOOOQOOO OOOOO o o _ _ § § 00000008 OQpQOQpOOOOQOQQo "88888800000002 000000000000 _ S S2 S 3 0 o 000000 8OOOpOOCNONCNONCNCNCN 8882§2§|§|§gg ONONCNCNCNCNCNCNCNONCNCNC~ — ~ — - - CN ON CN C>N J N

OOOOOOOOOOOO 0000 0000000 0000 OOOOOOO O O Q O O OO 8888 OOOOO

-C—" C M-C K N- C-C - CLf N- CNO - C -3CO" -CCN? O-C . —UN C MUN N" U\N - UCN UUNN N OUNUNUNUNUNNOVOCOVOUJV f^CO CNO — CM KN-C m OO N O — CM UN"? ICNVO NCO 0\0 NO r-r^r-i^-r-r^-r^-o-r- r-co 00 LTN 000000000 0000000000 ooonooooo 0000000000 00000000to OOOtlOllOOllOltOOOOO OOOOO100000 OOOOO1000000000000 08 000000000 0000000000 oontooooo 0000000000 00000000to COOllOllOOtlOllOOOOO ooooo 100000 ooooo toooooooocooo 6Z. 000000000 0000000000 ootnooooo 0000000000 00000000 to OOOllOllOOllOllOOOOO OOOOO 100000 OOOOO1000000000000 8L 000000000 0000000000 0000100000 0000000000 00000000to OOOllOltOOllOtlOOOOO OOOOO100000 OOOOO1000000000000 LL 000000000 0000000000 OOlllOOOOO 0000000000 00000000to OOOllOllOOllOtlOOOOO OOOOO100000 ooooo1000000000000 9L tlllOtlll 0000001 too OOlllOlOlO toioiotoii iiimmo 00112112000000000000 00011200000 00001C200011112112 SZ. tacooim ocoooootoo 0001 to to to to to to 1000 llllioooto 00000112000000000000 00031200000 OOCO 19200112112112 7Z. 1000°1100 COOOCQltOQ 0011too©to to to to toco llllliooio 00990I12009909000990 0096000090? COCO 1929 9090112112 6Z. toooottnn OOP"!?? I too0011100 0 to to to to too? ntlltmto 000001120999900C009C 00969099900 00001' '299000011009 2Z tutotm icocoottoo oomototo to to to 1000 tmmoio 000001 tsooooocoooooo 09069900000 0000l02c0pf,c0llll2 tZ. lllioilll toocoottoo oonto to to to to to to to llllllioto 00990112003CCOOOOOCO 00069909000 neon IC2'0000112112 CZ. tltOOUll OOOOOllOC CO 111000 to to to to to to llllllioio 00000112009:90000000 OOO6COOCOCC 0090losoono?112112 69 oooooocco C00009U09 OOlllOtOlO to to to 1000 nmoooto cooooooooooooooocooo 00060000:00 0909192000111 let 12 89 00000 IOC? CCOCCCllOl llllioioio to to to 1000 tilitiooic oooooItscncroooooooo OOO6OCOCOO9 COCO 192^0911112112 £9 090009000 coooccttoo octtioicio tc tc totooc 11 mono to coocoooooooooooooooo COOGOCXCCO 000010209011112112 99 000001111 00000 t to tllttllll l ItlllllOlO lltttlioto ooooo t teooooooocoox 00091200000 009010290090 112112 $9 000001110 0000000000 ooootmi ntllllOOO llltllioio 00000112000000000000 00001200000 0000102C00C9011011 79 OCOCOttll 0000001to I imittti timnoto lltlltioio OOOOO112000000000000 000012000C0 000010200009112112 £9 111001000 ooccoot too OOlltltll lllllllOlQ mutton OCCOO112000000000000 00001200000 000010200000000000 29 moon to 0000000000 ooootim iimnoio iminoto OOCOCI12000000000000 00001209000 00001020C900000000 19 1 iooo mo cocci it too lllllOOlt nttioiooo llttliooio OOOCOI12000C00000000 00001200000 000010200000112112 09 1100011 to OOOOllltCO niiioon III 1101000 ttttitooto ocooo 112000000000000 00001200000 000610200000112112 66 moot t toooooootoc o 00to I to too tototoicoo lllltiooio OOOOO 112000000000000 00001200000 000010200000112112 85 110001000 0000001000 00 to I too 10to to to 1000 tmnooio OOOOO 112000000000000 0000120C000 COOO10200000112112 Z.S 11000 too" oooooo1000 00101 too to to to to toco tmnooto OOOOO112^00000000000 00001200009 OCOO10200000 U2112 96 t tooooooo000000 1 too oomoioto toioiotoio mtinoio OOOOO 11200000C000009 00001200000 000110200000112112 65 iioocoooo ooooocotco 0000 to to to to to to to to mmtoic 00000 112000000000000 00001209010 000110200011112112 +?5 IIOOCOOOO 0000000 too oocoto 10 to to to to to tomiiuoi o OOOOO112090000000000 00001200010 000110200011112112 65 Itoooooco oooooootoo 0000 to to to to 10 to to tolllllliot o OOCOOOOCOOOOOOOOOOOO 900C 1200009 000110200011112112 25 moot too 0000000 too 00 mo to to to to to to to inimoto oocoooiioeuoroooooo 10060009 ICO CO 1 tootcooooo 11011 IS it too I tooccoooo o too OOlllOlOlO to to to to to tilt mo to 0000001 too 1100000000 10060000 ICO 001100100009011011 OS I toooooco 0000001 too OCCO100000 0000000000 00000000 to OOOOO112000000000000 10001200100 001 too 100000112011 gtj 11toooooc 000^001 too OOCO100000 ooocoooooo 00000000 to OOCOCI1200C000000000 100C1200 ICO 001 too 100000112011 87 II too 1000ooooocooo o 000010 to to to to to 1000lltlliooi o cocooooooooocooooooo 10060000100 COllCOtOQOOOOttOll b] lllccooco 0000000000 ocoooo to to to to to to to 00000000 to oooooonoooooocoocoo 10060000100 001 ICO too 112112112 97 tocoooooo oooooottoo OOlltOOOOl 0101 to tooo initnoio OCOOOI1200C0C0000000 00001200000 000110200112112112 57 tliootlll ooccoootoo 0001 to to tot o to to to to llllllioio ooooooitoonoooooooo 00001200000 000110209000112112 tft t10001000 oooooot too OOlllOlOlO to to to to to mtmoio OOOOO 11200000000C/-00 00001200000 000110200000112112 6*7 111100000 0000000000 OOOOllOlll ItlOlOlOlO nooooooio OOOOO112000000000000 00001200000 000110200000112112 27 it too 1000roooo o t tooooiimi n lltotoilio iimnoio OOOOO112000000000000 00001200000 000110200000112112 17 U U I I I I t—I I I I I I I I—I I—I \LZ 192 192 1*2 IC2 'B ITLLI 's 122 9 VZ HZ 'soz' 'tOZZOZiOZ 18 19 20 21 31 36 39 41 43 50 51 56 61 65 71 73 78 81 83 r7==! 81 G. dakotensls 0019999010100 00000901000000000 210110000000 000000000000000 19000000000000 001000000190 000000000000 0000011101000 000000000900 82 G. duche sirens Is 0019999010100 00000901000000000 210110000000 000000000000000 19000000000000 001000000190 000000000000 0000011101000 000000100900 83 G. thomomyus 0019999010100 00000901000000000 210110000000 000000000000000 19000000000000 001000000190 000000000000 0000011101000 000000100900 84 G. (T)wardl 0000000000100 00000901000000000 210110000000 000000000000000 19000000000000 000000000000 000000000000 0000011101000 000000100900 85 G. (Thomoraydoecus) 0019999200100 00000901000000000 210110010000 000000000010000 19000000000000 000000000000 000000000000 0000011101000 000000100900 86 N. mlnutus 0019999010100 00000901201000000 210011000000 000000000000000 19000000000000 000000000000 000000000000 0000001101000 000000000000 87 N. osbornl 0019999010100 00000901201000000 210011000000 000000000000000 19000000000000 000000000000 000000000000 0000001101000 000000100000 88 N. thoraclcus 0019999010100 00000901201000000 210011000000 000000000000000 19000000000000 000000000000 000000000000 0000001101000 000000100000 89 H. mephitidis 0019999010100 00000901201000000 210011000000 000000000000000 19000000000000 000000000000 000000000000 0000001101000 000000100000 90 N. woirfhuegell 0019999010100 00000901201000000 210011000000 000000000000000 19000000000000 000000000000 000000000999 9999901199999 999990999999 91 H. semlstrlatus 0019999010100 00000901000000000 210010000000 000000000000000 19000000000000 000000000000 000000000010 0000001111000 101100000900 92 N. arlzonae 0019999010100 00000901000000000 210010000000 000000000000000 19000000000000 000000000000 000000000010 0000001111000 101100000900 93 N. chilensIs 0019999010100 00000901000000000 210010000000 000000000000000 19000000000000 000000000000 000000000010 0000001111000 102110000900 94 N. lnterruptofasciatus 0019999010100 00000901000000000 210010000000 000000000000000 19000000000000 000000000000 000000000010 0000001111000 102110000900 95 N. cuinmlngsl 0019999010100 0000C901000000000 210010000000 000000000000000 19000000000000 000000000000 000000000010 0000010011010 100000000900 96 N. gastrodes 0019999010100 00000901000000000 210010000000 000000000000000 19000000000000 000000000000 000000000010 0000010011010 100000000900 97 N. pallldus 0019999010100 00000901000000000 210010000000 000000000000000 19000000000000 000000000000 000000000000 0000000912101 1000000000900 98 N. barbarae 0019999010100 00000901000000000 210110000000 000000000000000 19000000000000 000000000000 000000000000 0000000900000 000000000100 99 E. paradoxus 0019999200000 00000910201000000 000009000000 000000000000000 19201000000000 000000000000 000000000000 0000000900000 000000100900 100 C. machadoi 0019999200000 00000911000000000 000009000000 000000000000000 19110000000000 000000000000 000000000000 0000000900000 000000100900 101 0. (D) dendrohyracls 0011100200000 00000910110000000 000009000000 000000000100000 19000010000000 000000000000 000000000000 0000010110000 000001000900 102 D. (D) gulneensis 0001100200000 00000911110000000 000009000000 000000000100000 19201010000000 000000000000 000000000000 0000010110000 000001000900 103 D. (D) hopkinsi 0001100200000 00000910201000000 000009000000 000000000100000 19201010000000 000000000000 000000000000 0000010110000 000001000900 0019999000000 104 D. (D) minor 00000911201000000 000009000000 000000000100000 19201010000000 000000000000 000000000000 0000010110000 000001000900 0001100200000 105 D.(D ovalis 00000910000000000 110009000000 000000000100000 19201010000000 000000000000 000000000000 0000010110000 010000000900 0019999200000 106 D.(D smallwoodae 00000910201000000 000009000000 000000000100000 19201010000000 000001100000 000000000000 0000099999909 090009000900 0011100200000 107 D.(D ugandensls. 00000911201000000 110009000000 000000000100000 19201010000000 000000000000 000000000000 0000010110000 000001000900 108 D.(D validus 0001100200000 00000911201000000 110009000000 000000000100000 19201010000000 000000000000 000000000000 0000010110000 000001000900 0019999200000 109 D.(N caponsis 00000910201000000 110009000000 000000000100000 19201010000000 000000000000 OOOOOOOOOGOO 0000010110000 010000000900 0019999200000 110 D;(N diacanthus 00000910201000000 1100G900C000 000000000100000 19201010000000 000000000000 000000000000 0000010110000 010000000900 0001100200000 111 D.(N nalroblcnsis 00000910201000000 ooooosoooooo 000000000100000 19201010000000 000000000000 ooocoooooooo 0000010110000 010000000900 0011100200000 112 D.(M ) ruficep3 00000910201000000 000009000000 000000000100000 19201010000000 000000000000 000000000000 0000010110000 010000000900 0011100200000 113 D.(M ) transvaalensis 00000910201oooocc 000009000000 000000000100000 19000010000000 000000000000 000000000000 0000010110000 010000000900 0019999200000 114 D.(N ) waterburgensis 00000910000000000 000009000000 000000000100000 19201010000000 000000000000 000000000000 C000010110000 010000000900 0000000000001 115 P. (M) africanus 00000910000000000 000000000000000 19201GOOOOCOOO 000000000000 000000000011 0000011110000 000000000900 0U00000000001 210009000000 116 P.(M ) aneolensls 0J000910201000000 210009000000 000000000000000 19201000000000 000000003000 000000000011 0000011110000 000000000900 0000000000001 ) colobi 19201000000000 000000000000 000000000011 0000011110000 000000000900 117 P.(h 0019995000001 00000910000000000 110009000000 000000000000000 } congoensis 19201000000000 000000000000 0000011110000 000000000900 118 P.(M 0000000000001 00000910000000000 110009000000 000000000000000 000000000011 ) muesebeckl 19201000000000 000001100000 0000011110000 000000000900 119 P.(M 0000000000001 00000910000000000 110009000000 000000000000000 000000000011 120 P.(M ) baculatus 00000911201000000 110009000000 000000000000000 19201000000000 000000000000 000000000011 0000011110000 000000000900 vn CO VJ1 91 | 99 [ ^ 101 111 121 131 141 151 161 164^ 166| <168| 171 174 < 176[ 181 t 18^ < 1HH t 191 194f 81 1010000yXJv 5399399999939 9900000001 1000095900 1111101011 0090999009 9010099011 9995990oTn 2,00,1 0 a'ooooo'ooo'oooo'oc^joi 0 OOOcSxjOOl 0*00^09 ooododi 11111 82 101000099999 9999999999999 9900000001 1000099900 1111101011 0090999009 9010090011 9999990011120010 200000000000000000010 000000001000009 000000111111 83 101000099599 9539999999999 9900000001 1000059900 1111101011 0090999009 9010090011 9999990011120010 200000000000000000010 000000001000009 000000111111 84 100000099999 9999399999999 9900000001:1000099901.1209901911 0090999009 9010090011 9999990011120010 200000000000000000010 000000001000009 000000111111 85 100000099999 9999999999999 9900000001 1000099901 1209901911 0090999009 9010090011 9999990011120010 200000000000000000010 000000001000009 000000111111 86 090000101101 0000099009900 0000000001 1000099900 1111101011 1090999009 9010090011 9999990011120010 200000000000001100000 000000001000009 000000111111 87 090000101101 0000099009900 0000000001 1000099900 1111101011 1090999009 9010090011 9959990011120010 200000000000001100000 000000001000009 000000111111 88 090000101100 0000099009900 0000000001 1000099900 1111101011 1090999009 9010090011 9599990011120010 200000000000001100000 000000001000009 000000111111 89 090000101101 0000099009900 0000000001 1000099900 1111101011 1090999009 9010090011 9999590011120010 200000003001001100000 000000001000009 000000111111 90 999393935999 9999999959999 9999999999 999C059900 1111101011 1090999009 9010090011 9599990011120010 200000003001001100000 000000001000009 000000111111 91 090000099999 9999999999999 9900100001 1100099900 1111101011 1090999009 9010090011 9399990011120010 200000003001001100000 000000001000009 000000111111 92 090000099999 9999999999999 9900100001 1100099900 1111101011 1090999009 9010090011 9999990011120010 200000003001001100000 00000000100C009 000000111111 93 090000099999 9999999999999 9900000001 1100099900 1111101011 1090999009 9010090011 9999990011120010 200000003001001100000 000000001000009 000000111111 94 090000099999 9999999999999 9900000001 1100099300 1111101011 1090999009 9010090011 9999990011120010 200000003001001100000 000000001000009 000000111111 95 100000099999 9999999999993 9900000001 1100099900 1111101011 1109000990 9010090011 9999990011120010 200000002011001100000 000000001000009 000000111111 96 ; 100000099999 9999999999999 9900000001 1100.099900 1111101011 .1190999009 9010090011 9999990011120010 200000002011001100000 000000001000009 000000111111 97 090000099399 9999999999999 9900000001 1100099900 1111101011 1090999009 9010090011 9999990011120010 200000003001001100000 000000001000009 000000111111 98 100000110009 0000099121000 0100000000 1000099900 1109901911 1090999009 9010090011 9999990011120010 200000003001001100000 000000001000009 000000111111 99 101100099999 9999999999999 9900000000 0000110001 0111101010 9991011009 9011100110 1112110200120010 200000000000001100010 001211001000009 000000000000 100 101100099999 9999999999999 9900000000 0000110012 1011101010 9991011009 9011100110 1112110200120010 200000000000000000010 001211001000009 000000000000 101 090100099999 9999999999999 9900000000 0000110010 1111101011 0101010009 9011100112 1112110200120010 200000000000001100000 001110001210109 000000000000 102 090100099999 9999999999999 9900000000 0000110000 1111101011 0111011009 9011100112 1110000200120010 200000000000001100000 001OOOOOI210109 000000000000 103 090100099999 9999999999999 9900000000 0000110000 1111001011 0101011009 9011100112 1111100900120010 200000000000001100000 001000001210109 000000000000 104 090100099999 9999999999999 9900000000 0000100000 1111101001 0101011009 9011100112 1110000200120010 200000000000001100000 001000001210109 100000000000 105 090100099999 9959999999999 9900000000 0000101010 1111101011 0101011009 9011100112 1111100200120010 200000000000001100000 001000001210109 10000000Q000 106 030100099959 9999999999999 9900000000 0000110012 1111101011 0101011009 9011100112 1111100200120010 200000000000001100000 001000001210109 000000000000 107 090100059999 9999999999999 9900000000 0000110010 1111101011 0101010009 9011100112 1112110200120010 200000000000001100000 001000001210109 100000000000 108 090100099999 9999999999999 9900000000 0000110010 1111101011 0101010009 9011100112 1112110200120010 200000000000001100000 001000001210109 100000000000 109 030100099999 3999999999999 9900000000 0000110010 1111101011 0111011009 9011100112 1112110200120010 200000000000001100000 001000001111009 000000000000 110 090100099599 9999999999999 9900000000 0000110010 1111101011 0111011009 9011100112 1112110200120010 200000000000001100000 001000001111009 000000000000 111 090100059939 9995559999999 9900000000 0000110000 1111101011 0111011009 9011100112 1112110200120010 200000000000001100000 001000001111009 000000000000 112 090100099999 5999999999999 9900000000 0000310010 1111101011 0111011009 9011100112 1112110200120010 200000000000001100000 001000001111009 000000000000 113 090100099999 9999999999999 9900000000 0000110010 1111101011 0101011009 9011100112 1112110200120010 200000000000001100000 001000001111009 000000000000 114 090100099999 9999999999999 9900000000 0000110010 1111101011 0111011009 9011100111 1110000200120010 200000000000001100000 001000001111009 000000000000 115 101100099999 9999999999999 9900000000 0000110012 1111101010 9991011005 9011100112 1111100200120010 200000000000001101000 000000001000011 000110000000 116 101100039999 9999599359999 9900000000 0000110012 1111101010 9991011009 9011100112 11111002G0120010 200003000000001101000 000000001000011 000110000000 117 101100099999 9999999999999 9900000000 0000110012 1111101010 ^991011009 9011100112 1111100200120010 200000000000001101000 000000-.01000011 000110000000 118 101100099999 9999999999999 9900000000 0000110012 1111101010 9991011009 9011100112 1111100200120010 200000000000001101000 C00000001OOCOI1 000110000000 119 101100039999 9995999999999 9900000000 0000110012 1111101010 9991011009 9011100112 1111100200120010 20000000Q000001101000 000000001000011 000110000000 120 030101099999 9999999999999 9900000000 0000110012 1111101010 9991011009 9011100112 1111100200120010 200000000000001101000 000000001000011 000110000000 vji 00 ON ^1^202 [203| 206[ 211 216 221 |223|[224[)225| 227(|226| 231 241 251 261 271 , 8.1 000000000000100000 00000100000 000S2l72U 02lT2l700 0100000000 0000000000 0000011100 0111000000 000000000 82 000000000000100000 00000100000 00000211211021121100 0100000000 0000000000 0000011100 0111000000 000000000 83 000000000000100000 00000100000 00000211211021121100 0100000000 0000000000 0000011100 0111000100 000000000 84 000000000000100000 00010100000 00000211211021121100 0100000000 0000000000 0000011100 0011000000 000000000 85 000000000000100000 00010100000 00000211211021121100 0100000000 0101010000 0000011100 0011000000 000000000 86 000000000000100000 00000100100 OOOOOO00000000000000 0100000000 0000000000 0000000000 0000000000 000000000 87 000000000000100000 00000100100 00000000000000000000 0100000000 0000000000 0000000000 0000000000 000000000 88 C000000000001ooooo 00000100100 00000000000000000000 0100000000 0000000000 0000000000 0000000000 000000000 89 0000000000001ooooo 00000100100 00000000000011000000 0100000000 0000000000 0000000000 0000000000 000000000 90 0000000000001ooooo 00000100100 00000000000000000000 0100000000 0000000000 0000000000 0000000000 000000000 91 0000000000001ooooo 00000100100 00000000000000000000 0100000000 0000000000 0000000000 0000000000 000000000 92 ocooooooooooiooooo 00000100100 00000000000000000000 0100000000 0000000000 0000000000 0000000000 000000000 00000100100 00000000000000000000 0100000000 0000000000 0000000000 93 000000000000100000 0000000000 000000000 00000100100 00000000000000000000 0100000000 0000000000 0000000000 000000000 94 000000000000100000 0000000000 00000100100 00000000000000000000 0100000000 0000000000 0000011100 000000000 95 000000000001oooooo 0110000000 96 000000000031000000 00000100100 00000000000000000000 0100000000 0000000000 0000011100 0110000000 000000000 97 000000000000100000 00000100100 00000000000000000000 0100000000 0000000000 0000000000 0000000000 ooooocooo 98 OOOOOOOCOOOOIooooo 00000100100 00000000000000000000 Ot10000000 1000000000 0000000000 0000000000 OOOOOoOOO 99 000000000002010000 00000210010 00000000000021121110 0101111111 0111111111 1111011111 11111111 00000210010 00000000000021121110 0101111111 0101111111 1101011111 100 000000000002010000 01111100 101 000000000002010000 00000210000 00000000000021121100 0101111111 0101111111 1111011111 111111 00000210000 00000000000021121100 0101111111 0101111111 1111011111 102 000000000002010000 111111 103 000000000002010000 00000210000 00000000000021121100 0101111111 0101011111 1111011111 111111 00000210000 00000000000021121100 0100111100 0101010101 0101011100 io4 000000000002010000 001111 00000210000 00000000000021121100 0101111111 0101111111 1111011111 105 0C00000000020100G0 111111 C0000210000 00000000000021121100 0101111111 0101111111 1111011111i 106 000000000002010000 111111 00000210000 OOOOOOOOCOOG21121100 0101111111 0101111111 1111011111 107 00OOOOOG0002010000 111111 0000021OCOO 0C0000000C0021121100 0101111111 0101111111 1111011111 000000000002010000 111111 108 0000021COOO 0C000000000021121100 0101111111 0101010111 1111011111 00U0CC0GC001OOuOOO 111111 109 00000210000 00000000000021121100 0101111111 0101010101 1111011111 0CC000000G01oooooo 111111 110 00000210000 G0CC0000GC00211211uO 0101111111 0101011111 1111011111 111 000000000001000000 111111 112 OOCCGGCG0002010000 00000210000 00000000000021121100 0101111111 01010101 el 1111111111 111111 113 G00C0C00G002G1COOO 00000210000 000000000000211211 GO 0101111111 0101011111 1111011111 111111 114- 000000000002010000 00G00210000 00000000000021121100 0101111111 0101011111 1111011111 111111 115 000000000002010000 0000021OCOC 00C0QQ0G0QG021121100 0101111111 0101010101 1111011111 111111 116 00000000000201OooO 00000210000 00000000000021121100 0101111111 0101010101 1111011111 111111 117 00000000000201ocoo GOOOC210000 00000000000021121100 0101111111 0101010101 0101011111 111111 118 000000000002010000 00000210000 00000000000021121100 0101111111 0101010101 0101011111 111111 119 000000000002010000 00000210000 00000000000021121100 0101111111 0101010101 1101011111 111111 120 000000000002010000 00000210000 00000000000021121100 01011111^1 0101010101 0101011111 111111 —0]0 oo oo IfN 006000000000 00000 UOOOOOO 000000000000 061000000000 00000100000001 000000000000000 000000600000 000000000111llOOO 00000000 llOOO sc'aoipjoo *3 091 006000000000 ooooo uoooooo 000000000000 061000000000 00000100000001 000000000000000 00000060C0C0 000000000till 1000 0000000011000 SUBQJUIOO '3 651 006000000000 ocooo 1 tooooooOOOOCOOOOOO O 061000000000 oooco tooocoooI ocooooooooooooo 0000C0600000 ooocococou utoco 00000000ttooo saqotoojao '3 g£i OO6OCOCOOOOO OOOOO11oooooo cooooooooooo 061000000000 oooootooooocot 000000000000000 000000600000 OOOCOOOOO lilt toco 0000000011000 smroojxaui *3 Z.SI 006000000000 ooooottcooooo 000000000000 061000000000 oocootooooooot 000000000000000 000000600000 000000000u I llOOO 0000000011000 snqaqoi '3 961 016000000000 ooooo 11000000 000000000000 061000000000 00000 100000001 000000000000000 0CC000600000 OOOOOOOCOtlU1000 0000000011000 snnSjxa '3 551 016000000000 0001060000000 OOOOCOOOOOCO 061000000000 00000100000001 000000000000000 C00000600000 000000000lilt tooo 0000000011000 JOUIUI *3 tjgi 006000000000 00G0C60000000 COOOOOOOOOOO 061000000000 00000000000001 000000000000000 000000600000 oooocoooo tin 1000 000000001 tooo snsoqas *3 fSt 016000000000 tooo 160000000 OCOOOOOOOOOO 061000000000 00000000010201 000000000000000 000000600000 0000000000111 tooo 0000020011000 sniJ3Jtqx9 '0. 25t 006000000000 OOOOOIICOOOOO cooooooooooo 061000000000 00000000110261 coooooooocooooo 000000600012 00000000001600000 0000020011000 snqaiurejnas t£t 006000000000 oooootioooooo 000000000000 061000000000 00000000110261 000000000000000 000000600012 00000000001600000 0000026666000 snqsare *o oSl 006000000000 OCOOO uoooooo 000000000000 061000000000 00000000 U0201 000000000000000 000000600000 0000000000 to 11000 0000026666100 juoQjoqa (d)'d 6*7t 006000000000 OOOOO uoooooo 000000000000 061000000000 00000000110201 000000000000000 000000600000 0000000000 to 11000 0000026666100 sjsuaivimi (d)'d 871 006000000000 OOOOO uoooooo 000000000000 061000000000 00000000U020I 000000000000000 000000600000 0000000000 to 11000 0000026666100 jeaaajjoui (d)'d £tfl 006000000000 OOOOO UOOOOOO ooooooocoooo 061000000000 00000000110201 000000000000000 000000600000 00000000001011000 0000006666100 aqaujgjsnia (d)*d 9171 006000000000 OOOOO UOOOOOO 000000000000 061000000000 00000000110201 000000000000000 000000600000 0000000000 to 11000 0000026666100 MamojB (d)'d 571 006000000000 OOOOO UOOOOOO 000000000000 061000000000 00000000110201 000000000000000 000000600000 0000000000tottooo 0000026666100 061000000000 taoiuq (d)*d *7*7l 006000000060 6O666666OOCOO 000000000000 00000000110201 000000000000000 000000600000 00000066601666000 0000006666 ICO saauwas (d)»d £tjt 061000000000 006000000000 OOOOO UOOOOOO 000000000000 00000000110201 000000000000000 000000600000 0000000000 to 11000 00000200 u too sjsuDTaoqajd (d)*d stji 006000000000 OOOOO UOOOOOO 000000000000 061000000000 00000000t1020I 000000000000000 000000600000 00000000001011000!0000020011 too 061000000000 snupiA (d)'d 1*71 006000000060 6066666600000 000000000000 00000000110261 000000000000000 000000600000 00000010201600000 0000006666100 (d)'d 061000000000 sojnj'BAjred 071 016000000010 OOOOlllOOOOOO 010000000000 00000000110201 000000000000000 000000600000 0000000000 to I tooo 00000200it too isadox (d)#d 6£l ootooooooooo 006000000000 OOOOO uoooooo 000000000000 00000000110201 000000000000000 000000600000 0000000000 III tooo 0000026666100 sTujjr0 (d)'d 8£t 000000000000 006000000000 0000060000000 001000001000 00000010010211 000000000000000 000000600000 0000000000 to 1 tooo 0000020011 too IPI8IJUIT (0)*d Aft 000000000000 106100000000 0000060000000 00000000tooo 00000010010211 000000000000000 000000600000 0000000000 to 11000 0000026666100 isuwdoq (o)*d 951 000000000000 006100000000 ooooo utooooo 00000000tooo 00000010010211 000000000000000 000000600000 0000000000 to 11000 0000026666100 snjoijssjp (o)'d S£t 000000000000 006100000000 OOOOO UIOOOOO 000000001000 00000010010211 000000000000000 000000600000 0000000000 to 1 tooo 0000026666100 ipjojpaq (D)*d 7£t 000000000000 006000000000 ooooo10100000 010000000000 00000001010261 OOOOOI000000000 0000006000U 00000010201600000 0000026666100 tsajaqoj 'd ££l 000000000000 006000000000 ooooototooooo 010000000000 00000001010261 oooootoooooocoo 0000006000 u 000000102 U600000 0000026666100 tsiajvq'd 2ft 000000000000 006000000000 OOOOO10100000 010000000000 00000001010261 ooooo1000000000 0000006000It 00000010201600000 0000026666100 lapum 'J'd t£l 000000000000 006000000000 00000001010261 ooooo1000000000 000000600011 00000010201600000 0000026666100 ISJJJ3J *d Oft OOOOO10100000 010000000000 000000000000 006000000000 otoooooooooo 0000000to1026I ooooo1000000000 000000600011 00000010201600000 0000026666100 sapioinireaS -j'd 621 ooooo10100000 000000000000 too 00000001010261 000000600000 00000000001600000 0000006666100 snqt?inj0j3 *d 821 006000000000 ooooo10100000 010000000 000000000000 oooootooooooooo 00000001010261 ooooo1000000000 000000600000 00000000001600000 0000006666100 smqnp *d Z.21 006000000000 OOOOO10100000 010000000100 000000000000 00000001010261 ooooo1000000000 000000600000 00000010201600000 0000026666100 snotaaaaas (w)#d 921 006000000000 OOOOO10100000 llOOOOOOOOOO 000000000000 006000000000 OOOOO10100000 110000000000 00000001010261 ooooo 1000000000 000000600000 00000010201600000 0000026666100 staojaios (U)*d 521 000000000000 00000000010261 006000000000 OOOOOUOOOOOO 100000000000 000000000000000 000000600012 OOOOOO102tl600000 0000016666100 jajtanos (H)*d 721 000000000000 00000000010261 006000000000 OOOOO uoooooo 100000000000 000000000000000 000000600012 00000010211600000 0000026666100 Jireixiof (M)*d £21 000000000000 00000000010261 0060C0000000 0000I11100000 110000000000 000000000000000 000000600011 00000010201600000 1000006666100 lOJiapuaq (w)'d 221 coooooooooco 00000000010261 0060000C0000 0000111100C00 110000000000 000000000000000 000000600011 OOOOCO10211600000 1000006666100 luueranau (n)*d 121 t±l ti ce is 8 L e l u S9 19 9S IS OS I* 6C 9G IC IZ OZ 61 91 W U 91 ^ 99 ^ ^ 101 111 121 131 141 151 161 ^ 164^ 166[ 168 171 174 ^ 176< 181 184^ > 1Q8 • 191 194|

121 s£l591999999999 5900000000 0000110012 1111101010 9991011009 9011100112 111lTOo25ol1209999S9599959 9900000001 0000101010 1110100990 9990999009 9011100011 1190000011120011 200000000000001100000 000000001000009 010000000000 142 050100099595 9999999999999 9900000099 0000101010 1110100990 9990999009 9011100011 1190000011120011 200000000000001100000 000000001000009 010000000000 143 090100099999 9999599993999 9900000001 0000101010 1110100990 9990999009 9011100012 1111100999120011 200000000000001100000 000000001000009 010000000000 144 090100053999 9999999999939 9900000001 0000110000 0111101010 9S91110009 9011100112 1111100200120011 200000C00000001100000 000000001000009 010000000000 145 090100099995 9999999999395 9900000001 0000110000 0111101010 9991110009 S011100112 1111100200120011 200000000000001100000 000000001000009 010000000000 146 030100099599 9999999999959 9900000001 0000110000 0111101010 9591110009 9011100112 1111100110120011 200C00000000001100000 000000001000009 010000000000 147 090100099599 9999999999999 9900000000 0000110000 0111101010 9991110009 9011100112 1111100011120011 200000000000001100000 000000001000009 010000000000 148 090100095995 9993999939999 9900000000 C000110000 0111101010 9991110009 9011100112 1111100011120011 200000000000001100000 000C00001000009 010000000000 149 090100059999 9999999999999 9900000000 0000110000 0111101010 5991110009 9011100112 1111100110120011 200000000003001100000 000000001000005 010000000000 150 091100099999 9999999999999 9900000000 0000110010 1111101011 0111011010 0011100011 1190000200120010 2000000C0000001100000 000000001000009 001000000000 151 091100099599 9999999999539 9900000000 0000110010 1111101011 0111011010 0011100011 1190000200120010 200000000000001100000 000000001000005 001000000000 152 101000099999 9995995999999 9900000099 0000110010 1111101011 0111011009 9011100111 1010000011120010 2000C0000000001100000 000000001000009 001000000000 153 09000X99999 5599999999599 990000X00 000011X00 1110101110 9991111009 90111X011 1190X011012X11 2000000000X00110X01 000000101000009 OOOOOOOXOX 154 09X00099999 3999959995999 9900000000 000011X00 1111101011 0011011X9 9011100112 111000X11120011 2XXXOOXOOOI10X01 000000101000009 OOOCOCOOOOX 155 090X0099999 5959999595599 9900000000 00X110X0 1111X1111 0111011009 90111X112 111XX01112X11 2000000X000X11XXI 000000101000009 OOOOOOOCOOX 156 09000C095999 99S9S99S9S999 5900000000 0000110000 1111101011 0011011005 9011100112 1111100011120011 2X0000X0X00110X01 000000101000009 0000000X000 157 090X0099599 9999999999999 99X000000 00X110X0 1111101011 0011011009 90111X112 111110X11120011 2X000000000X1100X1 X00001010X005 XXOOOOOXO 158 09000X99995 9599999959999 99X000X1 0000110000 1111001111 0111011009 9011100112 1111100011120011 2X00X000000011X001 000000101000009 XOOCOXOOOO 139 090000099999 9999959999959 99XX0000 0X011X00 1111001111 0111011009 9011100112 11111X011120011 2X00X00000001100001 0000X1010X009 OOOOOOOXOX 160 090000059999 9595999999599 9900000000 0000110X0 1111X1111 .0111011009 9011100112 1111100011120011 200000X0000001*100X1 OOXXIOIOOOOO9 OOOOOXOOOX ^ vo ^201^202^203^ 206^ 211 216 221 223^224^2^ 227 228 231 241 251 261 271 121 000000000(X)20?0000 00000210000 OOOOOOOOOOOO2T12T?00 010 0101010101 01010111 1 1111111111 122 000000000002010000 00000210000 00000000000021121100 010 0101010101 01010111 1 1111111111 123 000000000002010000 00000210000 00000000000021121100 010 0111111111 11110111 1 1111111111 12k 000000000002010000 00000210000 00000000000021121100 010 0101010111 11111111 1 1111111111 125 000000000002010000 00000210000 00000000000021121100 010 0001111111 11110111 1 1111111111 126 000000000002010000 00000210000 00000000000021121100 010 0001010101 11000111 1 1111111111 127 000000000002010000 00000210000 00000000000021121100 010 0101010101 11111111 1 1111110011 128 000000000002010000 00000210000 00000000000021121100 010 0101010101 11111111 1 1111110011 129 000000000002010000 00000210000 00000000000021121100 010 0101010101 01011111 1 1111111111 130 000000000002010000 00000210000 00000000000021121100 010 0101010101 01011111 1 1111111111 131 000000000002010000 00000210000 00000000000021121100 010 0101010101 01011111 1 1111111111 132 000000000002010000 00000210000 00000000000021121100 010 0101010101 01111111 1 1111111111 133 000000000002010000 00000210000 00000000000021121100 010 0101010101 01011111 1 1111111111 13k 000000000002010000 00000211000 00000000000021121100 011 1111111111 11111111 1 1111111111 135 000000000C02010000 00000211000 00000000000021121100 011 1111111111 11111111 1 1111111111 136 000000000002010000 00000211000 00000000000021121100 011 1111111111 11111111 1 1111111111 137 000000000002010000 00000211000 00000000000021121100 011 1111111111 11111111 1 1111111111 138 000000000002010000 00000210000 00000000000021121100 010 0101011111 11110111 1 1111111111 139 000000000002010000 00000210000 00000000000021121100 010 0101111111 11110111 1 1111111111 Iko 000000000002010000 00000210000 00000000000021121100 010 0101111111 11110111 1 1111111111 tkl 000000000002010000 00000210000 00000000000021121100 010 0101011111 11110111 1 1111111111 142 000000000002010000 00000210000 00000000000021121100 010 0101111111 11110111 1 1111111111 Ik3 000000000002010000 00000210000 00000000000021121100 010 0101011111 11110111 1 1111111111 ikk 000000000002010000 00000210000 00000000000021121100 010 0111111111 11110111 1 1111111111 Ik5 000000000002010000 00000210000 00c00000000021121100 010 0111111111 11110111 1 1111111111 Ik6 000000000002010000 00000210000 00000000000021121100 010 0111111111 11110111 1 1111111111 1k7 000000000002010000 00000210000 00000000000021121100 010 0111111111 11110111 1 1111111111 1k8 000000000002010000 00000210000 00000000000021121100 010 0101011111 11010111 1 1111111111 1k9 000000000002010000 00000210000 00000000000021121100 010 0101111111 11110111 1 1111111111 150 000000000001000000 00000211000 00000000000000021100 010 0101111111 11110111 1 1111111100 151 000000000002010000 00000210000 00000000000000021100 010 0101111111 11110111 1 1111111100 152 000000000002010000 00000210000 00000000000000021100 011 0101010111 11110111 1 1111111111 153 000000000011000000 00000210000 00000000000000000000 010 0101010111 11110111 1 1111111111 15k 000000000011oocooo 00000210000 ooocoooooooooooooooo 010 0101010101 01010111 1 1111111111 155 000000000011000000 00000210000 00000000000000000000 010 0101010111 11110111 1 1111111111 156 000000000011000000 00000210000 00000000000000000000 010 0111111111 11110111 1 1111111111 157 000000000011000000 00000210000 oooooooooocooooooooo 010 0101010101 11110111 1 1111111111 158 000000000011000000 00000210000 00000000000000000000 010 0101111111 11010111 1 1111111111 159 OOOOCOOOOOI1000000 00000210000 oooooooooooocooooooo 010 0101111111 11110111 1 1111111111 160 000000000011000G00 00000210000 00000000000000000000 010 0111111111 11110111 1 1111111111 591

o o o o o o Oo oO oO o o o CM CM CM CT\cn CTcn\ G"\ con cn on

H s

0) CD 03 3 •o § a u o01 23 01 h> w oi — hi C o to 01 01 01 iH 01 3 a **> 01 —4 o 0) >> s) <0 01 |H «p c c |H G 0 a 0} "O a O h> 0) 01 a hi 01 —4 G o r>> o G G I A 2 S 01 > TJr Hr H 01 01 o O 23 01 c o « 01 X3 41 |H oi 0) G 01 *P a tP a ® 23 23 XI rH 03 O rH G 41 w iH u x: ai rH *P a C hi G o — l-s _ hi h> 01 > H H h. 3 Ol —4 IP c 44 h. h. G G o 01 c iH a > a C. CJ • > 2 X O hi > iH « a 3 h et o XI rH 01 G 41 o 01 X hi OC o a iH G rH 23 O •H O — G g 3 CI 03 01 cr 3 o O 01 3 hi h. .G 01 01 h GG x: h. ^ hi 01 a) xi o o x: a « • 23 41 C O N N XI U ui c o JAx: o o IP O •5 a G s a a r-» O CM CM • • • • • . , . . , . u u co ca ca ca CO CO c C 3 3 SS335 3 ca CO A Q a a a o Q Q a a a o O a O on vo CO cn o CVJ an vo r-CO cn o Tot ao COr CO an cn cn cn CTVcr v CTVc n o C—O C —O C—O — — C —O —CTV —OV r— — — CM 91 [ 101 111 121 131 141 151 . 161 |164| 166 [ 168 171 174 176< 181 184< 188 _ 191 194^ 161 090000099999 9999999999999 9900000000 0000110000 1111101011 0011011009 9011100112 1111^00111^11 ^OOOOO'OOO'OOOO'OITSJGOI 00005(?101000^09 000000000000 162 090000099999 9999999999999 9900000000 0000110000 1111001111 0111011009 9011100112 1111100011120011 200000000000001100001 0C0000101000009 000000000000 163 090000099999 9999999999999 9900000000 0000110000 1111001111 0111011009 9011190112 1111100011120011 200000000000001100001 000000101000009 000000000000 164 090000099999 9999999999999 9900000010 0000199900 0011001010 9991011010 0011190112 1110000011120009 200000000000000000000 000000001000009 000000000000 165 090000099999 9999999999999 9900000010 0000099900 0011001010 9991011010 0011090111 1919990011120009 200000000000000000000 000000001000009 000000000000 166 090000099999 9999959999999 9900000010 0000101000 0011001110 9991011010 0011090112 1919990011120009 200000000000000000000 000000001000009 000000000000 167 090000099999 9999999999999 9900000010 0000110001 0011001010 9991011010 0011190111 1110000011120009 200000000000000000000 00000000)000009 000000000000 168 09C000093999 9599599999999 9900000010 0000111000 1011001011 0000999010 0011090012 1999990011120009 200000000000000000000 000000001000009 000000000000 169 090000039999 5599999999999 9900000010 0000110010 1111001011 0001011010 0011090112 1919990011120009 200000000000000000000 000000001000009 000000000000 170 090000099999 9999999999999 9900000001 0000110010 1111001011 0001011010 0011100111 1010000011120009 200000000000000000000 000000001000009 000000000000 171 090000093SS9 9999999999999 9900000001 0001310010 1110101011 0111011009 9011190110 1110000011120009 300100000000000000000 000000001000009 000000000000 172 030000059999 5999999999999 9900000001 0001999999 9599999999 9999999999 9999999999 5999990011199999 999999999999990000000 000000001000009 000000000000 173 050000099999 9999999999999 9900000001 0001999959 9999999999 9999999999 9999999999 999999P011199999 999999999999990000000 000000001000009 000000000000 174 090000099999 9999999999999 9900000000 0000110001 0210101010 5991011010 0011090112 1919990011120009 200000000000000000000 001000001000009 000000000000 175 090000099999 9999999999999 9900000000 0000010001 0210101010 9991011010 0011090112 1919990011120009 200000000000000000000 001000001000009 000000000000 176 090000099999 9993999999999 9900000001 0000310001 0011101111 0011011010 1011190110 1110000011120009 300100000000000000000 000000001000009 000000000000 177 090000099999 9999995995999 9900000001 0000210001 0011101111 0011011010 1011100110 1010000011120009 300100000000000000000 000000001000009 000000000000 178 O3OOOOO99999 9999993999999 9900000001 0000101000 1111101011 0101111010 0011090010 1999990011120009 200000000000000000000 000000001000009 000000000000 179 090000059999 9999999999999 9900000000 0000110000 1111001011 0101111010 0011090110 1019990011120009 200000000000000000000 000000001000009 000000000000' 180 090000059359 9999999999999 9900000099 0000111000 1111101011 0101111010 0011100010 0099990011120009 200000000000000000000 000000001000009 000000000000 181 090000099999 9999999999999 9900000000 0000111000 1111101011 0101111010 0011150110 1110000011120009 200000000000000000000 000000001000009 000000000000 182 090000099599 9599995999999 9900000000 0001999S99 9999999599 9999999999 9999999999.9999990011199999 999999999999990000000 000000001000009 000000000000 183 090000095939 9999999999999 9900000000 0000111000 1111101011 0101011010 0011090110 1019990011120009 200000000000000000000 000000001000009 000000000000 184 090000099999 9999999999999 9900000000 0000111000 1111101011 0101011010 0011090010 1999990011120009 200000000000000000000 000000001000009 000000000000 185 090000093399 9999999999999 9900000000 0001099912 1009901911 0190999009 9011090010 199999001112000S 200000000000001101000 000000001000009 0000000C0000 186 090000099959 9999999999999 9900000000 0000110000 1111001010 9990999009 9011050011 1099990011120009 200000000000000000000"000000001000009 000000000000 187 090000099999 9999999999999i9900010000 0010311001 1211101011 0011011011 0011190112 1111100011120009 200000000000002111000 000000011000009 000000000000 188 090000099999 9999999999999 9900010000 0010110001 1211101011 0001011011 0011190112 1111100011120009 200000000000002111000 000000011000009 000000000000 189 090000093999 9999999599999 9900010000 0010310001 1211101011 0001011011 0011190112 1111100011120009 200000000000002111000 000000011030009 000000000000 190 090000055999 9999995999999 9900010010 0000111001 1010101011 0111011011 0011190111 1110000011120009 200000000000002111000 000000011000009 000300000000 191 090000099999 9999993999999 9900010000 0000111001 1211101011 0111011011 0011190112 1111100011120009 200000000000002111000 000000011030009 000000000000 192 030000G99939 9593995999999 9900000000 0000111001 1211001011 0901011011 0011190112 111110001112C009 200000000000002111000 000000001000009 000003000000 193 090000099999 9999999999999 9900000000 0000111001 1011101011 0111011011 0011190111 1110000011120009 200000000000002111000 000000001000009 000000000000 194 090000093939 999SS99999999 9900000099 0000311011 1011101011 1011111011 0011190112 1110000011120009 200000000000002111000 000000001000009 000000000000 195 090000095999 9959999599999 9900000000 0000099910 1011101011 1011011009 9011090111 1919990011120009 200000000000002111000 000000301000009 OOOOOOOOOOCO 196 090000099999 9999999999999 9900000003 C000099910 1011101011 1011011009 9011090111 1919990011120009 200000000000002111000 000000001000009 000000000000 197 0S0000099999 9999999999999 9900000003 0000099910 1011101011 1011011009 9011050111 1919990011120009 200000000000002111000 000000031000009 000000000000 198 050000099999 9939999999999 9901000099 0000110059 9999999999 9999999009 9091190112 1110000011120009 2003C0000000002111000 C00000001000009 000000000000 199 090000099999 9999999999999 9901000000 0000310001 1011001111 0101111009 9011190012 1101100011120009 200000000000002111000 000030001C00009 000000000000 200 C90000099999 9999999399999 990)000000 00003)000) 0011001111 0101011009 9011190112 1112000011120009 200000000000002111000 000000001000009 000000000000., ^ ro o ID 00 -J CsUljr-U4 t\J - o vo CO -vjCT\ Vn V* TO O VO OO-JovVn^VxtV)—'OVO 00 -J ON Vn CifO--

- - o ooooooo—o 2-8 22^^22^2-° o — vo\Q voQB ——• — ->• o o o ooooooooo § © \Z> vo o o 222222^22222® —• VO VD — — o o - ooooooooo 2£2-225fl2£222° O O VO VD O O o o - coo-^oovooocooo VO VO — — —O -O» VvDo \£v>o O-» o — -o» o— vco vo o— —. --2::£2iS22£2" -O O- Vfvovl VcD - O -- — — — oOO o Oo — — — — —-.- -»4 o c vo vo o — —'OOOO-'OOOO-' — Zt2c22\S2222~*Ii o c vfi vo - oooooooo o 0-»0 0 00VD0000 — -»

— o VJVDD -—» -4. —» o vo —• —• —' co VO — — — © — — o o o 2 — § — 22vd — 2 —

s^igr — vo vo — _ VO VvDo ——

e6£ o o *o— *o— o o 88 oOO o O _rr__o o o 3L LOo o -o 5_ o o

to 3 O (0 3 3 3 1-1 —H 3 01 01 3 c —4 3 3 X rH to ei 3 01 01 3 3 —• rH •a 3 3 —4 01 01 o 6fl i-H 3 cu rH c oi o XI g 63 3 rH 3 3 c 3 4 o 3 3 O 3 3 "2 a c •o 3 o r-4 O •a g 3 S 3 3 rH o 3 o c 63 o o 44 3 •o u «-4 3 •O 3 c XI o 3 x 3 >» o o H O 3 3 3 u 3 C 3 o o. c >» C T3 3 3 S-4 c o. 4-> a. 4-> u u —4 XI g Q. g •a o O •a 3 3 3 u 3 3 {* 3 o 3 3 c 3 o. o 3 c 3 rH O O aj r&H to A S u li u 3 t-> rH C 3 3 o G. 0] •3 13 rH c. rH G, 3 § () > O e XI • • • • . . 4 . . • . • • • , Q a Q o Q a a a a a a Q a a a O a o Q a a a q a a Q a a a a a a CM in vo r- co o 4- CM KN-3" m VO r- CO ON o CM m-3 in vo co ON o o CM o O € o o o o 8 CM CM CM CM CM CM CM CM CM KN MN CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM SI CM CM CM CM CM CM CM CM CM CM a CM 91 99 101 ^ 121 131 141 151 161 164 ^ 166 ^ 168 ^ 171 174 176 181 1841 188 ^ 191 194 — 201 090000099999 9959999993999 9900000000 00G0110011 0211101110 9991011010 001 1190111 111 c£000T?12(3009 20000000000X02111000 0000^0010^09 OQCOOcSx)GOX 202 090C00C99S99 9999999959999 991X01000 0000110001 1011011111 0111011010 101 1190112 11100000111200X 20X0000000X02111000 OCC0000010G0009 00000X00000 203 090000099999 9999999999999 9910001 OX 0000110001 0111111011 1001011010 101 1190112 111000X11120009 2X0000X00X02111X0 000000001X0009 0000X0X000 204 090000099993 9999999999999 991X01000 0000110X1 0111011111 0111011010 001119011 1 11100000111200X 2000000000000021110X 00000000100c009 00X0000X00 205 090000099999 9959599999999 991x01 OX 0000110001 0011011011 0111011010 001110011 1 1110000011120009 200000000000002111XO X00000C1000009 COOOOOOGOOOO 206 090000099999 9999995959999 9910001000 X00110C01 X11011011 0111011010 001119011 1 1100XX11120009 200000000000002111000 000000001000009 0C000000X00 207 (>30000099999 9999599999959 9910001 ox 0000110000 0111010991 10010000X 901 1190111 11111000111200X 20000000000X021110X 000000001000009 0000X000000 200 09C0000S9S99 9999999999999 9910001 ox 0000110000 011101X91 1001000009 901119011 1 11111000111200X 200000000000002111000 000000X1000009 ooooooxxoo 209 090000099959 9999999999999 9910001000 0000110000 011101X91 10010000X 901 1190111 11111000111200X 2000X00000X02111 XO OQOCOC001G00009 000000000000 210 090X0099999 9999999999999 9910001000 0000110000 011101X91 1001X00X 901 1190111 1111100011120009 200X00000000021110X ooooooooi oxoos 000000000000 211 090000099599 9999999999999 9910001000 0000310001 0111001110 9991011010 001119011 2 111000X111200X 200000000000002111000 000000001000009 00X00000000 212 090000095999 9999999999999 9910001001 :oooo3ioooi 0011101110 9991111010 001119011 1 111211X111200X 200000000000002111 XO oooocoooi000009 00000X00X0 213 090000039999 9999999999999 9910001001 0000310001 X11101010 9991101010 001119011 1 11121100111200X 2000000X000002111000 00000X01000009 oooooooooox 214 090000110X9 0000099009900 cooooooooo 0000110000 0111001111 01X999010 011 1100111 100000001112x09 200000000000002111000 00000X01XX09 oooooxoxx 215 090000110009 000X590X900 00X000000 X001100X 0111101111 0101011010 011119011 2 1110XX111200X 2X000X00000021110X 00X0X01000009 ooxooooxx 216 090000110009 0000099101000 ocoooooooo 00X110000 0111001111 0111011010 011119011 1 11100000111200X 200X000000X02111 XO xoooxoi 000009 oxooooooox 217 090000110009 0000X9122000 000X00000 00001100X 0111101111 0111011010 011119011 1 11100000111200C9 2OOXOOOOXXO2111000 ooooooooi000009 oooooxoxx 218 090000110009 000009910X00 0000000X1 00X110000 0111001111 0101011010 011119011 1 111X0X11120009 20000000000X021110X 0X00000100X09 000000X0000 219 050000110009 00000991020X 0000000001 0000110000 0111101111 0101011010 011119011 1 1110XX1112x09 200000000000002111000 000X0001000009 000000000000 220 090000110009 0X0X90X900 0000000001 0000111 ox 01X911011 1101111010 011119001 1 1150X00111200X 3001000000000021110X 00000000100X09 000000000X0 221 050000100009 0000X90X9X 000X00001 X001100X 0111111011 1001111010 001119001 1 .11900000111200X 3X10000X00002111000 ooooooooi 000009 xoooxoxoo 222 090000100009 0000X9102000 0000000001 0000110000 0111111111 1111111010 001119011 1 1110000011120009 300100000000002111 XO 000000001000009 00000X00000 223 oyooooi oox(9 0X0X9102000 000X0X01 00C0110000 0111111111 0111111010 001119011 1 111X000111200X 3001000000X002111 ox 0X0000010X009 xooooooxoo 224 0900C0100009 0000X9029900 000000X01 00X110000 0111011011 0111111010 001119011 1 1110000011120009 20000000000X02111 XO 000X0001000009 00000X0X00 225 090000100009 00XX90X9X 000X00001 0000110000 0111011011 0111111010 011119011 1 11111000111200X 20000X0000X02111000 00000X01000009 00000000X00 0111011010 011 226 090000100009 0X0X9029900 000X00001 X00110000 0111011111 1190112 1111100011120009 200X0000000002111 XO 000000001000009 xxoxooox 227 090C00100009 1100X90X900 0000X0000 0000110000 0111011111 10110110X 901 1190112 11100000111200X 2X0000000X002111 XO 0XCC0001000009 0X000000000 228 05'00001C0009 00000991220X 0000X0001 0000110000 0111011111 1111011010 001119011 2 111000X111200X 200000X000X021110X 000000001X0009 00000X00000 229 090000100009 0XGX90299X 0000000C01 0000110000 0111011111 0111011010 001119011 2 11100000111200X 2000000000X0021110X 0X000001000009 oooooooooox 0111011010 001 230 0X0X99999X 00000X099 0000110000 0111011111 1190112 11111000111200X 2000X0XX0X21110X xooxoooox 090x0100009 99999990X 9X 0000x001000009 231 090x0100009 1100X90X900 0000000000 00001100X 0199919999 1190112 111110X11120009 2000X0000000021110X 000000001000009 0000X00X00 232 1100099019900 oxxooxo 0111011111 0111011010 001 200000000000002111 XO 03000010x09 0000110000 0101011010 X1 1190112 1111100011120009 0000x001000009 000000000000 233 030000099909 9999999999999 950X00010 0000110000 0011X1011 1190112 1110000011120009 2000000000000X000000 OXOC 0X1000009 oooooooooox

VJ1 vo VJ1 v^j ro — o vo a> -J o\uiir-i>i ro — o vo co ->i o\ ui ro — o Co co cr»Cn£-ix ro

ooooo ooooo — — -»-»oo — oo —

OOOOOOOOOO — — OOOOO — OOOOO — — — — ©O — OOO £ — — OOOOOOO— — — — — OOOOOOOO — ©OOOOOO — — O — — — OOO — OO — — — — — — O — O — OOO — — — — — OOO — — O — —— OOO— OO — — — — — — — — — — OOO — — — — — — o — — — — — — — OOO — OO — — — — — — — — — — O — — — — — — — — — — —— — — -»— ooo—o — — — — — — — — — — — o — — — — — — — — — — — —— K — — — ooo — o — — — — — — o — — o — o — — — — — — — — — — — — — 22222222222229S22222---02--0200-"'

9 597

7.2. APPENDIX B - GLOSSARY

Abdominal flecks - small pits to either side of abdominal terga III - VII (at least) of male Trichodectidae. (=abdominal lateral flecks) Aedeagus - structure of male genitalia formed by fusion of the two mesomeres (q.v.), and through which everts the endophallus (q.v.). Not present in Psocodea. Anterior setae — setae positioned anterior to posterior setal row (q.v.) on pronotum, pteronotum and abdominal terga> sterna and pleura. Apomorphic - pertaining to the derived state of a character (apomorphy), that indicates by its distribution a holophyletic group. Arolium - membranous or partly-sclerotised median lobe of the pretarsus, arising between the claws. Autapomorphic - pertaining to an apomorphy characteristic of a holophyletic group; c.f. apomorphic, synapomorphic. 'b.a.l.s.1 - see 'Lateral struts of basal apodeme'. Basal apodeme - ventral anteriad projection of phallobase (q.v.). In Psocodea the phallobase is absent, and the basal apodeme projects anteriad from the bases of the parameres and mesomeres. (= 'phallobase', basal plate, genital apodeme, internal plate)

Basiparameral sclerites — sclerites of the male genitalia of Trichodectidae . positioned on the ventral wall of the endophallus just anterior to the parameres and probably formed by detachment of part of the parameres (Fig. 33 )• Clade - a branch of a cladogram; a holophyletic group of species. Conus - prolongation of the lateral margin of the head of Ischnocera immediately anterior to the antennal socket, of which it forms the anterior margin. (= 'trabecula', zapfen, clavus) Clypeal marginal carina — internal carina marking preantennal margin of phthirapteran head (in dorsal or ventral aspect). Dermecos - the environment produced by the skin and fur or feathers of the host, and utilised by the ectoparasite. Dorsal preantennal sulcus - dorsal antero-median interruption of the clypeal marginal carina (q.v.). Empodium - spine or lobulate process arising from the unguitractor or planar sclerite distal to the unguitractor. Endophallus - eversible or partially eversible structure of male genitalia in Phalloneoptera, distal to the gonopore (q.v.), and continuous with the male genital chamber walls (Fig.25 ). (='endotheca', genital sac, internal sac, mesosome, preputial sac, vesica, vesica penis, vesicula penis) 598

Endotheca - eversible part of lining of phallotheca (q.v.). Not present in Psocodea. Flagellomeres - annulations of the antennal flagellum distal to the pedicel (q.v.). Genital chamber (female) - invagination of the integument between sterna VIII and IX, from which the common oviduct leads anteriorly. In Phthiraptera the chamber is oval, dorso-ventrally compressed, and may be lightly-sclerotised or bear sclerotised spines or scales. Genital chamber (male) - invagination of the integument posterior to sternum IX, in which are the male genitalia (Fig. 19 )• Genital opening (male) - opening of the male genital chamber, through which the genitalia extrude during copulation. Gonopore (male) - the external opening of the median ejaculatory duct (ductus ejaculatorius)• In Phalloneoptera the gonopore opens into the endophallus (q.v.)• Gonapophysis lobe - expansion of the ventral margin of gonapophysis VIII in female Trichodectidae, sometimes defining by its presence a distal spur. Holophyletic - pertaining to a group of species or other taxa that comprises a single ancestral species (known or inferred) and all of its descendants. Homoplastic - pertaining to non-homologous similarity; including convergence, parallelism and character—state reversal. Homologous (=phyletically homologous as defined in section 1.3.2. when used in this study without qualification) - the relationship which pertains between features (or states of features) present in two or more organisms, that may be traced back to the same feature (or state) in the immediate common ancestor of those organisms. Host class - animals that are potentially suitable for parasitism by a member of the parasite group under consideration. Lateral rods of the subgenital plate (male) - the two longitudinal elements of the male subgenital plate (Fig. 19 )• Abbreviated to 's.g.p.r.' in this study. Lateral struts of the basal apodeme - lateral margins of the basal apodeme (q.v.), generally strongly sclerotised and with posterior ends articulating with, or fused to, parameres and mesomeres (Fig. 26 ). Abbreviated to 'b.a.l.s.' in this study. Median gap - break, at the mid-line, of posterior setal row (q.v.) on thoracic or abdominal segments; generally dorsal. 599

Median setal group — setae of posterior setal row on abdominal terga closest to centre of tergum, separated from more lateral setae ('lateral setal group') by pronounced gap, and from median setal group on other side of tergum by median gap. Mesomeral arch - product of apically-fused mesomeres (q.v.) in many Psocodea. May he in the form of a simple ring, or have a more or less developed median extension. (=pseudopenis, sclerotised ring, parameral arch, mesosome arch, ventral apodeme, epimeris, lanceolate band) Mesomeres - the median pair of primary phallic lobes, and the sclerites derived from them in the adult insect. Apically fused into the 'mesomeral arch' (q.v.) in many Psocodea. Frequently misidentified as parameres in Phthiraptera. (='parameres', internal parameres) Monophyletic - pertaining to a group of species or other taxa that comprises a single ancestral species (known or inferred) and some or all of its descendants. Monoxenous — pertaining to parasite taxa that utilise only one host species. Nidicolous - nest-dwelling. Occipital ring - internal carina of the head underlying the postoccipital sulcus in Phthiraptera. Osculum — median anterior indentation of trichodectid head (in dorsal or ventral aspect). Parameres - the lateral pair of primary phallic lobes, and the sclerites derived from them in the adult insect. In Psocodea the sclerites lie basally ventral or median to mesomeres or mesomeral arch (q.v.), articulating with or fused to b.a.l.s. (q.v.) and/or mesomeres. May fuse ventrally to form a single median plate ('parameral plate'). (=endomeres, external parameres, endomeral plate) Paraphyletic - pertaining to a group of species or other taxa that comprises a single ancestral species (known or inferred) and some but not all of its descendants. Paraprocts - sclerites of abdominal segment XII in many insects, corresponding to the two lateral sclerites of the telson. Pedicel - basal annulation of the antennal flagellum, characterised in the Thysanura-Fterygota by the presence of Johnston's Organ. Penis - in Psocodea a sclerotised structure surrounding the gonopore (q.v.) and derived from the endophallus (q.v.). 600

Perisetal gaps - unsclerotised areas of the phthirapteran male subgenital plate (q.v.) surrounding setae, particularly the posterior setal rows of the sterna concerned. Phallobase - common base of the parameres and mesomeres, formed by lack of full separation of these two structures in ontogeny, and which may be provided with a basal apodeme (q.v.). Absent from Psocodea. Fhallotheca - tubular extension of phallobase enclosing aedeagus (q.v.). Absent from Psocodea. Phragmata — transverse infoldings of the intersegmental sclerites (postnota) on the insect thorax, providing attachment sites for the dorsal longitudinal muscles. Plantula - ventral process of tarsomere. (=euplantula) Plesiomorphic - pertaining to the primitive or ancestral state of a character (plesiomorphy), that does not necessarily indicate by its distribution a holophyletic group. 'p.l.s.' - see JPostero-lateral seta'. Polyphyletic - pertaining to an assemblage of species or other taxa that does not include the most recent common ancestor (known or inferred) of all species in the assemblage. Postcoxale - semicircular sclerite lying posterior to the coxa (used in this study with particular reference to the third leg). Posterior setal row - row of setae running around each abdominal segment and dorsally on the prothorax and pterothorax, generally placed posteriorly on each segment. Postero-lateral seta — small seta on the posterior margin of abdominal terga II-V or VI close to the pleurum. Abbreviated to fp.l.s.' in this study. Post-genital pleurites — pair of lateral sclerites on female abdominal segment IX + X; may be homologous with paraprocts. Abbreviated to 'P^-g-P1 in this study. Post-genital sclerite (male Trichodectidae) - sclerite lying between male genital opening and sternite IX (Fig. 19 ). Post-vulval sclerite - sternal sclerite on segment IX + X in female Trichodectidae; may be single or medially divided. Frimary host - host that is associated with its parasites because of inheritance from an ancestral host species that was infested with the ancestors of the parasites. 601

Pseudostyli - extensions of the posterolateral angles of the male subgenital plate on segment IX in some Trichodectidae. Pulvillus - lobulate process arising from auxillae of pretarsus, usually paired. Pulvinus - membranous bilobed structure to the anterior of the labrum in Ischnocera, formed from the clypeo-labral suture. Scape - basal segment of the antenna. Secondary absence - the absence of parasites from a host taxon that might, from its presumed ancestry, be expected to have parasites of that taxon. Secondary host — a host taxon that has acquired the parasite taxon under discussion from another taxon of the host class by a process of secondary infestation (q.v.). Secondary infestation - n. The association of a para-site taxon and a secondary host (q.v.). - v. The transfer and establishment of parasites to a host taxon not previously infested. Sister-groups - two (or, exceptionally, more) species or higher holophyletic groups that are postulated to be each other's closest genealogical relatives exclusive of their common ancestor; holophyletic groups connected on the cladogram at a single node, and thus sharing a common ancestor not shared by other groups. Sitophore sclerite - cup-shaped sclerite on the ventral face of the sitophore in Psocodea. (= cibarial sclerite) Subgenital lobe — median expansion of the posterior margin of the female subgenital plate (sternum VIII)• Subgenital plate (female) — the sclerite formed from sternites VII and VIII, or from VIII alone. Subgenital plate (male) - the sclerite formed from sternites VII, VIII and IX, or from only some of these. 's.g.p.r.' - see 'Lateral rods of the subgenital plate'. Synapomorphic — pertaining to an apomorphy shared by members of a holophyletic group; c.f. apomorphic, autapomorphic. Tergocentral microsetae — very small setae present in the centre of the posterior setal row (q.v.) of the terga of male Neotrichodectes (Fig. 240). Trabecula — conical structure arising in ~schnocera from the antennal socket, the posterior margin of which is continuous with the anterior margin of the scape. Restricted in fully-developed form to Philopterus and some related genera. 602

Type—host — the host from which the type specimen of a louse species was collected. Ventral preantennal sulcus — ventral antero-median interruption of clypeal marginal carina in Phthiraptera. Vulva — posterior opening of the female genital chamber between sterna VIII and IX.