THE SYSTEMATICS, ECOLOGY, AND ZOOGEOGRAPHY OF THE AFRICAN GERBILS, TATERILLUS (RODENTIA: CRICETIDAE)
Item Type text; Dissertation-Reproduction (electronic)
Authors Robbins, Charles Brian, 1940-
Publisher The University of Arizona.
Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
Download date 07/10/2021 19:11:22
Link to Item http://hdl.handle.net/10150/289257 INFORMATION TO USERS
This material was produced from a microfilm copy of the original document. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the original submitted.
The following explanation of techniques is provided to help you understand markings or patterns which may appear on this reproduction.
1. The sign or "target" for pages apparently lacking from the document photographed is "Missing Page(s)". If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting thru an image and duplicating adjacent pages to insure you complete continuity.
2. When an image on the film is obliterated with a large round black mark, it is an indication that the photographer suspected that the copy may have moved during exposure and thus cause a blurred image. You will find a good image of the page in the adjacent frame.
3. When a map, drawing or chart, etc., was part of the material being photographed the photographer followed a definite method in "sectioning" the material. It is customary to begin photoing at the upper left hand corner of a large sheet and to continue photoing from left to right in equal sections with a small overlap. If necessary, sectioning is continued again — beginning below the first row and continuing on until complete.
4. The majority of users indicate that the textual content is of greatest value, however, a somewhat higher quality reproduction could be made from "photographs" if essential to the understanding of the dissertation. Silver prints of "photographs" may be ordered at additional charge by writing the Order Department, giving the catalog number, title, author and specific pages you wish reproduced.
5. PLEASE NOTE: Some pages may have indistinct print. Filmed as received.
Xerox University Microfilms 300 North Zeeb Road Ann Arbor, Michigan 48106 75-26,935 ROBBINS, Charles Brian, 1940- THE SYSTEMATICS, ECOLOGY, AND ZOOGEOGRAPHY OF THE AFRICAN GERBILS, TATERILLUS (RODENTIA: CRICETIDAE). The University of Arizona, Ph.D., 1975 Zoology
Xerox University Microfilms, Ann Arbor, Michigan 48106
THIS DISSERTATION HAS BEEN MICROFILMED EXACTLY AS RECEIVED. THE SYSTEMATICS, ECOLOGY, AND ZOOGEOGRAPHY OF THE AFRICAN
GERBILS, TATERILLUS (RODENTIA: CRICETIDAE)
by
Charles Brian Robbins
A Dissertation Submitted to the Faculty of the
DEPARTMENT OF BIOLOGICAL SCIENCES
In Partial Fulfillment of the Requirements For the Degree of
DOCTOR OF PHILOSOPHY WITH A MAJOR IN ZOOLOGY
In the Graduate College
THE UNIVERSITY OF ARIZONA
19 7 5 THE UNIVERSITY OF ARIZONA
GRADUATE COLLEGE
I hereby recommend that this dissertation prepared under my direction by Charles Brian Robbins entitled The Systematica, Ecology, and Zoogeography of
the African gerbils, Taterillus (RodentiatCricetidae) be accepted as fulfilling the dissertation requirement of the degree of Doctor of Philosophy
CscJLyunx 31 147$ Dissertation Director Date
After inspection of the final copy of the dissertation, the following members of the Final Examination Committee concur in its approval and recommend its acceptance:-'
QuMjeXV* O-frQ-M—- // in 5 // /)oh/ /?7J~
J j 7 (\pr'J
This approval and acceptance is contingent on the candidate's adequate performance and defense of this dissertation at the final oral examination. The inclusion of this sheet bound into the library copy of the dissertation is evidence of satisfactory performance at the final examination. STATEMENT BY AUTHOR
This dissertation has been submitted in partial fulfillment of requirements for an advanced degree at The University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.
Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his judgment the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author.
SIGNED: S/Uq^ ACKNOWLEDGMENTS
I extend my deepest gratitude to my major professor,
Dr. E. Lendeli Cockrum, for his guidance as chairman of my graduate committee, for his advice and encouragement during
this study, and for his critical reading of the earlier drafts. Special thanks also go to Dr. Henry W. Setzer, for
providing me with the African Mammal Project specimens, for allowing me space in the Division of Mammals, National
Museum of Natural History, Washington, D. C., to conduct
the major part of this research, for arranging for my
participation in the African Mammal Project as field
collector in Mauritania, Senegal, Ghana, Togo, and Dahomey,
and for his many helpful suggestions during this research.
My sincere appreciation is extended to the following
staff members and technical personnel of the Division of
Mammals, National Museum of Natural History. Charles 0.
Handley, Jr., Richard Thorington, Duane A. Schlitter, John
H. Miles, Jr., David Harvey, Arthur Brigida, Frank Greenwell,
and Helen Hutchinson.
For permission to study specimens in their collec
tions and for the loan of comparative material, I thank Dr.
G. C. Corbett and I. R. Bishop of the Mammal Room, British
Museum of Natural History, London; Dr. Francis Petter,
Zoologie, Mammals and Fish, Museum National d'Histoire
iii iv
Naturelle, Paris; Dr. Sydney Anderson, Division of Mammals,
American Museum of Natural History, New York; Dr. Barbara
Lawrence and Charles W. Mack, Mammal Department, Museum of
Comparative Zoology, Harvard University, Boston, Massa chusetts; Dr. Louis de la Torre, Division of Mammals, Field
Museum of Natural History, Chicago, Illinois; Dr. Donald
Patten and Dean Harvey, Division of Mammals, Los Angeles
County Museum, Los Angeles, California; Dr. Heinz Felten,
Zoologie, Natur-Museum Senckenberg, Frankfurt, Federal
Republic of Germany; Dr. R. Angermann, Museum fur Naturkunde
der Humboldt-Universitat zu Berlin, Democratic Republic of
Germany; Dr. Xavier Misonne, Department of Zoology, Institut
Royal des Sciences Naturelles de Belgique, Brussels,
Belgium; Dr. Walter H. Verheyen, Laboratorium von Alegemene
Dierkunde, Rijsuniversitair Centrum von Antwerpe, Antwerp,
Belgium; Professor M, Poll and Dr. Thys Van Den Audenaerde,
Musee Royal de l'Afrique Centrale, Tervuren, Belgium; Dr.
E, Tortonese, Director, Museo Civico de Storia Naturale
"Giacomo Doria," Genoa, Italy; Dr. F. W. Braestrup,
Universitets Zoologiske Museum, Copenhagen, Denmark; Dr.
Greta Vestergren, Naturhistoriska Riksmuseet Sektionen for
Vertebratzoologie, Stockholm, Sweden; and Mr. Aggundy I. R.,
Curator of Mammals, The National Museums of Kenya, Nairobi,
Kenya,
I would also like to acknowledge the time and
efforts of the following members of my graduate committee. V
Drs. Charles H. Lowe, Robert Chiasson, Russell Davis, Paul
S. Martin, and Everett H. Lindsay.
I am grateful to Daniel Piecesi, Charles Roberts,
Neil Roth, and Douglas Lorenz of the Information Systems
Division of the Smithsonian Institution. They provided me with statistical assistance and computer programs involved in this study.
Advice and assistance in multivariate statistics were given to me by Don E. Wilson and Michael A. Bogan of the National Bird and Mammal Labs, U. S, Department of
Interior Fish and Wildlife Service, Smithsonian Institution.
They also provided me with computer time on the Department of Interior computer for the NT-SYS programs. I am also grateful to Alfred L. Gardner for assistance with system- atics and especially cytogenetic interpretations,
To my wife Norrie, I am especially grateful for her understanding patience and constant encouragement during the preparation of this dissertation, for the preparation of the illustrations which appear in the text, and for the thankless job of recording innumerable skull measurements.
Finally, I would like to acknowledge the interest, encouragement, and support of my parents, Charles J. and
Virginia Robbins, throughout my academic career.
Travel to foreign and domestic museums was supported by National Science Foundation Grant GB 35143 to myself,
E, L, Cockrum, and The University of Arizona, Field work vi in Kenya was also supported by National Science Foundation
Grant GB 35143 to myself, Duane A. Schlitter, The University of Arizona, and the University of Maryland. Duane
Schlitter's field assistance in Kenya is gratefully acknowl edged. Without the above support the completion of this reaearch would not have been possible. TABLE OF CONTENTS
Page
LIST OF ILLUSTRATIONS ix
LIST OF TABLES . . . . xii
ABSTRACT xiv
CHAPTER
1. INTRODUCTION 1
2. HISTORY OF TATERILLUS NOMENCLATURE ...... 11
3. MATERIALS AND METHODS 17
Plan of Treatment 18
4, SYSTEMATIC ACCOUNTS 25
Key to the Species of Taterillus 27 Taterillus arenarius Robbins ... 29 Taterillus congicus Thomas .... 35 Taterillus emini (Thomas) 42 Taterillus gracilis (Thomas) 50 Taterillus gracilis angelus Thomas and Hinton 55 Taterillus gracilis gracilis (Thomas) 60 Taterillus gracilis meridionalis, New Subspecies 65 Taterillus gracilis nigeriae Thomas . . 70 Taterillus harringtoni (Thomas) 7 5 Taterillus lacustris (Thomas and Wroughton) 84 Taterillus pygargus (F. Cuvier) 90
5. SPECIFIC RELATIONSHIPS 100
Cranial Morphology 100 West Africa 101 East Africa 135 Total Distribution 153 Comparative Karyology , ..... 161
vii viii TABLE OF CONTENTS—Continued
Page
Distribution 171 Future Studies 177
6. EVOLUTIONARY AND ZOOGEOGRAPHICAL RELATIONSHIPS 179
Evolutionary and Systematic Relationships 179 Zoogeography 188
7. SUMMARY 19 2
APPENDIX A. GAZETTEER 194
APPENDIX B. MEASUREMENTS 205
APPENDIX C. STATISTICAL TECHNIQUES . , 210
APPENDIX D. SKULL MEASUREMENT TABLES ... 214
REFERENCES 241 LIST OF ILLUSTRATIONS
Figure Page
1. Distribution of the genus Taterillus in Africa 4
2. Diagram of auditory bullae for Taterillus ... 28
3, Distribution and known localities of Taterillus arenarius 30
4. Distribution and known localities of Taterillus congicus 36
5. Distribution and known localities of Taterillus emini 44
6. Distribution and known localities of Taterillus gracilis including subspecies ... 51
7, Distribution and known localities of Taterillus harringtoni 77
8. Distribution and known localities of Taterillus lacustris 85
9. Distribution and known localities of Taterillus pygargus 91
10r Locality numbers used in statistical analyses of West African Taterillus 102
11, Two-dimensional projections of the first three principal components of Senegal Taterillus with known karyotype 103
12. Projection of the first two canonical variates for three West African species of Taterillus 107
13. Projection of the first two canonical variates for geographic samples of T. g. gracilis , , . 112
14, Projection of the first two canonical variates for geographic samples of Senegal T. pygargus 114
ix X
LIST OF ILLUSTRATIONS—Continued
Figure Page
15. Projection of the first two canonical variates for population samples of T. g. angelus, T. g. rneridionalis, and T. g. nigeriae .... 116
16. Projection of the first two canonical variates for geographical samples of T. g. rneridionalis 118
17. Distance phenogram of population sample means for West African species and subspecies of Taterillus 120
18, Projection of the first three principal components of population sample means for West African species and subspecies of Taterillus 124
19. Projection of three-dimensional MDSCALE of population sample means for West African species and subspecies of Taterillus 132
20. Locality numbers used in statistical analyses of East African Taterillus 136
21. Two-dimensional projections of the first three principal components of Zaire and Uganda Taterillus 138
22. Projection of the first two canonical variates for three, species of East African Taterillus 142
23. Distance phenogram of population sample means and holotypes of East African Taterillus . , . 147
24. Projection of the first three principal components of sample means and holotypes of East African Taterillus , 149
25, Projection of three-dimensional MDSCALE of population sample means and holotypes of East African Taterillus 151 xi
LIST OF ILLUSTRATIONS—Continued
Figure Page
26. Distance phenogram of population sample means for geographic samples of Taterillus 154
27. Projection of the first three principal components of population means for geographic samples of Taterillus . 156
28. Projection of three-dimensional MDSCALE of population means for geographic samples of Taterillus ...... 160
29. Karyotypes of six species of Taterillus .... 162
30. Hypothetical pathways for chromosomal evolution in six species of Taterillus .... 169
31. Distribution of the four West African species of Taterillus . 172
32. Distribution of the three East African species of Taterillus 176
33. Distance phenogram for the seven species of Taterillus 182
34. First three principal components for the seven species of Taterillus ...... 184
35. Three-dimensional MDSCALE projection for the seven species of Taterillus ...... 187 LIST OF TABLES
Table Page
1. Loadings of the most important cranial measurements on the first three principal components for individual karyotyped specimens of T. g. gracilis and T. pygargus from Senegal 104
2, Discriminant coefficients for characters most useful in separating three West African species of Taterillus 108
3. Means and standard error for the seven cranial characters in their order of inclusion in the discriminant analysis , . . . 109
4. Loadings of the cranial measurements on the first three components of West African Taterillus 125
5. Loadings of the cranial measurements on the first three components for individual specimens of Zaire and Uganda Taterillus 139
6. Discriminant coefficients for the five cranial characters used to separate the three East African species of Taterillus 143
7, Means and standard error for the five cranial characters in their order of inclusion in the discriminant analysis of East African Taterillus 144
8, Loadings of the cranial measurements on the first three components of populations and holotypes of East African Taterillus . . . 150
9, Loadings of the cranial measurements on the first three components of population means of geographic samples of Taterillus . , 157
xii xiii
LIST OF TABLES—Continued
Table Page
10. Somatic chromosome number and morphological types of six species and two subspecies of Taterillus 163
11. Distance matrix for the seven species of Taterillus 181
12. Loadings of the cranial measurements on the first three components of means of the seven species of Taterillus 185
I ABSTRACT
The Rodent genus Taterillus (Family Cricetidae;
Subfamily Gerbillinae) occurs in the Sahel Savanna and the
Sudan and Guinea Woodlands of Africa, between the Sahara
Desert and the high forest. It is found west to Senegal and Mauritania, east to Somalia, and south to northern
Tanzania. A single species can occupy all of the above vegetation zones.
The systematic arrangement of Taterillus presented is a result of examination of the taxonomic relationships of the species in this genus. The taxonomy is a result of:
(1) a study of the history and use of the many named taxa;
(2) analysis of geographic variation using univariate and multivariate statistical techniques; and (3) examination of evolutionary and zoogeographic relationships.
The available data suggest only seven species rather than the 19 species indicated by original descrip
tions. In the past, small population samples from widely separated areas were described as different taxa using characteristics which were neither clearcut nor consistent,
Examination of over 3000 specimens from 21 countries,
including many not available to previous workers, and all
of the holotypes, shows that minor variation in size,
shape, color, and other features used as diagnostic of
xiv XV taxa in Taterillus are not necessarily indicative of taxonomic relationships.
The species I recognize are: Taterillus arenarius
Robbins, 1974; T. congicus Thomas, 1915; T. emini (Thomas,
1892); T. gracilis (Thomas, 1892); T. harringtoni (Thomas,
1906); T. lacustris (Thomas and Wroughton, 1907); and T. pygargus (F. Cuvier, 1838). The descriptions of these species are based on cranial morphology, karyotypic differences, distribution, and possible habitat selection.
Typical data show that five species described from Ethiopia,
Kenya, Somalia, and the Sudan, with only minor geographic variation in cranial morphology and the same karyotype
(2N=44), are in fact a single species (T. harringtoni).
A total of six chromosomal forms have been found.
Only minor population variation occurs within a species in the number of their major chromosome arms (fundamental number), Several localities in both West and East Africa have two different chromosomal forms occurring together with no hybridization yet found. Based on diploid and funda mental number, plus chromosome and cranial morphology, the six karyotyped species can be divided into three groups.
These are: (1) 2N=54, FN=64-66 (T. congicus) ; (.2) 2N=44,
FN=62 (T, harringtoni); and (3) the West African taxa
(2N=22/23, FN=36-4 0 - T. pygargus; 2N=28, FN=44 - T. lacustris; 2N=30, FN=36 - T. arenarius; 2N=36/37, FN=42-
44 - T. gracilis. xvi
Most differences in cranial morphology involve age variation. Arbitrary age classes were established based on molar tooth morphology and only a single adult age class was used to eliminate morphological variation due to age.
Secondary sexual differences were not found.
A major problem in Taterillus systematics concerns distinguishing between sympatric species which are morphologically similar, and allopatric species which are closely related. This report shows the importance of using all available data when studying the biology and systematics of an}, mammalian group. It can also be seen that compu terized statistical procedures are necessary when working with many specimens from a large geographic area.
This study has resulted in a systematic arrangement based on the biological relationships of members of the genus Taterillus so that future studies on ecological relationships can be carried out to give more meaningful results, CHAPTER 1
INTRODUCTION
The purpose of this study was first to examine the taxonomic and biological relationships of the species in the . genus Taterillus. After forming a new systematic arrange ment of the taxa, an examination of the evolutionary rela tionships, ecological differences, and zoogeography of the species was made. This study is divided into three sections
—nomenclature, geographic variation, and evolutionary and zoogeographical relationships. The section on nomenclature examines the history and use of the many named taxa in the genus. The study of geographic variation is included in sections of species accounts and specific relationships using both univariate and multivariate statistics, I will show that only five of the 10 species recognized prior to this study are distinct. These species are Taterillus congicus
(monotypic), Taterillus emini (monotypic), Taterillus gracilis (four subspecies), Taterillus lacustris (monotypic), and Taterillus pygargus (monotypic). In addition to these five species, one monotypic taxon is elevated to specific status (Taterillus harringtoni) and another recently described monotypic species, Taterillus arenarius (Robbins,
1 2
1974b), is recognized, resulting in a total of seven species within the genus.
Because a fossil record of the genus is lacking, I have analyzed the evolutionary relationships of the seven species to each other by comparing their external and cranial morphology, karyotypes, habitats, and distribution.
The section on specific relationships also should be used as an appendix to the section on systematic accounts. It includes the characters which were used in determining the differences between the species. The placement of the specific characters in a single section makes comparisons easier, and is in addition to the comparisons between adjacent taxa within the accounts of individual species,
The genus Taterillus (Rodentia: Cricetidae) is a member of the Old World subfamily Gerbillinae, which includes the genera Ammodillus, Desmodilliscus, Desmodillus,
Dipodillus, Gerbillus, Gerbillurus, Meriones, Microdillus,
Pachyuromys, Psammomys, Sekeetamys, and Tatera. Taterillus occurs in Africa south of the Sahara desert, from Senegal and Mauritania, east to Somalia, and south to northern
Tanzania (Wroughton, 1910; De Beaux, 1922; Rosevear, 1969;
Jullien and Petter, 1969; Petter, 1970; Hubbard, 1973),
Apparently, members of this genus occur only between the high forest in the south and the desert in the north
(Dekeyser, 1955; Booth, 1960; Rosevear, 1969; Petter, 1970).
The result is a distributional belt across Africa, between 19 degrees north latitude and six degrees south latitude
(Fig. 1). Suitable habitat is restricted to the Sahel
Savanna and Sudan and Guinea Woodlands. For a detailed description of these vegetational zones and their typical vegetation see Rosevear (1953, 1965) and Hopkins (1965).
A following description of one species in the genus Taterillus (T. gracilis) has been summarized from
Rosevear (1969). It is small and lightly built, its head and body is about four and one-half inches long and its pencilled tail averages about twenty-five per cent longer.
The entire pelage is soft and fine. The dry zone forms are of a characteristic sandy hue. This pale coloration lies solely in the ends of the hairs while the bases are slate.
Even the darker Guinea Woodland forms show some tendency to sandiness in the pelage. The underparts are pure white; the backs of both fore and hind feet are white; the ears are fairly large. The tail is pale sandy-brown above to white below for most of its length and the terminal part has long dark hairs forming the pencil. The hindfeet and legs are longer than average for rats and by habit they are somewhat saltatorial. The first and fifth digits of the hindfeet are much shorter than the remaining three. In the forefeet the first digit is rudimentary but clawed; the fifth digit is reduced.
In an attempt to synthesize the knowledge of African rodents, researchers held a conference in Brussels in 1964. 4
ALGERIA
MAURITANIA
SENEGAL^ ,1 3 ^•S 2\GUINEAY ^SOMALI/ cr >1 VOR Y\
[COASTA xxxx. CENTRA^ ETHIOPIA <0^ ^ ^AF^REPS v CAM E ROON EQUATORIAL GUINEA-^T GABON( C?J UGANDA/ 1 AFARS AND ISSAS ZAIRE \K E N Y A 2 PORT. GUINEA & 3 TOGO 4 DAHOMEY TANZANIA 5 RWANDA 6 BURUNDI 7 SWAZILAND ANGOLA 8 LESOTHO MALAWI ZAMBIA
MALAGASY) SOUTH-WEST RHODESIA'VV AFRICA REPUBLIC/
500 1000 i Mi le 500 1000 Ki lometers SOUTH <^> AFRICA
Fig. 1. Distribution of the genus Taterillus in Africa. 5
At that time the existing study collections of African rodents were estimated to contain 80,000 specimens (Davis,
1966a). As Davis pointed out, most of the material had already been studied. Most had been reported on a regional basis, resulting in a changing of nomenclature from place to place. Few attempts had been made to determine whether a species recorded under one name in one region was the same as a second species recorded from another region.
Since the Proceedings of the Colloquium on African
Rodents (1966), the Smithsonian Institution sponsored an
African Mammal Project which was directed by Dr. H. W.
Setzer, National Museum of Natural History, Washington,
D. C. That project added approximately 90,000 additional specimens of small mammals to the previous total. These new specimens make it possible to determine distributions and species relationships i~ many genera, The knowledge of the distribution of Taterillus as well as the study of its geographic variation has reached the point where synthesis is now possible.
Allen (1939) and Ellerman (1941) listed 16 species and 20 subspecies of Taterillus in their compilation of the known taxa in the genus, Setzer (1956) in a study related only to the Sudan, combined 10 taxa into two species,
Rosevear (1969) reduced the three West African species to two, Petter (1975), in his checklist and key to the
Gerbillinae, retained 10 species and 19 subspecies in 6
Taterillus. Petter did not attempt to discuss the validity
of the species and subspecies he recognized. Also, he did
not include all of the known taxa previously described as a
part of Taterillus. Petter (1975) stated that the genus
Taterillus needed further revision.
Ellerman (1941) and Rosevear (1969) suggested that
the described West African species of Taterillus were a
single polymorphic species. Ellerman also stated that there
was probably only one valid species in the genus. Recent
cytogenetical data (Matthey, 1969; Matthey and Petter, 1970;
Matthey and Jotterand, 1972; Genest and Petter, 1973;
Tranier, Hubert, and Petter, 1974; Robbins, 1972, 1974a;
Petter, 1972) indicate that there are four species in West
Africa and at least six species in the genus.
In the past, the morphological species concept was
applied to what were thought to be allopatric populations
(for example, see the discussion by Rosevear [1966J on the
description of T, nigeriae by Thomas [1911]). In many
instances, only small population samples were available from
geographically separated areas. Some of these samples were
described as distinct taxa using characteristics which were
neither clearcut nor consistent (refer to sample sizes and
characteristics used in species descriptions by Thomas and
Hinton [1923J ; Wettstein [1916]; and Wroughton [1907, 1910]).
Often they were simply expressions of minor variations in
size, shape, and color. Examination of the many additional 7 examples of Taterillus not available to previous workers, clearly shows that various features used as diagnostic of taxa in this genus in the past are not necessarily indica tive of taxonomic relationships. Age, seasonal, geographic, and soil color differences influence pelage color and are well documented in many mammals (Ranck, 1968; Robbins, 1973;
Hubert, 1973; Poulet, 1972a, 1972b, 1974).
Taterillus gracilis, for example, occurs in all vegetation zones in which the genus is known to occur
(Robbins, 1974b). Although karyotype comparisons between population samples of this species in the different habitats show them to be almost identical, there are great differ ences in pelage color.
Various age groups within a single, large population sample of Taterillus show significant statistical differ ences for most of the external and cranial characters measured (Robbins, 1973). My subsequent analyses have shown that these age differences occur in all of the taxa and populations of Taterillus examined. Age differences within a population are, in most cases, greater than differences
between animals of the same age in different populations.
My examination of the small population samples used by
previous authors to describe a taxon in Taterillus has shown
that some of these samples are composed of mainly young animals, while other populations are of old animals
(Rosevear, 1969), I have found that differences in pelage 8 color as a result of age and soil color as well as variation in size due to age, encompasses many of the populations originally thought to be separate species,
Taterillus is a complex genus, able to adapt to a wide variety of habitats. At least two species occur sympatrically in several localities in both West and East
Africa (Matthey and Jotterand, 1972; Petter et al., 1972;
Genest and Petter, 1973; Hubert, 1973; Robbins, 1974b).
Sympatric species present a greater problem and appear to be more common in Taterillus than previously thought.
Sympatric populations of species that are difficult to distinguish morphologically, but are reproductively isolated, are usually called sibling species. Their morphological similarities make them difficult to detect or differentiate. Closer study, including cytogenetic com parisons, usually show some morphological differences
(Mayr, 1970), Originally, all of the Taterillus in Senegal were thought to be T. gracilis. Cytogenetic studies by
Matthey and Jotterand (1972) and Petter et al. (1972), indicated that two sympatric species (T. gracilis and T. pygargus) were present, Matthey, Jotterand, and Petter were able to distinguish the two species only by karyotype or serum protein analysis. They were not able to find any measurable external or cranial characters to separate the two species, Petter et al, (1972) also showed that 9 breeding studies indicated that no viable offspring resulted by crossing the two species.
The primary criterion of species rank in a natural population is reproductive isolation. Any morphological differences displayed by a natural population is a secondary by-product of genetic divergence resulting from reproductive isolation (Meester, 1963; Mayr, 1970). The degree of morphological similarity in sibling species is an indica tion not only of genetic similarity, but also of develop mental homeostasis, A reconstruction or changing of the genotype, resulting in reproductive isolation of two species, can take place without visible effect on the morphology of the phenotype (Mayr, 1970).
Developmental homeostasis has been demonstrated in
Taterillus gracilis and T. pygargus. They apparently share the same habitat, food supply, and distribution in Senegal
(Hubert, 1973; Poulet, 1972a, 1972b, 1973, 1974). Neverthe less, the distinctiveness of the above two species has been demonstrated even though morphological distinctiveness is difficult to observe.
With new and old cytogenetical data, as well as detailed morphological analyses based on univariate and multivariate statistical techniques, geographic variation, color variation, and differences between sympatric species will be analyzed, My examination of most of the known 10 specimens of Taterillus makes it possible for me to deter mine the distribution of the taxa.
This study will analyze the evolution, ecology, zoogeography', and environmental influences to present a clearer understanding of the biological organization of the genus Taterillus. This organization is possible by giving a more detailed description of the species including a redefinition of these species and subspecies based on morphological characters, karyotypic differences, distribu tion, and habitat selection. CHAPTER 2
' HISTORY OF TATERILLUS NOMENCLATURE
The genus Taterillus was described as distinct from both Gerbillus and Tatera by Thomas (1910). He designated
T. emini as the type of the genus and included gracilis, harringtoni, and lacustris as a part of Taterillus. With the description of two additional species (T. butleri and
T. osgoodi), Wroughton (1910) arranged the first key to the six described species of Taterillus. Subsequent descrip tions discussed briefly the relationships of the new species or subspecies to previous names. Although the genus had become widely known and used, Wettstein (1916) assigned many gerbilline genera, including Taterillus, as subgenera of
Gerbillus. His three new species, subsequently included in
Tater illus, were placed in separate subgenera. These were
Gerbillus (Tatera) rufa, G, (Taterillus) kadugliensis, and his newly named subgenus Taterina, with G. (Taterina) lorenzi.
Taterillus pygargus (F. Cuvier, 1838) was the first species described that is now considered to be a member of this genus. As originally described, it was thought to be a Gerbillus, and assigned most often to G. pyramidum
(Anderson, 1902). Petter (1952) reexamined the holotype and
11 12 found that due to some mixup in the collection, the skin was of an animal in one genus, the skull, another. He desig nated the skull as the holotype. The skull drawings pub lished by FCuvier (1838) show that it is a Taterillus.
Since pygargus was not recognized as a Taterillus until 1952, the two species first assigned to Taterillus were T. emini and T. gracilis (Thomas, 1892). These were also originally described as being a part of the genus
Gerbillus• Subsequently, they were included as a part of the genus 'I'atera by Wroughton (1906) as were T. harringtoni
(Thomas, 1906) and T. lacustris (Thomas and Wroughton, 1907).
The first attempt at intergredation of the rapidly increasing species names was that of Hatt (1934). He suggested that congicus, butleri, clivosus, and lacustris belong to T, emini and also described a new subspecies (T. e. anthonyi).
With many species and subspecies names available,
Allen (1939) gave the first listing of all of the nominal taxa that he recognized as Taterillus. Ellerman (1941) followed the arrangement of Allen. The list was not an attempt to revise the genus, but rather only a compilation of the known taxa, as follows:
Taterillus Thomas, 1910
Taterillus Thomas, 1910 (genotype Gerbillus emini Thomas) Taterina Wettstein, 1916, subgenus of Gerbillus, type Gerbillus lorenzi Wettstein Taterillus butleri Wroughton, 1910 (incl. Taterillus kadugliensis (Wettstein, 1916) and Taterillus lorenzi (Wettstein, 1916)
Taterillus clivosus Thomas and Hinton, 1923
Taterillus congicus Thomas, 1915
Taterillus emini emini (Thomas, 1892) Taterillus emini anthonyi Hatt, 1934 Taterillus emini zammarani De Beaux, 1922
Taterillus gracilis gracilis (Thomas, 1892) Taterillus gracilis angelus Thomas and Hinton, 1920
Taterillus gyas Thomas, 1918
Taterillus harringtoni (Thomas, 1906)
Taterillus lacustris (Thomas and Wroughton, 1907)
Taterillus lowei Dollman, 1914
Taterillus melanops G. M. Allen, 1912
Taterillus nigeriae Thomas, 1911
Taterillus nubilus nubilus Dollman, 1911 Taterillus nubilus illus tris Dollman, 1911
Taterillus osgoodi Wroughton, 1910
Taterillus perluteus Thomas and Hinton, 1923
Taterillus rufus (Wettstein, 1916)
Taterillus tenebricus Dollman, 1911
Since Ellerman (1941), two additional taxa have recognized or described as being a part of Taterillus.
Taterillus pygargus (F, Cuvier, 1838), see Petter (1952) and Petter et al. (1972)
Taterillus nubilus meneghettii Toschi, 1946 14
A summary of the known distribution of the East
African species and subspecies of Taterillus accompanied the new subspecies description by Toschi (1946).
A revision of the taxa occurring in the Sudan was made by Setzer (1956). Two species were retained and the taxa allocated as follows.
Taterillus congicus
Taterillus emini emini T. e. butleri (incl. T. kadugliensis) T. e. clivosus T. e. anthonyi T. e. gyas T. e. perluteus T. e. rufus (incl. T. lorenzi)
The West African forms were examined by Rosevear
(1969). He recognized the following two species, although he was apparently not aware of T. pygargus as discussed by
Petter (1952).
Taterillus gracilis T. g. angelus T. g. lacustris
Taterillus nigeriae
The latest listing is in a checklist and key by
Petter (1975). He rearranged all of the species and sub species as follows,
Taterillus congicus
Taterillus emini emini T. e, anthonyi T, e. butleri (incl, kadugliensis and lorenzi) T. e, osgoodi T, e, zammarani Taterillus gracilis gracilis T. g. angelus
Taterillus gyas
Taterillus lacustris lacustris T. 1_. per luteus
Taterillus lowei lowei T. 1_. harringtoni
Taterillus nigeriae nigeriae T. n. clivosus
Taterillus nubilus nubilus T. n. illus tris
Taterillus pygargus
Taterillus tenebricus
The following resume of the known karyotypic information for Taterillus was compiled from Matthey (1969),
Matthey and Petter (1970), Matthey and Jotterand (1972),
Genest and Potter (1973), Tranier et al. (1974), Robbins
(1972, 1974a), and Petter (1972).
The fundamental number (FN) given below, includes only the number of major autosomal arms, and follows most
American authors. They have been revised from the number given by Matthey, Petter, Genest, Jotterand, and Tranier, which included the arms of the sex chromosomes.
Taterillus congicus--from Central African Republic and Chad, is 2N=54, FN=64-66,
Taterillus emini--from Central African Republic,
Ethiopia, Kenya, and Somalia, is 2N=44, FN=62-64 (equals
T, harringtoni, this report), 16
Taterillus gracilis—from Senegal, Upper Volta,
Ivory Coast, and Ghana, is 2N=36/37, FN-42-44.
Taterillus lacustris--from Cameroon, is 2N=28,
FN=4 4.
Taterillus nigeriae--from Mauritania, is 2N=30,
FN=3 6 (equals T. arenarius; see Robbins [1974b] and this report).
Taterillus pygargus--from Senegal, is 2N=22/23,
FN=40-44.
The above karyotypic data are useful in the follow ing attempt to clarify the systematic status, species dif ferences, and evolutionary and zoogeographic relationships among the many named taxa in the genus Taterillus. CHAPTER 3
MATERIALS AND METHODS
A total of 3315 specimens of Taterillus were examined during the course of this study. Most specimens were standard museum skins and skulls accompanied by the appropriate field data. Additional material consisted of skins without skulls (SO), skulls without skins (S), and fluid preserved specimens (Fl). All of the holotypes of the nominal taxa of Taterillus were examined except those of T. rufus, T. kadugliensis, and T. lorenzi. These were examined for me in the Vienna Museum by Duane A. Schlitter.
The skulls of these holotypes were figured by Wettstein
(1917). For the purpose of this study, only adult specimens of age class 4 were utilized for statistical comparisons.
The criteria for recognizing this age class in Taterillus was taken from Robbins (1973) as follows, I found that the lower molar series was most useful in defining age classes.
The stage of eruption, and comparisons of the depths of the valleys between the lophs to determine the degree of wear of the entire molar series was used but the condition of the third molars was most useful. In age class 4, the lower third molar shows some signs of wear and is usually above
17 the level of the occlusal surface between the first and second molars. The auditory bullae are ossified as shown by their being semi-transparent. Wear is pronounced in the first and second molars, although all of the lophs are visible and closure of the valleys has not occurred.
Plan of Treatment
In the chapter on systematic accounts the genus
Taterillus is introduced by a comparison with other gerbilline genera. This is followed by an alphabetical listing of the species recognized by me. Within each species recognized, subspecies are listed alphabetically.
The following subdivisions are used in the treatment of each taxon.
Scientific Name: The currently accepted (by this author) binomial name combination in agreement with the
International Code of Zoological Nomenclature (Stoll et al.,
1961) is followed on the same line by the name of the author. The scientific name is succeeded by the original name of the taxon, the name of the author (s) , a reference to the original account, the date on which it was published, and the type locality as given by the original author.
Modifying information, if given, is included in parentheses.
This is followed by a partial synonomy,
Holotype: The sex, age, type of museum preparation, museum in which it is housed, museum number, collector, date 19 of collection, and collector number are given. When any of
the above information is not given it is because it is not available or not found by me.
General Distribution of the Species: The known distribution of the species is given. A distribution map
is included indicating localities from which I have examined specimens. These are indicated by solid black circles. The shaded portions of the maps indicate the range of a given
species or subspecies as judged from record of occurrence.
Obviously, no species or subspecies occupies all parts of
the shaded area, but is limited to those areas of suitable
habitat.
Distribution of the Subspecies; When subspecies are
recognized, they are arranged alphabetically and their
range briefly outlined. These are also included on distribu
tion maps.
Specimens Examined'. Specimens examined, unless
otherwise indicated, are skins and skulls. The name of the
country is given first, followed by the total number from
that country, the name of the province or district (when
known), the exact locality of capture, and the exact number
from that locality. Unless otherwise indicated, the speci
mens are in the collection of the Division of Mammals,
National Museum of Natural History, Smithsonian Institution,
Washington, D, C, (USNM), The following museums are
included and their abbreviations indicated: The American 20
Museum of Natural History, New York (AMNH); Los Angeles
County Natural History Museum, Los Angeles, California (LA);
Field Museum of Natural History, Chicago (FM); Museum of
Comparative Zoology, Harvard University, Cambridge,
Massachusetts (MCZ); Museum National d'Histoire Naturelle,
Paris, France (MNI-IP) ; Musee Royal de l'Afrique Centrale,
Tervuren, Belgium (RGMC); University of Antwerp, Department of Biology, Antwerp, Belgium (AT); Institut Royal des
Sciences Naturelles de Belgique, Brussels, Belgium (BR);
Museo Civico Di Storia Naturelle "Giacomo Doria," Genoa,
Italy (MSNG); Naturhistorisches Museum Zoologische Abeilung,
Vienna, Austria (VM); British Museum (Natural History),
London, England (BM); Museum fur Naturkunde der Humboldt-
Universetat zu Berlin, Democratic Republic of Germany (DR);
Natur-Museum Senckenberg, Frankfurt, Federal Republic of
Germany (FS); Universitets Zoologiske Museum, Copenhagen,
Denmark (MD); and National Museums of Kenya, Nairobi, Kenya
(NA).
The countries are listed alphabetically, and within the provinces or districts the collecting localities are arranged in sequence from north to south; those having approximately the same latitudinal coordinates are arranged from east to west.
Underlined localities are those that I have not been able to find in any reference source listed in Appendix A.
Localities with modifiers are listed in the Gazetteer 21
(Appendix A) if specific locality coordinates were given by the original collector. In other instances, only the coordinates for the town or village are given. All localities, except for those underlined, are alphabetically listed in their country of occurrence, with coordinates, in
Appendix A.
Published Records: In this section, all known records of specimens from the literature, not examined by me, are summarized, This includes only those specimens that
I consider to be included in the species under discussion.
Description: A general description of the skin and skull of the holotype is given as follows.
1. Skin. Capitalized color terms are those of Ridgway
(1912). Non-capitalized color terms are used in
those instances when a detailed color description
was unnecessary. Other characters of the skin are
included.
2, Skull. Detailed descriptions involving gross
features of the skull are given.
Measurements; Measurements are of the holotype and a representative sample of the taxon of age class 4 animals.
For the age class 4 specimens used in this study, five external body measurements and weight were recorded from the specimen labels. The mammals used in this study that are housed at the National Museum of Natural History, were 22 all prepared under field conditions as museum study skins following the procedures given by Setzer (1968), with the exception of those preserved in fluid. Twenty-four differ ent measurements were made of various parameters of the skull and its parts. The list of measurements with abbreviations used in the text are given in Appendix B, as well as definitions and a skull diagram showing the reference points, between which the measurements were taken. The skull measurements were taken with the aid of dial calipers and recorded to the nearest 0.01 millimeter.
Considering that the last decimal place was only an estimate, for statistical purposes the measured values were rounded off to the nearest significant figure (0.1 millimeter).
I rounded to the nearest significant figure according to the method discussed by Simpson, Roe, and Lewontin (1960).
Comparisons; Comparisons are limited to those species which are easily confused with each other.
Taxonomic Conclusions: The taxonomic history of the taxon is covered in this section, and problems and events relating to systematics and dispersal are discussed.
Ecological Observations; This section deals with the habitat requirements of the taxon and discusses other species of rodents which share the same habitat.
Description of New Taxa; The same procedure is used under "Scientific Name" except that the newly proposed name is followed by designation of a holotype. The data on the holotype is the same as that given under "Holotype" above.
The new taxon is included alphabetically by species, or by
subspecies within a species. If subspecies are recognized,
the above subdivisions are included in its treatment.
Analysis of Variation and Diagnostic Characters of
the Species or Subspecies: This subdivision is included
under "Cranial Morphology" in the chapter on specific
relationships. The populations analyzed consisted of
specimens selected from localities having a sufficiently
large sample size or from closely adjacent localities of a
particular taxon and were treated statistically. Statis
tical procedures requiring computer analyses were performed
on the Honeywell 120 computer at the Information Systems
Division, Smithsonian Institution, on the CDC 6600 utilized
by the Smithsonian, and the Department of Interior IBM 360.
Univariate analyses of both geographic and intra-
population variation were performed using a program (DSTAT)
written by the Information Systems Division, Smithsonian
Institution, Multivariate analyses were performed with the
Smithsonian-written program MAINPG for principal components
analysis, the BMD07M program for step-wise discriminant
function analysis, and the NT-SYS programs for cluster
analysis, principal components analysis, nonmetric multi
dimensional scaling analysis, and single linkage cluster
matrix. 24
When using the NT-SYS programs (developed at the
University of Kansas by F. J. Rohlf, R. Bartcher, and J.
Kishpaugh), the samples used (sample equals OTU or Opera tional Taxonomic Unit) are either single or grouped localities discussed in their appropriate sections later, and the value for each variable within a sample is the mean for the measurement. For a complete discussion of the statistical procedures used, see Appendix C. CHAPTER 4
SYSTEMATIC ACCOUNTS
The subfamily Gerbillinae is represented by many genera and subgenera. Of these, I have found that
Taterillus can be confused with the genus Dipodillus, the genus Gerbillus and its subgenera Gerbillus and
Hendacapleura, the genus Gerbillurus, and the genus Tatera.
For generic status of the Gerbillus group see Wassif,
Lutfy, and Wassif (1969).
Externally, Dipodillus can be separated from the others by its small size and a tail length less than its head and body length. Cranial comparisons show that two morphological groups can be recognized in the remaining genera: (1) the mastoidal portion of the bullae are markedly inflated in Gerbillurus (part), Gerbillus (Gerbillus), and
Gerbillus (Hendacapleura); and (2) the mastoidal portion of the bullae is not inflated in Gerbillurus (part), Tatera, and Taterillus.
Dividing Gerbillurus into two groups based on the mastoids usually separates the G. peaba group (mastoids reduced or absent) from the G. valinus group (mastoids inflated), It is probably that there are two separate genera in the Gerbillurus group and they are more closely
25 related to Tatera and Taterillus (Petter, 1974) than to the
East and North African Gerbillus. Since G. peaba and G. valinus both occur in Southern Africa, their range does not overlap that of Taterillus and further distinctions are not needed here.
In the genera sympatric with Taterillus and of similar morphology, Tatera can be distinguished by its larger overall body size, larger and more robust skull, and in animals of similar size, by the much larger molar teeth of Tatera. Tatera is also distinguishable in that they have small posterior palatine foramina which do not usually extend anteriorly to the first molar. The posterior
palatine foramina in Taterillus are long and always reach
the first molar.
Species of Gerbillus (Gerbillus) sympatric with
Taterillus can be distinguished by the plantar surfaces of
the hind feet being well haired. Taterillus has at most
only a narrow band of hairs in the ankle region. Gerbillus
(Hendacapleura) has naked plantar surfaces on their hind
feet, but both subgenera of Gerbillus can be distinguished
from Taterillus by their greatly inflated mastoidal bullae.
In Taterillus, the mastoidal bullae are reduced.
The following placement of the genus Taterillus
within the class Mammalia follows Simpson (1945) and Petter
(1975). 27
Order Rodentia
Family Cricetidae
Subfamily Gerbillinae
Genus Taterillus Thomas
Taterillus Thomas, Ann. Mag. Nat. Hist., Ser. 6, 8:222, August, 1910. Type species Taterillus emini (Thomas). Tatera, Wroughton, Ann. Mag. Nat. Hist., Ser. 7, 17:474-499 (part), May, 1906. Gerbillus, Wettstein, Anz. Kais. Akad. Wiss . Wien, 53 (14):151-154 , 1916.
Key to the Species of Taterillus
The following key is suitable only for age class 4 animals. Obviously, karyotype and some bone shape features will separate younger or older animals into these species.
For comparisons of auditory bullae shapes and size, see
Fig. 2.
1 A. Skull long and broad; greatest length of skull greater than 3 6.0 nun; zygomatic breadth greater than 18 mm; incisors long.
1 B, Skull short; greatest length of skull less than 36.0 mm; zygomatic breadth less than 18 mm; incisors short,
2 A. Alveolar length of first upper molar greater than 2.8 mm; length of posterior palatine foramina usually short and narrow—less than 3.9 mm. T. emini
2 B, Alveolar length of first upper molar less than 2,8 mm; posterior palatine foramina long and wide—greater than 3.9 mm in length; 2N=54, T, congicus
3 A, Auditory bullae narrow, uninflated and small; skull small; incisors short; 2N=28, T, lacustris Fig. 2. Diagram of auditory bullae for Taterillus -- Shown is the ventral view of the left bulla; anterior is at top of figure; lateral is to the right. T. arenarius, USNM 401919 (holotype) from Tiguent, Mauritania; T. pygargus, MNHP 496 from Fete-Ole, Senegal; T. g. gracilis, MNHP 514 from Saboya, Senegal; T. lacustris, MNHP 7 85 from Mora, Cameroon; T. congicus, MD 22 66 (topotype) from Poko, Zaire; T. emini, AMNH 89677 from Rhino Camp, Uganda; T. harringtoni, USNM 484001 from Voi, Kenya; T. g. meridionalis, USNM 435359 (holotype) from Wulasi, Ghana; T. g. angelus, USNM 403387 (topotype) from Panisau, Nigeria; T. g. nigeriae, USNM 377310 (topotype) from 1 mi S Kabwir, Nigeria. 28
T. arenarius T. pygargus T. g. graci li s T. lacustris
T. congicus T. emini T. harringtoni
T T. g. angelus T. g. nigeriae
Fig, 2. Diagram of auditory bullae for Taterillus. 29
3 B. Auditory bullae relatively broad, inflated, and medium-sized to large; karyotype other than 2N=28. 4
4 A. Braincase broad; posterior palatine foramina long; auditory bullae large and greatly inflated; occurs only in East Africa; 2N=44. T. harringtoni
4 B. Braincase relatively narrow; posterior palatine foramina short; auditory bullae not as large nor as inflated; occurs only in West Africa; 2N=22/23, 30, or 36/37. 5
5 A, Auditory bullae inflated anteriorly and antero- laterally; skull robust; 2N-22/23 or 30. 6
5 B. Auditory bullae not as inflated anteriorly or antero-laterally; skull relatively large but gracile or narrow; rostrum long; 2N=:36/37. T. gracilis.
6 A. Skull short but robust; greatest length of skull usually less than 35.0 mm; rostrum short; 2N-22/23. T. pygargus
6 B. Skull long and robust; greatest length of skull usually greater than 35.0 mm; rostrum long; 2N=30. T. arenarius
Taterillus arenarius Robbins
Taterillus arenarius Robbins, Proc, Zool, Soc, Wash,, 87:399, December, 1974, Type locality, Tiguent, Trarza Region, Mauritania. Taterillus nigeriae, Matthey, Mammalia, 33:522-538, 1969 .
Holotype: Adult male, skin and skull, USNM 401919, collected by C. B, Robbins, 9 April, 1967, original number
799, General Distribution of the Species; Mauritania,
Niger, and presumably Mali (Fig, 3), 15°W 10°W 5° W
MAURITANIA
MALI NIGER
•15°N
SENEGAL
GAMBIA
PORT/ UPPER VOLTA "/GUINEA
GUINEA
SIERRA NIGERIA LEONE IVORY GHANA COAST
300 CAMEROON •5°N 0 500 Kilometers
T~1 3 Distribution and known localities of Taterillus arenarius Specimens Examined; Mauritania (97). Inchiri
Region: Bou Rjeimat, 1 (MNHP); Trarza Region: 11 km N
Nouakchott, 5; 6 km E Nouakchott, 3; Tiguent, 48; Garak, 10;
Brakna Region: 3 km S Aleg, 14; Gorgol Region: Kaedi, 10;
Guidimaka Region: Passe de Soufa, 6. Niger (15). Agadez
Region: 5 km NE Agadez, 7; 30 km S In-Gall, 3; Tahoua
Region: 120 km S In-Gall, 5.
Description: Interauricular, interorbital, and rostral areas same color as dorsum and varying from Sayal
Brown to Snuff Brown; circumorbital region, postauricular
patches, mystacial and pectoral areas, fore- and hind limbs,
and entire underparts white; cheeks and sides Cinnamon-Buff; dorsal hairs plumbeous basally, the brown-pigmented portion
only 2-3 mm in length, and some hairs finely tipped with
dark brown; fore- and hind limbs have five digits with
claws, plantar surfaces naked except for a narrow band of
white hairs in the ankle region; pinna of ear long and
almost naked, color almost the same as the dorsum, and
anterior margin with short buff-colored hairs; vibrissae
long and composed of both white and dark brown hairs; tail
long and uniformly Cinnamon-Buff basally with the dorsal
hairs interspersed with darker brown hairs grading to a
terminal Mummy Brown pencil, yet ventrally, the Cinnamon-
Buff color grades to white in the region of the pencil.
Skull—relatively large for the genus and moderately
robust; zygoma heavy; lachrymals large; molariform teeth 32 medium-sized; auditory bullae large and inflated; para- pterygoid fossae large and deep but not markedly flared; rostrum relatively long and slightly expanded anterior to the infraorbital shield; nasals relatively broad and long; braincase flattened.
Measurements: Measurements of the holotype (age class 5; see Robbins [1973J) followed by averages and extremes of eight adults (age class 4) from the type- locality are, respectively: TOT 289, 279 (269-300); HB 128,
117 (111-124); TAL 161, 163 (158-176); HF 33, 33 (32-34);
E 22, 21 (19-22); W 59, 46 (42-52); GLS 36.9, 35.6 (34.9-
3 6.3); OCNL 3 5.7, 34.3 (33.4-35.3); BAL 26.9, 25.7 (25.2-
26.5); ZB 18.2, 17.6 (17.4-17.8); CBAL 28.8, 27.5 (27.1-
28.2); BBC 14.6, 14.4 (13.6-14.7); IOC 6.0, 6.1 (5.8-6.6);
BR 4.5, 4.6 (4.4-4.8); GBAB 13.9, 13.4 (12.8-13.8); BrMlMl
7.3, 7.1 (6.7-7.3); DAL 9.0, 8.9 (8.6-9.6); PAL 15.3, 15.0
(14.6-15.3); PPAL 11.3, 10.6 (10.0-11.0); LAPF 6.2, 5.9
(5.5-6.5); LPB 7.4, 7.3 (7.0-7.7); LAB 10.1, 9.4 (9.0-9.7);
BAB 6.2, 5,8 (5.6-6.1); LN 15.0, 14.8 (14.1-15.3); LR 13.4,
12.9 (12.7-13.6); FNL 28.2, 27.2 (26.4-28.3); DC 13.7, 13.6
(13,0-13.8); LPPF 3.8, 3.7 (3.2-4.2); ALMl 2.8, 2.7 (2.4-
3,1); ALM 5.1, 5.1 (5.0-5.4).
Comparisons: Taterillus arenarius can be dis tinguished from T. pygargus, where the two occur together, by its markedly longer and more robust skull; flatter, less rounded braincase; broader rostrum; slightly larger 33 incisors; more bulbous bullae; slightly smaller cheek teeth; parapterygoid fossae larger (deeper) but not so widely flaring; and paler dorsal coloration. At Garak, Mauritania, where T. arenarius is sympatric with T. pygargus, the latter are darker dorsally and laterally--Snuff Brown and Tawny-
Olive respectively.
Taterillus arenarius, when compared to T. gracilis, has a broader, more robust skull, although the skulls are similar in length. The bullae of T. arenarius are more inflated anteriorly, antero-laterally, and ventrally. The pelage color of T. arenarius is conspicuously paler.
Taterillus arenarius closely resembles T, lacustris, from which it can be distinguished by its somewhat narrower nasals, flatter braincase, and more inflated bullae. The pelage color is similar.
Taxonomic Conclusions: A specimen of Taterillus from Bou Rjeimat, Mauritania, has a reported karyotype of
2N=30, FN=36 (Matthey, 1969) and was identified by Matthey
(1969) and Petter (1970, 1975) as T. nigeriae. This allocation was made because at that time, only two species of Taterillus were recognized by Rosevear (1969) as occurring in West Africa. Of these two species (T. gracilis and T, nigeriae) Matthey (1969) attributed a karyotype of
2N=22/23 to T. gracilis. However, my examination of available specimens of T. nigeriae, including the holotype, demonstrated that the Mauritania specimen was distinct. 34
With additional specimens available from Mauritania, I named a new species as Taterillus arenarius (Robbins, 1974b).
In this report I consider T. nigeriae to be a part of the
T. gracilis complex, and it is compared to and distinguished from T. arenarius in the section on T. gracilis nigeriae.
Ecological Observations: Taterillus arenarius occurs only in the vegetated sandy habitats of the northern
Sahel Savanna and Sub-Desert of West Africa. My examination of specimens from south of the Senegal River indicates that
T. arenarius does not occur in Senegal. North of the
Senegal River in southern Mauritania, I have found T. arenarius to be sympatric with T. pygargus at two localities.
Taterillus arenarius was found on the sandy substrates with scattered vegetation, while immediately adjacent to the area occupied by T. arenarius, T. pygargus was found. The habitat where T. pygargus occurred was composed of a hard clay-like substrate and tall grasses.
East of the range of T. pygargus, T. arenarius was found in the northern Sahel Savanna of Niger, while in the southern Sahel Savanna and further southward, I am only able to identify T. gracilis. East of the distribution of T. arenarius, T. iacustris is found.
Other rodents that occur within the range of T. arenarius are: Arvicanthis niloticus, Desmodilliscus braueri, Euxerus ery thropus , Ger'billus (G.) gerbillus , G.
(G.) pyramidum, G, (G,) nigeriae, G, (H.) mauritaniae, G. 35
(H.) amoenus, Jaculus jaculus, J. deserti, Tatera qambiana,
T. guinea, T. kempi, T. wellmani, and Taterillus pygargus.
Taterillus congicus Thomas
Taterillus congicus Thomas, Ann. Mag. Nat, Hist., Ser. 8, 16:147, August, 1915. Type locality, Poko, Upper Uelle, Zaire, Taterillus clivosus Thomas and Hinton, Proc, Zool. Soc. London, p. 258, 6 July, 1923. Type locality, Jebel Marra, south, Sudan. Taterillus emini congicus, Hatt, Amer. Mus. Novitates, No. 708, p. 2, April, 1934. Taterillus emini clivosus, Hatt, Amer. Mus. Novitates, No. 708, p. 3, April, 1934. Taterillus emini emini, Jeannin, Encyc. Biol. 16:155, 1936 (part). Taterillus nigeriae clivosus, Petter, African Identifi- cation Manual, part 6.3, Gerbillinae, 1975.
Holotype: Adult male, skin and skull, BM 19.5.8.81, collected by C. Christy on 6 August, 1914, original number
1188.
General Distribution of the Species: Cameroon,
Central African Republic, Chad, Sudan, and Uganda (Fig. 4).
Specimens Examined: Cameroon (17), Sir, 17 (RGMC,
1—SO). Central African Republic (37). Bangui, 2 (MNHP);
Gordil, 7 (MNHP); Ippy, 1 (MNHP); Koumbala, 21 (MNHP); Lac
Mamoun, 5 (MNHP); Maboke, 1 (MNHP). Chad (36). Aouk River,
1 (LA); Bekao, 1 (MNHP); Deli, 6 (MNHP); Fort-Archambault,
7 (MNHP); Fort-Lamy, 12 (MNHP); Moundou, 9 (MNHP). Sudan
(20). Delami, 1 (BM-S); Dibbis, 2 (FM); 40 mi W El Fasher,
1 (BM); Jebel Kadero, 3 (BM-2; FM-1); Jebel Marra, 1 (BM) ;
Jebel Marra, south, 2 (BM); Kulme, 2 (BM); Kurra, 4 (BM);
Terrokakka, 1 (BM); Tuksueina, 2 (FS); Yei, 1 (BM). Uganda CHAD
SUDAN
ETHIOPIA CENTRAL AFRICAN REPUBLIC
SOMALIA
ZAIRE KENYA UGANDA
RWANDA,
BURUNDI
TANZANIA
Fig. 4. Distribution and known localities of Taterillus congicus. u> 37
(1). Ngal, 1 (NA). Zaire (109). Bafuka, 2 (BM); Bagbele,
2 (RGMC, 1-S); Faradja, 32 (AMNH-29, 2-S; RGMC-3, 1-SO);
Gangala, 4 (BR-2; RGMC-2); Garamba, 1 (AMNH); Kabalo, 1
(RGMC-S); Mauda, 2 (RGMC); Mbanga, 3 (RGMC, 1-SO, 1-S);
Nagero, 2 (BR-SO); Niangara, 15 (AMNH-13, 5-SO; RGMC-2);
Poko, 39 (BM-7; MD-3, 1-SO; RGMC-29, 6-SO); I/O/I, 4 (RGMC);
II/gd/4, g (RGMC).
Description: Dorsum, interauricular, interorbital, and rostral areas varying from Prout's Brown to Mummy Brown; postauricular patches, mystacial and pectoral areas, inner sides of fore- and hind limbs, and entire underparts white; cheeks and sides Snuff Brown; dorsal hairs plumbeous basally, the brown colored portion only 3-5 mm in length, and a few hairs finely tipped with dark brown; pinna of ear long and almost naked, color almost the same as dorsum, and basally, the anterior portion with anterior portion has short buff-colored hairs; vibrussae long and composed of both white and dark brown hairs; tail moderately long and uniformly Tawney-Olive basally with the dorsal portion considerably darker due to the interspersed darker brown hairs, and grading to a Mummy Brown pencil.
Skull—large in size; braincase rounded; zygoma moderately strong; lachrymals large; molariforra teeth large; auditory bullae moderately large but constricted posteriorly and narrow, slightly expanded anteriorly, and not bulbous; nasals broad and long; rostrum long; posterior palatine 38 foramina long and wide; anterior palatine foramina long and wide; anterior palatine, foramina long and constricted anteriorly; parapterygoid fossae not markedly flared but deep; and pterygoid processes short, stout, and divergent.
Measurements; Measurements of the holotype (age class 4) followed by averages and extremes of 11 adults (age class 4) from the type locality, are, respectively: TOT 305,
292 (285-306); HB 135, 133 (125-138); TAL 170, 160 (150-
168); HF 33, 31 (29-33); E 20, 19 (18-20); W -, - ( -); GLS
36.3, 36.4 (35.4-37.5); OCNL 35.1, 35.4 (34.2-36.4); BAL
26.3, 26.7 (25.6-27.6); ZB -, 18.2 (17.1-18.7); CBAL 27.8,
28.4 (27.4-29.5); BBC 14.7, 15.2 (14.7-16.0); IOC 6.3, 6.5
(6.2-6.8); BR 5.0, 4.9 (4.6-5.4); GBAB 13.7, 13.9 (13.3-
14.5); BrMlMl 7.2, 7,4 (7.1-7.8); DAL 8.5, 8.9 (8.4-9.7);
PAL 15.2, 15.6 (14.8-16.4); PPAL 10.8, 10.9 (10,5-11.3);
LAPF 6.5, 6.6 (6.0-7.1); LPB 7.5, 7.7 (7,1-8.2); LAB 9.4,
9.7 (9.2-10,6); BAB 5,2, 5,5 (5,2-6,0); LN 15,1, 15.0
(14,6-15.5); LR 13,4,113,2 (12,6-14.0); FNL 26,9, 27,3
(26,6-28.2); DC 14.2, 14,3 (13.8-14.8); LPPF 4.2, 4.2 (3,7-
4,5); ALM1 2.7, 2.8 (2,6-2.9); ALM 5,2, 5,5 (5.2-5.8),
Comparisons; Taterillus congicus is characterized by its karyotype of 2N=54, FN=64-66, It can be distinguished from T. emini, the only other similar species as follows.
Taterillus congicus has slightly smaller molar teeth; a slightly larger and more robust skull; less inflated 39 auditory bullae; and longer and wider posterior palatine foramina.
Taterillus congicus is easily distinguished from other species in areas of sympatry. With T. harringtoni,
T. congicus differs by having a larger skull, longer upper incisors, larger molars, and less inflated bullae. For additional distinctions, see "Comparisons" under T. emini,
T. harringtoni, and T. lacustris, as well as Chapter 5.
Taxonomic Conclusions: The first specimens of
Taterillus congicus collected were those reported by
Pousargues (1896). His specimens from Poste de la Mission,
River Kemo, Central African Republic, were identified as
!£• emini • Although I have not seen these specimens, recently collected specimens from southwestern Central
African Republic discussed by Genest and Petter (1973) have a karyotype of 2N=54 and have been identified as T. congicus.
The taxonomic status of Taterillus congicus was not investigated nor questioned until Hatt (1934) described a new subspecies of T. emini (T. e. anthonyi). In this he included comparisons with other Taterillus species from
Nigeria (T, lacustris), the Sudan (T. butleri and T. clivosus), and Uganda (T. emini). Hatt (1934) designated all of the above taxa as subspecies of T. emini. His subsequent analysis (Hatt, 1940) of the rodents of the
Congo (=Zaire) included newly collected specimens of 40
Taterillus. He designated all of these as T. emini congicus.
Although Allen (1939) and Ellerman (1941) had recognized T. congicus as a valid species, Schouteden
(1943, 1944, 1948) designated the Zaire specimens he collected as T. emini congicus. His (Schouteden, 1944) identification of a skin only specimen from Kabalo, Zaire
(06°061S.,26°321E.) as a Taterillus is doubtful. It is south of the known distribution of Taterillus, in the high forest, and is probably a Tatera.
Research on the mammals of Cameroon by Jeannin
(1936) and followed by Good (1947) included Taterillus.
Both authors report the species from northern Cameroon as
T. e. emini. It is possible that T. emini occurs in
Cameroon, but to date I am only able to identify T, congicus and T. lacustris in Cameroon.
In a discussion of the fauna of French Central
Africa, Malbrant (1936) identified specimens from near
Fort-Lamy, Chad, as T, emini. Specimens from Fort-Lamy have a reported karyotype of 2N=54, FN=64-66 (Tranier et al., 1974); that of T, congicus.
Setzer (1956) in his revision of the mammals of the
Sudan, recognized T, congicus as a valid species and identified specimens from some localities as being both
T, congicus and T. emini. Taterillus clivosus, after its original description by Thomas and Hinton (1923) was not 41 compared to other species of Taterillus until Hatt (1934).
He designated it as a subspecies of T. emini. That alloca tion was affirmed by Setzer (1956). Petter (1975) reassigned this taxon (T. clivosus) to T. nigeriae (T. n. clivosus). My comparisons of the holotype of T. clivosus to other Taterillus taxa shows that it belongs as a part of
T. congicus.
Cytogenetic studies have revealed different karyotypes for Taterillus in East Africa. A 2N=54, FN=44-
46 form from Cameroon and Chad (Tranier et al., 1974) and
Central African Republic (Matthey and Petter, 197 0; Genest and Petter, 1973) were allocated to T. congicus. Petter
(1975) also designated T. congicus as having a diploid number of 54. My comparisons of the holotype and topotypi- cal specimens of T. congicus to the specimens with a karyotype of 2N=54 confirms the identifications of them being T. congicus.
Ecological Observations: Taterillus congicus occurs in the Guinea Woodland near its border with the high forest in Zaire and Central African Republic (Hatt,
1940; Genest and Petter, 1973), It has also been found in the arid Sudan Woodland and Sahel Savanna of Cameroon,
Central African Republic, Chad, and the Sudan, It is sympatric with T. harringtoni at Gordil, Central African
Republic, although reportedly occupying the drier habitats
(Genest and Petter, 1973). It is also sympatric with T. 42
harrinqtoni in the Nuba Mountains, the Jebel Marra Range, and the Equatoria Provinc e of the Sudan. Taterillus c congicus is sympatric with T. emini in several areas in the
Sudan and in northeastern Zaire. Any ecological or micro-
habitat differences between T. congicus and either T.
harringtoni or T. emini are unknown at present.
Other rodents which occur within the range of T.
congicus are: Acomys lowed, A. wilsoni, A. percivali,
Aethornys kaiseri, Arvicanthis niloticus, Aolomys goslingi,
Cricetomys gambianus, Euxerus erythropus, Gerbillus (G.)
gerbillus, G. (G.) nigeriae, Dipodillus lowei, Grammomys
dolichurus, Heliosciurus gambianus, Lemniscomys barbarus,
L. macculus, L. striatus, Malacomys longipes, Lophuromys
sikapusi, Mastomys natalensis, Mylomys dybowskyi, Mus
bellus, M. musculoides, Oenomys hypoxanthus, Praomys
jacksoni, Steatomys opinus, Tatera dichrura, T. benvenuta,
T. robusta, Taterillus emini, T. harringtoni, and Thryonomys
swinderianus.
Taterillus emini (Thomas)
Gerbillus emini Thomas, Ann. Mag, Nat. Hist,, Ser, 6, 9:78, January, 1892. Type-locality, Wadelai, Uganda. Taterillus emini, Thomas, Ann. Mag. Nat. Hist., Ser. 8, 6:222, August, 1910. Tatera emini, Wroughton, Ann. Mag. Nat. Hist., Ser. 1, 17:477, May, 1906. Taterillus butleri Wroughton, Ann. Mag. Nat. Hist., Ser. 8, 6:294, September, 1910. Type-locality, Dug-dug, Sudan. 43
Taterillus gyas Thomas, Ann. Mag. Nat. Hist., Ser. 9, 2:150, August, 1918. Type-locality, Kamisa, Dinder River, Sudan. Taterillus emini anthonyi Hatt, Amer. Mus. Novitates, No. 708, p. 2, April, 1934. Type-locality, 20 mi S Jebelein, Sudan.
I-Iolotype: Adult (sex unknown), skin and skull,
BM 87.12.1.50, collected by Emin Pasha on 24 May, 1886.
General Distribution of the Species. Northwestern
Kenya, Sudan, Uganda, and northeastern Zaire (Fig. 5).
Published Records: Nabilatuk, Uganda (Delany,
1964).
Specimens Examined: Kenya (3). Lokichoggio, 3
(NA). Sudan (166). Blue Nile Province: Akona, 1 (AMNH);
20 mi S Jebelein, 1 (AMNH); Kamisa, 8 (BM); Darfur Province: near El Fasher, 3 (BM, 1-SO); 10 mi W El Fasher, 1 (BM);
40 mi W El Fasher, 1 (BM); 35 mi WSW El Fasher, 2 (BM); 35 mi N El Fasher, 1 (BM); 40 mi WSW El Fasher, 2 (BM); 60 mi
WSW El Fasher, 2 (BM); Kurra, 3 (BM, 1-SO); NE Jebel Marra,
1 (BM); Jebel Marra, 7 (BM-SO); foothills, south Jebel
Marra, 1 (BM); Jebel Maidob, 2 (BM); Kulme, 3 (BM-2; MSNG-
1); Kordofan Province: Delami, 3 (BM); Nuba Mountains, 2
(BM); Augur, 27 (BM-6; DR-21,S); 60 mi N El Obeid, 1 (BM) ;
20 mi W El Obeid, 1 (BM); 50 mi W El Obeid, 6 (FM-1; BM-5,
SO) ; 5_5 mi_ E Hamza, 1 (BM) ; Um Dona, 2 (BM) ; Angolo, 1 (FS) ;
Bahr A1 Ghazal Province: Dugdug, 4 (BM); Wau, 2 (BM); Katta,
1 (BM); Rafille, 3 (BM); Equatoria Province: Mongalla, 2
(BM); Juba, 1 (BM); Obbo, 4 (FM-2, 1-S, 1-SO); Opari, —S]5°E
CHAD
SUDAN
ETHIOPIA CENTRAL AFRICAN REPUBLIC
SOMALIA
CONGO ZAIRE KENYA UGANDA
J RWANDA
I 1 1 1 BURUNDI Miles 0 500 / TANZANIA 'III J 1 Kilometersi Fig. 5. Distribution and known localities of Taterillus emini, 45 2 (MD); Kagelu, 1 (AMNH); Torit, 40 (FM-4, 1-SO; MCZ-8,
1-SO; MD-23; NA-2); Ikoto, 3 (MD-2); Imurok, 1 (MD) ;
Logoforok, 2 (MD); Moli, 1 (MCZ); Upper Nile Province:
Malek, 15 (BM); Upper Nile, 1 (BM). Uganda (95). Karamoja
District: Kolido, 1 (BM); Kamchuru, 9 (BM); south of Keem,
2 (AMNH); Moroto, 1 (NA); Lobome, 7 (NA); Amudat, 3
(BM—1, SO; 1-S, 1-SO); Acholi District: Falabek, 1 (BM);
Pabbo, 1 (BM); Pabboi, 1 (BM, SO); Awaich, 3 (BM); south of
Lacho, 37 (AMNH); West Nile District: Rhino Camp, 11 (AMNH-
7, 1-SO; BM-4, 1-SO); Wadelai, 2 (BM); Ngal, 6 (AMNH-2;
BM-3, 1-SO; NA-1); Teso District: Aido, 1 (BM); Ajeluk, 7
(AMNH-2; BM-5, 1-SO); Serere, 2 (AMNH-1; BM-1). Zaire (9).
Gangala, 2 (RGMC); Garamba, 1 (AMNH); Mbanga, 1 (RGMC);
I/O/I, 4 (RGMC); II/gd/4, 1 (RGMC).
Description: Dorsum, interauricular, interorbital, and rostral areas varying from Prout's Brown to Cinnamon-
Brown; posauricular patches, mystacial, and pectoral areas, fore- and hind limbs, and entire undersides white; cheeks and sides Clay Color; dorsal hairs plumbeous basally, the brown pigmented portion only 2-3 mm in length, with some hairs finely tipped with dark brown; pinna of ear long and almost naked, color almost the same as dorsum, and anterior and basal margins with short light brown hairs; vibrussae long and composed of both white and dark brown hairs; tail medium in length and uniformly Tawny-Olive ventrally with 46 dorsal hairs interspersed with dark brown hairs, grading to a terminal Mummy Brown pencil.
Skull—large in size; braincase slightly rounded; zygoma missing in the holotype but are weak in other specimens; lachrymals also missing but probably narrow and long; molariform teeth largest of any species; auditory bullae moderately large but constricted posteriorly and narrow, slightly expanded anteriorly, and moderately bulbous; nasals also missing but probably medium in width and long; rostrum long; posterior palatine foramina long but narrow; anterior palatine foramina slit-like and medium in length; parapterygoid fossae not markedly flared.
Measurements: Measurements of the holotype (age class 3) followed by averages and extremes of 11 adults (age class 4) from Rhino Camp, Uganda, are, respectively:
TOT 292, - ( - ); HB 130, - ( - ); TAL 162, - ( - ); HF 33,
- ( - ); E 18, - ( - ); W - ( - ); GLS -, 35.9 (34.7-
36.6); OCNL 34.8 (33.4-35.5); BAL -, 26.4 (24.5-27.3);
ZB 18.4 (17.9-18,8); CBAL 28.7, 28,1 (26.7-28.9); BBC -,
15,0 (14.6-15.6); IOC 5.8, 6.3 (5.9-6.7); BR 4.6, 4.7 (4.2-
5.1); GBAB 12.8, 13.9 (13.5-14.4); BrMlMl 7.4, 7.5 (7.1-
7.9); DAL 8.9, 8.5 (7.7-9.2); PAL 15.5, 15.2 (14.0-16.1);
PPAL -, 10.9 (10,4-11.4); LAPF 6.4, 6.3 (5.4-7.0); LPB 7.4,
7.1 (5.7-7.9); LAB 9.6, 9.7 (9.1-10.2); BAB 6.0, 5.7 (5.3-
6.1); LN -, 14.5 (13.8-14,8); LR -, 13,0 (12.4-13.8); FNL -, 47
27.0 (26.2-28.2); DC 13.7, 14.3 (13.8-15.0); LPPF 4.2, 3.6
(3.3-3.9); ALM1 2.7, 3.0 (2.7-3.2); ALM 5.4, 5.4 (5.2-6.0).
Comparisons: The karyotype of this species is not known. Taterillus emini can be distinguished from T. congicus by a larger first upper molar and longer molar tooth row; a slightly smaller and more gracile skull; slightly more antero-lateral inflation of the auditory bullae; a shorter and narrower posterior palatine foramina; flared but shallow parapterygoid fossae; and thin pterygoid
processes,
Taterillus emini is easily distinguished from other
East African species of Taterillus. From T. harringtoni,
T. emini has a larger skull; longer upper incisors; larger molars; and less inflated bullae. From T. lacustris, T.
emini is distinguished by its much larger skull and slightly
larger and more inflated auditory bullae.
Taxonomic Conclusions: The original description of
Taterillus emini (Thomas, 1892), along with T. gracilis,
was placed in the genus Gerbillus. Subsequently, Wroughton
(1906) placed both emini and gracilis in the genus Tatera.
When Thomas (1910) named the genus Taterillus as distinct
from both Gerbillus and Tatera, he chose T, emini to be
the genotype of his newly described genus.
Specimens of Taterillus collected at Dugdug, Sudan,
were identified by Wroughton (1907) as T, emini. Wroughton
(1910) then used these specimens as the basis for his 48 naming a new species, T. butleri. Taterillus butleri is shown by my analysis to be conspecific with T. emini. My allocation is in agreement with previous authors (Hatt,
1934; Setzer, 1956; Petter, 1975).
In a subsequent paper by Wroughton (1911), he identified specimens from Mongala, Sudan, as T. emini, but allocated specimens from Rafille and Wau, Sudan, to T. butleri. My examination of those specimens shows that all of them are T* emini with the exception of those from Wau.
They comprise both T. emini and T. harringtoni.
The taxon T. emini zammarani (De Beaux, 1922) is shown by this report to be T. harringtoni. In his revision of the Taterillus from the Sudan by Setzer (1956), all taxa except for T. congicus were included under T. emini. As discussed under other species in this report, I consider
Setzer's (1956) T. kadugliensis (equals T. e. butleri) , T. emini perluteus, and T. e. rufus (including T, lorenzi) to
be conspecific with T, harringtoni. I also consider T. clivosus to be conspecific with T. congicus rather than
T. emini, while his (Setzer, 1956) allocations of T. emini
anthonyi and T. gyas to T. emini agree with my analysis.
Specimens from Torit, Sudan, and nearby localities, were
identified by Setzer (1956) as T. e. emini. Although some
of these specimens are nearly identical morphologically to
typical T, emini, my comparisons have shown that others
are clearly T, harringtoni. 49
Petter (.1975) separated T. gyas from the emini complex and recognized it as a full species. The holotype of T. gyas is the largest known specimen of Taterillus that
I have examined. Comparisons with both T. emini and T. congicus show T. gyas to be a part of the emini complex.
Its large size is insufficient to warrant either specific or subspecific status at this time. Petter (1975) designated both T, osgoodi and T. emini zammarani as being T. emini.
Morphological comparisons including comparative karyology has resulted in my placing these two taxa with T. harringtoni. For further discussion of these taxa see the following section on T. harringtoni.
The specimens of Taterillus having a karyotype of
2N=44, FN-62, previously reported by Matthey (1969), Genest and Petter (1973), and Robbins (1974a) to be T. emini, are shown by this analysis to be T. harringtoni and distinct from T, emini, See "Taxonomic Conclusions" under T. harringtoni for further discussion. The karyotype of T. emini is unknown at present.
Ecological Observations: Taterillus emini occurs in the moist Guinea Woodland in central and eastern Uganda, as well as in the arid Sudan Woodland and Sahel Savanna of northwestern Kenya and the Sudan. It is sympatric with T. harringtoni in northern Uganda, and the Sudan, and with
T, congicus in several localities in the Sudan and northeastern Zaire. Any ecological differences or habitat 50 segregation between T. emini and other species of
Taterillus, where sympatric, are unknown.
Other rodents which occur within the range of T. emini are: Acornys hystrella, A. percivali, A. wilsoni,
Aethomys kaiseri, Arvicanthis niloticus, Cricetomys gambianus, Dasymys incomtus, Dendromus melanotis, D, mesomelas, Euxerus erythropus, Gerbillus (G.) gerbillus,
Grammomys dolichurus, Heliosciurus gambianus, Lemniscomys
barbarus, L. macculus, L. striatus, Mastomys natalensis,
Mylomys dybowskyi, Myomyscus fumatus, Mus bellus, M. musculoides, M. tenellus, Saccostomus campestris, Tatera
benvenuta, T. macropus, T. robusta, Taterillus congicus, and T. harringtoni.
Taterillus gracilis (Thomas)
Gerbillus gracilis Thomas, Ann. Mag. Nat. Hist., Ser. 6, 9:77, January, 1892. Type-locality, Gambia. Taterillus gracilis, Thomas, Ann. Mag. Nat. Hist., Ser. 8, 6:222, August, 1910. Tatera gracilis, Wroughton, Ann. Mag. Nat. Hist., Ser. 77 17:477, May, 1906. Taterillus gracilis, Rosevear, The Rodents of West Africa, 1969 (part). Taterillus gracilis, Matthey, Mammalia, 33:522-528, 1969 (part).
General Distribution of the Species: Senegal, Mali,
Ivory Coast, Upper Volta, Ghana, Niger, Togo, Dahomey,
Nigeria, and possibly Cameroon, Portuguese Guinea, and
Guinea (Fig. 6). I 5fw
T. G. MERIDIONALIS
MAURITANIA T. G. NIGERIAE
T. G. GRACILIS
T. G. ANGELUS
MALI •• / /V. rwr. NIGER
A I / / S S S // /// // // // // // // // // s / / / jUINLOAL \sssss''\'//////////////S// /(V '/''///////// -O>'//// / sAs//////V///////s/// /\///ss/y///////// © © • j '/ // ^/\/ / / / / A/ / / / // PORT u w UPPER VOLTA-FE GUINEA~;;;;~///////////// o //////////////s// e>/ra GUINEA^'"
SIERRA M Q/ 1 GAMBIA LEONE
IVORY GHANA COAST NIGERIA CAMEROON
K11 ometer s
Fia. 6. Distribution and known localities of Taterillus gracilis including subspecies. 52
Distribution of Subspecies;
Taterillus gracilis angelus—the northern Sudan
Woodland and Sahel Savanna in Niger and Nigeria, and
possibly Chad and Cameroon.
Taterillus gracilis gracilis--Gambia, western Mali,
Senegal, and probably Portuguese Guinea and Guinea,
Taterillus gracilis meridionalis--the Guinea and
Sudan Woodlands and Sahel Savanna of Dahomey, Ghana, Ivory
Coast, southwestern Mali, southwestern Niger, western
Nigeria, Togo, and Upper Volta.
Taterillus gracilis nigeriae--the southern Sudan
Woodland and the Guinea Woodland east of the Niger River in
Nigeria and possibly Cameroon.
Published Records: Sanaferedougou, Ivory Coast
(Bellier and Gautun, 1967; T. g. meridionalis); Foro-Foro,
Ivory Coast (Gautun and Petter, 1973; T. g. meridionalis).
Comparisons; This species differs from T. arenarius
by its more slender and gracile skull, although they are
similar in length. The bullae of T. gracilis are less
inflated anteriorly, antero-laterally, and ventrally. The
pelage color of T. gracilis is usually darker except for
specimens of T. £. angelus and T. g. meridionalis occurring
in the Sahel Savanna.
Taterillus gracilis is distinguished from T.
pygargus by having a longer, more gracile skull and a longer
rostrum in relation to the length of the skull. Its 53 auditory bullae are more constricted anteriorly and less inflated, especially in areas of sympatry. For additional distinctions see "Comparisons" under T. pygargus, T. arenarius, and T. lacustris.
Although bullar shape separates T. gracilis from T. pygargus in areas of sympatry, the bullae of T. gracilis change shape in the remaining part of its range. Their bullae become more inflated, approaching that of T. pygargus.
This added inflation distinguishes T. g. angelus and T. g. nigeriae from T. lacustris.
Comparisons of the specimens with a karyotype of
2N=36/37 from Senegal to the holotype of T. g. gracilis, show that they are almost identical. The characters which best distinguish T. gracilis and T. pygargus in Senegal hold when comparing specimens of known karyotype to the holotypes.
Taxonomic Conclusions: Taterillus gracilis, along with T, emini, was one of the first species allocated to the genus Taterillus (Thomas, 1910). Prior to the discovery of
T, pygargus (F. Cuvier, 1838) as outlined by Petter (1952), all specimens of Taterillus from West Africa, with the exception of those from central Nigeria (T. nigeriae), had been assigned to T. gracilis.
Ellerman (1941), although retaining T. gracilis,
T. nigeriae, and T, lacustris, as valid species in West
Africa, thought that they were possibly all part of a single 54 polymorphic species. The status of many of the West
African specimens of Taterillus was unclear even after
Rosevear (1969) revised the taxa occurring in that area.
Rosevear (1969) was apparently not aware of the existence or availability of T. pygargus in Senegal in his revision of Taterillus. He synonymized T. lacustris with T. gracilis and retained T. nigeriae as a full species, although he expressed doubt as to its species status. I consider nigeriae to be a part of T. gracilis.
Matthey (1969) originally designated the 2N=23 karyotypic form from Senegal as T. gracilis. A subsequent discovery of a 2N=;36 form in Upper Volta (Matthey and Petter,
1970) did not change the species allocation. Taterillus gracilis was attributed with a karyotype of both 2N=23 and
2N-3 6, This remained an unsolved problem until the discovery by Matthey and Jotterand (1972) of both chromosomal forms occurring sympatrically in Senegal.
Petter et al. (1972) considered the form with 23 chromosomes as T. pygargus and the one with 3 6 chromosomes as T. gracilis.
Those allocations remained tentative until Robbins
(1974b) and this report because no morphological features other than karyotype or serum protein differences had been found that could separate the two species. Therefore, comparisons of the holotypes of these two species (T. gracilis and T. pygargus) to the karyotyped specimens had 55 not been discussed. As was mentioned previously under
"Comparisons" for T. gracilis, and following under
"Comparisons" for T. pygargus, comparisons of the holotypes to the specimens karyotypically allocated to the above two species, shows that the 2N=36 specimens are T. gracilis and the 2N=23 specimens are T. pygargus.
I have shown in the section concerning T. arenarius, that the chromosome number of 2N=3 0, previously associated with T. nigeriae (Matthey, 1969; Petter, 1970,
1975) was in error. I consider Taterillus nigeriae to be a subspecies of T. gracilis (T. £. nigeriae).
Taterillus gracilis angelus Thomas and Hinton
Taterillus gracilis angelus Thomas and Hinton, Novitates Zoologicae, 27:317, 15 June, 1920. Type-locality, Farniso (=Panisau), near Kano, Nigeria.
Holotype: Adult male, skin and skull, BM 21.2.11.36, collected by A. Buchanan on 24 December, 1919, original number 51.
Specimens Examined; Niger (5), 45 km NW Niamey,
2; 25 km S Tahoua, 2; Garari, 1 (MNHP). Nigeria (105).
Northern Region: Farniso, 10 (BM); Panisau, 31; Karaduwa,
52; Tangaza, 6; 12 mi N Sokoto, 1; Kano, 5 (BM),
Measurements•. Measurements of the holotype (age class 3) and averages and extremes of 8 adults including
paratypes and topotypes(age class 4), are, respectively: 56
TOT 284 (273-292); HB 113, 126 (118-131); TAL 148, 154
(145-161); HF 29, 33 (32-34); E 19, 21 (20-22); W -, 53
(49-60); GLS 33.2, 35.5 (34.8-36.2); OCNL 32.0, 34.2 (33.7-
34.9); BAL 24.1, 25.7 (25.1-26.6); ZB -, 17.5 (17.4-17.6);
CBAL 26.1 27.5 (27.0-28.4); BBC 14.1, 14.4 (13.9-14.9);
IOC 5.9, 6.0 (5.6-6.5); BR 4.2, 4.4 (4.1-4.9); GBAB 12.9,
13.6 (12.9-14.4); BrMlMl 6.7, 7.1 (6.9-7.4); DAL 7.8, 8.4
(8.1-8.6); PAL 13.9, 14.5 (14.0-15.0); PPAL 9.8, 10.8
(10.2-11.2); LAPF 5.8, 5.8 (5.5-6.0); LPB 6.5, 7.2 (6.9-
7.8); LAB 9.2, 9.5 (9.0-9.8); BAB 5.4, 5.5 (5.1-5.9); LN
12.9, 14.7 (14.0-15.2); LR 11.8, 12.9 (12.4-13.5); FNL 25.0,
26.7 (26.1-27.4); DC 13.3, 13.8 (13.1-14.2); LPPF 2.9, 3.7
(3.3-4.0); ALM1 2.4, 2.8 (2.6-3.0); ALM 5.0, 5.3 (5.1-5.8).
Description: Interauricular, interorbital, and rostral areas same color as dorsum and varying from
Saccardo1s Umber to Tawny-Olive; circumorbital region, postauricular patches, mystacial, and pectoral areas, fore- and hind limbs, and entire underparts white; cheeks and sides Tawny-Olive; dorsal hairs plumbeous basally, the brown pigmented portion only 2-3 mm in length and some hairs finely tipped with dark brown; pinna of ear long and almost naked, color almost the same as the dorsum, and anterior margin with short buff- and white-colored hairs; vibrussae long and composed of both white and dark brown hairs; tail moderately long and uniformly Pinkish Buff basally with the 57 dorsal surface interspersed with darker brown hairs, grading to a terminal Fuscous pencil.
Skull--relatively large for the species and gracile in appearance; zygoma weak; lachrymals medium; nasals and rostrum long and narrow; molariform teeth medium-sized; auditory bullae large and inflated but not bulbous; para- pterygoid fossae deep but not flared; braincase rounded.
Comparisons t Taterillus gracilis angelus can be distinguished from T. g. nigeriae by its longer nasals, narrower rostrum, narrower braincase, and longer and more inflated auditory bullae. It is also distinguished by its lighter pelage color except for some localities in the western part of its range.
Taterillus gracilis angelus differs from T. g. meridionalis by its longer bullae in relation to breadth, longer nasals and rostrum, larger skull, and a narrow rostrum which reflects its much narrower nasals as compared to the short but broad nasals in T. g. meridionalis.
Although T. g, angelus is separated both by geography and by physiography from T. g. gracilis, they differ morphologically in that T. g. angelus has a slightly larger and more robust skull; greater breadth across the bullae as reflected by its bullae being longer and more inflated; slightly larger molars; and longer nasals.
Taxonomic Conclusions: The taxonomic status of T. g. angelus as a subspecies of T. gracilis (Thomas and 58
Hinton, 1920) remained until 1953. Although Rosevear
(1953) apparently recognized T. g. angelus as distinct from
T. g. gracilis, a contradiction appeared. His distribution maps show that these two subspecies occupy the same geographical area in Nigeria. According to Mayr (1969), it is misleading in such a case to say that two subspecies overlap in a given area. Taterillus gracilis is repre sented in this area of Nigeria only by a single population, no matter how variable. Rosevear (1953) did not discuss differences or similarities between the two subspecies. In his revision of the rodents of West Africa, Rosevear
(1969) did not distinguish subspecies of T. gracilis.
Rather, he stated that fine distinctions based on color would proliferate names. Rosevear (1969) also considered
T. lacustris to be conspecific with the Nigerian T. gracilis.
Comparisons using the specimens available to
Rosevear and additional specimens from the type-locality of T. g. angelus as well as Cameroon specimens of T. lacustris, have shown me that T. g. angelus is a distinct subspecies of T. gracilis and that T. lacustris should be retained as a full species. I consider specimens from northeastern Nigeria, previously allocated to T. gracilis by Rosevear (1969), to be T. lacustris.
Ecological Observations: Taterillus gracilis angelus occurs in Nigeria and Niger in the northern Sudan
Woodland and Sahel Savanna, I have not found it to be 59 sympatric with T. arenarius to the north in Niger nor with
T. lacustris in eastern Nigeria and Cameroon. Its western districution is apparently influenced by the Niger River.
I consider the Niger River to be the boundary between T. g. angelus and T. c£. meridionalis as indicated by morphological distinctiveness observed when comparing specimens from either side of the river.
Ecological differences separate this subspecies from
T. cj. nigeriae. Taterillus 2.. angelus occupies the sandier habitats in the Sudan Woodland while T. g. nigeriae prefers the harder substrates in the Sudan Woodland. This is demonstrated by comparing the mammalian faunas of several
Nigerian localities in the Sudan Woodland. Taterillus g. angelus is found at Karaduwa, although its pelage color is darker than specimens from the type-locality. It occurs there with Gorbillus (G.) nigeriae and Desmodilliscus braueri. These two gerbilline taxa represent a northern desert fauna as evidenced by their occurring mainly in the
Sahel Savanna and Subdesert, At Dada and Iella, Nigeria,
T. g. nigeriae occurs, Neither Gerbillus nor Desmodilliscus occurs at these two localities, but rather rodents which are indicative of more southern faunas in the southern Sudan and Guinea Woodlands,
Other rodents that occur within the range of T. g. angelus are: Acomys cahirinus, Arvicanthis niloticus,
Cricetomys gambianus, Desmodilliscus braueri, Euxerus 60 erythropus, Graphiurus murinus, Gerbillus (G.) nigeriae,
Lemniscomys barbarus, Mastomys natalensis, Mus hausa, Myomys daltoni, Steatomys cuppedius, and Tatera kempi.
Taterillus gracilis gracilis (Thomas)
Gerbillus gracilis Thomas, Ann. Mag. Nat. Hist., Ser. 6, 9:77, January, 1892. Type-locality, Gambia (no specific locality given). Taterillus gracilis, Thomas, Ann. Mag. Nat. Hist., Ser, 8, 6:222, August, 1910. Taterillus gracilis gracilis, Rosevear, Checklist and Atlas of Nigeria Mammals, 1953 (part). Tatera gracilis, Wroughton, Ann. Mag. Nat. Hist., Ser. 7, 17:477, May, 1906.
Holotype: Subadult male in alcohol with skull removed, BM 85.2.2.1, collected by C. A. Maloney.
Specimens Examined: Gambia (9). MacCarthy Island
Division: Kudang, 1; 1 mi E Kuntaur, 2 (BM); Zemaiyatta, 2
(BM); Upper River Region: Sun Kunda, 2 (BM); Tambasensan, 1
(BM); Gambia, no specific locality, 1 (BM). Mali (6).
Bamako, 4 (MNIIP) ; Mopti, 2 (MNHP). Senegal (198). River
Region: Richard Toll, 6; Cascas, 2; Fete-Ole, 13 (MNHP);
Galoya, 1 (MNHP); 10 km SE St, Louis, 8; Ogo, 1; Ranerou, 11;
Diourbel Region: 8 km E Louga, 24; Linguere, 19 (MNHP-9);
Ndoulo, 36; Thies Region: 10 km W Thies, 2; Bandia, 2
(MNHP); Oriental Region: Cuabo, 1 (BM); 5 km S Bakel, 19;
Goudiry, 3; Kotiari Naounde, 6; Koussanar, 13; Gemenjullah, 1
(BM); Sine-Saloum Region: 15 km N Kaffrine, 10; 6 km E
Kaolach, 2; Koungheul, 3; Tubakuta, 2 (BM); Saboya, 4 61
(MNHP); Cap Vert Region: Dakar, 7; Casaraance Region: 5 km
N Kolda, 2.
Measurements: Measurements of the holotype (age
class 3) followed by averages and extremes of 10 adults
(age class r) from Ndoulo, Senegal, are, respectively:
TOT -, 270 (259-277); HB -, 122 (113-129); Tal -, 148 (142-
155); HF -, 31 (30-32); E -, 20 (20-21); W -, 56 (46-63);
GLS 31.6, 35.6 (35.0-36.4); OCNL 30.9, 34.3 (33.4-35.2);
BAL 22.6, 25.4 (24.3-26.2); ZB -, 17.5 (16.8-18.2); CBAL
24.5, 26.9 (25.9-27.4); BBC 13.7, 14.1 (13.6-14.6); IOC 5.9,
6.0 (5.6-6.6); BR 4.3, 4.5 (4.3-4.7); GBAB 12.2, 13.4 (13.0-
14.0); BrMlMl 6.8, 7.1 (6.8-7.3); DAL 6.9, 8.3 (7.9-8.6);
PAL 12.9, 14.5 (14.2-14.7); PPAL 9.4, 10.5 (10.0-11.0); LAPF
5.0, 5.8 (5.3-6.2); LPB 6.6, 7.0 (6.8-7.4); LAB 8.6, 9.4
(9.1-9.7); BAB 5.2, 5.3 (5.1-5.5); LN 12.5, 14.5 (13.6-
15.0); LR 10.5, 13.2 (12.5-13.6); FNL 22.2, 26.7 (26.0-27.3);
DC 12.3, 13.8 (13.4-14.2); LPPF 3.3, 3.8 (3.5-4.3); ALMl
2.3, 2.8 (2,7-2.9); ALM 5.2, 5.3 (5.1-5.5).
Description: Dorsum, interauricular, interorbital,
and rostral areas varying from Olive Brown to Buffy Brown;
postauricular patches, mystacial, and pectoral areas,
fore- and hind limbs, and entire underparts white; cheeks
and sides Tawny-Olive; dorsal hairs plumbeous basally, the
brown pigmented portion being only 3-4 mm in length, and
a few hairs finely tipped with dark brown; pinna of ear
long and almost naked, color almost the same as the dorsum, 62
and basally the anterior portion with short buff-colored
hairs; vibrussae long and composed of both white and dark
brown hairs; tail long and uniformly Light Ochraceous-Buff
basally with the dorsal portion considerably darker due to
the presence of scattered dark brown hairs; terminal
pencil Mummy Brown.
Skull--Medium-sized for the species and gracile in
appearance; zygoma relatively weak; lachrymals large;
nasals and rostrum long and narrow; molariform teeth
medium-sized; auditory bullae medium and not markedly
inflated nor bulbous; parapterygoid fossae medium in depth
and not flared; braincase slightly rounded.
Comparisons: Taterillus gracilis gracilis can be
distinguished from the other three subspecies of T. gracilis
by the shape of its auditory bullae. When compared to T. g.
meridionalis, the bullae of T. g. gracilis are not as
inflated anteriorly and anterolaterally, nor are they as
' bulbous, This is reflected by the measurement of greatest
breadth across the bullae being smaller. Also, the nasals ! and rostrum of T. cj. gracilis are longer and narrower.
Taterillus gracilis gracilis differs from both T. f g. angelus and T. g, nigeriae by having a smaller, more
gracile skull, and less inflated auditory bullae.
Taxonomic Conclusions: Taterillus gracilis gracilis
was the first species described in the genus prior to the
discovery of T, pygargus (Petter, 1952), Although T. 63 gracilis has page preference to T. emini, Thomas (1910) chose the latter as the genotype for the genus Taterillus.
As previously mentioned in "Taxonomic Conclusions" under T. gracilis, this species now comprising the subspecies gracilis, was described as being a Gerbillus (Thomas, 1892).
Subsequently it was placed in the genus Tatera by Wroughton
(1906).
Taterillus gracilis was monotypic until the description of T. c[. angelus (Thomas and Hinton, 1920). The two subspecies (T. g. angelus and T. g. gracilis) were distinguished on the basis of pelage color. No further discussion concerning geographic differences in T. gracilis occurred until Rosevear (1953). As previously stated in
"Taxonomic Conclusions" under T. g. angelus, Rosevear
(1953) showed that the distribution of these two subspecies overlapped in northern Nigeria. His (Rosevear, 1969) subsequent discussion of the genus Taterillus did not include any subspecific designations.
Senegal specimens of T. gracilis were originally thought to have a karyotype of 2N=23, FN=40 (Matthey, 1969).
A later discovery of a 2N=3 6 form in Upper Volta (Matthey and Petter, 1970) gave T. gracilis two different karyotypes.
The discovery of the 2N=23 and 2N=36 forms sympatric in
Senegal (Matthey and Jotterand, 197 2) resulted in the designation of the 2N=23 forms as being T. pygargus (Petter 64 et al., 1972), while T. gracilis was assigned the 2N=36 karyotype.
Ecological Observations: This subspecies occupies the three main vegetation zones (Guinea and Sudan Woodland and Sahel Savanna) in which the genus Taterillus occurs.
Although this obviates pelage color differences because of the variation in substrate color between these zones, when
T. g. gracilis is compared to the other subspecies of T. gracilis, cranial differences are evident. I believe that this is due to the geographic isolation of T. g. gracilis from the other subspecies. Taterillus gracilis gracilis is found only in extreme western Africa. Morphological differences between T. £. gracilis and T. g. meridionalis
(the subspecies that is found east of T. g. gracilis) are found when comparing specimens of T. gracilis on both sides of the Niger River in Mali and Upper Volta,
Taterillus gracilis gracilis has not been found in the more arid, sandy areas of Senegal and Mali. These sandy habitats are occupied by either T. pygargus or T. arenarius. This is in contrast to T. g. meridionalis in
Upper Volta and southern Niger, which is found in the arid, sandy regions of the Sahel Savanna. For further discussion of T. g, meridionalis see "Ecological Observations" under that subspecies. In areas of sympatry between T. g. gracilis and T. pygargus, T. g. gracilis apparently prefers 65 the harder, clay-like substrates, whilt T. pygargus is found on the sandy substrates (Hubert, 1973).
Other rodents that occur within the range of T. g. gracilis in Senegal are: Arvicanthis niloticus, Cricetomys gambianus, Desmodilliscus braueri, Euxerus erythropus,
Gerbillus (G.) pyramidum, Lemniscomys barbarus, Mastomys natalensis, Mus minutoides, Mus sp., Steatomys sp., Tatera gambiana, T. sp., Taterillus pygargus, and Uranomys ruddi.
Taterillus gracilis meridionalis, New Subspecies
Taterillus gracilis gracilis, Rosevear, Checklist and Atlas of Nigerian Mammals, 1953 (part)
Holotype; Adult female, skin and skull, USNM
435359, collected by J. C. Geest, original number 7288, on 17 May, 1968.
Type-Locality: Wulasi, Northern Region, Ghana,
08° 3 91N. ,00° 0 0 1 G . M,
Distribution: Dahomey, Ghana, Ivory Coast, Togo, western Nigeria, and Upper Volta (Fig, 6),
Specimens Examined; Dahomey (26 2). Guene, 2 2
(7-F1); Banikoara, 10 (5-F1); Porga, 1; Segbana, 55 (18-F1);
Kouande, 37 (4-F1); Bimbereke, 1 (Fl); Nikki, 19 (1-F1);
Parakou, 1 (Fl); Soubroukou, 66; Diho, 25 (8-F1); Zizonkame,
25 (1—F1). Ghana (222). Bangwon, 10; Shishe, 32; Gambaga,
38; Navarro, 1 (BM-SO); Pulima, 4; Wa, 1 (BM); Lawra, 19
(BM, 7-SO); Pirisi, 23; Pong Tamale, 2 (BM); Nabogo, 14; 66
Damongo, 32 (1-F1); Sakpa, 15; Wulasi, 28; Yabrasso, 3.
Ivory Coast (48). Yama, 19 (2-F1); Sienso, 1; Bouna, 12
(2-F1); Kong, 2; Beoumi, 14 (BM). Nigeria (46). Afon, 13;
Upper Ogun, 2 (BM); Felele, 28; Ibaden, 3. Togo (125).
Namoundjoga, 8 (AT); Dapango, 74 (1-F1); Borgou, 26 (AT);
Padori, 2; Aledjo, 4 (AT); Pewa, 6; Pagala, 4 (1-F1);
Kamina, 1 (AT). Upper Volta (994). Petoye, 7; 17 km E
Goradji, 30; Tatarko, 25; 6 km SE Seguenega, 226; 9 km NE
Barga, 108; Dio, 106; Boussouma, 158; Konankira, 57; Goden,
24; 3 km SE Nayoure, 13; 1 km N Cella, 69; 8 km S Dana, 1;
Oulo, 1; Fo, 81; Arly, 4; Natiaboani, 22; 9 mi S Nobere, 3;
Founzan, 10; 8 km NE Satiri, 2; Kjipologo, 3; Nasso, 4
(MNHP); Bobo Dioulasso, 18 (MNHP); 27 km ENE Orodara, 4;
5 km SE Koutoura, 1; Sideradougou, 17.
Measurements: Measurements of the holotype (age class 4) followed by averages and extremes of five adults
(age class 4) from the type locality, are, respectively:
TOT 265, 269 (258-286); HB 109, 119 (109-126); TAL 146, 148
(139-160); HF 32, 31 (31-32); E 21, 21 (21); W 43, 47 (43-
50); GLS 34.3, 34.3 (34.0-34.7); OCNL 32.9, 33.1 (32.7-
33.7); BAL 25.1, 24.6 (23.9-25.3); ZB 17.2, 17.0 (16.5-
17.3); CBAL 26.3, 26,0 (25.4-26.6); BBC 14.7, 14.4 (14.1-
14.8); IOC 5.9, 6.0 (5.8-6.4); BR 4.5, 4.4 (4.2-4.6); GBAB
13.6, 13.2 (12.7-13.6); BrMlMl 7.1, 7.0 (6.7-7.6); DAL 8.2,
7.9 (7.5-8.3); PAL 14.3, 14.2 (13.7-14.7); PPAL 10.0, 10.1
(9.7-10.4); LAPF 5.9, 5.6 (5.4-5,9); LPB 7,1, 7.2 (7.0-7.5); 67
LAB 9.1, 9.1 (8.9-9.3); BAB 5.5, 5.4 (5.3-5.6); LN 13.6,
13.5 (13.2-13.8); LR 12.1, 12.2 (12.1-12.5); FNL 25.4, 25.4
(24.8-25.8); DC 14.0, 13.9 (13.4-14.2); LPPF 3.6, 3.8 (3.6-
4.1); ALM1 2.6, 2.7 (2.6-2.8); ALM 4.9, 5.2 (4.9-5.4).
Description: Interauricular, interorbital, circumorbital, and rostral areas same color as dorsum and varying from Blister to Sepia; mystacial and pectoral areas, fore- and hind limbs, and entire underparts white; cheeks and sides Clay Color; dorsal hairs plumbeous basally, the brown pigmented portion only 2-3 mm in length, some hairs finely tipped with dark brown, and some guard hairs entirely dark brown; fore-limb with four digits with claws and a reduced thumb; hind limbs with five digits with claws; and plantar surface naked except for an almost minute band of white hairs in the ankle region; pinna of ear long and almost naked, color almost the same as the dorsum, and anterior basal margin with short light brown hairs; vibrussae long and composed of both white and black hairs; tail long and uniformly Cinnamon-Buff with the dorsal hairs interspersed with dark brown hairs grading to a terminal
Clove Brown pencil; yet ventrally, the Cinnamon-Buff grades to near white in the region of the pencil.
Skull--medium-sized for the species and moderately robust; lachrymals medium; molariform teeth small; auditory bullae short, broad and bulbous; parapterygoid fossae deep 68 and moderately flared; nasals and rostrum short and broad; braincase slightly rounded.
Comparisons: Taterillus gracilis meridionalis can be distinguished from T. £. angelus by its short and wide nasals; a more rounded braincase; slightly smaller molars; markedly more bulbous bullae which generally are shorter and more inflated antero-laterally; and the pterygoid processer are shorter.
Taterillus gracilis meridionalis can be distinguished from T. g. nigeriae by its smaller skull size; weaker zygoma; smaller molariform teeth; shorter but more bulbous auditory bullae; and parapterygoid fossae which are not as flared.
Comparisons between T. g. meridionalis and T. g. gracilis show that the former have a more robust skull; shorter and broader nasals and rostrum; more rounded braincase; and more inflated auditory bullae,
Taxonomic Conclusions: Subspecific designation has not been previously given to specimens now included within the range of this subspecies. However, Rosevear (1953) did show in a distribution map of Nigerian Taterillus, the range of T, g. gracilis. The range was shown to be on both sides of the Niger River in Dahomey, Niger, Nigeria, and Upper
Volta. That geographical area is shown by this report to include only T. g. angelus and T. g. meridionalis (Pig. 6).
Petter (1975) recognized two subspecies of T. gracilis as T. g. angelus and T. g. gracilis. He restricted 69 the former to northern Nigeria and the latter to Gambia.
Specimens available to Petter from Ivory Coast, Mali, and
Upper Volta were not given subspecific consideration.
Ecological Observations: Morphological comparisons
using specimens I have assigned to T. gracilis show geographical differences. One morphologically similar group of specimens, assigned to T. g. meridionalis, differs
morphologically from specimens which were examined from east
of the Niger River in Niger and Nigeria (assigned to
T. g. angelus and T. g. nigeriae). The T. g. meridionalis specimens also differ morphologically from specimens of T.
gracilis that I have examined from Mali and Senegal (T. g.
gracilis). For these reasons, I believe that the Niger
River is at least a partial barrier to the dispersal of
T. gracilis, resulting in morphologically distinct groups,
Taterillus gracilis meridionalis, as well as T. g.
gracilis, is found in all of the vegetation zones in which
the genus Taterillus occurs. In contrast to the substrate
preference found for T, g. gracilis, T. g. meridionalis
has been found on the sandy soils in the Sahel Savanna of
Upper Volta (Vaden, 1969) as well as on the harder clay
like substrates in the Guinea and Sudan Woodlands of
Ghana, Togo, Dahomey, and Ivory Coast (L. W. Robbins,
1973-1975; C. B, Robbins, 1972). I have not been able to
identify any other species of Taterillus occurring with
T, g. meridionalis, which may account for a lack in its 70 substrate or habitat restriction. Habitat use and differ ences between T. g. meridionalis and other species in another genus are not known.
The locality from which the holotype was chosen is
in the southern part of the distribution of this subspecies
in the Guinea Woodland. As discussed previously, pelage color variation is a result of soil color and vegetation density. As a result, the pelage color of the holotype is
not the same as many other specimens in this subspecies. It
is distinguishable from the other subspecies by cranial
morphology.
Other rodents that occur within the range of this
subspecies are: Acomys cahirinus, A. dimidiatus, Arvicanthis
niloticus, Cryptomys foxi, Cricetomys gambianus, Dasymys
incomtus, Dendromus melanotus, Desmodilliscus braueri,
Euxerus erythropus, Grammomys rutilans, Graphiurus murinus,
Heliosciurus gambianus, H. rufobrachium, Lemniscornys
barbarus, L. striatus, Mastomys natalensis, Mus minutoides,
M. mattheyi, M. musculoides, Myomys daltoni, Steatomys sp.,
Tatera guinea, T. kempi, T. robusta, and Uranomys ruddi.
Taterillus gracilis nigeriae Thomas
Taterillus nigeriae Thomas, Ann. Mag. Nat. Hist., Ser. 8, 7:459-460, May, 1911. Type-locality, Kabir (=Kawbir), Nigeria. Taterillus gracilis angelus, Rosevear, Checklist and Atlas of Nigerian Mammals, 1953 (part). Taterillus gracilis gracilis, Rosevear, Checklist and Atlas of Nigerian Mammals, 1953 (part). 71
Holotype: Young adult male, skin and skull,
BM 11.3.24.14, collected by G. T. Fox, on 7 June, 1910.
Specimens Examined: Nigeria (184). Anara Forest
Reserve, 2 (BM); Dada, 103; Kabwir, 11 (BM); 1 mi S Kabwir,
1; Iella, 2 mi E Bahinde, 16; Tsanchaga, 8 mi E Bida, 7;
Mada River, 3 mi E Gudi, 31; Kudu, 3; Panyam, 10 (BM,
1-SO).
Measurements: Measurements of the holotype (age class 3) and averages and extremes of 10 adults (age class
4) from Kabwir and Panyam, Nigeria, are, respectively:
TOT -, - (-); HB 125, 111 (99-121); TAL 190, 156 (145-169);
HF 34, 30 (28-33); E 22, 20 (19-22); W -, - (-); GLS 36.7,
35.1 (33.2-36.4); OCNL 35.3, 33.9 (32.0-35.6); BAL 26.7,
25.5 (23.8-27.0); ZB 17.8, 17.0 (16.1-17.4); CBAL 28.8,
27.1 (25.6-28.0); BBC 14.8, 14.4 (13.9-15.1); IOC 6.3, 6.2
(5.8-6.7); BR 5.8, 4.7 (4.4-5.0); GBAB 14.2, 13.1 (12.7-
13.9); BrMlMl 7.5, 7.3 (7.0-7.6); DAL 9.0, 8.2 (7.4-9.2);
PAL 15.7, 14.8 (13.7-16.3); PPAL 10.9, 10.4 (9.8-10.8);
LAPF 6.8, 5.9 (5.4-6.7); LPB 7.3, 7.0 (6.5-7.7); LAB 10.1,
9.2 (8.7-9.4); BAB 5.7, 5.2 (4.9-5.4); LN 14.5, 13.8 (12.6-
14.6); LR 13.7, 12.9 (11.8-13.3); FNL 27.4, 26.3 (24.9-
27.2); DC 14.4, 13.7 (13.1-14.6); LPPF 4.0, 3.6 (3.4-3.8);
ALM1 2.7, 2.8 (2,6-2.9); ALM 5,6, 5.4 (5,2-5,8).
Description; Dorsum, interauricular interorbital, and rostral areas varying from Prout's Brown to Cinnamon-
Brown; posauricular patches indistinct but slightly lighter 72 in color than the dorsum; mystacial and pectoral areas, fore- and hind limbs, and entrie underparts white; cheeks and sides Pinkish Buff; dorsal hairs plumbeous basally, the brown pigmented portion only 2-4 mm in length, and a few hairs finely tipped with dark brown; pinna of ear long and almost the same color as the dorsum; and basally, the anterior portion with short buff-colored hairs; vibrussae long and composed of both white and dark brown or black hairs; tail long and uniformly Clay Color; the dorsal part darker due to interspersed dark brown hairs, and grading to a Blister pencil.
Skull — large and robust; zygoma moderately strong; lachrymals large; molariform teeth large; auditory bullae large, bulbous, and not markedly inflated laterally; nasals relatively long and broad; parapterygoid fossae deep and widely flaring; pterygoid processes short and slightly divergent; braincase moderately rounded.
Comparisons: Taterillus gracilis nigeriae differs from T. g, angelus by its lighter pelage color; slightly smaller but more robust skull; wider braincase; auditory
bullae less inflated laterally and more bulbous; shorter and broader nasals and rostrum. For differences with other subspecies see "Comparisons" under T. g. gracilis and T. g. meridionalis.
It is possible that T. cj. nigeriae may be found sympatric with T. lacustris. If so, it can be distinguished 73 by a larger skull and large inflated bullae, as opposed to the generally small skull and weakly inflated bullae of
T. lacus tris.
Taxonomic Conclusions: The first specimen assigned to this taxon, other than those in central Nigeria by
Thomas (1911), was by Matthey (1969) and Petter (1970, 1975).
That specimen is from Bou Rjeimat, Mauritania, and I have included it under the new species T. arenarius (Robbins,
1974b).
Rosevear (1953) showed that the distribution of
T. c[. nigeriae in Nigeria was only in the Guinea Woodland.
In a more complete discussion, Rosevear (1966) stated that the only specimen of Taterillus different from all other
West African Taterillus was the holotype of T, nigeriae.
As my examination has found and as Rosevear (1966) stated, the skin and skull of the T. nigeriae holotype are extremely large for an animal with relatively unworn teeth. Never theless, Rosevear (1966) discounted these attributes but further stated that the large hind-foot indicated specific independence.
As can be seen in a following section on "Cranial
Morphology" in Chapter 5, all of the large measurements of the holotype of T. nigeriae can be found in specimens assigned to this subspecies. Admittedly, these are usually very old individuals. Nevertheless, individual variation in all specific and subspecific categories within this 74 genus is great. Specimens that are unusually small or very large within a given age class are noticeable only when a large series is available from a single locality. Indi viduals within a population that are slightly smaller or larger than average, should not be considered specifically distinct unless other supporting evidence is available.
Ecological Observations: Taterillus gracilis nigeriae occurs in the Guinea and southern Sudan Woodlands in Nigeria. It does not occur further north in the sandier soil of the Sahel Savanna, which is occupied by T. g. angelus. This subspecies does not occur in those areas in which Desmodilliscus braueri and Gerbillus (G.) nigeriae are found. The differences, in fauna and habitat between T. g. angelus and T. g, nigeriae were discussed previously under
T. g. angelus.
Taterillus gracilis nigeriae is apparently restricted in distribution to the west by the Niger River. It has not
been found to the east where T, lacustris or T. congicus occur, or to the south where it is restricted in distribu
tion by the Benue River and the high forest. The unique
type-locality on the Jos Plateau, Nigeria, at Kabwir, is
apparently not a "mountain island" or restrictive to the distribution of this taxon. This is supported by the fact
that other taxa described from the Jos Plateau have been
identified in other parts of Nigeria, 75
Other rodents which occur within the range of this subspecies are: Acomys cahirinus, Aethornys stannarius,
Arvicanthis niloticus, Cricetomys gambianus, Cryptomys foxi,
Dasymys incomtus, Dendromus melanotis, Euxerus erythropus,
Graphiurus murinus, Ileliosciurus gambianus, Lemniscomys barbarus, L. striatus, Mastomys natalensis, Mus musculoides,
Myomys daltoni, Praomys jacksoni, P. tulbergi, Steatomys caurinus, Tatera kempi, Thryonomys swinderianus, and
Uranomys ruddi.
Taterillus harringtoni (Thomas)
Ta tera harr ingtoni Thomas, Ann. Mag. Nat. Hist., Ser. 7, 13:303, October, 1906. Type-locality, Mutti Galeb (north of Lake Rudalf), Ethiopia. Taterillus harringtoni, Thomas, Ann. Mag. Nat. Hist., Ser. 8, 6:222, August, 1910. Taterillus osgoodi Wroughton, Ann, Mag. Nat. Hist., Ser. 8, 6:293, September, 1910. Type-locality, Voi, Kenya. Taterillus tenebricus Dollman, Ann. Mag. Nat. Hist., Ser,. 8, 7:520, May, 1911. Type-locality, Nyama Nyango, Eusso Nyiro, Kenya. Taterillus nubilus nubilus Dollman, Ann. Mag. Nat. Hist., Ser. 8, 8:656, November, 1911. Type- locality, Orr Valley, Mt. Nyiro, Kenya. Taterillus nubilus illustris Dollman, Ann. Mag. Nat. Hist., Ser. 8, 8:656, November, 1911. Type- locality, 12 mi N of the Eusso Nyiro, Kenya. Taterillus melanops Allen, Bull. Mus. Comp. Zool., 54:446, April, 1912. Type-locality, Meru River, northern Guaso Nyiro, Kenya. Taterillus lowei Dollman, Proc. Zool. Soc. London, p~i 312, 16 June, 1914. Type-locality, 10 mi W Ngamatak Hills, Turkwell River, Kenya. Taterillus rufus (Wettstein), Anz. K. Akad. Wiss., 53:141, 25 May, 1916. Type-locality, El Obeid, Sudan. Taterillus kadugliensis (Wettstein), Anz. K. Akad. Wiss., 53:152, 25 May, 1916. Type-locality, Kadugli, Sudan. 76
Taterillus lorenzi (Wettstein), Anz. K. Akad. Wiss., 53:153, 25 May, 1916. Type-locality, El Obeid, Sudan. Taterillus emini zammarani De Beaux, Atti. Soc. Ital. Sci. Nat. e Mus. Civ. Stor. Nat. Milano, 61:27, February, 1922. Type-locality, Dolo, Somalia. Taterillus perluteus Thomas and Hinton, Proc. Zool. Soc. London, p. 259, 6 July, 1923. Type-locality, Um Kedada, 100 mi E El Fasher, Sudan. Taterillus nubilus meneghettii Toschi, J. East Af. Nat. Hist. Soc., 18:146, January, 1946. Type- locality, Olorgesailie, 20 mi north of Magadi, Kenya. Taterillus emini emini, Setzer, Proc. U. S. Nat, Mus., 106:498, 1956 (part). Taterillus emini, Matthey, Mammalia, 33:525, 1969. Taterillus emini, Robbins, Mammalia, 37:644, 1974 (part). Taterillus gracilis osgoodi, Hubbard, Zool. Afr., 7:445, 1973.
Holotype: Subadult female, skin and skull, BM
6.11.1.28, collected by P. C. Zaphiro, 26 July, 1905.
General Distribution of the Species; Central
African Republic, Ethiopia, Kenya, Somalia, Sudan, Tanzania,
Uganda, and possibly Chad (Fig. 7).
Published Records: 60 mi N Harar and north bank of Awash River, Ethiopia (Corbet and Yalden, 1972).
Specimens Examined; Central African Republic (1),
Gordil, 1 (MNIIP) . Ethiopia (2). Mutti Galeb, 1 (BM) ;
Vallee de l'Omo, 1 (MNHP); Kenya (91). North-Eastern
Region: Murri, 2 (BM) Ijara, 2 (BM) ; Wenje, 1 (BM) ;
Eastern Region: Orr Valley, 2 (BM-1; NA-1, SO); 8 mi N
South Horr, 2; Mt, Nyiru, West base, 4 (LA); Meru River, 1
(MCZ); 6 mi N Isiolo, 1; Ngare Ndare, 5; Ucase, 6 (AMNH,
3-SO); Rift Valley Region: Lodwar, 2 (BM); Turkwell River, BH85HBP
15"E 30 °E 40°E 45 °E 15°N
CHAD
SUDAN
ETHIOPIA AFRICAN CENTRAL REPUBLIC 5°N
SOMALIA
ZAIRE KENYA CONGO UGANDA
RWANDA,
300 BURUNDI
500 TANZANIA
K ilometers
Distribution and known localities of Taterillus harringtoni. 78 10 mi W Ngamatak Hills, 1 (BM); Lokori, 3 (NA); Kangatet,
4 (BM, 1—SO); 59 mi N Kapedo, 4; Marali, 1 (BM); Marsabit
Road, 1 (BM); 12 mi N Eusso Nyiro, 3 (BM-2; NA-1, SO);
12 mi N Archer's Post, 5; Nyama Nyango, 2 (BM, 1-S);
Samburu Game Lodge, 5; Baringo, 1 (NA, SO); Hot Springs,
Baringo Road, 2 (BM); Lake Hannington, 1 (NA, SO);
Olorgesailie, 5 (NA-4, 2-SO); Coast Region: 6 mi N Masabubu,
2 (AMNIi) ; 0.5 mi E Masabubu, 1 (AMNH); Merifano, 3 (NA) ; Mt.
Mbololo, 1 (HA); Voi, 12 (BM-5; MCZ-3; NA-3, 2-SO); Taveta,
1 (BM); Takaungu, 1 (BM); Malingo, 1 (MD-F1); Engen
Oplysningen, 1 (MD); Kisigau, 1 (NA-SO); Kauriro, 1 (NA-
SO). Somalia (12). Dolo, 7 (MSNG); Lugh, 1 (BM); Arenga, 1
(MSNG); Afmadu, 2 (MNHP); Chauli, 1 (MNHP). Sudan (60).
Darfur Province: near Tagbo Hills, 1 (BM); Madu, 80 mi NE
El Fasher, 1 (BM); 100 mi E El Fasher, 4 (BM-3; FM-1); 110 mi E El Fasher, 1 (BM); Um Kedada, 2 (BM); Kordofan Province:
190 mi E El Fasher, 1 (BM); El Obeid, 2 (VM); Nuba Mts.,
2 (BM); 2 5 mi W Nahud, 1 (BM); 3 5 mi E Nahud, 1 (BM);
Kadugli, 2 (VM); Bahr A1 Ghazal Province: Wau, 1 (BM); 10 mi S Wau, 1 (BM); Equatoria Province: Latome, 3 (MD);
Torit, 29 (FM-3; MCZ-5; MD-15); Idoto, 1 (FM); Moli, 2
(MCZ); Loferika, 1 (MD); Gondokoro,, 4 (1-Fl), Tanzania
(1). Kisima, 1 (BM). Uganda (9), Karamoja District:
Lotome, 8 (AMNH); Lorengileipi, 1 (NA).
Description: Interauricular, interorbital, and rostral regions same color as dorsum and varying from 79
Saccardo's Umber to Snuff Brown; mystacial and pectoral areas, postauricular patches, fore- and hind limbs, and entire underparts white; cheeks and sides Tawny-Olive; dorsal hairs plumbeous basally, the brown pigmented portion only 2-3 mm in length, with most hairs finely tipped with dark brown; pinna or ear long and almost naked, color almost the same as the dorsum, and anterior and basal margin with short buff-colored hairs; vibrussae long and composed of both white and dark brown hairs; tail long and uniformly
Pinkish Buff ventrally with dorsal hairs interspersed with darker brown hairs, grading to a terminal Clove Brown pencil.
Skull--medium size and moderately robust; zygoma weak; lachrymals relatively large; molariform teeth and incisors small; auditory bullae large, bulbous, and greatly inflated anteriorly and antero-laterally; parapterygoid fossae flared but not deep; rostrum medium in length and narrow; nasals long and straight sided; braincase rounded.
Measurements: Measurements of the holotype (age class 2) followed by averages and extremes of five adults
(age class 4) from northern Kenya, are, respectively:
TOT -, 280 (270-290); HB 96, 119 (115-125); TAL 132, 161
(155-165); HF 28, 30 (29-31); E 19, 19 (18-20); W -, 51
(49-55); GLS 30.7, 35,0 (34.4-36.0); OCNL -, 33.8 (33.1-
34.7); BAL 22.4, 25.2 (24.6-25.8); ZB -, 17.7 (17.3-18.2); 80
CBAL 24.5, 26.9 (26.3-27.6); BBC 14.0, 14.9 (14.2-15.5);
IOC 6.0, 6.4 (6.2-6.7); BR 4.1, 4.4 (4.1-4.7); GBAB -, 13.9
(13.4-14.5); BrMlMl 6.4, 7.0 (6.9-7.1); DAL 7.3, 7.9 (7.4-
8.2); PAL 13.1, 14.1 (13.6-14.6); PPAL 9.1, 10.8 (10.2-
11.2); LAPF 5.1, 5.7 (5,5-5.9); LPB 6.4, 6.9 (6.6-7.1);
LAB 9.0, 9.7 (9.4-10.0); BAB 5.7, 5.6 (5.3-5.9); LN -,
13.9 (13.5-14.2); LR -, 12.7 (12.4-12.9); FNL 26.2
(25.6-26.6); DC 12.6, 13.7 (13.3-14.0); LPPF 3.0, 3.4 (3.0-
3.8); ALM1 2.2, 2.5 (2.2-2.7); ALM 4.6, 5.1 (4.9-5.3).
Comparisons: Taterillus harringtoni can be dis
tinguished from both T. congicus and T. emini by its rela
tively larger and more bulbous auditory bullae which are
more inflated anteriorly and antero-laterally. It is also distinguished by its smaller skull size, shorter molariform
teeth, and smaller incisors. Taterillus harringtoni is
distinguished from T. lacustris by its smaller molariform
teeth and greatly inflated bullae (T. lacustris has larger
molars and weakly inflated bullae).
Taterillus harringtoni is characterized by its
karyotype of 2N=44, FN=62. When comparing the holotype to
specimens reported with the 2N=4 4 karyotype, they are
identical,
Taxonomic Conclusions: The type-locality of T,
harringtoni, listed by Thomas (1906) as Mutti Galeb,
Ethiopia, has proven difficult to trace. The modifier
"east of Lake Rudolf" further confused the problem. 81
Ogilvie-Grant (1913) listed the itinerary and showed a route map of the collecting trip of P. C. Zaphiro, the collector of T. harringtoni. Since the paper by Ogilvie-
Grant was on the birds collected in Ethiopia by Zaphiro, it has apparently remained unknown to mammalogists until this time. Although Mutti Galeb is not shown nor mentioned, localities are given on days just prior to and after the date that Zaphiro was at Mutti Galeb. This enables me to pinpoint Mutti Galeb as being between Baku and Kerre (approximately 05°40'N., 36°20'E.), almost due north of Lake Rudolf in the Omo River Valley (rather than east of Lake Rudolf).
The taxonomic status of T. harringtoni and its relationship to other taxa in Uganda, Somalia, and northern
Kenya was not investigated until this report. A specimen of Taterillus was collected in the Omo Valley, Ethiopia, with a karytoype of 2N=44, FN=62 (Matthey, 1969). It was assigned to T. emini although its relationship to the nearby type-locality and holotype of T. harringtoni was not dis cussed; possibly because the location of the type-locality was unknown. Petter (1975) reassigned T. harringtoni
(Thomas, 1906) as a subspecies of T. lowei Dollman (1914), apparently ignoring his own specimen previously identified as T, emini, and that T. harringtoni is an older name which has priority. Genest and Petter (1973) also designated the
Omo Valley specimen as T, emini which would make it the same 82 species as that from the type-locality of T. e_. emini from
Uganda and T. e. zammarani from Somalia.
Specimens of Taterillus with a karyotype of 2N=44,
FN=62, from Central African Republic, Ethiopia, and Somalia, were also assigned to T. emini (Genest and Petter, 1973).
They (Genest and Petter) did not discuss the relationship of T. emini to T. lowei or T. harringtoni. My study
(Robbins, 197 4a) of the many named taxa of Taterillus in
Kenya determined that they should all be placed in T. emini, based on their karyotype of 2N=44 being assigned to T. emini by Matthey (1969) and Genest and Petter (1973). I have subsequently reanalyzed the specimens with a chromosome number of 44. The morphology of these specimens changes the above species designation. As indicated in the
"Comparisons" section, morphological comparisons of the specimens with the 2N=44 karyotype to those from near the type-locality of T. e. emini, including the holotype, show many differences. These morphological differences have allowed a reassessment of the taxa previously thought to be a part of the T. emini complex and are responsible for the synonomy under T. harringtoni.
A revision of the Taterillus of the Sudan by Setzer
(1956) included specimens from Torit and nearby localities.
These were allocated to T. e. emini, Comparisons of these specimens to those I have assigned to T. emini from Uganda and those of T. harringtoni from Kenya, show that both species are present. Also, Setzer (1956) assigned T. kadugliensis, T. lorenzi, T. perluteus, and T. rufus as subspecies of T. emini. My comparisons of these taxa to
T. emini and T. harringtoni show that they should be con sidered conspecific with T. harringtoni.
The single specimen of T. osgoodi collected by
C. A, Hubbard, was reported by Davis (1966b) to be the first record of the genus Taterillus in Tanzania. Hubbard (1973) subsequently reported on that specimen and one from a different locality in Tanzania. I have examined a specimen from Tanga, Tanzania, identified by Hubbard as a
Taterillus. It is a Tatera and makes the above identifica tions doubtful. However, there is a specimen in the British
Museum of Natural History, collected in 1965 by G. S. Child from Kisima, Tanzania that I have identified as a Taterillus.
It is referable to T. harringtoni and is possibly the only known specimen of Taterillus from Tanzania.
Ecological Observations: Taterillus harringtoni occurs in the Sudan Woodland and Sahel Savanna regions of
East Africa. The only species of Taterillus that I have identified as occurring in Tanzania, Kenya, Somalia, and southern Ethiopia is T. harringtoni, Taterillus harringtoni is sympatric with T. emini in northern Uganda and the Sudan and with T. congicus in Central African
Republic and the Sudan, Any ecological differences or 84 habitat segregation between T. harringtoni, T. congicus, and
T. emini are unknown at present.
Other rodents that occur within the range of T. harringtoni are: Acomys dimidiatus, A. hystrella, A. intermedius, A. subspinosus, A. silsoni, Aethornys kaiseri,
Arvicanthis niloticus, Desmodilliscus braueri, Gerbillus
(G.) cosensi, G. (G.) dunni, G. (G.) gerbillus, G. (G.) pulvinatus, G. (H,) pusillus, Lemniscomys barbarus, L. macculus, L. striatus, Mastomys natalensis, Mus bellus, M. musculoides, M. tenellus, Saccostomus campestris, Tatera benvenuta, T. kempi, T. nigricauda, T. robusta, Taterillus congicus, T. emini, and Xerus rutulus.
Taterillus lacustris (Thomas and Wroughton)
Tatera lacustris Thomas and Wroughton, Ann. Mag. Nat. Hist., Ser. 7, 19:37, May, 1907, Type-locality, Lake Chad (=Kaddai), Nigeria. Taterillus lacustris, Thomas, Ann. Mag, Nat. Hist,, Ser, 8, 6:222, August, 1910. Taterillus nigeriae, Thomas, Ann, Mag. Nat. Hist,, Ser. 8, 7:460, May, 1911 (part). Taterillus emini lacustris, Hatt, Am. Mus. Novitates, No, 789, p, 3, 4 April, 1934. Taterillus gracilis, Rosevear, The Rodents of West Africa, 1969 (part).
Holotype: Adult male, skin and skull, BM 7.7.8,129, collected by G. B. Gosling, on 9 February, 1905,
General Distribution of the Species; Northeastern
Nigeria, Cameroon, and possibly Chad and Niger (Fig, 8).
Specimens Examined: Cameroon (35). 10 km E Fort-
Foreau, 4 (AMNH); 35 km S Garoua, 1 (AMNH); 50 km S Maroua, 10°E 15°E
MAURITANIA
MALI NIGER
15 N SENEGAL
PORTT^ ^/GUINEA UPPER VOLTA
GUINEA 10°N
I L SIERRA 1 GAMBIAI LEONE IVORY NIGERIA GHANA COAST
300 CAMEROON 500
K i Jometers
Fig. 8. Distribution and known localities of Taterillus lacustris. CO 1 (AMNH); Waza, 10 (AMNH); Yagoua, 17 (AT); Mora, 2 (MNHP).
Nigeria (23). Kaddai, 10 (BM); 22 mi S Maiduguri, 3;
Maiduguri, 1 (BM); 31 mi NE Dikwa, 6; Bama, 2 (BM); Yola, 1
(BM, SO).
Description: Dorsum interauricular, interorbital, and rostral areas varying from Buckthorn Brown to Ochraceous-
Tawny; postauricular patches, circumorbital region, mystacial and pectoral areas, inner sides of fore- and hind limbs, and entire underparts white; cheeks and sides Pinkish Buff; dorsal hairs plumbeous basally, the buff-colored portion
3-4 mm in length, and some hairs tipped with dark brown; pinna of ear long and almost naked, color almost the same as dorsum, and basally the anterior portion with short buff-colored hairs; vibrussae long and composed of both white and dark-brown hairs; tail moderately long and uniformly Ochraceous-Tawny basally with the dorsal hairs interspersed with darker brown hairs grading to a mummy
Brown pencil; yet ventrally, the tail color is lighter in appearance due to the absence of the dark hairs, the
Ochraceous-Tawny color grades to nearly white in the region of the pencil.
Measurements: Measurements of the holotype (age class 5), followed by a single individual (age class 5) of known karyotype from Mora, Cameroon, are, respectively:
TOT 260; I1B 119, 115; TAL 156, 145; HF 31, 32; E 19, 21;
W -, GLS 3 6.2, 35.2; OCNL 35.2, 34.2; BAL 26.9, 26.1; 87
ZB -, 17.3; CBAL 28.7, 27.5; BBC 14.4, 14.1; IOC 6.1, 5.9;
BR 4.5, 4.4; GBAB -, 13.6; BrMlMl 7.6, 7.2; DAL 8.5, 8.5;
PAL 15.1, 14.7; PPAL 11.4, 10.5; LAPF 5.8, 6.3; LPB 7.4,
6.7; LAB 9.8, 8.9; BAB 5.8, 5.1; LN 14.5, 14.3, LR 13.2,
13.1; FNL 17.2, 26.3; DC 14.1, 13.5; LPPF 4.1, 3.2; ALMl
2.9, 2.7; ALM 5.4, 5.3.
Comparisons: Taterillus lacustris can be dis tinguished from all other West African species of Taterillus
(T. arenarius, T. gracilis, and T. pygargus) by its narrow and weakly inflated auditory bullae. Its bullae are similar in shape to both T. emini and T. congicus, but is distinguishable from these two species on the basis of its smaller cranial measurements. Although T. lacustris is similar in skull size to T. harringtoni, it can be dis tinguished by its larger molariform teeth and weakly inflated bullae (T. harringtoni has greatly inflated bullae).
Taterillus lacustris is karyotypically distinct from all other species of Taterillus with a 2N-28 and FN=44
(.Tranier et al., 1974). My comparisons of the holotype of
T, lacustris with specimens reported with the 2N:=28 karyotype show that they are identical in every respect.
Taxonomic Conclusions; In the original description of T. lacus tris by Thomas and Wroughton (1907) , the type- locality was listed as Lake Chad, It had been assumed by some (Petter, 1975) that the distribution of this species 88 and its type-locality was in Chad. Rosevear (1953) listed the type-locality as northeastern Nigeria. I determined that the type-locality was Kaddai, Nigeria by examining the publication by Alexander (1907). He gave the route maps and itinerary of the Gosling expedition and pinpointed the camp of G. B. Gosling as Kaddai, when the holotype and type-series were collected,
The designation by Rosevear (1969) of T. lacustris as a subspecies of T. gracilis was due to a lack of compara tive material and chromosomal data. He mentioned that the type-series was composed of all old animals, making com parisons to other specimens unreliable. My comparisons using the specimens available to Rosevear and the many additional specimens of Taterillus now available from West
Africa, has shown the T. lacus tris and T. gracilis are morphologically distinct. Also, these two species differ chromosomally with T. gracilis having 36 or 37 chromosomes and T, lacus tris having 28,
The inclusion of T, perluteus as a subspecies of
T, lacustris by Petter (1975) does not agree with my analysis. My comparisons of the holotype of T, perluteus to specimens allocatable to other taxa has shown it to be a part of the T. harringtoni complex. It can be dis tinguished from T. lacustris as given under "Comparisons" above. For additional remarks on T. perluteus and T. harring toni, see the section on T. harringtoni. 89
During their study on the Taterillus from Central
African Republic, Genest and Petter (1973) speculated that
T. congicus might become synonymous with T. lacustris.
Recently available karyotypes of specimens from Fort-Lamy,
Chad, showed them to have 54 chromosomes, that assigned to
T. congicus, while specimens from Mora, Cameroon had a karyotype of 2N-28. The 2N=28 karyotype was assigned to
T. lacustris, indicating the distinctness of T. congicus and T. lacustris (Tranier et al., 1974). Even so, they
(Tranier et al.) were not able to find any cranial or dental features that would distinguish the two species. They did speculate that differences in length of cranium might differentiate the two species providing animals with equivalent tooth wear (the same age class) were compared.
Ecological Observations: This species seems to be restricted to the very arid Sahel Savanna of northeastern
Nigeria and northern Cameroon, Although I have not been able to identify T. lacustris as occurring with T. gracilis in Nigeria or Cameroon, with T. emini or T. congicus in
Cameroon or Chad, or with T. arenarius in Niger, the possibility seems likely.
Other rodents which occur within the range of T. lacustris are: Acomys johannis, Arvicanthis niloticus,
Cricetomys gambianus, Desmodilliscus braueri, Euxerus erythropus, Gerbillus (G.) gerbillus, G. (G.) nigeriae,
Lemniscomys barbarus, Mastomys natalensis, Mus hausa, Tatera 90 wellmani, and probably Taterillus arenarius, T. congicus,
T. emini, and T. gracilis.
Taterillus pygargus (F. Cuvier)
Gerbillus pygargus F. Cuvier, Trans. Zool. Soc. London, 2(10):142, pi. 25, Figs. 10-14, 4 May, 1838, Type- locality, Senegal (probably St. Louis). Taterillus pygargus, F. Petter, Mammalia, 16:37-39, 1952. Gerbillus senegalensis G. Cuvier, Nouv. Ann. Mus. d'Hist. Nat., 2:444, 1833, nomen nudum. Taterillus gracilis, Matthey, Mammalia, 33:522-528, 1969 (part). Taterillus gracilis, Rosevear, The Rodents of West Africa, 1969 (part).
Holotype: Subadult, sex unknown, skull only
(.without mandible), MNHP A2658 (370), collector and collec tion date unknown.
General Distribution of the Species: Gambia, western Mali, southern Mauritania, and Senegal (Fig. 9).
Specimens Examined: Gambia (3). Lower River
Division: Toniataba, 2; Upper River Division: Sun Kunda,
1 (BM). Mali (3). Bamako, 3 (MNHP), Mauritania (19).
Trarza Region: Garak, 10; Guidimaka Region: Passe de Soufa,
9. Senegal (250). River Region: Podor, 1; Debi, 1 (MNHP);
Richard Toll, 12 (MNHP-1); Madina Ndiatebe, 1 (MNHP); Pete-
Ole, 25 (MNHP); Galoya, 1 (MNHP); 10 km SE St. Louis, 14;
Senegal, probably St, Louis, 1 (MNHP, SO); Ogo, 2; Ranerou,
20; Thies Region: 10 km E Thies, 2; Bandia, 3 (MNHP);
Oriental Region: 5 km S Bakel, 46; Goudiry, 4; Koussanar,
13; Kotiari Naounde, 5; Gamon, 2 (BM); Diakaba, 1 (BM); MAURITANIA
MALI NIGER
.SENEGAL
PORT-! UPPER VOLTA GUINEA
GUINEA
SIERRA NIGERIA 1 GAMBIA 1 LEONE IVORY GHANA COAST
CAMEROON Miles
K i lometers
Fig. 9. Distribution and known localities of Taterillus pygargus.
H 92 Diourbel Region: 8 km E Louga, 17; 17 km NE N'Doulo, 33;
Linguere, 30 (MNHP-9); Sine-Saloum Region: 15 km N Kaffrine,
8; Koungheul, 4; Saboya, 3 (MNHP); Tubakuta, 1 (BM).
Description: Specimens from 10 km SE St, Louis are dorsally Tawny-Olive; preorbital region has grey hairs; mystacial and pectoral areas, postauricular patches, fore- and hind limbs, and entire underparts are white; cheeks and sides Cinnamon-Buff; terminal part of the rostrum has dis tinct dark brown hairs; the interauricular area is the same color as the dorsum; dorsal hairs plumbeous basally, the buff-colored portion being 2-3 mm in length with some hairs finely tipped with brown; tail long and both dorsally and ventrally is Pinkish-Buff, the dorsal part is interspersed with dark brown hairs grading to a terminal pencil which is
Fuscous, and ventrally the Pinkish-Buff grades to white in the region of the pencil.
Specimens that I have identified as T. pygargus from the Guinea Woodland in southern Senegal, dorsally approach
Prout's Brown and are very nearly the same color as specimens of T, gracilis occurring in the same localities
(see "Description" under T. g. gracilis).
Skull--moderate in size; zygoma weak; lachrymals large; molariform teeth relatively large; auditory bullae well inflated anteriorly and antero-laterally; rostrum relatively short and narrow; nasals broad and short; brain- case slightly rounded. 93
Measurements: Measurements of the holotype (age class 2) followed by averages and extremes of 13 adults
(age class 4) are, respectively: TOT -, 266 (266-267); HB
114 (108-120); TAL -, 153 (140-160); HF -, 30 (29-32); E -,
19 (19-21); W -, 45 (44-48); GLS 32.1, 34.9 (34.0-36.3);
OCNL 30.8, 33.6 (32.4-35.2); BAL 22.9, 25.2 (23.8-26.6);
ZB -, 17.4 (16.6-18.3); CBAL 24.7, 26.9 (25.6-28.1); BBC
14.3, 14.1 (13.7-14.7); IOC 5.7, 6.1 (5.5-6.4); BR 4.4, 4.4
(4.1-4.6); GBAB -, 13.6 (13.0-14.3); BrMlMl 6.9, 7.2 (6.8-
7.7); DAL 7.3, 8.1 (7.4-9.0); PAL 13.4, 14.3 (13.7-15.2);
PPAL 9.3, 10.5 (9.7-11.3); LAPF 5.4, 5.7 (5.3-6.2); LPB 6.2,
6.9 (6.3-7.3); LAB 8.7, 9.3 (8.7-9.7); BAB 4.8, 5.5 (5.2-
5.9); LN 12.3, 14.2 (13.7-14.9); LR 11.0, 12.8 (12.2-13.5);
FNL 24.2, 26.4 (25.2-27.1); DC -, 13.6 (13.2-14.1); LPPF
3.7, 3.7 (3.2-4.1); ALM1 2.7, 2.8 (2.6-3.0); ALM 5.0, 5.3
(4.9-5.6).
Comparisons: This species differs from T. gracilis as follows: in areas of sympatry, T. pygargus usually has a shorter, more robust skull, a shorter rostrum in relation to length of the skull, and markedly inflated bullae, especially anteriorly and antero-laterally. The bullae in
T. gracilis are narrow anteriorly and not nearly as inflated.
The short rostrum in T. pygargus (compared to T. gracilis) does not hold in specimens of age class 3 or younger.
Bullar shape is the only reliable method for distinguishing these two species for all age classes. 94
For differences between T. pygargus and T. arenarius, see "Comparisons" under T. arenarius.
Taxonomic Conclusions: The original listing of
Gerbillus pygargus was by F. Cuvier (1825), The name as given is nomen nudum in that no holotype was designated nor was there any description of the species. A subsequent description of Meriones gerbillus by Ruppell (1830) was reported by F. Cuvier (1838) to be a synonym of G. pygargus in his formal description of the latter species. The allocation of Meriones gerbillus as being G. pygargus by F.
Cuvier (1838), as was done later by Anderson (1902), was in error. Anderson (1902) also designated Ruppell's
(1830) original specimen of Meriones gerbillus as a co-type of G. pygargus. The specimen of Ruppell (1830) was from
Upper Egypt, and probably a G. pyramidum (Petter, 1952).
The holotype of G. pygargus is from Senegal (labeled
"Gerbil du Senegal"), about 6000 km west of Egypt.
The designation of a co-type of G. pygargus by
Anderson (1902) has probably confused subsequent workers.
Allen (1939) listed G. pygargus of F. Cuvier (1838), but incorrectly gave its type-locality as Upper Egypt. Ellerman
(1941) synonomyzed G. pygargus under G. pyramidum, when in fact he was not dealing with a name in proper usage. All of Ellerman's specimens were Egyptian Gerbillus.
Further confusion could have resulted from the list of specimens in the Paris Museum by G. Cuvier (1833). He 95 designated the specimen "Gerbile du Senegal" as Gerbillus senegalensis, This also is a nomen nudum for and probably the same specimen as the holotype of G. pygargus (=Taterillus pygargus).
It is also possible that workers prior to Petter
(1952), upon examination of the holotype of G. pygargus in Paris, utilized the mounted skin for their comparisons.
As Petter (1952) stated and as I have confirmed, the skin of G. pygargus is a G, pyramidum-type of animal. But, the associated skull, which agrees in every aspect with the figures in the original description of F. Cuvier (1838), is a Taterillus. Petter (1952) designated the skull as the holotype of T. pygargus with its type-locality in Senegal, probably near the port of Saint Louis (Petter et alt, 1972).
Originally, a karyotype of 2N-23 was reported on specimens of Taterillus from Senegal and allocated to
T. gracilis (Matthey, 1969), Then, specimens with a karyotype of 2N=3 6 were found in Upper Volta, These were also designated as T. gracilis (Matthey and Petter, 1970).
Subsequent discovery of sympatric forms exhibiting both of the above karyotypes were found in Senegal (Matthey and
Jotterand, 1972) and resulted in the designation of the
2N-22/23 forms as T. pygargus (Petter et al., 1972).
Further distinction between T. pygargus and T. gracilis has been achieved by serum-protein electrophoresis techniques (Hubert and Baron, 1973; Baron et al., 1974; 96
Tranier et al., 197 4). The lack of intermediate karyotypes combined with serum-protein distinctness eliminates the possibility that these two species could be a single species with extremely polymorphic karyotypes. In addition, Matthey and Jotterand (1972) have shown that breeding a male T. gracilis (2N=37) with a female T. pygargus (2N=:22) produced a sterile offspring with a diploid number of 30. The karyotype of the sterile hybrid differs greatly from the
2N=30 karyotype of T. arenarius.
My morphological comparisons of the holotype of
T. pygargus to specimens with a karyotype of 2N-22/23, shows that they agree in every respect, and correspond to those characters mentioned as descriptive of T. pygargus as mentioned under "Description" and "Comparisons" previously.
Ecological Observations; The habitat requirements of T. pygargus range from the moist Guinea Woodland in southern Senegal to a very arid environment in the Sahel
Savanna in northern Senegal and southern Mauritania.
Taterillus pygargus is found with T. gracilis in the main
part of its distribution and with T. arenarius at its
northern limits in southern Mauritania, In areas of sympatry with T» gracilis, T. pygargus was found on the more sandy substrates while T. gracilis preferred the harder, clay-like substrates (Hubert, 1973). Its soil preference is reversed in areas of sympatry with T. arenarius, with T.
pygargus occurring on the harder substrates in tall grasses 97 and T. arenarius occurring on the sandy substrate with sparse vegetation (Robbins, 1974b).
Petter et al. (1972) compared population densities of T. pygargus and T. gracilis in the Sahel Savanna and the
Sudan and Guinea Woodlands of Senegal. Taterillus pygargus comprised about 9 0 per cent of the total numbers of
Taterillus in the Sahel Savanna, 60 per cent in the Sudan
Woodland, and less than 5 0 per cent in the Guinea Woodland.
Because of these results, Poulet (1972a, 1972b, 1973) stated that T. pygargus preferred more arid conditions than did T. gracilis. I believe that it also indicates a greater predominance of sandy soils in the more arid vegetation zones in Senegal, which would account for the abundance of
T. pygargus in relation to T. gracilis in the more arid habitats.
Poulet (1973) has studied the ecological relation ships of the rodents in northwestern Senegal. Although both
T. pygargus and T. gracilis occur in the vicinity of his study plots, his paper dealt only with T. pygargus. Pre sumably the area chosen for the study is only suitable for
T. pygargus. At the time of Poulet's study (.1969-1971) , the only methods known for separating T, gracilis from T. pygargus were by karyotype or serum-protein analysis. All of the animals on the study plot were not tested with these techniques, so it is possible that both species occur there. 98
If this is true, the ecological findings of Poulet (1973) attributed to T. pygargus should be revised.
One of the important observations by Poulet (1973) concerns reproduction and life span. Female Taterillus reached sexual maturity at age two months while males were not sexually mature until three months of age. This could be an effective mechanism for decreasing the chances of inbreeding and probably contributes to the dispersal of subadult animals. Poulet also found that no marked animals over a year old were recaptured.
Although Poulet (1973) stated that reproduction only occurred from November to March, his data show that some percentage of the population of Taterillus was reproducing year round. With a gestation period of three weeks and an additional three week period for rearing the young, it is possible for a female to have a litter every six weeks. It is apparent from Poulet's data that reproduc tion is closely tied in with the previous year's rainfall.
His first year's study was conducted when the previous year's rainfall had been normal. During his first year's study he found a high population density of Taterillus. The next year's data show a sharp decline in animal numbers. This followed a year with below normal rainfall. Also, according to Poulet (1974), the 1972-73 drought had affected the mammalian populations at Fete-Ole, Senegal. Desert species like Desmodilliscus braueri increased in number. While 99
T. pygargus decreased sharply, they continued to breed.
Poulet attributed this to a surplus of seeds remaining from
the previous year.
Other rodents which occur within the range of T.
pygargus in Senegal are: Arvicanthis niloticus, Cricetomys
gambianus, Desmodilliscus braueri, Euxerus erythropus,
Gerbillus (G.) pyramidum, Lemniscomys barbarus, Mastomys
natalensis, Mus minutoides, Steatomys sp., Tatera gambiana,
T. sp., Taterillus gracilis, and Uranomys ruddi. CHAPTER 5
SPECIFIC RELATIONSHIPS
Relationships among the seven species of Taterillus recognized in the previous chapter, including subspecies, are divided into four sections: cranial morphology, comparative karyology, pelage color, and distribution.
Within each of these sections, the taxa will be compared by analyzing those occurring in West Africa, East Africa, and the total distribution of the genus Taterillus.
Cranial Morphology
The cranial morphology of the taxa of Taterillus have been compared using principal components analysis, discriminant function analysis, phenetic distances as shown in distance phenograms, nonmetric multidimensional scaling, and minimum spanning tree analysis. These multivariate statistical techniques are explained in Appendix C.
Comparisons of the species have been made for those that were judged to be closely related systematically, as well as for all seven species to show the relationships for all of the taxa in the genus Taterillus. The species which are closely related can be divided into two geographical areas:
West Africa, from Senegal and Mauritania to Cameroon; and
100 101
East Africa, from Cameroon, east to Somalia, and south to northern Tanzania.
West Africa
The West African samples used for statistical analyses of species and intrapopulational variation are shown by their locality numbers in Fig. 10. The exact localities of these samples are given in Appendix D along with standard statistics for the 24 cranial measurements used in the analyses.
Recognition and differentiation of the West African taxa involved two major problems. First, cranial comparisons and differentiation of the two taxa present in Senegal (T. gracilis and T. pygargus); and second, analysis of the geographic variation, including possible subspecific differences in T. gracilis. Once these two problems were analyzed, comparisons between the four West African species
(T. arenarius, T. gracilis, T. lacustris, T. pygargus) were possible.
Taterillus gracilis and Taterillus pygargus. Morpho logical distinctions except for karyotypic differences between T. gracilis and T. pygargus in Senegal, have been discussed briefly by Robbins (1974b). This report presents more detailed discussion of these distinctions. My first attempt at separating these two taxa was by means of a principal components analysis (Fig. 11, Table 1). In order MAURITANIA 28' ® NIGER
SENEGAL 19 ® ®21 o UPPER ©30 7(1 ® OZU O VOLTA JO ® ® © GUINEA S
G GAMBIA 5® IVORY • GHANA NIGER A COAST CAMEROON
0 MHeS 500 I J K i lometers J Fig. 10. Locality numbers used in statistical analyses of West African Taterillus. oto 103
11
14 G 9 •r* '10 13 p p *19 H 18*
•?
17 IE Fig. 11. Two-dimensional projections of the first three principal components of Senegal Taterillus with known karyotype — Individual specimens designated with a "G" are karyotyped specimens of T. gracilis, and those with a "P" are T. pygargus. 104
Table 1. Loadings of the most important cranial measure ments on the first three principal components for individual karyotyped specimens of T. g. gracilis and T. pygargus from Senegal -- For explanation of abbreviations see Appendix B. An asterisk indicates those measurements which contribute the most to a given component.
Component
Measurement I II III
GLS . 866* -. 334 -.028 OCNL .887* -.339 - . 000 BAL . 872* -.341 . 065 ZB .389 .400 .382 CBAL .888* - . 248 -.240 BBC . 599 . 252 .438 GBAB . 437 .598* - .385 BrMlMl -.097 . 059 .913* PAL .651* - .290 .373 PPAL .687* -.247 . 042 LAPF -.138 . 238 .600* LAB .645* . 407 -.215 BAB . 299 .663* -.422* LN . 070 -.366 -.466* LR .406 -.530* -.480* DC .333 . 592* .007 ALM1 .204 .705* - .152 ALM . 456 . 552* . 185 105 to determine possible cranial differences only specimens of known karyotype were used representing both species. For
T. gracilis, localities 22, 23, and 27 are represented, and for T. pygargus, localities 23, 26, and 27 (see Fig. 10 and
Appendix D).
Individual specimen placement on the first and second components as well as on the second and third components is shown in Fig. 11. It is the first component that separates the two species. The first component, accounting for 26.5 per cent of the variation in the 23 cranial measurements used, is influenced by most of the cranial size measurements (Table 1). This component indicates that for the individual specimens used and previ ously allocated to either T. gracilis or T. pygargus by karyotype, that T. gracilis has a longer skull and longer auditory bullae.
The second component, accounting for 17.7 per cent of the variation, is influenced by breadth across the bullae
(_GBAB) , breadth of auditory bullae (BAB) , depth of cranium
(DC), molar toothrow measurements (ALMl, ALM), and nega tively, length of rostrum. A slight trend can be seen in
Fig. 11, showing that some separation is achieved by the second component, Taterillus gracilis seems to have a smaller breadth across the bullae, narrower bullae, smaller molars, and a longer rostrum. 106 The third component, accounting for 13.7 per cent of the variation, is influenced by breadth across the molars, length of anterior palatine foramina, and nega tively, length of nasals, length of rostrum, and breadth of bullae. This component does not contribute to the separa tion of the two species.
Using most of the specimens of T. gracilis and T. pygargus from the above analysis, a discriminant analysis was made to compare the Taterillus taxa occurring in extreme western Africa. The discriminant analysis included specimens of T. arenarjus from its type-locality in western
Mauritania (locality 28, Fig. 10). Figure 12 and Tables 2 and 3 show the results of this discriminant analysis. Of major concern at this time are the differences between T. gracilis and T, pygargus. Taterillus arenarius will be discussed in the last part of this section on the cranial differences for all of the West African taxa.
All of the variation in the 23 cranial measurements used is accounted for in the two canonical variates (Fig.
12), The discriminant coefficients in Table 2 can be used for allocating unknown specimens or populations to one of these three species. Using the coefficients given for separating two of the species as shown, the values (means) for the particular measurement shown are needed. Multiply ing the coefficient by the measurement value and summing them gives a discriminant score for a given species. The LAB
PPJ/.
3-
oo ALS11 LPPF rsi 2- T. GRACILIS GBAB LU
•< Di £ 1 < u o 0- z < u
-1 • T ARENARIUS
T. PYGARGUS -2-
i— —r~ —i— -5 -4 -3 -2-10 1 2
CANONICAL VARIATE 1 (80.7ro) Fig. 12. Projection of the first two canonical variates for three West African species of Taterillus. 108
Table 2. Discriminant coefficients for characters most useful in separating three West African species of Taterillus — For explanation of cranial measurement abbreviations see Appendix B. A = T. arenarius, G = T. gracilis, P = T. pygargus.
Character A vs . P A vs . G P vs. G
BAB 9.287 11.552 2. 265
GBAB -6.197 -0.334 5. 863
PPAL -11.921 -20.116 -8.195
GLS 4.758 -7 .975 -12.715
ALM1 14.857 22.794 8.237
LAB -6 .624 -2.408 4.216
LPPF -14.407 -13.813 0.594 109
Table 3. Means and standard error for the seven cranial characters in their order of inclusion in the discriminant analysis -- For explanation of abbreviations see Appendix B.
T. arenarius T. pygargus T. gracilis (n=15) (n=9) (n=7)
Standard Standard Standard Character Mean Error Mean Error Mean Error
1. BAB 5. 70 0. 084 5.50 0.087 5. 30 0.117
2. GBAB 13 . 40 0. 109 13 . 80 0.137 13. 30 0.157
3. PPAL 10. 50 0. 149 10. 50 0.070 10.90 0. 146
4. GLS 35. 40 0.190 35,00 0.102 35. 30 0 .261
5. ALMl 2. 66 0. 081 2. 83 0.041 2.77 0. 047
6. LAB 9. 40 0.082 9.30 0.087 9. 40 0.084
7, LPPF 3,60 0.121 3 . 60 0.095 3, 50 0. 094 110 discriminant scores are used as a base value for a species.
Using other specimens or populations with these same coefficients will give a value which can be compared with those from the discriminant analysis. Specimens of unknown identity can then be identified.
The mean and standard error for seven cranial characters in their order of inclusion in the step-wise discriminant analysis are shown in Table 3. These seven measurements were selected by the program as the minimum necessary for discriminating the three species for two reasons: (1) reliability in the characters selected by the discriminant analysis is in proportion to the sample size of the smallest group; and (2) by step seven, the F-matrix in
the analysis showed significant differences between the three species (groups) at the 95 per cent level.
The length and direction of the cranial vectors
(Fig. 12) and comparisons of the means of these characters
(Table 3) indicate that T, gracilis differs from T. pygargus
as follows: Taterillus gracilis has a slightly longer skull
(GLS), longer auditory bullae (LAB), longer postpalatal
length (PPAL), shorter posterior palatine foramina (LPPF),
a narrower breadth across the bullae (GBAB), narrower bullae
(BAB), and a smaller first upper molar (ALMl). As discussed
in "Comparisons" under T. g, gracilis, its more gracile
skull is reflected by being longer and narrower. Ill
A combination of nonmensurable differences such as bullar size and shape (see "Comparisons" under T. g. gracilis and T. pygargus) and the mensurable characters found in the principal components and discriminant analyses, have allowed me to allocate all of the Senegal specimens of Taterillus to either T. gracilis or T. pygargus. Once these allocations were made it was possible to examine geographic variation in Senegal for the two taxa.
Geographic variation in T. cj_. gracilis (this sub- specific designation is as given in the previous chapter for
Senegal and western Mali T. gracilis) is shown in Fig, 13.
The Ndoulo population, according to the cranial vectors as generated by the discriminant analysis, has a longer and deeper skull (GLS--greatest length of skull; DC--depth of cranium) and a narrower breadth across the bullae (GBAB) than the other locality samples used in the analysis.
Examination of the standard statistics for all 24 cranial measurements in these population samples (Appendix B) shows some variation but little significant differences for most measurements. For the population samples excluding Ndoulo, the more northern ones have a longer rostrum (LR), longer palatal bridge (LPB), and a shorter molar toothrow (ALM).
From east to west in Senegal, the differences shown are slight but indicate that the eastern populations have a shorter skull (GLS), narrower bullae (BAB), and smaller depth of cranium (DC), Analysis of these trends by Fig. 13. Projection of the first two canonical variates for geographic samples of T. g. gracilis — Localities as follows (see Fig. 10 and Appendix D) W = western Senegal, #22; NW = northwestern Senegal, #23; Ndoulo, #24; E = eastern Senegal, #25; SW - southwestern Senegal, #27; NE = north eastern Senegal, #29. For explanation of cranial measurement abbrevia tions, see Appendix B. 3- LR
BAB 2- CHAL ALM rsiSO 1- HLU < e 5 ( N' Doulo | C£ < f pCD ) > 0- v_y < U (N W J z o z < u -2 o
-3-
-4 -3 -2 -1 0 12 3 4 CANONICAL VARIATE 1 (52.6%) Fig. 13, Projection of the first two canonical variates for geographic samples of T, g, gracilis, comparing all of the specimens available have not shown these differences to be consistent. In the following section, comparisons of the specimens assigned to T. g. gracilis with other T. gracilis population samples show that the cranial differences within T. g. gracilis are not as great as that between the recognized subspecies of
T. gracilis.
Geographic variation in T. pygargus is shown by the discriminant analysis (Fig. 14) of four Senegal popula tion samples. Although some of these samples are from localities 200 kilometers apart, little differences are found. Differences between the eastern and southwestern
Senegal samples are most noticeable. As indicated by the cranial vectors, eastern Senegal samples differ by having a larger first upper molar (ALM1) yet shorter molar toothrow (ALM), longer anterior palatine foramina (LAPF), longer palatal bridge (LPB), shorter length of skull
(GLS), and a smaller depth of cranium (DC). Values for these cranial measurements are in Appendix D.
Comparisons between north and south or east and west
Senegal show some overlap of individual specimens. Although
92,6 per cent of the variation in the 23 cranial measure ments (.zygomatic breadth excluded) are accounted for in the two canonical variates, individual variation appears to be as great as that reflected by the four samples. Distinc
tions based on nonmensurable and measurable cranial Fig. 14. Projection of the first two canonical variates for geographic samples of Senegal T. pygargus — For explanation of cranial measurement abbreviations, see Appendix B. Individual specimens and their localities (Fig. 10 and Appendix D), enclosed within the polygons, are: G = E Senegal, #25; L = W Senegal, #26; P = NW Senegal, #23; B = SW Senegal, #27. ALS11 LAPF
Seneaal
Senegal
Senegal
W Senegal
-3-2-101234 CANONICAL VARIATE 1 (81.4%) Fig. 14, Projection of the first two canonical variates for geographic samples of H' Senegal T, pygargus. H1 115 characters do not warrant subspecific recognition in T.
pygargus.
Geographic variation in Taterillus gracilis. Once
geographic variation in cranial morphology was determined for the most western populations of T. gracilis, a dis
criminant analysis on several populations representing the
remaining range of T. gracilis was made. This is shown in
Fig. 15. The population sample from Panisau, Nigeria (#8,
Fig. 10), is the type-locality of T. g. angelus. Kabwir,
Nigeria (#9, Fig. 10), is the type-locality of T. nigeriae
(=T. g. nigeriae). Analysis of the length and direction
of the cranial vectors (Fig. 15) shows that T, g. angelus
differs from T. g. nigeriae (Dada and Kabwir) by its having
longer nasals (LN), narrower rostrum (BR), and a narrower
braincase (BBC). The T. g. angelus sample differs from
those of T. g. meridionalis (Ghana, Togo, Dahomey, Upper
Volta, and Felele, Nigeria) mainly by its having longer
nasals (LN) and- narrower rostrum (BR) , The T. g.
meridionalis samples differ from those of T. cj. nigeriae
by their having broader but shorter bullae (BAB, LAB) and
a shorter length of skull (OCNL).
Geographic variation in T. gracilis is most
noticeable when comparing the Nigerian samples in Fig. 15.
As mentioned previously under "Ecological Observations"
under T. g, angelus and T, g, nigeriae in Chapter 4, Fig. 15. Projection of the first two canonical variates for population samples of T. g. angelus, T. g. meridionalis, and T. g. nigeriae -- Explanation of cranial measurement abbreviations are in Appendix B. The subspecies are represented by the following localities, including their locality numbers as shown in Fig, 10, T. g. angelus, #8--Panisau, Nigeria; T. g. meridionalis, #10—Felele, Nigeria; #'s 13 and 14—N and C Dahomey; #'s 15 and 17--S Ghana and S Dahomey; #'s 16 and 18--N Ghana and N Togo; #19—NW Upper Volta; #30—C Upper Volta; and T. g. nigeriae, #9—Kabwir, Nigeria; and #11—Dada, Nigeria. The circles represent the 90 per cent confidence limits around the mean for each sample, computed according to Seal (1964), 3- BAB
BR 2- %-^.BBC
LN' N Ghana-Togo LAB 0CSL fe? rv. NW U.V c*CNI ,1- SjGhana= Dahomey csi LU
Q£ 0- < Felele, C Dahomey > _1 < U -1-
< Pani sau, u Nigeria
-2- Dada, Nigeria
-3-
~T~" —r~ ~r -5 -4 -3 -2 -1 0 1 2 CANONICAL VARIATE 1 (35.1%) Fig. 15. Projection of the first two canonical variates for population samples of H T. g. angelus, T. g. meridionalis, and T. g. nigeriae. H 117 substrate differences resulting in faunal changes separate these two taxa. Substrate differences between the Felele,
Nigeria sample and other Nigerian samples do not account for their differences reflected by cranial morphology. The only probable reason for the differences between the Felele sample (as well as the Ghana, Togo, Dahomey, and Upper Volta samples) and those representing T. g. angelus and T. g. nigeriae, is apparently geographic isolation imposed by the influence of the Niger River.
The largest sample of a taxon in this analysis is of T. g. meridionalis (1697 of the 3315 total specimens examined; equals 51 per cent). Also, for many of the population samples, the sample size is very large (up to
230), With a resulting increase in the reliability of the statistical tests, geographic variation in this subspecies should be easier to detect. Geographic variation was analyzed using discriminant analysis (Fig. 16). Although I used specimens from the southern parts of the distribution in the analysis, several observations can be made. It is apparent that the more northern samples occupy an area of the figure distinct from the southern samples. It can be seen that 16 of the 24 cranial measurements used in the analysis are shown as vectors. By step 16 in the dis criminant analysis, the F-matrix does not indicate significant differences between the samples. Also, only
6,2 per cent of the total variation in the 24 measurements Fig, 16. Projection of the first two canonical variates for geographical samples of T. g. meridionalis — Locality numbers (see Fig. 10 and Appendix D) as Follows. N Dahomey, #13; C Dahomey, #14; S Dahomey, #15; N Togo, #16; SE Ghana, #17; NE Ghana, #18; NW Ghana, #32. Explanation of cranial character abbreviations are in Appendix B. N j Dahomey °N N W Ghana rocsi rsi N Togo LU
C IDahomey <
N E Ghana
U S E Ghana S Dahomey -2-
-3-
3 2 0 2 3 CANONICAL VARIATE 1 (38.3%)
Fig. 16. Projection of the first two canonical variates for geographical samples H of T. g. meridionalis. H is accounted for in the first two canonical variates. I have noticed that when significant differences are found between samples used in a multivariate analysis, the varia tion accounted for is at least 7 5 per cent. Therefore, the differences shown are more apparent than real. This is also noticeable when examining the individual measurement differences for the population samples used in the analysis in Appendix D. This analysis shows the homogeneity of T. g. meridionalis which has resulted in its distinctness from the other taxa of T. gracilis.
Relationships of all West African Taxa. The final analysis to determine the relationships among the West
African taxa was made using the NT-SYS multivariate pro grams. Population samples were chosen that represented the four species (T. arenarius, T. gracilis, T, lacustris, and T. pygargus) as well as samples of the four subspecies of T. gracilis (T. g. angelus, T. £. gracilis, T. g. meridionalis, and T, g. nigeriae). The NT-SYS routines used were the cluster analysis as reflected by the distance phenogram, principal components, nonmetric multidimensional scaling, and minimum spanning tree analyses.
The results given by the distance phenogram are shown in Fig. 17. The grouped localities of T. pygargus
(#'s 26, 27) are clustered with the Senegal samples of T. g, gracilis (#'s 22-25), which shows these two taxa to be Fig. 17. Distance phenogram of population sample means for West African species and subspecies of Taterillus -- The large values in the scale indicate greater taxonomic distance. Population numbers shown correspond to the following sample localities (see also Fig. 10 and Appendix D). #8, Panisau, Nigeria; #9, Kabwir, Nigeria; #10, Felele, Nigeria; #11, Dada, Nigeria; #12, Karaduwa, Nigeria; #13, NE Dahomey; #14, C Dahomey; #15, S Dahomey; #17, SE Ghana; #18, NE Ghana; #19, Barga, Upper Volta; #20, Fo, Upper Volta; #'s 22-25, Senegal; #'s 26-27, Senegal; #28, Tiguent, Mauritania; #31, NW Ghana; #32, SW Ghana; #33, Mora, Cameroon. 28 T. arenarius
26,27 T. pygargus
22-25 T. g. gracilis
8 ) T. g. angelus 12 J
9 T. g. nigerioe
19 "
32 —c 17 14
10 > T. g. meridiona 20
18
31
13
15 _
11 T. g. nigeriae
33 T. lacustris
1.85 1.60 1.35 1.10 0.85 0.60
Fig. 17. Distance phenogram of population sample means for West African species and subspecies of Taterillus. to o closer in phenetic distance than either are to the other taxa. Relationships and distinctness between the taxa of
T. gracilis are also shown (Fig. 17). The top group of clusters shows that the T. g. gracilis and T. g. angelus samples are distinct yet closer in phenetic distance than to the other T. gracilis taxa. The type-locality of T. g. nigeriae (Kabwir, Nigeria, #9), although clustered with
T. g. angelus and T. g. gracilis, is quite distinct.
Differences between T. g. meridionalis and the other subspecies of T. gracilis are even more evident. Samples of
T. g. meridionalis are in a major cluster distinct from the other taxa except for the T. g. nigeriae sample from Dada,
Nigeria (#11). Dada is just east of the Niger River with a possibility of intergradation across the river with animals assigned to T. g. meridionalis. The distinctness of the three Nigerian subspecies of T; gracilis is demonstrated by the placement and phenetic distances between localities
8, 9, 10, 11, and 12. The Felele Sample (#10), representing
T. g, meridionalis, is clearly distinct from the samples of
T, £. angelus and T. g, nigeriae, including the Dada (#11)
sample,
The distance phenogram (Fig. 17) also shows the distinctness of the two remaining West African species.
Taterillus arenarius, represented by the sample of cranial measurement means from its type-locality of Tiguent,
Mauritania (#28) , is shown as distinct from but a part of 122 the top cluster. The top cluster includes samples with known karyotyped specimens of T. pygargus and T. g. gracilis, as well as the eastern Nigeria subspecies of T. gracilis (T, g. angelus and T. g. nigeriae).
The cranial measurements which distinguish T. arenarius from T. pygargus were shown previously in Fig. 12 and Tables 2 and 3, as a part of a discriminant analysis.
The cranial measurement vectors indicate that T. arenarius has broader bullae (BAB), narrower breadth across the bullae
(GBAB), longer length of skull (GLS), and slightly shorter first upper molar (ALMl). The discriminant analysis also showed that T. arenarius differs from T. g, gracilis mainly by its auditory bullae dimensions; T. arenarius having longer
(LAB) and broader (BAB) bullae than T. g. gracilis.
It should be noted that three of the cranial characters selected by the discriminant analysis program
(Fig. 12, Table 3) reflect auditory bullae dimensions
(length of bullae, LAB; breadth of bullae, BAB; and greatest breadth across the bullae, GBAB). These measurements, as well as bullar shape, are the most reliable methods of distinguishing between T. arenarius, T. g. gracilis, and
!£• pygargus, especially in areas where two of the three species are sympatric (T. arenarius with T. pygargus and
T. g, gracilis with T, pygargus). It is also apparent that although T. arenarius has bullae which are slightly broader than T. pygargus (Table 3), the measurement of greatest 123 breadth across the bullae (GBAB) does not reflect this difference. Taterillus pygargus has a measurable longer distance across the bullae. Examination of the cranial measurement breadth of braincase (BBC) in Appendix D shows that the breadth across the bullae is influenced by the breadth of the braincase, which is greater in the T. pygargus sample used.
Taterillus lacustris, as shown by the distance phenogram (Fig, 17), is quite distinct from all of the other
West African taxa. The sample of T. lacustris is only a single age class 4 animal with known karyotype. It was not possible to use additional specimens as the other population samples that I have assigned to T. lacustris were either very small in number or composed of mainly young or very old individuals. Although the specimen is not from the type-locality, when compared to the holotype and type- series, it agreed in all morphological cranial characters which distinguish T. lacustris from the other West and East
African taxa. This single specimen is also the only age class 4 specimen with a known karyotype which has been assigned to T. lacustris.
Further clarification of the relationships of the
West African taxa is shown by the NT-SYS principal compo nents analysis (Fig, 18) and the cranial measurement loadings (Table 4). The same population samples were used as in the distance phenogram (Fig. 17), Analysis of the 9
22-25 26-27 12
I Fig. 18. Projection of the first three principal components of population sample means for West African species and subspecies of Taterillus — The first component accounts for 42.5 per cent of the variation in the 23 cranial measurements used, the second component, 13.2 per cent, and the third, 9,9 per cent. For sample numbers refer to Fig. 17 and Appendix D. 125
Table 4. Loadings of the cranial measurements on the first three components of West African Taterillus -- Explanation of abbreviations is given in Appendix B. The asterisks indicate those measurements which contribute the most to a given component.
Component
Measurement I II III
GLS .936* - .185 .150 OCNL . 961* -.042 .122 BAL .938* .162 .005 CBAL .953* -.005 . 072 BBC -.385 -.331 .630* IOC .167 .063 .471* BR -.375 .144 .486* GBAB .116 - .302 -.356 BrMlMl .142 .735* .360 DAL .856* -.109 -.063 PAL .781 .049 .371 PPAL .847 -.197 -.048 LAPF .712 . 449 -.038 LPB -.190 -.503 . 461* LAB .472 -.643* -.012 BAB .022 -.617* .072 LN . 897* -.187 -.104 LR .940* . 089 .005 FNL .916* -.197 .149 DC -.013 -.298 .489* LPPF -.588 -.588* . 298 ALM1 ,289 . 269 .511* ALM -.225 , 557* .382 126 relationships and differences between the samples in Fig.
18 requires the cranial measurement loadings which influ ence the sample separation in each of the three dimensions
(components). The first component, accounting for 42.5 per cent of the variation in the 23 cranial measurements used, is influenced by the major cranial length measurements which include greatest length of skull (GLS), occipitonasal length (OCNL), basilar length (BAL), condylobasilar length
(CBAL), length of diastema (DAL), length of nasals (LN), length of rostrum (LR), and frontonasal length (FNL). The second component, accounting for 13,2 per cent of the variation, is influenced by the length and width of the auditory bullae (LAB, BAB), as well as length of posterior palatine foreimina (LPPF) , breadth across the molars
(BrMlMl), and alveolar length of the molar tooth row (ALM).
The first three (LAB, BAB, LPPF) are shown as negative loadings which indicates that the samples where these measurements are large are located near the bottom of the figure. The length of the vertical lines represent the third component, which accounts for 9.9 per cent of the variation. The height of the lines are influenced by breadth of braincase (BBC), interorbital constriction
(IOC), breadth of rostrum (BR), length of palatal bridge
(LPB), depth of cranium (DC), and alveolar length of the first upper molar (ALMl). 127 Although the Senegal samples of T. g. gracilis and
T. pygargus were closely linked in the distance phenogram
(Fig. 17), the principal components analysis (Fig. 18,
Table 4) shows that these samples are closer to other population samples than to each other. The explanation for the differences in relationship between the samples when comparing the phenogram to the principal components is that representing the original multidimensional distance matrix
by a two-dimensional phenogram introduces some distortion in
the relationships between the samples. Adding a third dimension reduces the distortion in the spatial relation ship of the samples.
Differences between T. g. gracilis and T. pygargus
(samples 22-25 and 26, 27 respectively) are similar to those
previously discussed in conjunction with Fig. 11.
Taterillus cj_. gracilis is separated from T. pygargus by the
first component indicating that T, g. gracilis has a longer
skull. Also, some difference is shown by the second
component, This is due to the more inflated bullae (LAB,
BAB) and longer posterior palatine foramina (LPPF) of T.
pygargus.
Relationships and differences between the four
subspecies of T. gracilis are better shown by the principal
components than by the distance phenogram. In the first
component, T. g. gracilis (#22-25) is slightly longer
cranially than T. g. angelus (#'s 8 and 12), Taterillus 128 g. nigeriae (#'s 9 and 11) are slightly shorter cranially than T. g, angelus and longer than the samples of T. g. meridionalis (#'s 10, 13, 14, 15, 17, 18, 19, 20, 31, and
32).
The samples of T. g. angelus are further separated from those of T. g_. gracilis and T. g. nigeriae by the second component. Taterillus g. angelus has larger bullae, longer posterior palatine foramina, shorter molar tothrow, and a narrower breadth across the molars than does T, g. gracilis or T. g. nigeriae. Although the holotype of T. g. nigeriae is an extremely large individual, the placement of the samples of T- 2.- nigeriae shows that they are intermediate in cranial length between T. £, angelus and T, g. meridionalis. The principal components analysis also indicates, by the second component, that the more northern sample of T, g, nigeriae (Dada, #11) has longer and broader bullae and a smaller first upper molar than does the T. g. nigeriae sample from the type-locality of Kabwir (#9). It is probable that the bullar size and inflation is in response to more saltatorial habits in the sandier substrate of the Sudan Woodland (Dada) as opposed to the Guinea
Woodland habitat at Kabwir.
Possible intergradation between T. g. meridionalis and T. cj. nigeriae was previously mentioned in the discus sion of the phenogram (Fig. 17). The Dada sample of T. g. nigeriae was clustered with the samples assigned to T, 129 meridionalis. The principal components analysis shows that two-dimensional distortion accounted for the apparent close ness of the Dada sample to the T. g. meridionalis samples.
In the principal components analysis (Fig. 18), the Dada sample (#11) is closest to the sample from the type- locality of T. g. nigeriae (#9).
Comparisons between the sample from the type- locality of T. £. nigeriae (Kabwir, #9) and the type- locality of T. cj_. angelus (Panisau, #8) show differences as indicated by the cranial measurement loadings on the first, second, and third components (Table 4). Taterillus g. nigeriae has shorter nasals (LN), shorter length of skull (GLS), less inflated and smaller bullae (BAB, LAB), larger molars (ALM1, ALM), greater breadth of braincase
(BBC), greater depth of cranium (DC), and a broader rostrum
(BR). Differences in cranial size may not be apparent when comparing individual specimens of the two taxa but the measurable characters mentioned above reflect nonmensurable differences in that T. g. nigeriae has a very robust skull in comparison to T. g_. angelus.
The principal components analysis gives an indica tion of the measurable characters responsible for separating
T, g. meridionalis from the other three subspecies of T. gracilis. The T. cj_, meridionalis samples are in a group at the left of the figure (Fig. 18). This indicates that T. g. meridionalis has the shortest skull of the T. gracilis 130 subspecies. Geographic variation within the samples of
T. g. meridionalis is also evident. The measurements involved in the variation are those which load heavily on
the second and third components. Although the variation
noted is between population samples, no patterns are evident which give me an indication that the samples of
T. £. meridionalis could be further subdivided geographically or morphologically.
Taterillus arenarius (sample #28) differs from the
other West African taxa by both the first and second components. Taterillus arenarius is not only the largest
(skull length) of the West African species, but has the
largest and broadest bullae of the population samples used
in the analysis.
The principal components analysis (Fig. 18) gives a
better representation of the relationship of the T.
lacustris sample to the others than is shown by its great
phenetic distance from them in the phenogram (Fig. 17).
The first component places T. lacustris with the samples
with the largest or longest skulls. Those closest to T.
lacustris are Senegal samples of T. g. gracilis (#22-25),
the Panisau sample of T. g. angelus (#8), and the sample
of T. arenarius (#28). Taterillus lacustris is separated
from these samples by the second and third components. The
second component indicates that T. lacustris has shorter
and narrower bullae, shorter posterior palatine foramina, 131
longer molar toothrow, and a greater breadth across the molars. The third component shows T. lacustris to have a
narrower braincase, smaller interorbital constriction,
narrower rostrum, a small depth of cranium, and smaller first upper molar than the samples mentioned above.
The principal components analysis grouping (Fig. 18)
is similar to the taxa and samples with which T. pygargus
is clustered in the distance phenogram (Fig, 17). In the first component, T. pygargus is closest in skull size (see
Table 4) to the Karaduwa sample (#12) of T. g. angelus and
the Kabwir sample (#9) of T. g. nigeriae. It is separated
from these two samples by both the second and third compo
nents. As compared to the Karaduwa sample, T. pygargus differs by having a greater breadth across the molars
(BrMlMl), shorter and narrower bullae (LAB, BAB), a narrower
braincase (BBC), narrower rostrum (BR), a smaller depth of
skull (DC), and a shorter first upper molar (ALM1).
Taterillus pygargus differs from T. g. nigeriae (#9) by a
narrower breadth across the molars (BrMlMl), longer and
broader bullae (LAB, BAB), smaller molar toothrow (ALM),
narrower braincase, rostrum, and interorbital constriction
(.BBC, PR, IOC) , shorter depth of skull (DC) , and a smaller
first upper molar (ALMl),
Morphological relationships among the West African
taxa are shown in Fig, 19, This nonmetric multidimensional
scaling analysis (MDSCALE) was computed from the original 8
26,27
22-25,
I Fig. 19. Projection of three-dimensional MDSCALE of population sample means for West African species and subspecies of Taterillus -- Locality numbers refer to Fig. 17 and Appendix D. Lines connecting the samples are the minimum spanning tree. Stress is 0.163.
t—1 u> NJ 133 distance matrix. The figure also includes lines connecting
the population samples. These lines are the result of the
minimum spanning tree program and indicates those samples
between which the morphological (cranial) distance (phenetic
distance) is the smallest.
The T. arenarius sample (#28) is closest to the
Panisau sample (#8) of T. g. angelus. Taterillus arenarius
is separated from all of the other West African taxa mainly
by the first dimension.
Within the T. gracilis complex, the T. g. angelus
sample from Panisau (#8) is linked to the Karaduwa sample
(#12) of T. g. angelus, to the T. pygargus sample from
Senegal (#26, 27), and T. arenarius (#28). The Karaduwa
sample of T. g. angelus is linked to the Dada sample (#11)
of T. cj. nigoriae. Although the Senegal sample of T. g.
gracilis (#22-25) is connected to the sample of T. pygargus
(#26, 27), it is morphologically closer to the karyotyped
specimen of T. lacustris (#33). The spatial arrangement
in Fig. 19 shows that the T, £. gracilis sample is separate
from other samples of T, gracilis by the first, somewhat by
the second, and by the third dimension. An interesting
feature appears in the MDSCALE projection with T. g.
meridionalis. The lines connecting the samples radiate from
the Wulasi, Ghana sample (#17) to most of the other samples
of T. cj. mer idionalis, This sample (#17) is from the type-
locality of T, £. meridionalis and one of the two localities 134
in its range from which karyotypes are known. The sample
localities connected to Wulasi are either in the Guinea or
Sudan Woodlands of Ghana, Togo, Dahomey, Nigeria and Upper
Volta. Only one Sudan Woodland sample from NE Ghana (#18) and one from the Sahel Savanna of Upper Volta (#20), assigned to T, g_. meridionalis, are not directly connected
to Wulasi. The MDSCALE projection does not clearly separate
T. g. nigeriae from either T. g, meridionalis or T. g.
angelus. The Kabwir sample (#9) is removed from the other
samples to a certain extent, but connected by the minimum spanning tree to NE Ghana (#18). The Dada sample (#11) of
T. g. nigeriae is connected to both NE Ghana (T. g. meridionalis) and Karaduwa (#12, T. g. angelus).
The T. lacustris sample (#33) is linked to the
grouped Senegal samples (#22-25) of T. g. gracilis. It differs from that sample by its placement in the second and
third dimensions. The third dimension, as indicated by the
length of the vertical lines, readily distinguishes T.
lacustris from the other samples in the analysis.
Taterillus pygargus (#26, 27) is connected to both
T. g, angelus (#8) and T. g. gracilis (#22-25). Although
its position in three dimensions seems closer to Dada (#11,
T. g. nigeriae) and Karaduwa (#12, T. g. angelus), examina
tion of the original distance matrix which was used to
construct the minimum spanning tree indicates that the T.
pygargus sample is closer in taxonomic distance to T. g. 135 angelus (#8) and T. g. gracilis (#22-25). It differs from these two samples in the first dimension.
East Africa
The East African samples used for statistical analyses of species and intrapopulation variation in cranial morphology are shown by their locality numbers in Fig. 20.
The exact localities of these samples are given in Appendix
D along with standard statistics for the 24 cranial measure ments used in the analyses,
Recognition and differentiation of the East African taxa presented problems that differed from those for West
Africa. Nineteen taxa have been described from East Africa including one each from Ethiopia, Somalia, Uganda, and Zaire, seven from Kenya, and eight from Sudan. The latest listing of the species of Taterillus by Petter (1975) included one from Zaire (T. congicus), one from Sudan (T. gyas), three from Kenya (T. lowei lowei and T. !.• harringtoni from
Ethiopia; T. tenebricus; and T. nubilus nubilus and T. n. illustris), and one from Uganda (T. emini including T. e. anthonyi and T, e, butleri from Sudan; T. e, osgoodi from
Kenya; and T, e. zammarani from Somalia). Petter also included T. perluteus from the Sudan as a subspecies of T. lacustris. Problems concerning the use of and priority of the above names was discussed in Chapters 3 and 4. o® ®
SUDAN
ETHIOPIA CENTRAL AFRICAN
REPL BLIC
© RWANDA BURUNDI TANZANIA K i lometers I Fig. 20. Locality numbers used in statistical analyses of East African Taterillus• u> 137 Multivariate statistical analyses resulted in the four species I recognize as occurring in East Africa (T. congicus, T. emini, T. harringtoni, and T. lacus tris). Three steps were required to interpret the relationships between the East African taxa. First, cranial comparisons between the two taxa described from Zaire and Uganda (T. congicus and T. emini) ; second, comparisons of; specimens representing the Kenya taxa and the Sudan taxa to these first two taxa; and third, once morphological groups attributable to species status were recognized, to allocate the described taxa to them. Once these problems were analyzed, relationships and comparisons of the East African, species were possible. Taterillus congicus and Taterillus emini. The multivariate ordination technique used for finding possible differences between specimens from Uganda and Zaire was a principal components analysis (Fig. 21 and Table 5) . Individual specimens from Poko, Zaire (#1, Fig. 20, equals T. congicus) and selected specimens from various Uganda localities (T. emini), The first component, comprising 51.9 per cent of the variation in the 24 cranial measurements used, includes most of the major size measurements. The second component, accounting for 15,0 per cent of the variation, includes breadth measurements, bullar size, and molar toothrow size. Fig. 21. Two-dimensional projections of the first three principal components of Zaire and Uganda Taterillus — Individual specimens plotted are identified as follows: E = T. emini, C = T. congicus, and H = T. harringtoni. Numbers 1-10 are specimens from Zaire, and numbers 11-25 are from Uganda. 138 '11 '17 '24 '14 13 '25 E *23 E '19 »22 E ®20 21 '18 •C 7 C H C •C *5 *12 C 9 *4 C • '10 •16 '15 Fig. 21. Two-dimensional projections of the first three principal components of Zaire and Uganda Taterillus. 139 Table 5. Loadings of the cranial measurements on the first three components for individual specimens of Zaire and Uganda Taterillus -- Explanation of abbreviations are given in Appendix B. The asterisks indicate those measurements which contribute the most to a given component. Component Measurement I II III GLS .901* -.152 -.041 OCNL .940* -.042 .187 BAL .869* . 185 .354 ZB .452 - , 430* -.018 CBAL . 858* . 287 .347 BBC .299 -.659* .135 IOC .269 -.576* -.472* BR .061 . 273 -.540* GBAB - . 083 - . 484* .300 BrMlMl .085 . 511* -.167 PAL .797* .328 -.155 PPAL .389 -.167 .719* LAB .053 -. 470* .271 BAB -.124 -.470* .411* LN .779* -.146 -.310 FNL .799* - . 084 -.163 DC .449 .491* .327 ALM1 .077 .701* . 250 ALM . 008 , 652* . 221 140 Component three, accounting for 9.5 per cent of the varia tion, is influenced by postpalatal length (PPAL), breadth of auditory bulla (BAB), and negatively by interorbital constriction (IOC), length of nasals (LN), and frontonasal length (FNL), It is the second component which separated the specimens into three distinct morphological groups. It shows that the Zaire specimens representing T. congicus have a broader skull (BBC), slightly larger bullae (LAB, BAB), and smaller molars (ALM1, ALM) than do the specimens from Uganda representing T. emini. The specimen (H-15) which is not placed either with T. congicus or T. emini has been subsequently identified as a T. harringtoni which is readily distinguishable from both of the above species. Geographical Relationships. Relationships of the specimens used in the previous principal components analysis to specimens from Kenya and the Sudan were made using a discriminant analysis. The three groups used for compara tive purposes were: specimens of T. congicus from its type- locality at Poko, Zaire (#1, Fig. 20); selected Uganda specimens representing T. emini; and, grouped Kenya speci mens representing T. harringtoni (#'s 3, 4, 5; Fig. 20). In addition, specimens of unknown identity and specimens from localities with a small sample size were included in the analysis as unknowns. The program then allocated these unknown specimens to one of the above three groups or could 141 have indicated that they differed from them. For ease of analysis, I have figured only the three known groups in Fig. 22. Separation of the three groups (species) was achieved using only five cranial characters. In this case they were not selected because of sample size but rather the F-matrix showed significant differences between the groups by step five in the discriminant analysis. Dis criminant coefficients for separating the groups and means of the five characters are given in Tables 6 and 7. The cranial measurement vectors shown in Fig. 22 represent the relative contributions of the five measurements in separating the groups. These vectors are enhanced by observing the means of the variables (Table 7) shown in their order of inclusion in the discriminant analysis. The discriminant analysis shows that the group assigned to T. congicus differs from the Uganda T. emini by their longer posterior palatine foramina (LPPF), smaller first upper molars (ALM1) , and slightly broader bullae (BAB), The large first upper molars of T. emini best distinguish it from T. congicus. The mensurably broader bullae of T. congicus do not reflect shape. Actually, the bullae of T, emini are more inflated anteriorly, antero lateral^ and ventrally. Inspection of the cranial vectors (Fig. 22) and the means for these measurements (Table 7) indicates that each ALMl LPB T. EMINI LAPF BAB LU LPPF H H u T. HARRINGTONI o T. CONGICUS u CC -1- -2- H -4 -3 -2 -1 0 1 2 3 4 5 6 CANONICAL VARIATE 1 (87.4%) Fig. 22, Projection of the first two canonical variates for three species of East African Taterillus, w 143 Table 6. Discriminant coefficients for the five cranial characters used to separate the three East African species of Taterillus -- C = T. congicus from Poko, Zaire; E = T. emini from Uganda; H = T. harringtoni from Kenya. For explanation of cranial measurement abbreviations see text and Appendix B. Character C vs . E C vs. H E vs. H LAPF 4.06 20.87 16.81 LPB .98 12.20 11.22 ALM1 -11.35 2. 12 13 .47 LPPF 7.40 4. 55 -2.85 1 — 0 I BAB 4. 62 -5.79 1 144 Table 7. Means and standard error for the five cranial characters in their order of inclusion in the discriminant analysis of East African Taterillus -- For explanation of abbreviations see text and Appendix B. T. congicus T. harringtoni T. emini (n=17) (n=20) (n=10) Standard Standard Standard Character Mean Error Mean Error Mean Error LAPF 6. 84 0.097 5. 63 0. 042 6.55 0.083 LPB 7. 48 0. 103 6 .97 0.052 7.38 0.072 ALM1 2. 86 0. 042 2. 56 0. 037 3. 05 0.055 LPPF 4 . 14 0.054 3 . 47 0.066 3 .73 0.080 BAB 5. 68 0. 067 5.72 0.059 5.35 0.058 145 of the five variables will distinguish the specimens allocated to T. harringtoni from Kenya from both the T. congicus and T. emini groups. The characters which best distinguish T. harringtoni from the other two species are breadth of bullae (BAB), alveolar length of the first upper molar (ALM1), and length of the anterior palatine foramina (LAPP). Although the bullae of T. harringtoni are shorter than those of T, congicus and T. emini (see Appendix D), they are slightly broader. This indicates the greater inflation of the bullae as noted as characteristic of T. harringtoni in Chapter 4. Although the holotype of T. emini is a young animal (age class 3), my comparisons between it and the Uganda specimens used in the analysis have shown them to be conspecific. Also, my comparisons of the holotypes of the described taxa from Kenya, Somalia, and Ethiopia have shown that they are morphologically the same taxon. The unknown specimens used in the discriminant analysis were allocated by the program to one of the three species groups used. These allocations were in part responsible for the synonomies in Chapter 4. East African Species Relationships. The final analysis of the relationships between the species status of the East African taxa was made by using the NT-SYS pro grams. Means of 23 cranial measurements (zygomatic breadth 146 omitted) for population samples were used. These samples were from Poko, Zaire (#1, T. congicus), Uganda (#2, T. emini), Faradje, Zaire (#34, T. congicus), Chad (#35, T. congicus), Sir, Cameroon (#6, T. congicus), north, central, and southern Kenya (#'s 3, 4, 5, T. harringtoni) , the karyotyped specimen of T. lacustris (#33), and holotypes of T. congicus and six of the seven taxa described from Kenya. The distance phenogram (Fig. 23) has two major clusters. The top cluster contains population samples assigned to T. congicus, the holotype of T. congicus, and the Uganda sample of T. emini. The holotype of T. congicus is placed between the 2N=54 karyotyped specimens from Chad and the Faradje, Zaire sample. My allocation of the Sir, Cameroon sample to T. congicus, rather than to T. emini was because of this analysis by its being clustered with the Poko, Zaire sample (type-locality of T. congicus). The placement of the Uganda sample of T. emini with those of T. congicus in the top cluster demonstrates the cranial similarities of these two species. It can be seen that T. congicus and T. emini are distinct from the other two species occurring in East Africa (T. harringtoni and T, lacustris). The second major cluster links the three samples from north, central, and southern Kenya with the six Kenya holotypes, and the karyotyped specimen of T. lacustris. The Fig. 23. Distance phenogram of population sample means and holotypes of East African Taterillus — Lower case identifiers indicate the population sample localities as shown in Fig. 20 and Appendix D. Species allocations are: 1, 6, 34, 35, T. congicus; 2, T. emini; 3, 4, 5, T. harringtoni. Upper case scientific names are the holotypes and their names as originally described. The cophenetic correlation is 0.847. — 1 Poko, Zaire — 6 Sir, Cameroon '2 Uganda 34 Faradje, Zaire T. CONGICUS 35 Chad [-3 N Kenya 1—4 Q Kenya ' 5 S Kenya r- T. N. MENEGHETTII T. LOWEI -T. N. ILLUSTRIS T. N. NUBILUS -33 T. LACUSTRIS -T. MELANOPS T. OSGOODI l l l l I i 1.68 1.43 1.18 0.93 0.68 0.43 Fig. 23. Distance phenogram of population sample means and holotypes of East African Taterillus. 148 Kenya specimens and holotypes, as well as the described taxa from Ethiopia and Somalia have been allocated to T. harringtoni. Holotypes of the remaining East African taxa were omitted from the analysis because of their being young or very old individuals. The distance phenogram demon strates the distinctness of T. harringtoni. The dis criminant analysis discussed previously and the NT-SYS programs included specimens with a karyotype of 2N=44, and are a part of the T. harringtoni complex. The distance phenogram (Fig. 23) placed the T. lacustris specimen with those of T. harringtoni, although its taxonomic distance from the T. n. nubilus holotype is moderately great. That this is another case in which the two-dimensional distortion of the original multidimensional distance matrix is demonstrated by the following principal components and MDSCALE analyses. The principal components analysis (Fig. 24, Table 8) and the MDSCALE projection (Fig. 25) show that the samples of both T. congicus and T. emini are separated from the other samples and holotypes by the first component or dimension, Loadings of the 23 cranial measurements on the three components of Fig. 24 (Table 8) indicate that the first component is influenced by most of the cranial size measurements, Both T. congicus and T. emini are much larger cranially than the other taxa, It can be seen in both the principal components and MDSCALE analyses that the T. congicus T. lowei T. n. meneghettii -r i T. osqoodi I. melanops —r I Fig. 24. Projection of the first three principal components of sample means and holotypes of East African Taterillus — Population sample numbers refer to Figs. 20 and 23 and Appendix D. The first component accounts for 58.8 per cent of the variation in the 23 cranial measurements used, the second component, 18.4 per cent, and the third, 8.2 per cent. 150 Table 8. Loadings of the cranial measurements on the first three components of populations and holotypes of East African Taterillus -- For explanation of cranial measurement abbreviations see text and Appendix B. An asterisk indicates those measure ments which contribute the most to a given component. Component Measurement I II III GLS . 947* -.221 -.152 OCNL .937* -. 204 -.218 BAL .958* - . 074 -.235 CBAL . 936* -.234 -.156 BBC .356 -.834* .246 IOC -.166 - .286 .671* BR . 874* .143 . 208 GBAB .108 -.889* -.383 BrMlMl .851* .277 . 076 DAL .953* - . 001 -.155 PAL .964* .071 .009 PPAL . 520 -.749* -.175 LAPF .816 .436 . 182 LPB .849* -.242 .079 LAB -.107 -.929* . 212 BAB -.211 -.381 .606* LN .868* . 293 , 154 LR , 865* . 227 -.275 FNL .868* -.318 -.129 DC .845* -.080 . 403 LPPF .773 -.094 .378 ALM1 .781 .416 . 057 ALM .761 . 404 .320 T. n. i11ustri s T. lowei T. n. meneghettii T. congicus G— T. osgoodi T. melanops I Fig. 25. Projection of three-dimensional MDSCALE of population sample means and holotypes of East African Taterillus — Locality sample numbers and identifiers refer to Figs. 20 and 23 and Appendix D. Lines connecting the samples are the minimum spanning tree. Stress is 0.097. 152 T. emini sample is in the grouping with the same samples as in the phenogram. The minimum spanning tree, computed from the original distance matrix, connects the Uganda sample to Sir, Cameroon. It indicates that although the T. congicus and T. emini samples all belong in the same cluster, that the T. emini sample is closest (in taxonomic distance) to the sample which is furthest from it geographically. Due to the use of other distinctive samples and species, the NT-SYS programs did not clearly separate T. congicus from T. emini as was achieved previously by a separate principal components analysis and discriminant analysis. The principal components and MDSCALE analyses show the relationships of the samples and holotypes I have included in T. harringtoni to those of T. congicus and T. emini. Taterillus harringtoni (Fig. 24) is separated from the other two species by the first component. Within the large group of T, harringtoni, which includes T. lacustris, the second and third components separate T. harringtoni from T. lacustris. The second component indicates (Table 8) that T. harringtoni has a broader braincase (BBC), greater breadth across the bullae (GBAB), a longer post- palatal length (PPAL), and longer and broader auditory bullae (LAB, BAB). The third component also readily separates these two species as indicated by the length of the vertical lines. Taterillus harringtoni has a greater 153 interorbital constriction (IOC) and broader bullae (BAB) than does T. lacustris. The MDSCALE projection (Fig. 25) further separates T. harringtoni and T. lacustris. In this case, T. lacustris (#33) is quite removed from those of T. harringtoni. Although T. lacus tris is connected by the minimum spanning tree to the holotype of T. n. nubilus, it is separated from it and the other samples of T. harringtoni by the first and third dimensions. Total Distribution Analysis of the cranial morphology relationships of population samples throughout the range of Taterillus was made using the NT-SYS programs. Since these analyses used sample means, the single specimen of T. lacustris was omitted. The other samples used include most of the same ones as discussed previously concerning the West and East | African taxa. The distance phenogram (Fig. 26) has four major clusters which indicate patterns of cranial morphology in I the genus Taterillus. The top cluster shows the sample of T. congicus from its type-locality linked with the Sir, Cameroon sample of T, congicus. The top cluster also separates the T. emini sample from those of T. congicus. Also, the placement of the Jebel Marra, Sudan sample between those of T. congicus and T. emini confirms Fig. 26. Distance phenogram of population sample means for geographic samples of Taterillus -- Sample numbers and/or scientific names refer to the following: 1, T. congicus, Poko, Zaire; 2, T. emini, Uganda; 3, 4, 5, T. harringtoni, north, central, and southern Kenya; 6, T. congicus, Sir, Cameroon; 7, T. congicus and T. emini, Jebel Marra, Sudan; 8, T. g. angelus, Panisau, Nigeria; 9, T. g. nigeriae, Kabwir, Nigeria; 10, T. g. meridionalis, Felele, Nigeria; 11, T. g. nigeriae, Dada, Nigeria; 12, T. g. angelus, Karaduwa, Nigeria; 13-21, T. g. meridionalis; 22-25, T. g. gracilis, Senegal; 26, 27, T. pygargus, Senegal; 28, T. arenarius, Tiguent, Mauritania. Large values indicate greater taxonomic distance. The cophenetic correlation is 0.918. 154 1 T. CONGICUS 6 Cameroon 7 J. Marra, Sudan 2 T. EMINI 3^ 4 >T. HARRINGTON! 5. 8 T. G. ANGELUS 12 N Nigeria 28 T. ARENARIUS 9 T. G. NIGERIAE 10^ S Nigeria 25 T. G. GRACILIS 22 •c 26 T. PYGARGUS 24 T. G. GRACILIS 27 T. PYGARGUS 23 T. G. GRACILIS 21 N Upper Volta 1 4 C Dahomey 20 W Upper Volta 19 NW Upper Volta 18 N Ghana —c 17 SE Ghana I5 S Dahomey 16 N 1ogo 13 NE Dahomey HZ II NW N igeria L 2.17 1.82 1.47 1.12 0.77 0.42 Fig. 26, Distance phenogram of population sample means for geographic samples of Taterillus• 155 my analysis using nonmensurable characters that both species are represented in the sample. It can be seen that T. congicus and T. emini are quite distinct from other popula tion samples and taxa of Taterillus. The second major cluster (Fig. 26) includes only the three Kenya samples of T. harringtoni. Although they are closer in taxonomic distance (cranially more similar) to the bottom clusters than to the top cluster (T. congicus and T. emini), they are very distinct from any of the samples and taxa used in the analysis. The bottom two clusters are not as distinctive as the top two. Included in the third cluster are two samples of T, g. angelus (#'s 8 and 12), T. arenarius (#28), and the type-locality of T. g. nigeriae (#9). The bottom cluster includes one sample of T. g, nigeriae (#11), Senegal samples of both T. g. gracilis and T„ pygargus (#'s 22-25 and 26, 27 respectively), and all of the samples used of T, g, meridionalis. Although some of the species of Taterillus are separated by the phenogram, this analysis mainly shows three major groups divisible by cranial morphology. These three groups include: first--T. congicus and T. emini; second—T. harringtoni; and third--all of the West African population samples and taxa. The principal components analysis (Fig. 27 and Table 9) further point out the cranial relationships and differences between the three major groups. Table 9 Fig. 27. Projection of the first three principal components of population means for geographic samples of Taterillus -- For population sample numbers and identifiers refer to Fig. 26, Figs. 10 and 20, and Appendix D. The first component accounts for 57.2 per cent of the variation in the 24 cranial measurements, the second component, 15.6 per cent, and the third, 8,8 per cent. 5 O o 9 7 o 21 o19 o 12 17 n i 18 16 o o 0 26 O 6 ^0 O 27 25io n 20 0 28 d uo 6° 00 22 9 14 P 9 24 T 23 a13 o 9 ? I Fig. 27. Principal components of geographic samples of Taterillus. U1 Table 9. Loadings of the cranial measurements on the first three components of population means of geographic samples of Taterillus -- Explanation of cranial measurement abbreviations are in Appendix B. Asterisks indicate the measurements which con tribute the most to a given component. Component Measurement I II III GLS .944* . 119 - . 229 OCNL .956* . 129 -.182 BAL . 971* . 061 .019 ZB . 819 .109 . 229 CBAL . 956* .110 -.009 BBC . 579 . 283 .681* IOC .606 . 344 .304 BR .741 -.489 . 250 GBAB .459 . 696* .350 BrMlMl . 710 -.473 . 297 DAL .857* -.053 -.177 PAL .913* -.278 - . 031 PPAL .795 .487 -.017 LAPP .943* -.209 -.009 LPB . 589 -.496 .158 LAB .650 .683* .017 BAB .296 . 627* . 051 LN .766 .045 -.576* LR .696 . 078 -.661* PNL .853* . 215 -.406 DC . 876* - . 064 .350 LPPF . 502 -.411 .092 ALM1 .487 -.710* -.100 ALM .640 -.588* . 129 158 indicates which of the 24 cranial measurements are responsible for the separation of these groups. The first component is influenced by greatest length of skull (GLS), occipitonasai length (OCNL), basilar length (BAL), condylobasilar length (CBAL), length of diastema (DAL), palatilar length (PAL), length of anterior palatine foramina (LAPF), frontonasal length (FNL), and depth of cranium (DC). It separates T. congicus and T. emini from the rest of the samples and taxa by the cranial size measurements. The second component is influenced by greatest breadth across the bullae (GBAB), length of bullae (LAB), breadth of bullae (BAB), and negatively, length of the first upper molar (ALMl) and length of molar toothrow (ALM). It is the second component that separates the three T, harringtoni samples (#'s 3, 4, 5) from the West African group of T, arenarius, T. gracilis, and T. pygargus. The measurements responsible show that T, harringtoni has a greater breadth across the bullae, longer and broader bullae, and smaller molars than the West African taxa. Within the large group of cranially similar West African taxa, some differences are observable which were discussed previously in the section on West Africa. The principal components analysis shows T. arenarius (#28) to have a slightly larger skull (first component), longer nasals and rostrum, and relatively narrower breadth of 159 braincase (third component). Taterillus arenarius was grouped in the phenogram (Pig. 26) with two samples of T. g. angelus (#'s 8 and 12) and one sample of T. g. nigeriae (#9). These same four samples are also slightly- separated from the other West African samples in the principal components analysis due to slightly larger skulls. The T. g. angelus samples are separated from the T. g. nigeriae sample by the second component indicating that T. g. angelus has a greater breadth across the bullae, longer and broader bullae, and smaller molars than T. g. nigeriae. The MDSCALE analysis (Fig. 28), including the minimum spanning tree, shows relationships similar to those mentioned previously for the distance phenogram (Fig, 26) and the principal components analysis (Fig. 27, Table 9). The three major groups are distinct. The minimum spanning tree lines, indicating taxonomic distance to the closest sample, are mainly confined within each of the three groups. Differences within the large West African group are not evident except for the slight separation by the third dimension of T. arenarius (tt'2'8; . As will be seen in the following sections on comparative karyology and distribu tion, the relationships and differences among the three major groups indicated by cranial morphology are consistent. I Fig, 28. Projection of three-dimensional MDSCALE of population means for geographic samples of Taterillus -- For population sample numbers and identifiers refer to Fig. 26, Figs. 10 and 20, and Appendix D. Stress is .0, 077. 161 Comparative Karyology Taterillus chromosome information from many localities within the range of the genus has appeared in the literature (references are included below under their appropriate taxa). Karyotypes are available for all species of Taterillus except T. emini. Examination of the sex-chromosomes in Taterillus shows that three species (T. congicus, T. harringtoni, T. lacustris) have the classic XX/XY system, but two (T. gracilis, T, pygargus) have an XX/XY^Y2 system. The sex- chromosome system of T. arenarius is not known because only a female has been karyotyped. Taterillus pygargus. This species is characterized by a diploid number of 22/23 and a fundamental number of from 36 to 40 (Fig. 29, Table 10). Chromosomes of T, pygargus are known from Fete-Ole, Bandia, and Saboya, Senegal (Matthey, 1969j Petter et al., 1972). Its karyotype includes one pair of large sub-telocentric, one pair of medium-sized to small metacentric and submetacentric autosomes. The X-chromosome is a large submetacentric. The Y-chromosomes differ in size; the Y^ is a small metacentric and the Y2 is a medium-sized submetacentric chromosome. The XX/XY-^J^ system accounts for the diploid number differ ence between the males and females. Chromosomal poly morphism, due to pericentric inversions in four of the autosomes, gives different fundamental numbers within the Fig. 29. Karyotypes of six species of Taterillus -- The taxa are as follows: 1, Taterillus pygargus; 2, Taterillus lacustris; 3, Taterillus arenarius; 4, Taterillus gracilis gracilis; 5, Taterillus gracilis meridionalis; 6, Taterillus harringtoni; 7, Taterillus congicus from Central African Republic; 8, Taterillus congicus from Chad. The autosomes are divided into the following four morphological groups: A, large submetacentric; B, medium-sized to small metacentric and sub metacentric; C, sub-telocentric; D, acrocentric. 162 B D xy,y2 AK XX Ax XX X* xx AX x A X * xy n kx xx XX XX M /\n A/1 /l/i ft XX An H(\ An /)n nn XK x x xx A/l A XX XX xx nA o/i /9/i m xx Qft A/-1 Ar> AA A/1 A A X ^r\ (\/\ rt/) OA XX A)( XX AA /\A M HA aa X A AA -A (10 /\0 AA AA Aa y V1/ M AX AK M xx. AX Xx FLF* ^A /\ r\f\ AA (l X X X XX /\n An ah xy xx xx AA /A A Af\ AA /1/1 XX X* **• X AA AA A Af\ ^ /W AA AA /^A /\a f\(] r\f\ AA XX XY XX XX XX XV AA AA /\f\ A(\ A 1x ^f) A A A A A A A/\ /I A AA A /"* Fig, 29. Karyotypes of six species of Taterillus. 163 Table 10. Somatic chromosome number and morphological types of six species and two subspecies of Taterillus — 2N = diploid number, FN = fundamental number, M = metacentric, SM = submetacentric (including sub-telocentric), A = acrocentric. C.A.R. is Central African Republic. Autosome Pairs Species 2N FN M/SM A X Y1 Y2 T. pygargus 22/23 40 10 0 Lrg Sml Med SM M SM T. lacustris 28 44 9 4 Lrg Sml SM SM T. arenarius 30 36 4 10 Lrg •p SM T. g. gracilis 36/37 44 5 12 Lrg Sml Med SM SM SM T. g. meridionalis 36/37 42 4 13 Lrg Sml Med SM SM SM T. harringtoni 44 62 10 11 Lrg Sml SM SM T. congicus (C.A.R,) 54 64 6 20 Lrg Sml SM SM T. congicus (Chad) 54 66 7 19 Lrg Sml SM SM 164 same individual and between individuals in the same popula tion (Matthey and Jotterand, 1972). These inversions result in from one to four of the metacentrics becoming acrocen trics. Individual cells can have a fundamental number from 36 to 40. The most common karyotype has all biarmed autosomes, giving a fundamental number, exclusive of the sex-chromosomes of 40. Taterillus lacustris. Chromosomes of this species are known only from a few specimens from Mora, Cameroon (locality 33, Figs. 10 and 20). Its diploid number, as reported by Tranier et al. (197 4) is 28 and the fundamental number is 44. It has the classic XX/XY system. Its karyotype includes two pairs of large submetacentric, seven pairs of medium-sized to small metacentric and submeta centric, one pair of large acrocentric, and three pairs of small acrocentric autosomes. The X-chromosome is a large submetacentric and the Y-chromosome a small submetacentric. Taterillus arenarius. The chromosomes of T. arenarius are only known from the female reported under the name T, nigeriae by Matthey (1969). Its diploid number is 30 and the fundamental number is 36 (Fig, 29, Table 10) , including one pair of large sub-telocentric, three pairs of medium-sized to small metacentric, and ten pairs of medium- sized acrocentric autosomes. The X-chromosome is a large submetacentric; the Y-chromosome is unknown. 165 Taterillus gracilis gracilis. The diploid number of this taxon from Senegal (Matthey and Jotterand, 197 2; Petter et al., 1972) is 36/37 and the fundamental number is 44 (Fig. 29, Table 10). It has the XX/XY-^Y^ system, Its karyotype includes one pair of large sub-telocentric, four pairs of medium-sized to small submetacentric, one pair of large acrocentric, and eleven pairs of medium-sized to small acrocentric autosomes. The X-chromosome is a large sub metacentric, while the Y-chromosomes differ in size. The Y^ is a small metacentric and the a medium-sized sub metacentric, accounting for the diploid number differing in the males and females, Taterillus gracilis meridionalis. Karyotyped specimens of this taxon are from Bobo Dioulasso, Upper Volta (Matthey and Petter, 1970), Wulasi, Ghana (Robbins, 1972), and northern Ivory Coast (Petter, 1972). It is characterized by a diploid number of 36/37, a fundamental number of 42, and an XX/XY-^2 system (Fig. 29, Table 10) including one pair of large sub-telocentric, three pairs of medium-sized to small submetacentric, one pair of large acrocentric, and 12 pairs of medium-sized to small acro centric autosomes. The X-chromosome is a large sub metacentric and the Y-chromosomes differ in size. The Y^ is a small metacentric and the Y^ a medium-sized submeta centric, giving the different diploid number for males and females. 166 Taterillus harringtoni. Karyotyped specimens of this species are from the Omo Valley, Ethiopia (Matthey, 1969; Matthey and Petter, 1970), Afmadu, Somalia (Genest and Petter, 1973), Gordil, Central African Republic (Genest and Petter, 1973), and Samburu Game Lodge, 12 mi N Archer's Post, 59 mi N Kapedo, and Voi, Kenya (Robbins, 1974a). This has a diploid number of 44 and a fundamental number of 62, with the classic XX/XY system (Fig. 29, Table 10), including two pairs of large submetacentric, eight pairs of medium- sized to small metacentric and submetacentric, and 11 pairs of medium-sized to small acrocentric autosomes. The X- chromosome is a large submetacentric and the Y-chromosome a small submetacentric. Taterillus congicu&. The chromosomes of this species are known from specimens from Bangui, Maboke, Gordil, and Lac Mamoun, Central African Republic (Matthey and Petter, 1970; Genest and Petter, 1973), and Fort-Lamy, Deli, and Moundou, Chad (Tranier et al,, 1974) f It has a diploid number of 54, In Central African Republic the fundamental number is 64, while a pericentric inversion in Chad specimens gives a fundamental number of 66 (Fig, 29, Table 10). Its karyotype (Central African Republic) includes six pairs of medium-sized to small metacentric and submetacentric and 20 pairs of medium-sized to small acrocentric autosomes. The X-chromosome is a large submetacentric and the Y-chromosome a small submetacentric. 167 The karyotyped specimens from Chad differ slightly from those described above as a result of a pericentric inver sion. The inversion in one of the pairs of acrocentric autosomes has resulted in seven pairs of medium-sized to small metacentric and submetacentric and 19 pairs of medium-sized to small acrocentric autosomes. Conclusions. Based on diploid and fundamental number, plus chromosome morphology, the six karyotyped species of Taterillus can be divided in to three groups. These are: (1) 2N=54 (T. congicus); (2) 2N=44 (T. harringtoni), and (3) the West African taxa with 2N of 22/23, 28, 30, and 36/37 (T. pygargus, T. lacustris, T, arenarius, and T. gracilis). Differences in chromosome morphology between T. congicus and T. harringtoni (Fig. 29) are as great as those found for their cranial morphology. Using the Tr congicus karyotype as the most probable ancestor for the other taxa in Taterillus, derivation of the remaining five species involves many chromosomal rearrangements, The derivation of the 2N=44, FN=62, XX/XY karyotype of T, harringtoni from T. congicus (2N=54, FN=64, XX/XY) is possible. Three fusions would result in a hypothetical intermediate with 2N=48, FN=62, XX/XY. From this 2N=4 8 intermediate, two additional Robertsonian fusions would give the 2N=44 karyotype of T, harringtoni. The derivation of T, harringtoni from T. 168 congicus would need a minimum of five major chromosomal rearrangements (Fig. 30). On skull morphology, the four West African species appear to be closely related. I had stated previously (Robbins, 1974b) that their different karyotypes did not affirm a common derivation. However, reanalysis of their karyotypic similarities indicates a possible common ancestor. Nevertheless, derivation of a karyotype ancestral to the West African species from either T. congicus or T. harringtoni would require at least twelve chromosomal rearrangements. Differences in the sex-chromosome systems of the West African taxa indicate that the most logical intermediate would have a 2N of 38, FN of 44, and the classic XX/XY system. Two possible pathways exist to form the 2N=38 form from the ancestral 2N=54 form of T. congicus. First, from 2N=54, eight Robertsonian fusions and four pericentric inversions would give the 2N=3 8 form. This pathway would require 12 steps. Second, from the 2N=48 intermediate dis cussed previously, six fusions and six inversions would also give the 2N=38 form. This pathway would require twelve steps from the 2N=4 8 form plus three steps from 2N=54 to 2N=48, giving a total of 15 steps (Fig. 30). It would require at least 17 steps to derive the 2N=38 form from the 2N=44 karyotype of T. harringtoni. The number of steps plus differences in chromosome morphology make this Fig. 30. Hypothetical pathways for chromosomal evolution in six species of Taterillus — This is the most parsimonious arrangement based on the assumption that Taterillus congicus has the most primitive karyotype. 169 2N=22/23 FN= 40 XX/XY,Y2 T. /lygtirfius 2N= 28 FN=44 4 fusions XX/XY T. lacustris 3 fusions 1 inversion 1 fu sioiv* 2N= 30 2N= 30 FN= 44 FN= 36 XX/XY XX/X? 7". aruticirius 1 fusion 1 fu si 2N= 32 3 inversions FN= 44 XX/XY 2N = 36/37 1 invoriion 2N=36/37 3 fusionsN FN = 42 FN=44 XX/XY,Y2 XX/XY,y2 7'. mcriclionalis 7\ £. gracilis ] fusion 2N= 38 FN=44 XX/XY -7 2N= 44 FN«62 / XX/XY 6 fusions / • 6 inversions 7\ harringtoni / 2 fusions^ 2N =48 / 8 fusions FN= 62 4 inversions XX/XY 3 fu s 2N = 54 2N= 54 FN= 64 FN=66 XX/XY 1 inversion XX/XY 7\ COMgfCMS T. congicus Pig. 30, Pathways for chromosomal evolution in Taterillus. 170 last pathway the least likely. In any case, chromosomal differences between the three groups are great. Derivation of the four chromosomally different ,species in West Africa from the 2N 38, FN=44, XX/XY hypothetical ancestor is possible. One fusion of an autosome with an X-chromosome would give the 2N=36/37, FN=42, XX/XY^Y2 karyotype of T. g. meridionalis. One additional pericentric inversion would give the 2N=36/37, FN=44, XX/XY^Y2 karyotype of T. g. gracilis. A different series of chromosomal rearrangements from the 2N=3 8 ancestor would give the remaining three species. First, three fusions would result in a hypothetical intermediate with a 2N of 32, FN of 44, XX/XY karyotype. From this 2N=32 form, one fusion and three inversions would give the 2N=3 0, FN=36 karyotype of T. arenarius, In another pathway from the 2N=32 form, one fusion would result in another hypothetical intermediate with a 2N of 3 0 and FN of 44. From this 2N=30 intermediate, one fusion would give the 2N=38, FN=44, XX/XY karyotype of T. lacustris. Also, from the 2N=30 intermediate, either four fusions or three fusions and one inversion would give the 2N-22/23, FN=40, XX/XY^Y^ karyotype of T, pygargus. Comparisons of serum-proteins (Tranier et al,, 1974) showed that T, lacustris was more closely related to T. pygargus than to T, gracilis. The derivation of both the T. lacustris and T. pygargus 171 karyotypes from the hypothetical intermediate 2N=32 and 2N=30 forms agrees with serum-protein analyses. The above analysis of the chromosomal relationships and evolution of the six species of Taterillus was made by considering diploid number, fundamental number, and morphology. For species to be closely related, certain stages (hypothetical intermediates) were needed in the sequence of changes necessary for deriving karyotypes that are similar in morphology. Consideration was also given to the minimum number of steps required to reach a common ancestor for a given number of karyotypes. Chromosomal banding pattern analysis and additional karyotyped popula tions would clarify the relationships among the taxa of Taterillus. Distribution Although the distribution and ecology of the various taxa of Taterillus were discussed in the chapter on systematic accounts, I have combined them here for compara tive purposes, Taterillus arenarius f This species occurs in southern Mauritania, southern Niger, and probably central Mali (Fig, 31), The vegetation in the Sahel Savanna and Subdesert areas where arenarius occurs is typical savanna vegetation of many species of grasses and scattered trees. As the Sahara Desert zone is approached, the vegetation 15° W 10 W 5 °W 5°E 10°E 15°E T. ARENARIUS T. LACUSTRIS •20 N MAURITANIA T. GRACILIS T. PYGARGUS NIGER MALI 15 N .©UPPER VOLTA J®-' GUINEA GUINEA •ION I \ SIERRA 1 GAMBIA ILEONE IVORY GHANA NIGERIA 300 S COAST Mi les CAMEROON 500 K ilometers Fig. 31, Distribution of the four West African species of Taterillus• f—• -J 173 becomes scattered small bushes and patchy areas of grass. Occasional trees of the Sahel zone do occur but always very localized in vavorable situations (Rosevear, 1953). Taterillus arenarius occurs sympatrically only with T. pygargus although it is near the range of T. gracilis and T. lacustris. In the area where arenarius and pygargus are sympatric, which is in southern Mauritania, arenarius is found in the drier areas with a sandy substrate and pygargus in the grassy areas with a harder substrate. Taterillus pygargus. This species is known from southern Mauritania, most of Senegal, Gambia, and western Mali (Fig. 31). Although it occurs in the Guinea and Sudan Woodlands, and Sahel Savanna, it prefers the drier areas. Taterillus pygargus occurs sympatrically with T. g. gracilis in Senegal, Gambia, and Mali, and with T. arenarius in southern Mauritania (see above). In those areas where pygargus and gracilis are sympatric, pygargus is found in the drier areas with a soft sandy substrate and gracilis in the moister areas with a harder substrate. Taterillus gracilis gracilis. This taxon is known from Senegal, Gambia, and western Mali, and probably occurs in Guinea and Portuguese Guinea (Fig. 6), It occurs sympatrically with T. pygargus (see above). The vegetation zones where T. cf. gracilis occurs are the Guinea and Sudan Woodlands and the Sahel Savanna, 174 Taterillus gracilis meridionalis. This taxon is known from Upper Volta, Ivory Coast, Ghana, Togo, Dahomey, and western Nigeria (Fig. 6). As with T. g. gracilis, it occurs in the Guinea and Sudan Woodlands and the Sahel Savanna. It has not been found to occur with any other taxa of Taterillus. Taterillus gracilis angelus. This subspecies of T. gracilis is known from southern Niger and northern Nigeria (Fig. 6). It is found in the northern Sudan Woodland and Sahel Savanna. It is apparently restricted in distribution to areas with a sandy substrate. The nearest taxa geographically and phylogenetically are T, g. nigeriae and T. g. meridionalis, It is geographically separated from T. g. meridionalis by the Niger River in western Nigeria and ecologically from T. cj_, nigeriae by its soil preference. Taterillus gracilis nigeriae. This taxon is known only from central and southern Nigeria. It occurs in the southern Sudan and the Guinea Woodlands. It is ecologically separated from T. g, angelus by its preference for harder substrates (T, g, angelus prefers sandy soils) and geo graphically from T. g, meridionalis as discussed above. Taterillus lacustris. This geographically restricted species is known only from eastern Nigeria and northern Cameroon (Fig, 31). This area is in the Sahel Savanna. The nearest taxa geographically are T. g. angelus and T. g. nigeriae in Nigeria, T. congicus in northern Cameroon and 175 Chad, and T. arenarius in Niger. It is not known to occur sympatrically with any of these taxa. Taterillus congicus. This species is known from northern Cameroon, southern Chad, Central African Republic, Zaire, western Uganda, and the Sudan (Fig. 32). It is known to occur sympatrically with T. harringtoni at Gordil, Central African Republic, and is probably sympatric with T. lacustris in Cameroon, with T. emini in Uganda and the Sudan and with T. harringtoni in Sudan. It occurs in the Guinea and Sudan Woodlands and Sahel Savanna. In areas of sympatry with harringtoni, lacustris, or'emini, no eco logical separation is known. Taterillus emini. This species is known from eastern Zaire, Uganda, Sudan, and northwestern Kenya (Fig, 32). This area includes the Guinea and Sudan Woodlands and Sahel Savanna, It occurs sympatrically with T. congicus in northeastern Zaire, northwestern Uganda, and Sudan, and with T, harringtoni in northern Uganda and the Sudan, In the areas of sympatry with congicus or harringtoni, no eco logical separation is known, Taterillus harringtoni. This species is known from northeastern Tanzania, Kenya, southwestern Somalia, southern Ethiopia, northern Uganda, Sudan, and northern Central African Republic. It occurs in the Sudan Woodland and Sahel Savanna, Ta terillus harringtoni is known to occur sympatri cally with T, emini in northern Uganda and Sudan, and with 15°E 20 °E 45°E 15 N ETHIOPIA CENTRAL:-; AFRICAN REPUBLIC OM ZAIRE SOMALIA tyr ® KENYA CONGO T. HARRINGTONI UGANDA T. CONGICUS RWANDA, Fig. 32. Distribution of the three East African species of Taterillus. (_> T. congicus in Central African Republic and Sudan. Eco logical separation between harringtoni and either congicus or emini in areas of sympatry is not known. Future Studies A great deal remains to be learned about the biology and systematics of species of Taterillus. One important aspect of the biology of these animals has been virtually unstudied--their ecology. Only the works of Hubert (1973) and Poulet (1972a, 1972b, 1973, 1974) on T. gracilis and T. pygargus in Senegal are the exceptions. Studies of any other species would be of real value, but of more interest would be studies of species in areas of actual or potential sympatry. The area in which the ecological relationships of Taterillus arenarius and T. pygargus would be studied is in southern Mauritania. The ecological rela tionships of T. g. gracilis and T. pygargus should be studied in greater detail in each of the three vegetation zones in which they are sympatric in Senegal. Studies in southern Niger should be undertaken to determine the rela tionships between T. g. angelus and T. arenarius. Also, ecological separation between T. g. angelus and T. g. nigeriae should be investigated, as well as determination of their differences with T. lacustris and T. congicus in eastern Nigeria and northern Cameroon. The ecological relationships between T. congicus, T. emini, and T. 178 harringtoni could be studied in many areas in northern Zaire, northern Uganda, Central African Republic, and the Sudan. The ecological preferences between T. lacustris and T. congicus could be studied in northern Cameroon and adjacent areas of Chad. Studies of ecological physiology would be useful in understanding habitat preferences of the species. More systematic work in areas of serology and karyology still remains to be done. A much more extensive survey of karyological relationships is needed before species relationships and their evolution can be fully unders tood. I hope that this study has resulted in a systematic arrangement based on the biological relationships of members of the genus Taterillus so that studies such as those described above can be carried out to give more meaningful results. CHAPTER 6 EVOLUTIONARY AND ZOOGEOGRAPHICAL RELATIONSHIPS This chapter is divided into two parts based on results from the previous chapters. The first part on evolutionary and systematic relstionships includes a statistical analysis of seven samples representative of the species in Taterillus. The results of this analysis is discussed in relation to the probable karyotypic relation ships discussed in the previous chapter. The second part on zoogeography discussed the distribution of the species of Taterillus in relation to their evolution, radiation, and dispersal, Evolutionary and Systematic Relationships The lack of a fossil record for the genus Taterillus eliminates the possibility of analyzing the evolutionary relationships within the genus based upon the paleontological record, Therefore, in order to obtain an objective estimate of the possible evolutionary relationships among the seven species of Taterillus, I have used a cluster analysis based on the distance matrix, a principal components analysis, and an MDSCALE analysis, which are a part of the NT-SYS pro grams. These multivariate statistical techniques are explained in Appendix C, The minimum spanning tree was also 179 180 computed which can give an estimation of the phenetic distance between the groups of samples (species) being analyzed. The minimum spanning tree can also indicate the probable cladistic and patristic center for the species concerned. The characters used were all of the cranial measure ments shown in Appendix B except for zygomatic breadth (ZB) which is present in less than 50 per cent of the individuals analyzed. The population samples selected (see Figs. 10 and 20 and Appendix D) for the species comparisons were: Tiguent, Mauritania for T. arenarius (locality 28); Poko, Zaire for T. congicus (locality 1); all age class 4 specimens assigned to T. emini from Uganda (locality 2); known karyotyped specimens of T. g. gracilis from Senegal (including localities 22-25); southern Kenya specimens of T. harringtoni (locality 5); the single karyotyped specimen of T. lacustris from Mora Cameroon (locality 33); and karyotyped specimens of T. pygargus from Senegal (localities 26 and 21) , The distance matrix (Table 11) and the distance phenogram (Fig, 33) based on this matrix, show the morpho logical of phenetic relationships between the species. Taterillus gracilis and T. pygargus form a small cluster which is connected by increasing distances to T. lacustris, T. arenarius, and T. harringtoni respectively. Taterillus harringtoni is in a cluster by itself as is T. arenarius. Table 11. Distance matrix for the seven species of Taterillus. T. £. T. 1_. T. a. T. g. T. h. T. c. T. e. T. pygargus 0.000 T. lacustris 0.881 0.000 T. arenarius 0.999 1.139 0.000 T. gracilis 0.694 0.820 0.933 0.000 T. harringtoni 1.033 1.392 1.390 1.138 0.000 T. congicus 1.835 1.919 1.411 1.725 1.978 0.000 T. emini 1.705 1.664 1.542 1.582 1.874 0.855 0.000 T. pygargus T. gracilis T. lacustris T. arenarius T. harringtoni T. congicus T. emini 1.72 1,52 1.32 1.12 0.92 0.72 Fig. 33, Distance phenogram for the seven species of Taterillus -- Larger numbers in the scale at the bottom of the figure indicate greater phenetic distance between the samples. The cophenetic correlation is 0.935. t—1 NJ00 183 The distance matrix (Table 11) shows that T. harringtoni and T. arenarius are not very close phenetically to any of the other species. Taterillus congicus and T. emini form a cluster which is distinct from the other species. Although T. congicus and T. emini are moderately close to each other, the phenogram and distance matrix show that they are very far from all of the other species. In the principal components analysis (Fig. 34 and Table 12), 89,2 per cent of the phenetic variation based on the matrix of correlation among the 23 characters was explained by the first three components as follows: component I, 59,1 per cent; component II, 17.6 per cent; component III, 12.5 per cent. The character loadings which indicate the measurements influencing a particular component the most are shown in Table 12. The first component, influenced by most of the major size measurements, separates T. congicus and T. emini from the rest of the species, and slightly separates T. arenarius from the others as well, The second component appears to distinguish all of the species which are grouped in the first component. The characters which load the highest on the second component are breadth of braincase (BBC), greatest breadth across the bullae (GBAB), postpalatal length (PPAL), length of bullae (LAB), breadth of bullae (BAB), and negatively, length of rostrum (LR), The third component distinguishes T. emini from T, congicus and the loadings indicate that T, emini has OT. horringtoni emini o ! S\l* o \o^ T. congicus O QT. gracil o T. arenarius I 34. First three principal components the seven species of Taterillus. 185 Table 12. Loadings of the cranial measurements on the first three components of means of the seven species of Taterillus — Explanation of the abbreviations is given in Fig. 2, The asterisks indicate those measurements which contribute the most to a given component. Component Measurement I II III GLS .988* .032 -.060 OCNL .968* -.020 . 046 BAL . 899* -.128 .362 CBAL .956* . 037 . 239 BBC .638 .705* -.001 IOC .611 . 106 -.415 BR .991* . 025 -.027 GBAB . 213 .784* . 344 BrMlMl .840 -.082 .489* DAL .761 -.350 -.422 PAL .977* -.145 -.049 PPAL .489 . 578* . 146 LAPF . 838 -.321 .245 LPB .901* . 232 -.344 LAB .449 .841* -.171 BAB -.091 .645* -.688* LN .767 -.490 -.374 LR .47 5 -.780* -.383 FNL .787 -.191 -.523* DC .890* .313 .265 LPPF .793 . 254 -.458* ALM1 , 660 -.231 -.446* ALM , 806 -.168 -.481* 186 a greater breadth across the molars (BrMlMl), larger molar toothrow measurements (ALMl, ALM) , and smaller breadth of bullae (BAB), frontonasal length (FNL), and posterior palatine foramina (LPPF) than does T. congicus. The MDSCALE projection of the seven species (Fig. 35) includes the minimum spanning tree which connects species which are phenetically closest. The minimum spanning tree links T. emini to T. congicus. As mentioned previ ously, comparisons of karyotypes of the species (unknown in T. emini) shows that all of the species of Taterillus are derivable from T. congicus. It is possible that T. congicus was derived from T. emini, or the reverse. Chromosomally, T. harringtoni is derivable from T. congicus, although in Fig. 35, T, congicus is linked to T. arenarius using cranial measurements. Cranial morphology shows T. harringtoni as being closest to T. pygargus. Chromosomally, derivation of any of the West African species (T. arenarius, T. gracilis, T. pygargus, and T. lacustris) from T. harringtoni would require a great number of chromosomal rearrangements. Without the complexity of chromosomal evolution, the minimum spanning tree indicates that T. arenarius, T. pygargus, and T. lacustris arose from T. gracilis. The linkage tree deviates only slightly from the hypothesized chromosomal species derivations. Chromosomally it is pos sible to derive all of the West African taxa from a hypothesized ancestor which is closest to T. gracilis. T. congicus T. emini T. orenarius T. harringtoni I Fig. 35. Three-dimensional MDSCALE projection for the seven species of Taterillus -- Lines connect the species which are closest in phenetic distance. Stress value is 0.002. 188 Indications of the cladistic or patristic center are those species in which the links connect one species to more than one of the others. One center shown is T. gracilis, and the others are T. arenarius and T. congicus. Chromosomally, T. congicus is closer to T, harringtoni than to T. arenarius. Linking T. congicus to T. arenarius by cranial morphology does not agree with comparisons of their chromosome morphology. Also,' it is unlikely that T, pygargus was derived from T, harringtoni. With the exception of T. harringtoni, the West African taxa are shown (Fig, 35) to be phenetically close. The phenetical closeness of the West African taxa agrees with the hypothesized pathway (Fig. 30) for chromosomal evolution in Taterillus. Zoogeography Taterillus is known only from the Recent epoch. It is restricted to the Ethiopian zoogeographic realm (Africa south of the Sahara desert) in a belt across Africa between the Sahara desert to the north and the high forest in the south, Taterillus should be included in both Saharan and tropical African faunas. Although the genus Taterillus occurs throughout the three sub-Saharan vegetation zones (Guinea and Sudan Woodlands, Sahel Savanna), most of its species also range throughout these three zones. Therefore, vegetation zones do not describe the geographic range of species in Taterillus. 189 Taterillus was described as a diminutive of and distinct from the genus Tatera by Thomas (1910). Petter (1974) discussed morphological features in common between Tatera and Taterillus. Tatera is known from the late Pliocene and Pleistocene from East Africa. Morphological similarities, including cranial and chromosome morphology show that evolution and radiation of the genus Tatera probably led to the genus Taterillus. Assuming that the species of Taterillus with the highest diploid number of chromosomes (T. congicus, 2N=54) is ancestral to the other six species, its occurrence only in East Africa leads me to believe that the genus Taterillus probably evolved from Tatera in East Africa. Assuming that Taterillus evolved during the late Pleistocene or early Recent in East Africa, it has dispersed into several different habitats throughout East and West Africa. Tatera now occurs throughout Africa and parts of Asia. In Africa, it occurs from parts of the high forest to the northern Sudan woodlands. Due to a scarcity of the known fossil record for Tatera and a lack of a fossil record for Taterillus, speculation as to the past habitat or origin of either genus is not possible. Both Taterillus and Tatera are now adapted to a wide variety of habitats and vegetation zones. This is probably due to the fluctuating climates throughout Africa during the late Pleistocene and Recent. Two factors, rainfall and 190 competition, seem to be important in the distribution and dispersal of the species of Taterillus. First, in areas of high moisture Taterillus is replaced in part by Tatera and other wet-adapted rodents. Taterillus is also limited by extreme aridity, because no species occurs within the low rainfall areas of the Sahara desert. Second, congeneric species appear to form signifi cant barriers to dispersal. Although areas of sympatry are quite large, no more than two species have been found to occur in the same area. Where sympatry occurs, the two species have subdivided the habitat to avoid competition. Species already present in an area would serve to block the entry of others. In areas where only one species of Taterillus occurs, other genera and species probably occupy the habitat in which.a second species of Taterillus might prefer. This could inhibit the entry of some species of Taterillus which would otherwise be well adapted to the area. In only one case does a physical barrier seem to be a factor in separating two taxa of Taterillus, This is the Niger River in western Nigeria and southwestern Niger. It serves to separate T, cj, meridionalis from both T, g. angelus and T. g, nigeriae, with resulting differences in their morphology. Two other taxa, T. g, angelus and T. g, nigeriae, although potentially sympatric, seem to be restricted in their north-south distribution by the presence or absence of suitable habitat. Specifically, T. g, angelus 191 prefers the more northern sandy substrates while T. cj. nigeriae prefers the southern clay-like substrates. CHAPTER 7 SUMMARY The genus Taterillus is found in a wide variety of habitats in a belt across sub-Saharan Africa. Individual variation has confused past investigations of the biology and systematics of the genus by obscuring population and species variation. Some workers have suggested that the West African forms were a single polymorphic species, and thought that this might be true for the entire genus. Others, using the 18 described species and 23 subspecies, have recognized as many as, 10 species and 19 subspecies. Cytogenetical data and an increase in available study specimens have shown that neither of the above is true. Six different chromosomal forms are now known withi the range of the genus. This report shows the importance of using all available data when studying the biology and systematics of any mammalian group. Based on a greater number of specimens, available karyotypic data, and a detailed morphological analysis using univariate and multi variate statistical techniques, I have come to conclusions that differ in some points from all previous workers. I recognize seven species in the genus Taterillus. These are Taterillus arenarius, Taterillus congicus, 192 193 Taterillus emini, Taterillus gracilis, Taterillus harringtoni, Taterillus lacustris, and Taterillus pygargus. Examination of specimens from many museums has allowed me to give a more complete distribution of these species. Detailed descriptions of these species is given along with subspecific considerations. The descriptions are based on cranial morphology, karyotypic differences, distribution, and possible habitat selection. APPENDIX A GAZETTEER Names of geographic and topographic features listed below are those used in the text of this report. The primary sources for spelling and coordinates of localities were the gazetteers of the United States Board on Geographic Names (prepared by the Office of Geography, Department of Interior) for Cameroon, 1962; Central African Republic, 1962; Chad, 1962; Dahomey, 1965; Ethiopia, Eritrea, and the Somalilands, 1950; Gambia, No. 107, 1968; Ghana, No. 102, 1967; Ivory Coast, 1965; Kenya, 1964; Mali, 1966; Mauritania, 1966; Niger, No. 99, 1966; Nigeria, 1971; Senegal, No. 88, 1965; Sudan, 1962; Tanzania, No. 92, 1965; Togo, No, 98, 1966; Uganda, 1964; Upper,Volta, No. 87, 1965; and Zaire (Republic of the Congo, Leopoldville), 1964, If place- names were not found in the above-listed gazetteers, or in the references listed below, other sources used to locate them included: (1) Army Map Service, Series 1301 GSGS 4646, World (Africa) 1:1,000,000; (2) Institut Geographique National-Paris, Carte de l'Afrique de l'Ouest au 1:500,000; and (3) USAE Aeronautical Chart and Information Service, World Aeronautical Chart Series 1:1,000,000. 194 195 The equal sign enclosed in parentheses denotes the gazetteer spelling or the current name of the locality as opposed to that given on the specimen label. Where coordinates differ slightly from that listed in the gazetteer, I listed the ones given by the collectors on the specimen label or field notes. Coordinates for a locality with distance modifiers, when given, are for the specific collecting -locality. Places that could not be located directly, but that were located indirectly with reference to some other known features, are designated with the abbreviation "ca." Other gazetteers and references responsible for some of the localities and coordinates are those of Alexander (1907); Allen (1931) ; Chapin (1954) ; Davis and Misonne (1964); Delany and Neal (1966); De Saeger (1954); De Witte (1933); Genest and Petter (1973) ; Hatt (1940) ; Moreau, Hopkins, and Hayman (1946); Ogilvie-Grant (1913); Petter (1970); Poulet (1972b); Robbins (l'974a, 1974b); Rosevear (1953, 1965); Saint-Leger (1937); Schouteden (1936, 1944); Setzer (1956); and Tranier et al. (1974). The following two maps show the localities listed in the gazetteer, 15°W 10°W 5°W 5°E 10°E 20°N MAURITANIA NIGER MALI 15°N .'SENEGALO O a O UPPER o _ 0 VOL TA • GUINEA 10°N e« I L SIERRA 1 GAMBIA ( LEONE VORY GHANA COAST NIGERIA 300 CAMEROON 500 5°N Ki Jometer Localities listed in the gazetteer for West African countries. VD CHAD SUDAN ETHIOPIA 9~ / CENTRAL AFRICAN REPUBLIC CONGO UGANDA RWANDA, BURUNDI TANZANIA Localities listed in the gazetteer for East 198 Cameroon Fort-Foureau, 10 km E 12° 051N.,14° 561E. Garoua, 35 km S 9° 001N.,13° 3 01E. Maroua, 50 km S 10°10 1N.,14° 331E. Mora 11° 0 3'N.,14° 09 1 E. Sir 10° 33'N.,13° 411E. Waza (=Vaza) 10° 211N.,14° 201E. Yagoua 10° 201N.,15°141E. Central African Republic Bangui 4° 2 2'N.,18° 3 51E. Gordil 9° 4 4'N.,21° 3 51E. IPPY 6°151N.,21°12'E. Koumbala 8° 50'N.,22° 40'E. Maboke 3° 54'N.,17° 53'E. Lac Mamoun 10 0 071N. ,21° 57'E. Chad Aouk River, opposite Golongoso 8 0 51'N.,18 0 531E. Bekao 7° 54'N. ,16° 021E. Deli 8°42'N.,15°53'E. Fort-Archambault 9° 091N.,18° 231E. Fort-Lamy 12° 07'N,,15° 03'E. Moundou 8° 3 4'N.,16° 0 51E. Dahomey Banikoara 11°18'N.,02° 26'E. Bimbereke 10°14 1N.,0 2° 40 1E. Diho 8° 0 5 1 N . , 02° 31' E. Guene 11° 44'N.,03°13'E. Kouande 10° 20'N.,01°41 1E. Nikki 9° 56'N.,02°131E. Parakou 9° 21'N.,02° 37'E. Porga 11° 02'N, ,00° 581E. Segbana 10° 56'N.,03° 421E. Soubroukou 9° 41'N.,01° 3 81E. Zizonkame 7° 551N.,02° 011E. Ethiopia Mutti Galeb ca. 5° 40 ' N.,36°20'E. Gambia Kudang 13° 4 0'N.,15° 03'W. Kuntaur, 1 mi E 13° 40'N. ,14° 52'W. Sun Kunda 13 ° 23 1 N . , 13° 51 'W. Tambasensan 13° 23'N.,14°10'W. Toniataba 13° 241N.,15°35'W. 199 Ghana Bangwon 10° 58 'N. ,02°41 'W. Damongo 9 ° 04 •N. ,01° 4 5 'W. Gambaga 10 0 31 •N. ,00° 28 'W. Lawra 10°39 'N.,02°52 •w. Nabogo 9 ° 4 5 'N. ,00° 4 9 •w. Navarro (=Navrongo) 10° 54 N. ,01° 06 'W. Pirisi 10° 07 N. ,02° 27 'W. Pong Tamale 9° 41 N. ,00° 49 "W. Pulima 10° 51 N. ,02°03 •w. Sakpa 8° 52 N. , 0 2° 21 •w. Shishe 10° 42 'N. ,00°13 •w. Wa 10° 03 N. ,02° 29 w. Wulasi 8° 3 9 N. ,00°00'G. Yabraso 8° 0 4 N. ,01° 48 •w. Ivory Coast Beoumi 7° 4 0 N. ,05°34 •w. Bouna 9° 16 N. ,03° 00 w. Kong 9° 09 N. ,04°37 w. Sienso 9° 2 3 N. ,07° 31 •w. Yaraa 9° 3 6 N. , 06 018 •w. Kenya Archer's Post, 12 mi N 00° 4 6 N. , 3 7 0 4 2'E. Baringo 00° 28 N. , 3 5° 5 8 'E. Eusso Nyi.ro, 12 mi N (see Archer's Post) I jara 01° 3 6 S. ,40°31 'E. Isiolo, 6 mi N (=Meru River) 0 0° 2 5 N. , 3 7° 3 6 E. Kangatet 01° 59 N. ,36° 06 •E. Kapedo, 59 mi N 01° 55 N. ,36°07 'E. Kisigau (=Kasigau) 03° 49 S. ,38°40 E. Lakaungu (=Takaungu) 03° 40 S. ,3 9 ° 53 E. Lake Hannington, 3 mi E, 3 mi S Maji Ya Moto 00° 13 N. , 36°06 •E. Lodwar 03° 07 N. ,3 5° 3 6 E. Lokichoggio (=Lokichokio) 04° 15 N. , 3 4° 2 0 E . Lokori 0 2° 01 N. , 36°08 E. Marali (=Maralal) 01° 06 N. , 3 6 ° 4 2 E. Masabubu 01° 13 S. , 40°00 'E. Merifano 0 2° 19 S . , 40°08 'E. Murri (=Malka Mari) 04° 16 N. , 4 0° 46 'E. Mt. Mbololo, Taita Hills 03° 17 S. , 38°28 'E. Mt. Nyiru, west base 02° 05 N, , 360 50 'E. Mweru (=Meru) River (see Isiolo) Ngare Ndare, 15 mi W Isiolo 0 0° 15 N. ,37°32 E, Nyama Nyango (=Samburu Game Lodge) 0 0 0 3 4 N. ,37 ° 33 'E. Olorgesailie (=01olkisalie) 01° 3 4 S. ,36 ° 27 'E. Orr Valley (=Horr Valley), Mt, Nyiru 0 2° 0 8 N. , 36°51'E. N Samburu Game Lodge (see Nyama Nyango) O o C 0 O South Horr, 8 mi N (=Orr Valley) C N. ,36° 51'E. 200 Kenya (cont'd) Takaungu (see Lakaungu) Taveta 03 ° 2 4 S . , 3 7° 41 E. Turkwell River, 10 mi W Ngamatak Hills 02° 40 N. ,35° 21 E. Voi 0 3° 2 4 S . , 38°34 E. Wen je 01° 47 S. ,40° 06 E. Mali Bamako 12° 3 9 N. ,08° 00 W. Mopti 14° 3 0 N. ,0 4°12 W. Mauritania Aleg, 3 km S 17° 02 N. ,13° 55 W. Bou Rjeimat 19 0 04 N. ,15° 08 W. Garak 16 0 33 N. ,15° 46 W. Kaedi 16° 09 N. ,13°30 W. Nouakchott, 6 km E 18° 09 N. ,15° 58 W. Nouakchott, 11 km N 18° 13 N. ,16° 01 W. Passe de Soufa 15° 56 N. ,12° 00 W. Tiguent 17° 16 N. ,16° 01 W. Niger Agadez, 5 km NE 17° 02 'N. ,08° 02 E. Garari (=Tessaoua) 13° 45 N. ,07° 59 E. In-Gall, 30 km S 16°33 N. ,06°52 E. In-Gall, 120 km S 15° 4 5 N. ,06° 3 6 E. Niamey, 4 5 km NW 13° 48 N. ,01° 43 E. Tahoua, 2 5 km S 14° 3 9 N. ,05°22 E. Niger ia Af on 0 8 0 19 'N. ,04°31 E. Anara Forest Reserve 10° 42'N. ,07° 37 E. Bama, 43 mi SE Maiduguri 110 31 N. ,13° 41 E. Dada 11° 3 4 N. ,04°29 E . Dikwa, 31 mi NE 12° 22 N. ,14°10 E. Farniso (see Panisau) Felele, 3,6 mi NW Lokoja (Filele) 0 7 0 51 'N. ,06°43 E. Ibaden 0 6° 3 2 N. ,02°46 W. Iella, 2 mi E Bahinde 11° 24 "N. ,04°14 E. Kabwir 09 ° 24 'N. ,09°34 E. Kaddai 13° 3 0 N. ,13° 3 4 E. Kano 12 0 0 0 N. , 0 8° 31 E. Karaduwa 12° 19 'N. ,07° 41 E. Kudu 09° 16 'N. ,05° 20 E. Mada River, 3 mi E Gudi 08° 54 'N. ,08°17 E. Maiduguri 11° 51'N, ,13 0 09 E. Maiduguri, 22 mi S 11° 33 'N. ,13°16 E. Panisau (=Farniso) 12° 0 6 'N. ,0 8° 3 0 E . Panyam 09° 25 "N. ,09 °13 E. Sokoto, 12 mi N 13° 0 8 'N. ,05° 09 'E. Tangaza 13° 2 0 'N. ,04°54 'E. 201 Nigeria (cont'd) Tsanchaga, 8 mi E Bida 09° 04 N. , 0 6 0 0 4 E. Upper Ogun (Forest Reserve) 08° 30 N. ,03° 50 E. Yola 09° 12 N. ,12° 29 E. Senegal Bakel, 5 km S 14° 51 N. ,12° 28 W. Bandia, Foret Classee de 14° 3 5 N. ,16 0 58 W. Cascas 16° 23 N. ,14°04 W. Cuabo 14° 59 N. ,12° 28 W. Dakar 14 0 4 0 N. , 17 0 26 W. Debi 16° 2 8 N. ,16°17 W. Diakaba 12° 49 N. ,120 04 W. Galoya (see Padoz) Gamon 13° 20 N. ,12 0 5 5 W. Gemenjulla (=Dienoundilla) 13° 12 N. ,13 0 0 7 W. Goudiry 14° 11 N. ,12° 43 W. Kaffrine, 15 km N 14° 12 N. ,15° 32 W. Kaolach, 6 km E 14° 10 N. ,16° 02 W. Kolda, 5 km N 12° 57 N. ,14 0 57 •w. Kotiare Naounde (=Kotiari Naoude) 13° 54 N. ,13 0 27 'W. Koungheul 13° 59 N. ,14° 48 w. Koussanar 13° 52 N. ,14° 05 w. Linguere 15° 24 N. ,15° 07 w. Louga, 8 km E 15° 37 N. ,16° 09 w. Madina Ndiatebe 16° 18 N. ,14 0 0 8 w. N'Doulo (=Ndoulo), 17 km NE Diourbel I4044 N. ,16° 07 •w. Ogo, 13 km SW Matam 15° 33 N. ,13 017 w. Padoz (=Galoya) 160 051N. ,13 0 51 'W. Pete Ole (-Fete Ole) 16° 10 N. ,15 0 0 5 w. Podor 16° 40 N. ,140 57 w. Ranerou 15° 18 N. ,13 0 5 8 w. Richard-Toll 16° 28 N. ,15° 41 'W. Saboya, Foret Classee de 13 0 3 9 N. ,16° 07 w. St. Louis, 10 km SE 15° 56 N. ,16° 26 •w. Thies, 10 km W 14 0 4 6 N. ,17 0 01 'W. Tubakuta (=Toubakouta) 13 0 4 3 'N. , 16 0 2 8 •w. Somalia Afmadu 0 0° 31 N. , 42°04'E. Arenaga 00° 32 N. ,42° 55'E. Dolo 0 4° 11'N. , 42°04 'E. Lugh (=Lugh Ferrandi) 0 3 0 4 8'N. ,42° 33 'E. Sudan Agur (Ajur or Djur) 11°3 5'N. ,30° 28'E. Akona 11° 30 'N. ,32° 41 E. Angolo (=Angola), Jebel Korongo 10° 31'N. ,29° 53 'E. Delami, Nuba Mts, 11° 52 'N., 30°28 'E. Dugdug (=Duqduq) 0 8° 04 "N., 2 8 0 3 4 'E. Sudan (cont'd) Dibbis, Jebel Marra 12° 34 N. ,24°13 E. El Fasher (=A1 Fashir) 13° 3 8 N. ,25° 21 E. El Fasher, 110 mi E (=Ura Kedada) 13° 3 6 N. , 2 6° 4 2 E. El Obeid (=A1 Ubayyid) 13° 11 N. ,30°13 E. Gondokoro 0 4° 54 N. ,31° 40 E. Ikoto 0 4° 0 5 N. ,33°06 E. Iraurok 0 4° 19 N. ,32° 24 E. Jebel Kadero 12° 0 8 N. ,30°15 E. Jebel Marra (Mt.) (=Jabal Marrah) ]- 3° 10 N. ,24°22 E. Jebel Marra (Range) 12°451-13°30'N. and 24°15 - 24° 4 5 1 E Jebel Maidob (^Jabal Maydub) 15° 14 N. , 26 ° 3 0 E. Jebelein (=A1 Jabalayn), 20 mi S 12° 18 N. , 3 2° 4 5 E. Juba 04° 51 N. , 31°37 E. Kadugli (=Kaduqli) 11°01 N. ,29° 43 E. Kagelu 0 4° 0 3 N. ,30° 36 E. Kamisa, Dinda (=Dinder) River 13° 05 N. ,34°13 E. Katta 07° 52 N. , 27° 53 E. Kulme, Wadi Aribo 12° 3 4 N. ,23° 3 7 E. Kurra, NE Jebel Marra 13° 17 N. ,24°30 E. Latome 0 4° 07 N. ,33°37 E. Logoforok 03° 58 N. ,33° 04 E. Logwi (-Nogwi) 04° 0 6 N. ,32° 43 E. Madu, 8 0 mi NE El Fasher 14° 3 7 N. , 2 6° 0 4 E. Malek 06° 04 N. , 31° 3 6 E. Moli, 3 5 mi N Nimule 04° 02 N. ,31° 57 E. Mongalla 05° 10 N. ,31° 50 E. Nahud (=An Nahud) 12° 42 N. ,28° 26 E. Nogwi (see Logwi) 0 4° 0 6 N. ,32° 43 E. Nuba Mountains 12 0 0 0 N. , 30 0 45 E. Obbo 0 4° 0 2 N. ,32°28 E. Opari 03 ° 5 6 N. ,32°03 E. Raffili 0 6° 51 N. ,27° 59 E. Tagbo Hills 14° 40 N. ,25°50 E. Terrokakka (=Terakeka) 05° 26 N. ,31° 45 E. Tor it 04° 24 N. ,32°34 E. Tuksueina (=Tukswana) 11° 08 N. ,29° 49 E. Um Dona (=Umm Doma) 13° 10 N. ,30° 03 E. Um Kedada (=Umm Keddada) 13° 3 6 N. , 26 ° 4 2 E. Wau (=Waw) 0 7° 4 2 N. ,28° 00 E. Yei 04° 07 N. ,30°40 E. Tanzania 0 o Kisima o 0 'S. ,38°33 'E. Togo Aledjo, Foret Classee de 09° 15 'N. ,01°12'E. Borgou 10° 4 6 'N. ,00°35 'E. Dapango 10° 52 'N. ,00°13 'E. Kamina 0 7° 31 N, ,01°11'E. 203 Togo (cont'd) Namoundjoga 10° 54 N. ,00° 24 E. Padori 10° 311N. ,00°25 E. Pagala 0 8° 11 N. , 0 0° 5 8 E. Pewa 09 ° 17 N. ,01°14 E. Uganda Aido 01° 40 N. ,33° 30 E. Ajeluk (^Ajeruk) 01° 3 3 N. ,33°49 E. Amudat 01° 57 N. ,34°57 E. Awaich (=Awach) 0 2° 57 N. ,32°25 E. Falabek (=Palabek), 45 mi E Nimule 0 3 ° 2 6 N. ,32°34 E. Kamchuru 020 40 N. , 3 3° 3 5 E . Keem (=Kem) 0 2° 41 N. , 33°46 E. Kolido (^Kotido) 0 3° 0 0 N. ,34° 06 E. Lobome (see Lotome) Lorengileipi (=Lorengikipi) 0 2° 2 0 N. ,3 3° 51 E . Lotome (=Lobome) 0 2° 2 2 N. , 3 4 0 3 2 E. Moroto 0 2° 3 2 N. , 34°39 E. Ngal (=Angal) 02° 26 N. , 31°11 E. Pabbo (=Pabo and Pabboi) 0 3° 0 0 N. , 3 2 0 09 E . Rhino Camp 02 0 58 N. ,31° 24 E. Serere 01° 31 N. ,33°27 E. Wadelai 0 2° 4 4 N. ,31° 24 E. Upper Volta Arly 11° 3 4 N. ,01° 26 'E. Barga, 9 km NE 13° 51 N. ,02°12 W. Bobo Dioulasso 11° 12 N. ,04°18 w. Boussouma 12° 57 N. ,01° 05 w. Cella, 1 km N 11° 3 8 N. ,00 ° 22 •w. Dana, 8 km S 11° 4 8 N. , 0 2° 0 8 •w. Dio 13 ° 2 0 N. , 02°38 'W, Djipologo 10 0 5 6 N. ,03° 07 'W. Po 11° 531N. , 04° 311W. Founzan 11° 27 N. ,0 3°14'W. Goden 12° 2 2 N. ,02°18 'W. Gorgadji, 17 km E 14° 02 N. , 0 0° 2 2 ' W . Konankira 12° 54 N. , 03°53 'W. Koutoura, 5 km SW 110 19 N. ,04° 53 'W. Nasso 11°13 N. ,04°26 'W. Natiaboani 11° 4 2 N. ,00°30 'E. Nayoure, 3 km SE 12° 15 N. ,00°16 1E. Nobere, 9 mi S 11° 261N. ,01°10 •w. Orodara, 27 km ENE 11° 0 4 "N. ,04°41'W. Oulo 11° 54 •N. ,02° 58 'W. Petoye 14° 3 5 •N. ,00° 22 •w. Satiri, 8 km NE 11° 30 'N. ,0 4° 0 0 •w. Seguenega, 6 km SE 13 ° 24 'N. , 01° 55 'W. Sideradougou 10° 4 0 'N. ,04°15 •w. Tatarko 13 0 29 'N. , 00° 20 •w. 204 Zaire Bafuka (=Bafuku) 04° 09 N.,27° 54 'E, Bagbale (=Bagbele) 04° 22 N.,29° 15 1 E. Faradj e 03 ° 4 6 N.,29° 42 'E, Gangala (=Gangala-na-Bodio) 0 3° 41 N.,29° 08 'E. Garamba 0 4° 11 N.,30°00 'E. Mbanga 03°51I-04°00IN. and 29° 18 -29° 25'E. Mauda 0 4° 0 5 , 27° 41'E, Nagero 03° 4 5 , 29°31 'E. Niangara 0 3 ° 4 2 , 27° 52'E, Poko 0 3° 0 8 ,26°53 'E, I/O/I 04° 21 , 29° 16 'E, II/gd/4 03° 57 , 29° 23 'E, APPENDIX B MEASUREMENTS External Body Measurements Preparation of museum study skins resulted in the following measurements. For definitions see Cockrum (1962) and Setzer (1968). . Total length (TOT). Tail length (TAL). Length of hind foot (HF). Length of ear (from notch) (E). Weight (in grams) (W) .. One additional measurement was derived in the laboratory from the above. Head and body length (HB). Cranial Measurements The following list of the twenty-four cranial measurements taken on skulls of Taterillus, and their definitions, was adapted from Cockrum (1962). The reference points, between which the measurement was taken, are shown in the accompanying skull diagram. Greatest Length of Skull (GLS). The overall length from the anterior tip of the nasals to the most posterior projection of the supraoccipital. 205 Diagram showing the reference points used for taking the cranial measurements on specimens of Taterillus. Numbers in the figure represent the following: (1) Greatest length of skull (GLS); (2) Occipitonasal length (OCNL); (3) Basilar length (BAL); (4) Zygomatic breadth (ZB); (5) Condylobasilar length (CBAL); (6) Cranial breadth (BBC); (7) Least interorbital constriction (IOC); (8) Rostral breadth (BR); (9) Greatest breadth across the bullae (GBAB); (10) Greatest width across the upper molars (BrMlMl); (11) Length of diastema (DAL); (12) Palatilar length (PAL); (13) Postpalatal length (PPAL); (14) Length of anterior palatine foramina (LAPF); (15) Length of palatal bridge (LPB); (16) Length of auditory bulla (LAB); (17) Width of auditory bulla (BAB); (18) Length of nasals (LN); (19) Length of rostrum (LR); (20) Frontonasal length (FNL); (21) Depth of cranium (DC); (22) Length of posterior palatine foramina (LPPF); (23) Alveolar length of the first molar (ALM1); (24) Length of maxillary cheek teeth (ALM). 206 207 Occipitonasal Length (OCNL). The length from the anterior tip of the nasals to the posterior margin of the condyles. Basilar Length (BAL). The distance along the midline of the skull, from the posterior margins of the alveoli of the upper incisors to the anterior inferior border of the foramen magnum. Zygomatic Breadth (ZB). The greatest distance across the zygomatic arches measured perpendicular to the midline of the skull. Condylobasilar Length (CBAL). The distance along the midline of the skull, from the posterior margins of the alveoli of the upper incisors to the posterior margin of the condyles. Cranial Breadth (BBC). The greatest distance across the braincase immediately posterior to the zygomatic arches. Least Interorbital Constriction (IOC). The least distance across the top of the skull between the orbits. Rostral Breadth (BR). Width of rostrum taken where the maxillary and premaxillary bones meet on the sides of the rostrum, just anterior to the maxillary processes. Greatest Breadth Across the Bullae (GBAB). The greatest distance across the widest portion of the bullae, excluding the projections of the external auditory meatus. 208 Greatest Width Across the Upper Molars (BrMlMl). The greatest distance from the buccal sides of the upper molars, taken with the face of the calipers between the second and third lophs of the first molar. Length of Diastema (DAL). The distance from the posteriormost margin of the alveolus of the upper incisors to the anteriormost margin of the alveolus of the first molar. Palatilar Length (PAL). The distance, in the mid- ventral line of the skull, from the anteriormost border of the parapterygoid fossae to a line connecting the posterior- most margins of the alveoli of the upper incisors. Postpalatal Length (PPAL). The distance, in the midventral line, from the anteriormost border of the parapterygoid fossae, to the anteriormost inferior border of the foramen magnum. Length of Anterior Palatine Foramina (LAPF). The greatest length of the anterior palatine foramina. Length of Palatal Bridge (LPB). The distance from a line connecting the posteriormost margins of the anterior palatine foramina to the anteriormost border of the parapterygoid fossae. Length of Auditory Bulla (LAB). The greatest length of the right (if possible) auditory bulla. Width of Auditory Bulla (BAB). The greatest width of the right (if possible) auditory bulla. 209 Length of Nasals (LN). The distance from the anteriormost tip of the nasals to a line connecting the posteriormost extension of the nasal bones. Length of Rostrum (LR). The distance from the anteriormost part of the nasals to a line joining the posteriormost superior part of the infraorbital canals. Frontonasal Length (FIML). The distance from the anteriormost tip of the nasals to the frontoparietal suture. Depth of Cranium (DC). The vertical distance from a line conn cting the tips of the upper incisors with the most ventral portion of the posterior part of the cranium, to the highest part of the cranium. This is taken by- placing the skull on a glass slide of known thickness, measuring the distance from the lower surface of the slide to the highest point of the skull, then subtracting the thickness of the slide. Length of Posterior Palatine Foramine (LPPF), The greatest length of the posterior palatine foramina. Alveolar Length of the First Molar (ALMl), The alveolar distance from the anterior border to the posterior border of the first upper molar. Length of Maxillary Cheek Teeth (ALM), Alveolar distance from the anterior border of the anterior molar to the posterior border of the posterior molar, APPENDIX C STATISTICAL TECHNIQUES The univariate statistics generated from the DSTAT program were mean, range, standard deviation, standard error of the mean, variance, and coefficient of variation. When two or more groups were being compared, the program employs a single-classification analysis of variance, students' "t" and a homogeneity of variance F-test. Cluster analysis, using the NT-SYS program was conducted using UPGMA (unweighted pair-group method using arithmetic averages) on the distance and correlation matrices, and a phenogram was generated from each matrix. The phenograms were compared against the original matrix and a coefficient of cophenetic correlation was computed. This comparison assesses the amount of distortion in the twoTdimensional phenogram in relation to the original matrix. I have figured only the distance phenograms because their coefficients of cophenetic correlation were larger than those for the correlation phenogram and the results of the distance phenogram agree more closely with results of other analyses, When using either MAINPG or the NT-SYS routines for principal component analysis (PCA), a matrix of correlations 210 211 among variables was computed using unweighted standardized data. The first three principal components were extracted and both two-dimensional and three-dimensional projections of the OTU1s on the components were made. Variable loadings on each component are shown in accompanying tables. The purpose of the PCA is to display the relationships among the OTU's with respect to a new set of variables which are considerably fewer in number than the original variables in the study. The new variables (components) are linear combinations of the original variables such that the pro portion of the variance explained by these new variables is as large as possible. Step-wise discriminant function analyses were performed using the BMD07M program (Dixon, 1973). This program uses variance-covariance mathematics to differen tially weight characters relative to their within-group variation. Any number of groups can be used in this analysis, and unknown specimens are allocated to one of the known groups used. In each analysis a discriminant multiplier is calculated for each variable using the known groups, and this can be multiplied by the value of the respective variable; all such values can be summed to give a discriminant score for a group or individual. It is possible to compute minimum and maximum discriminant score values, below or above which an individual will have a better than 50 per cent probability of being in a particular 212 group. Vectors were also computed to show the contribution of the most important variables for the first two canonical variates used in the discrimination. These vectors were determined by multiplying the pooled within-groups standard deviation by the coefficients for canonical variates one and two, for th'e most useful characters (Power and Tamsitt, 1973). These are figured together with the plot of OTU's on the first two canonical variates. Nonmetric multidimensional scaling analyses were performed using the MDSCALE subroutine of the NT-SYS system. This ordination technique is described by Kruskal (1964a, 1964b). The main difference between the results of MDSCALE and PCA are: (1) differences between close OTU's are in general shown much more accurately by MDSCALE; and (2) the shorter and the longer distances between OTU's are not necessarily shown to the same scale by MDSCALE (Rohlf, 1972). MDSCALE uses the matrix of average distance values between OTU's as was used to compute the distance phenogram, whereas PCA uses a correlation matrix as previously explained. The results of this analysis are shown as a three-dimensional projection of the OTU's with the points connected by a single linkage cluster matrix or minimum spanning tree (see Prim, 1957) for interpoint similarity. A stress value is given and the lower this value the better the representation of the original distance matrix in three dimensions. 213 Discussion of various aspects of multivariate and univariate analyses are found in Simpson et al. (1960), Sokal and Rohlf (1969), Jolicoeur and Mosimann (1960), Sokal and Sneath (1963), and Sneath and Sokal (1973). Choate (1970), Genoways and Jones (1971), Genoways and Choate (1972), and Genoways (1973) have used these types of analyses in studying other mammalian groups. APPENDIX D SKULL MEASUREMENT TABLES The following tables are arranged as follows: Each table is for one of the twenty-four cranial measurements taken and diagramed in Appendix B. Within each table, the following statistics are shown: sample size (N), mean, standard error of the mean (SE), coefficient of variation (CV), and range. Locality numbers, as used in this report and shown in the locality maps following the tables, represent the following localities: 1. Poko, Zaire 2. Uganda 3. N Kenya 4. C Kenya 5. S Kenya 6. Sir, Cameroon 7. Jebel Marra, Sudan 8. Panisau, Nigeria 9. Kabwir, Nigeria 1 — o 1 Felele, Nigeria 11. Dada, Nigeria 12. Karaduwa, Nigeria 13. NE Dahomey (Segbana 214 215 14. C Dahomey (Soubroukou, Kaounde) 15. S Dahomey (Zizonkame, Diho) 16. N Togo (Dapango, Borgou, Naumandjoga) 17. SE Ghana (Wulasi) 18. NE Ghana (Shishe, Gambaga) 19. Barga, Upper Volta 20. Fo, Upper Volta 21. Seguenega, Upper Volta 22. W Senegal (T. gracilis only) (Linguere, Kaffrine) 23A. NW Senegal (T. gracilis) (Pete-Ole, Louga) 23B. NW Senegal (T. pygargus) (Pete-Ole, Debi, Richard Toll, Louga) 24. N'doulo, Senegal (T. gracilis only) 25A. E Senegal (T. gracilis) (Goudiry, Bakel, Koungheul, Koussanar) 25B. E Senegal (T, pygargus) (Goudiry, Koungheul, Koussanar, Bakel, Ranerou) 26, W Senegal (T. pygargus only) (Linguere, Kaffrine) 27A, SW Senegal (T. gracilis) (Saboya, Bandia, Thies, Kaolach, Dakar) 27B, SW Senegal (T, pygargus) (Bandia, Ndoulo, Thies) 28. Tiguent, Mauritania 29. NE Senegal (T. gracilis only) (Cascas, Ogo, Ranerou) 30. Boussouma, Upper Volta (= C Upper Volta) 31. NW Ghana (Pirisi, Bangwon, Lawra, Wa) 32. SW Ghana (Sakpa, Yabrasso) 216 33. Mora, Cameroon (one specimen, 2N=28) 34. Faradje, Zaire 35. Chad, karyotyped 2N-54 (Deli, Moundou) 217 Greatest Length of Skull N Mean SE CV Range 1 11 36 . 4 .190 1.6 35.4-37. 5 2 13 36. 2 . 101 1.0 35.7-36.9 3 5 35.0 .321 1.8 34.4-36. 0 4 8 35.3 . 290 2 . 2 34.5-36. 6 5 9 35.2 .183 1.5 34.5-36. 1 6 6 36.3 . 234 1.4 35.4-36. 8 7 8 36.4 . 333 2.4 35.1-37. 8 8 8 35. 5 . 208 1. 6 34.8-36. 2 9 10 35.1 . 125 1.6 33.2-36. 4 10 11 34.5 .171 1.6 33.7-35. 2 11 16 35.1 .157 1.7 34.1-36. 6 12 7 35.5 .356 2.4 34.5-36. 6 13 12 34.8 . 149 1.4 34.0-35. 6 14 7 34.6 . 182 1.3 34.1-35. 3 15 6 34.4 .150 1.0 34.1-34.9 16 11 34.9 .192 1.7 34.1-36 .1 17 8 34.5 . 142 1.1 34.0-35.1 18 17 34.3 .150 1.8 33.4-35. 4 19 23 34. 4 . 101 1.4 33.6-35. 3 20 11 34.5 . 166 1.5 33.5-35. 2 21 43 34.7 . 077 1.4 33. 5-35. 8 22 ' 10 35.2 . 271 2.3 33 . 8-36. 2 2 3A 15 35.1 .140 1.5 34.3-36. 3 23B 13 34. 9 .182 1.8 34.0-36. 3 24 10 35.6 .140 1.2 35.0-36. 4 2 5A 8 34. 5 . 200 1.5 34.0-35.7 25B 10 34. 4 . 203 1.8 33.7-35. 3 26 9 34.9 .194 1.6 33.9-35. 6 2 7A 12 35.6 . 221 2.1 34*. 4-36. 7 27B 14 35. 1 .144 1.5 34,0-35. 6 28 8 35.6 . 191 1.4 34.9-36. 3 29 8 34,6 . 171 1.3 34.0-35. 2 30 12 34. 9 .195 1.5 33. 5-36, 3 31 10 34,4 . 136 1.1 33.5-35. 6 32 8 34.5 . 231 1.6 33.8-35. 4 33 1 35.2 — — -- 34 8 36.5 . 234 1.7 35.8-37,4 35 10 37.1 .135 1.1 36.3-37,6 218 Occipitonasal Length N Mean SE CV Range 1 11 35.4 .187 1.7 34. 2-37. 5 2 13 35.1 .114 1.1 34 . 4-35. 9 3 5 33 . 8 .328 1.9 33.1-34. 7 4 8 34.2 .327 2 . 5 33.0-35. 3 5 9 34.0 .187 1.6 33.2-34. 8 6 6 35.3 .293 1.9 34.0-35. 9 7 8 35.4 .382 2.9 33.9-37. 0 8 8 34.2 .148 1.1 33 .7-34. 9 9 10 33.9 . 293 2.5 32.0-34. 9 10 11 33,2 . 221 2.1 32.0-34.1 11 16 33.9 .140 1.6 32.7-34. 8 12 7 34.3 . 332 2.4 33.3-35. 3 13 12 33.5 .162 1.6 32.6-34. 2 14 7 33.2 . 215 1.6 32.6-34.0 15 6 33.3 .138 0.9 33.0-33. 8 16 11 33 . 6 .188 1.8 32.7-34. 7 17 8 33.1 .129 1.0 32.7-33. 7 18 17 32.9 .131 1.6 32.1-33. 8 19 23 33.1 .108 1.5 31.9-34. 0 20 11 33.3 .159 1.5 32.5-34. 3 21 43 33.5 . 085 1.6 32.4-34. 6 22 10 34.0 . 296 2 . 6 32.6-35. 2 2 3A 15 33 . 8 .164 1.8 33 . 2-35. 2 23B 13 33.6 . 211 2.2 32.4-35. 2 24 10 34.3 . 202 1.8 33.4-35. 2 2 5A 8 32.3 . 227 1.8 32.6-34. 5 25B 10 33. 2 . 209 1.9 32.5-34. 2 26 9 33.7 .198 1.7 32.6-34. 5 27A 12 34.5 .219 2.1 33 . 6-35. 5 27B 14 33, 9 .135 1.4 32.8-34. 4 28 8 34.3 .231 1.8 33 .4-35. 3 29 8 33,3 .165 1.3 32.6-34. 0 30 12 -3,9 .215 2.3 32.4-35. 4 31 10 33,1 .247 2.5 32.4-34. 6 32 8 33,1 .176 1.5 32,5-34, 0 33 1 34.2 — — — 34 8 35.2 .320 2.4 34.3-36. 3 35 10 36,0 .185 2.6 35.2-36. 4 219 Basilar Length N Mean SE CV Range 1 11 26. 7 . 188 2.2 25.6-27.6 2 13 26.9 . 120 1.5 26. 2-27.5 3 5 25.2 .221 1.8 24.6-25.8 4 8 25.5 . 235 2.4 24. 7-26.5 5 9 25.6 . 205 2.3 24.8-26.6 6 6 26.7 . 210 1.8 26.1-27.1 7 10 26.7 . 293 3.3 25.5-28.2 8 8 25.7 .164 1.7 25.1-26.6 9 9 25.5 . 295 3.4 23.8-27.0 10 11 24.7 .144 1.8 24.0-25.3 11 16 25.3 . 107 1. 6 24.1-25.8 12 7 25.5 . 234 2.3 24.8-26 . 2 13 12 25.2 . 155 2.0 24. 4-26 .1 14 7 24.8 . 230 2.3 24.1-25.8 15 6 25,1 . 178 1.6 24.3-25.4 16 11 25,1 . 242 3 . 0 24.1-26.5 17 8 24.6 .168 1.8 23.9-25.3 18 17 24 . 4 .107 1.8 23.8-25.2 19 23 24.8 .114 2.2 23.5-25.7 20 11 24.8 . 133 1.7 24.1-25.6 21 43 25.4 .068 1.7 24. 5-26 . 2 22 10 25.2 . 256 3.1 24.0-26 . 2 2 3A 15 25.1 . 211 3.2 24.1-26.6 2 3B 13 25.2 . 221 3 . 0 23.8-26.6 24 10 25,4 . 191 2.3 24. 3-26 . 2 2 5A 8 24. 8 .187 2 . 0 24. 3-25 . 2 2 5B 10 24.7 . 217 2.6 23 .9-25. 6 26 9 25,0 . 162 1.8 24.3-25.7 27A 12 25.6 . 225 2.9 24,7-26.7 27B 14 25. 1 .149 2.1 24. 1-26. 1 28 8 25.7 .192 2. 0 25.2-26.5 29 8 24.4 .197 2.1 23.8-25.4 30 12 25.6 . 206 2.7 24.8-26,6 31 9 25.0 . 241 3 .0 24. 3-26 . 4 32 7 24. 6 . 215 2.1 23.9-25.5 33 1 26.1 — — — 34 8 26.7 .244 2.4 25.9-27,7 35 10 27,3 .128 1.6 26 .6-27 . 8 220 Zygomatic Breadth N Mean SE CV Range 1 4 18.2 .425 4.1 17.1-18. 7 2 8 18.3 .130 2.2 17.4-19.0 3 4 17.7 . 233 3 . 3 17. 3-18. 2 4 7 17. 6 . 127 1.8 17.2-18. 2 5 4 17.4 .458 4.6 16.9-18. 6 6 4 18.6 .126 0.7 18.4-18. 7 7 8 18.1 . 232 3 . 4 16.8-18. 8 8 2 17. 5 .141 0.8 17.4-17. 6 9 5 17. 0 . 165 2.6 16.1-17. 4 10 8 16. 7 . 067 1.1 16. 5-17. 0 11 10 17.4 . 229 3.9 16.2-18. 3 12 4 17.7 . 175 3.2 17.1-18. 1 13 9 17. 7 .106 1.7 17.1-18.1 14 3 17.6 . 286 2.3 17.2-18. 0 15 2 17.1 .141 0.8 17.0-17. 2 16 10 17.4 . 147 2.5 16 . 7-18.0 17 8 16.9 .106 1.7 16.5-17. 3 18 15 17.1 .103 2.2 16.2-17. 6 19 17 17.3 . 072 1.7 16.7-17.9 20 11 17. 2 . 124 2.3 16.6-17.9 21 20 17. 4 . 068 1.7 16.9-18. 0 22 6 17. 1 . 245 2.4 16.5-17. 6 23A 10 17. 0 . 107 1.8 16.7-18. 3 23B 10 17. 4 . 208 3 . 6 16.6-18. 3 24 8 17. 5 .182 2. 8 16.8-18. 2 25A 7 17.1 .194 2.9 16.4-18. 0 25B 4 16.9 .159 1.6 16.6-17. 2 26 7 17,0 . 208 3. 0 16 . 2-17. 8 27A 11 17. 4 . 215 3.1 16.3-18. 7 27B 11 17.6 .180 3.2 16.5-18.4 28 8 17.6 . 071 1.1 17 . 4-17. 8 29 7 16. 8 .137 2.0 16.3-17. 2 30 6 17. 5 . 242 2.3 17.1-18. 0 31 6 17.4 .227 2.1 17.0-18. 0 32 5 17. 0 .137 1.5 16.7-17. 2 33 1 17.3 — — — 34 5 18. 2 . 216 2.4 17.5-18. 5 35 3 18.6 . 270 2.1 18.4-19.0 221 Condylobasilar Length Mean SE CV Range 28.4 .191 2.1 27.4-29.5 28.6 .137 1.7 27.9-29.3 26.9 .273 2.0 26.3-27.6 27.3 .219 2.1 26.4-28.1 27.4 .185 1.9 26.4-28.2 28.4 .187 1.5 27.8-28.8. 28.5 .280 3.0 27.1-30.0 27.5 .170 1.6 27.0-28.4 27.1 .180 1.8 26.4-28.0 26.5 .198 2.4 25.5-27.7 27.0 .119 1.7 25.7-27.5 27.3 .265 2.4 26.5-28.4 27.0 .149 1.8 26.2-27.9 26.5 .250 2.3 25.8-27.6 26.6 .183 1.5 25.9-27.0 26.7 .176 2.3 25.8-27.8 26.1 .152 1.5 25.4-26.7 26.2 .114 1.7 25.6-27.1 26.5 .106 1.9 25.3-27.4 26.4 .142 1.7 25.6-27.2 27.4 .073 1.7 26.6-28.3 27.0 .241 2.7 25.7-27.9 26.9 .178 2.5 26.1-28.2 26.9 .197 2.5 25.6-28.1 26.9 .151 1.7 25.9-27.4 26.6 .183 1.8 26.1-27.4 26.5 .228 2.6 25.6-27.4 26.8 .157 1.7 26.1-27.5 27.4 .229 2.8 26.4-28.5 26.8 .205 2.8 25.2-27.9 27.5 .151 1.5 27.1-28.2 26.3 .203 2.0 25.6-27.3 27.4 .233 2.6 26.1-28.5 26.7 ,247 2.8 25.7-28,1 26,3 ,225 2,1 25,4-27.1 27.5 28.6 .248 2.3 27.8-29.7 29.1 .120 1.2 28,5-29,6 222 Cranial Breadth N Mean SE CV Range 1 11 15.2 . 136 2.8 14.7-16.0 2 13 14.9 . 052 1.2 14.6-15.1 3 5 14.9 . 244 3.3 14.2-15.5 4 8 14. 8 .^41 2.5 14.4-15.4 5 9 15.1 . 108 2. 0 14.6-15.4 6 6 14. 8 .105 1.6 14.4-15.0 7 9 15.0 . 148 2.8 14.4-15.9 8 8 14. 4 . 116 2.1 13.9-14.9 9 10 14. 4 . 128 2.5 13.9-15.1 10 11 14.0 .057 1.3 13.6-14.2 11 16 14.7 .082 2.2 14.1-15.3 12 7 14.7 .135 2.2 14.3-15.1 13 12 14.7 . 090 2.0 14.2-15.2 14 7 14. 7 . 112 1.9 14.2-15.0 15 6 14. 6 .115 1.8 14.4-15.1 16 11 14. 4 .088 1.9 14.1-14.9 17 8 14. 5 .103 1.9 14.1-14.8 18 17 14. 4 .082 2.3 13.9-15.0 19 23 14. 5 .073 2.4 14.0-15.3 20 11 14. 4 . 061 1.3 14.1-14.7 21 43 14. 5 . 048 2.1 13.9-15.3 22 10 14. 0 . 081 1.7 13.6-14.3 23A 15 14.1 . 090 2.4 13.5-14.5 23B 13 14.1 . 091 2.2 13.7-14.7 24 10 14.1 .093 2.0 13.6-14,6 2 5A 8 13. 8 .117 2.2 13.4-14.3 25B 10 14.0 . 082 1.7 13. 5-14.4 26 9 14, 2 .117 2.3 13.7-14.6 21A 12 14. 3 .107 2.5 13.8-15.0 27B 14 14. 2 .092 2. 3 13.7-14.8 28 8 14.4 .144 2.7 13.6-14.7 29 8 14,0 . 110 2.1 13.5-14.4 30 12 14.5 . 096 2.1 14.1-14.9 31 9 14. 5 . 074 1.6 14.3-14.8 32 8 14. 4 .102 1.8 14.1-14.8 33 1 14.1 — — — 34 8 15.1 .144 2,5 14.6-15.7 35 10 14. 8 .083 1.8 14,6-15.3 223 Least Interorbital Constriction N Mean SE CV Range 1 11 6.5 . 063 3.1 6.2-6.8 2 13 6.1 . 030 1.7 6.0-6.3 3 5 6.4 . 117 3.7 6.2-6 . 7 4 8 6. 5 . 087 3.5 6.1-6.8 5 9 6.0 . 068 3. 2 5.8-6.4 6 6 6.3 . 089 3.2 6.0-6.5 7 9 6.4 .093 4.1 6.0-6 . 7 8 8 6.0 . 101 4.5 5.6-6.5 9 9 6.2 .095 4.8 5.8-6. 7 10 11 5. 8 .044 2.4 5.6-6.1 11 16 6.1 .089 5. 6 5.6-6.9 12 7 6.1 . 116 4.7 5.8-6.6 13 12 6. 0 . 058 3 . 2 5.7-6.3 14 7 6. 0 .128 5.2 5.4-6.4 15 6 6.0 .067 2.5 5.8-6.1 16 11 6.2 , 043 2.2 5.9-6.4 17 8 6.0 . 071 3.1 5.8-6.4 18 17 6.1 . 053 3 . 5 5.7-6.5 19 23 6.1 . 053 4.1 5.7-6.6 20 11 6 . 0 .052 2.7 5.8-6.3 21 43 6.2 . 035 3.7 5.7-6.9 22 10 6.0 . 092 4. 6 5.6-6.6 23A 15 6.1 . 065 4.0 5.6-6.6 23B 13 6.1 , 083 4.7 5.5-6.4 24 10 6.0 .098 4.9 5.6-6.6 2 5A 8 5.8 . 042 1.9 5.7-6.0 2 5B 10 5.9 . 082 4.1 5.5-6.4 26 9 6.0 . 116 5.5 5.4-6.4 27A 12 6.1 .085 4.6 5.7-6.7 27B 14 6.0 , 048 2.9 5.7-6.4 28 8 6,1 .114 5.0 5.8-6.6 29 8 5, 9 , 074 3. 3 5.5-6.1 30 12 6 . 2 . 082 3.9 5.8-6.6 31 10 6.2 .075 3.7 5.8-6.5 32 8 6.0 , 073 3.2 5.7-6.3 33 1 5.9 — -- — 34 8 6.4 . 102 4.2 6.1-6.8 35 10 6.0 . 083 3.5 5.7-6.4 224 Rostral Breadth N Mean SE CV Range 1 11 4. 9 .082 5.3 4.6-5.4 2 13 4.9 .035 2.5 4.7-5.1 3 5 4.4 .134 6.1 4.1-4.7 4 8 4.3 . 035 2.1 4.2-4.3 5 9 4.4 .054 3. 5 4.2-4.7 6 6 5.0 . 080 3.6 4.8-5.3 7 10 4. 7 . 088 5.7 4.2-5.1 8 8 4.4 .101 6. 0 4.1-4.9 9 10 4.7 . 075 3.6 4.4-5.0 10 11 4.5 .042 3 .0 4.3-4.7 11 16 4.5 .062 5,3 4.1-4.9 12 9 4.5 . 110 5 . 9 4.3-5.4 13 12 4.5 . 055 4.1 4.1-4.8 14 7 4.5 . 071 3 . 8 4.2-4.7 15 6 4.5 . 091 4.5 4.2-4 . 8 16 11 4.6 . 070 4.9 4.2-4.9 17 8 4.5 . 043 3 . 2 4.2-4.6 18 17 4.7 . 040 3 . 3 4.5-5.1 19 23 4.4 . 040 4.2 4.1-4. 9 20 11 4.6 .091 6.3 4.1-4.9 21 43 4.6 . 028 4.0 4.2-5.0 22 10 4.4 .077 5.3 4.1-4.7 23A 15 4.3 . 033 2.9 4.1-4.5 23B 13 4.4 . 045 3. 5 4.1-4.6 24 10 4. 5 . 051 3 . 4 4.3-4.7 2 5A 8 4.4 . 057 3.5 4.3-4.6 2 5B 10 4.4 . 054 3.7 4.1-4.7 26 9 4.4 . 082 5.3 4.0-4.7 2 7A 12 4.4 .061 4.6 4.1-4.6 27B 14 4.4 . 045 3.7 4.2-4.7 28 8 4.6 . 049 2.9 4.4-4.8 29 8 4.4 .083 5.0 4.1-4.7 30 12 4.7 .080 5.4 4.4-5.1 31 10 4.6 .078 5.3 4.3-4.9 32 8 4.6 .041 2.2 4.5-4. 7 33 1 4.4 — — — 34 8 4.9 .037 2.0 4.7-5.0 35 10 4.7 .065 4.2 4.5-5.0 225 Greatest Breadth Across the Bullae N Mean SE CV Range 1 11 13 .9 .121 2.7 13.3-14.5 2 13 13.9 . 076 1.9 13.5-14.0 3 5 13. 9 . 227 3 . 3 13.4-14.5 4 8 14.1 .184 3 . 5 13.6-14.8 5 9 14 . 2 . 185 3.7 13.5-15.0 6 6 13.6 . 145 2.4 13.2-14.0 7 10 14.1 .118 2.5 13.5-14.7 8 8 13.6 . 203 3 . 9 12.9-14.4 9 9 13 . 1 . 130 2. 6 12.7-13.9 10 11 13 . 5 .082 1.9 13.0-13.9 11 16 13.7 . 075 2.1 13,2-14.3 12 7 13 . 6 . 141 2.2 13.3-14.0 13 12 13. 5 .092 2.2 13.1-14.0 14 7 13 . 4 .077 1.4 13.1-13.7 15 6 13.2 .080 1.4 13.0-13.5 16 11 13. 4 . 137 3 . 2 12.7-14.0 17 8 13 . 4 . 157 3.1 12.7-14.1 18 17 13.5 . 078 2.3 12.8-14.2 19 23 13. 7 .056 1.9 13.1-14.2 20 11 13 . 3 .129 3.1 12.8-14.1 21 43 13. 8 .065 3.0 13.0-14.7 22 10 13.4 . 118 2.6 13.0-14.0 23A 15 13. 3 .080 2. 3 12.8-14.1 23B 13 13. 6 .122 3.1 13.0-14.3 24 10 13. 4 .103 2.3 13.0-14.0 2 5A 8 13.0 .118 2.4 12.5-13.5 2 5B 10 13. 4 . 079 1.8 13.0-13.7 26 9 13.6 .136 2.8 13.0-14.1 27 A 12 13 . 5 . 100 2.4 13.1-14.2 27B 14 13.4 . 108 2.9 12.7-14.1 28 8 13. 4 .109 2.2 12.8-13.8 29 8 13.4 .129 2.6 12.8-13.7 30 12 13. 8 .107 2.5 13.4-14.6 31 9 13. 1 . 120 2.6 12.7-13.6 32 6 13,8 . 090 1.7 13.5-14.1 33 1 13,6 — -- — 34 8 13.7 . 172 3.3 13.1-14.4 35 10 14.4 .120 2.7 14.0-15.0 226 Greatest Width Across the Upper Molars N - Mean SE CV Range 1 11 7.4 . 069 2.9 7.1-7.8 2 13 7.6 .066 3.0 7.5-7.9 3 5 7.0 .045 1.3 6 .9-7 .1 4 8 6.9 .059 2. 2 6.7-7.1 5 9 7.1 . 049 2. 0 6 . 8-7.2 6 6 7 .6 . 105 3 . 1 7.3-7 . 9 7 9 7.5 . 117 4.5 6.9-7.8 8 8 7.1 .056 2.1 6.9-7 . 4 9 8 7.3 . 085 2.8 7.0-7.6 10 11 7 . 1 . 049 2 . 2 6.8-7.3 11 16 7.2 .059 3 . 2 6.8-7.6 12 7 7 . 0 . 068 2.4 6.7-7.2 13 12 7 . 2 . 079 3 .6 6 . 8-7.6 14 7 7 . 1 . 122 4 . 2 6 . 9-7. 7 15 6 7.2 . 044 1.4 7.0-7. 3 16 11 7.2 . 055 1.8 7.0-7.4 17 8 7. 1 .103 3 . 8 6.7-7.6 18 17 7.2 . 052 2.9 6.6-7.5 19 23 7.1 . 041 2.7 6.8-7.6 20 11 7.2 .091 4.0 6. 8-7.6 21 43 7 . 2 .033 3.0 6.7 7.8 22 10 7.0 .072 3.1 6.7-7.4 23A 15 7.1 . 031 1.6 6.9-7. 3 23B 13 7.2 . 071 3 . 5 6.8-7.7 24 10 7.1 .064 2.7 6.8-7.3 2 5A 8 6 .9 . 064 2.4 6 . 7-7 . 1 25B 10 7.1 . 108 4 . 6 6.7-7.7 26 9 7.1 . 036 1.4 6.9-7.2 27A 12 7.2 . 073 3.4 6.6-7.5 27B 14 7.3 .052 2.6 7.0-7.6 28 8 7.1 . 076 2.8 6.7-7.3 29 8 6.9 .092 3.6 6.6-7.3 30 12 7.3 .071 3.2 6.9-7.7 31 9 7.1 .053 2. 2 6.9-7.5 32 8 7.0 , 103 4.0 6.5-7.3 33 1 7.2 -- — -- 34 8 7.7 .070 2. 4 7.5-7.9 35 10 7,6 .066 3. 0 7.4-8.1 227 Length of Diastema N Mean SE CV Range 1 11 8.9 . 122 4.3 8.4-9.7 2 13 8.7 .109 4.4 8.2-9.1 3 5 7.9 .150 3 . 8 7.4-8.2 4 8 8.1 .103 3 . 4 7.8-8.6 5 9 8. 2 .110 3.8 7.6-8.6 6 6 8. 8 . 077 2.0 8.6-9 . 0 7 10 8. 8 . 140 4 . 8 8.2-9.5 8 8 8,4 ,074 2.3 8.1-8.6 9 10 8. 2 . 120 3.9 7.4-9.2 10 11 8. 0 .083 3.3 7.7-8.6 11 16 8.1 .076 3.7 7.4-8.7 12 7 8.2 . 136 4 .1 7.8-8.7 13 12 8.1 .089 3 . 6 7.5-8.5 14 7 7 . 8 .128 4.0 7.2-8. 1 15 6 8.1 . 100 2.7 7.7-8.3 16 11 8.1 .088 3. 4 7.8-8. 6 17 8 8.1 . 140 4.6 7.5-8. 7 18 17 7 . 9 .067 3.4 7.4-8.3 19 23 8.1 .063 3.6 7.5-8.7 20 11 8.1 .059 2. 3 7.8-8.4 21 43 8.7 .043 3 . 2 8.2-9.3 22 10 8.3 .084 3. 0 8.0-8. 8 23A 15 8. 2 .097 4.5 7.7-8. 5 23B 13 8.1 . 104 4 . 4 7.4-9.0 24 10 8. 3 . 079 2.9 7.9-8.6 25A 8 8.1 .128 4.2 7.9-8.7 2 5B 10 8,1 .138 5.1 7.6-9.0 26 9 8.1 . 091 3. 2 7.9-8 , 7 27A 12 8 , 2 . 114 4 . 6 7.7-9.1 27B 14 8.1 ,110 4 . 9 7.6-8.9 28 8 8.9 .121 3 . 6 8,6-9.6 29 8 8.1 .127 4.1 7.7-8.6 30 12 8.4 ,081 3 . 2 8.0-8,8 31 10 8.1 , 102 3 , 0 7.7-8.7 32 8 8.1 ,138 4.5 7.7-8.9 33 1 8.5 — — 34 8 9,0 .156 4.6 8.7-9.9 35 10 9,2 ,067 1,9 9.0-9.4 228 Palatilar Length N Mean SE CV Range 1 11 15.6 . 178 3 . 6 14.8-16.4 2 13 15.5 .082 1.8 15.2-16.0 3 5 14.1 .187 2 . 7 13.6-14.6 4 7 14. 3 . 136 2.3 13.8-14.7 5 9 14.4 . 134 2.6 13.7-14.8 6 6 15. 5 . 096 1.4 15.3-15.8 7 10 14.9 . 266 5.3 13.4-16.2 8 8 14. 5 . 125 2.3 14.0-15.0 9 10 14.8 . 235 5.2 13.7-16.3 10 11 14.1 .091 2 .0 13.7-14.6 11 16 14.6 . 078 2.1 14.1-15.3 12 7 14.7 .235 3 . 9 13.8-15.6 13 12 14 . 5 . 127 2 . 9 13.8-15.3 14 7 14,1 . 182 3.2 13.6-14.9 15 6 14.7 .105 1.6 14.3-15.0 16 11 14.6 .130 2. 8 14.2-15.5 17 8 14.2 .154 2.9 13.7-14.8 18 17 14.1 .059 1.7 13.8-14.5 19 23 14.3 .069 2.3 13.6-14.8 20 11 14.3 .074 1.6 13.9-14.7 21 43 14.4 , 042 1.9 13.8-15.1 22 10 14.5 . 125 2.6 13.8-14.9 23A 15 14.4 .108 2.8 13.8-15.2 23B 13 14. 4 .117 2.8 13.7-15.2 24 10 14. 5 .064 1.3 14.2-14.7 2 5A 8 14. 4 .162 3 .0 14.1-15.3 2 5B 10 14.3 . 209 4.4 13.6-15.7 26 9 14.3 . 117 2.3 13.9-14.9 2 7A 12 14.6 . 180 4.1 13.8-16.0 27B 14 14.4 . 128 3.2 13.5-15.4 28 8 15.0 . 092 1.6 14.6-15.3 29 8 14.1 . 195 3.7 13.5-15.0 30 12 14.8 . 105 2.3 14.1-15.5 31 9 14.3 .121 2.4 13.9-15.1 32 8 14,2 .136 2.5 13.7-14.7 33 1 14.7 -- — — 34 8 15.6 . 138 2.3 15.2-16.0 35 10 15,8 .135 2.4 15.5-16.2 229 Postpalatal Length N Mean SE CV Range 1 11 10 .9 .086 2.5 10.5-11.3 2 13 10.9 . 092 2.9 10.4-11.4 3 5 10. 8 .185 3.4 10.2-11.2 4 7 10.9 .160 3 .6 10.4-11.6 5 9 10.9 .125 3 . 3 10.5-11.5 6 6 11.0 . 159 3. 2 10.5-11.4 7 10 11.3 .149 4. 0 10.6-11.9 8 8 10. 8 .111 2.7 10.2-11.2 9 9 10.4 .130 2.6 9.8-10.8 10 11 10. 4 .096 2.9 9.8-10.9 11 16 10.4 .086 3. 2 9.3-10.9 12 7 10. 8 . 087 2.3 10.5-11.1 13 12 10.3 .098 3.2 10.0-11.0 14 7 10.3 .064 1.5 10.1-10.6 15 6 10. 2 .097 2.1 9.9-10.5 16 11 10. 2 . 121 3.7 9.7-10.8 17 8 10.2 .090 2.4 9.7-10.4 18 17 10. 0 .065 2.6 9.6-10.5 19 23 10.3 . 067 3.1 9.6-10.8 20 11 10.3 .099 3.1 9.9-11.1 21 43 10. 2 . 049 3.1 9.7-10.9 22 10 10.4 .162 4.7 9.7-11.3 23A 15 10. 4 .128 4.6 9.4-11.3 23B 13 10.4 .117 3,9 9.7-11.3 24 10 10. 5 .107 3.1 10.0-11.0 25A 8 10.2 .080 2.1 9.9-10.5 2 5B 10 10,3 .142 4.1 9.7-11.0 26 9 10. 4 .113 3.1 10.0-10.9 27A 12 10.6 .110 3. 4 10.3-11.0 2 7B 14 10.4 .052 1.8 10.1-10.9 28 8 10.6 .149 3,7 10.0-11.0 29 8 10,2 .098 2. 6 9.7-10.4 30 12 10. 5 .123 3. 6 10.0-10.9 31 8 10.4 .136 3.8 9.7-11.2 32 7 10.1 .170 4.0 9.7-10.6 33 1 10.5 — -- __ 34 8 10. 8 .180 4.4 10.1-11.3 35 10 11. 0 .092 2,3 10.8-11,4 230 Length of Anterior Palatine Foramina N Mean SE CV Range 1 11 6 .6 .107 5.2 6.0-7.1 2 13 6.5 .078 4 . 1 6.3-7.0 3 5 5.7 . 079 2,8 5.5-5.9 4 8 5. 6 .079 3 . 8 5.2-5.8 5 9 5.6 . 066 3.3 5.3-5.9 6 6 6.5 .073 2.5 6.2-6.7 7 10 6.4 .121 5.7 5.9-6.8 8 8 5.8 .067 3.1 5.5-6.0 9 10 5.9 .135 5.8 5.4-6.7 10 11 5. 6 . 042 2.4 5.3-5.8 11 16 5.8 -073 4.9 5.4-6.4 12 7 5.8 .120 5,1 5.3-6.5 13 12 5.7 .076 4.4 5.3-6.0 14 7 5.5 .081 3. 6 5.2-5.8 15 6 5.7 .083 3.3 5.4-5.9 16 11 5.7 . 082 4. 6 5.3-6.2 17 8 5.6 .074 3. 5 5.4-5.9 18 17 5.6 .046 3. 3 5.2-5.9 19 23 5.7 . 048 4.1 5.2-6. 0 20 11 5.8 . 074 4.1 5,5-6.2 21 43 5.8 .036 4.0 5.2-6.3 22 10 5.8 . 060 3.1 5.5-6 . 1 23A 15 5.8 .061 4.0 5.4-6.2 23B 13 5.7 .084 5,1 5.3-6.2 24 10 5.8 . 100 5.2 5.3-6.2 25A 8 5. 6 . 082 3.9 5.3-6.0 2 5B 10 5.8 . 120 6.2 5.4-6.6 26 9 5.6 .061 3 .1 5.4-5.9 21A 12 5,7 . 072 4.2 5.3-6.2 27B 14 5.8 .077 4.8 5.3-6.2 28 8 5,9 .121 5.4 5,5-6. 5 29 8 5.6 ,161 7.6 5.0-6.2 30 12 5.7 , 078 4.4 5.2-6. 2 31 10 5.5 ,107 5,6 4,9-5.9 32 8 5,6 ,084 3 , 8 5.2-5.9 33 1 5,6 -- — — 34 8 7.1 ,090 3.3 6.9-7.6 35 10 6,5 ,086 4 , 5 6,2-7.0 231 Length of Palatal Bridge N Mean SE CV Range 1 11 7.7 .113 4.7 7.1-8.2 2 13 7.4 . 067 3,1 7.2-7,6 3 5 6,9 .108 3 . 2 6.6-7.1 4 7 6.9 .093 3.3 6.7-7.3 5 9 7.0 , 089 3. 6 6.7-7.4 6 6 7.6 .115 3.4 7,4-8.0 7 10 7.3 .138 5,7 6.8-8.2 8 8 7.2 .108 4 . 0 6.9-7.8 9 10 7.0 , 125 4.5 6.5-7,7 10 11 7.3 . 078 3.4 7.0-7.8 11 16 7.3 . 064 3.3 7.0-7. 9 12 7 7.4 . 133 4 . 4 6.7-7,7 13 12 7.3 .093 4.2 6.9-7,8 14 7 6.9 .072 2. 6 6.7-7.2 15 6 7.4 . 138 4.1 7.1-7.8 16 11 7.3 .085 3,7 6.8-7.7 17 8 7.2 , 070 2.6 6.9-7.5 18 17 7.2 , 045 2.5 6.9-7.6 19 23 7.1 , 042 2.8 6.7-7,4 20 11 7.1 . 075 3.3 6.6-7.4 21 43 7.1 , 038 3,4 6.5-7.7 22 10 7.1 .104 4.4 6.7-7.7 23A 15 6.8 .051 2,8 6.4-7.1 23B 13 6.9 .077 3.9 6.3-7.3 24 10 7.0 .060 2. 6 6.8-7.4 2 5A 8 7.2 .088 3.3 6.9-7.6 25B 10 7.1 .096 4.0 6.6-7,5 26 9 7.1 . 083 3.3 6,7-7.5 27A 12 7.2 . 089 4,1 6 . 8-7 .9 27B 14 6,9 .108 5.6 6.4-8.0 28 8 7.3 , 089 3 . 2 7.0-7,7 29 8 7.0 .086 3 , 2 6.6-7.4 30 12 7.3 .083 3.9 6.8-7.8 31 9 7.0 . 082 3 . 2 6.6-7.4 32 8 7.1 .087 3.1 6.8-7.5 33 1 7.0 __ — — 34 8 7.3 .135 4.9 6.9-7.8 35 10 7.5 .108 4.3 6.8-7.8 232 Length of Auditory Bulla N Mean SE CV Range 1 11 9.7 .126 4.1 9 . 2-10. 6 2 13 9. 6 . 069 2.5 9.1- 9. 8 3 5 9.7 . 147 3 .0 9.4-10. 0 4 8 9.9 .093 2. 5 9.6-10. 3 5 9 9.8 . 153 4.4 9.3-10. 5 6 6 9. 7 . 131 3.0 9.4-10. 2 7 10 10.1 . 106 3 . 2 9.6-10. 7 8 8 9.5 .099 2.7 9.0- 9.8 9 10 9.2 . 067 1.8 8.7- 9. 4 10 11 9.3 . 091 3.1 8.9- 9.7 11 16 9.3 .069 2.9 8.7- 9.7 12 7 9.3 . 109 2.9 8.8- 9. 6 13 12 9.0 .079 2.9 8.6- 9. 5 14 7 9.3 .105 2.8 8.9- 9. 5 15 6 9.0 .104 2.6 8.8- 9. 3 16 11 9.2 . 050 1.7 8.9- 9. 4 17 8 9.1 .070 2.0 8.8- 9. 3 18 17 9.0 .076 3.4 8.5- 9.5 19 23 9.3 .050 2.5 8.9- 9. 8 20 11 8.9 . 062 2.2 8.7- 9. 3 21 43 9.2 .039 2.8 8.5- 9.7 22 10 9.2 .077 2.5 8.8- 9. 5 23A 15 9.2 .070 2.8 8.8- 9.7 23B 13 9.3 . 085 3.2 8.7- 9.7 24 10 9.4 . 078 2.5 9.1- 9.7 2 5A 8 9.1 .097 2.8 8.8- 9.4 25B 10 9.2 . 121 3.9 8.6-10. 0 26 9 9.4 .089 2.7 8.9- 9.7 27A 12 9.3 .106 3. 8 8.5- 9.7 27B 14 9.3 .055 2,1 8.9- 9. 5 28 8 9.4 . 082 2.3 9.0- 9.7 29 8 9,3 .065 1.9 9.1- 9. 6 30 12 9.1 . 067 2.3 8.9- 9.4 31 10 9.0 .081 2.6 8.5- 9.3 32 8 9.1 . 077 2.1 8.8- 9.4 33 1 8.9 — -- — 34 8 9.7 . 123 3.4 9.1-10.1 35 10 9.4 . 074 2.4 9,1- 9. 8 233 Width of Auditory Bulla N Mean §1 CV Range 1 11 5.5 . 072 4.1 5.3-6.0 2 13 5.3 .054 3. 5 5.1-5.5 3 5 5. 6 .110 3.9 5.3-5.9 4 8 5.6 .109 5.1 5.3-6.1 5 9 5. 8 .094 4.5 5.3-6. 2 6 6 5.5 . 122 5 .0 5.2-5. 9 7 10 5.7 .075 3.9 5.4-6.0 8 8 5.5 .106 5.1 5.1-5.9 9 10 5.2 . 065 3. 0 4.9-5. 4 10 11 5.3 .038 2.3 5.0-5. 5 11 16 5.2 .044 3 . 3 5.0-5. 6 12 7 5.4 .084 3. 8 5.2-5. 8 13 12 5.2 .048 3.1 4,9-5. 5 14 7 5.4 . 096 4.3 4.9-5. 6 15 6 5.4 .059 2.5 5.2-5. 6 16 11 5.3 .052 3.1 5.0-5. 6 17 8 5.4 . 056 2.7 5.2-5. 7 18 17 5.4 .062 4.6 5.0-5.9 19 23 5.6 . 043 3. 6 5.2-6. 0 20 11 5.5 .055 3.2 5.2-5. 8 21 43 5.5 .041 4.9 5.0-6.1 22 10 5.4 .060 3.3 5.1-5. 7 23A 15 5.3 .057 4.1 4.9-5. 6 23B 13 5.5 .064 4.0 5.2-5. 9 24 10 5.3 .042 2.4 5.1-5. 5 25A 8 5.3 . 060 3 . 0 5.0-5. 5 25B 10 5.4 .076 4.2 5.1-5. 8 26 9 5.6 . 043 2.5 5.4-5.7 21A 12 5.4 .041 2.5 5.1-5. 6 27B 14 5.3 .059 4.0 5.0-5.7 28 8 5.8 .084 3.9 5.4-6.1 29 8 5.4 .045 2.2 5.3-5. 6 30 12 5.5 .040 2.5 5,3-5. 8 31 10 5.5 .067 3,1 5.3-5. 8 32 8 5.4 . 052 2.6 5.2-5. 6 33 1 5.1 — — — 34 8 6.0 .067 3.0 5.7-6. 2 35 10 5.4 .074 4.1 5.2-5.9 234 Length of Nasals N Mean SE CV Range 1 11 15. 0 .104 2.2 14.6-15.5 2 13 14.7 .123 2.9 14.0-15.5 3 5 13.9 . 144 2.1 13.5-14.2 4 8 14. 1 .149 2 , 8 13.6-14.6 5 9 13.7 .137 2. 8 13.2-14.2 6 6 14.8 .164 2.5 14.3-15.3 7 8 14. 3 . 270 5.0 13.2-15.3 8 8 14.7 .159 2.9 14.0-15.2 9 10 13. 8 .153 3 .4 12.6-14.6 10 11 13 . 8 .157 3.6 12.9-14.6 11 16 14.1 .066 1.8 13.7-14.5 12 7 14.1 .134 2. 6 13.6-14.3 13 12 13. 6 .093 2.3 13.0-14.0 14 7 13,3 .115 2.1 12.8-13.6 15 6 13. 4 . 220 3.7 12.6-14.1 16 11 13. 8 . 080 1.9 13.3-14.2 17 8 13.6 .108 2.1 13.2-14.0 18 17 13. 6 .074 2.2 13.1-14.2 19 23 13 . 7 .081 2,8 12.8-14.5 20 11 13. 9 .129 2.9 13.2-14.4 21 43 13. 4 . 057 2.7 12,7-14.2 22 10 14.3 . 207 4.3 13.5-15.2 23A 15 14. 1 .095 2.5 13.4-15.0 23B 13 14.2 .102 2. 5 13.7-14.9 24 10 14. 5 .149 3.1 13.6-15.0 25A 8 13. 8 .096 1.8 13.4-14.3 25B 10 13.9 .161 3. 5 13.0-14.4 26 9 14.1 .156 3.1 13.4-14.8 21A 12 14.4 .097 2.2 13,8-15.0 27B 14 14.1 .075 1.9 13.7-14.5 28 8 14. 8 .165 2.9 14.1-15.3 29 8 13,9 ,136 2.6 13,2-14.3 30 12 14.3 .096 2.2 , 13.7-14.9 31 10 13. 8 , 238 4.5 12.8-14.9 32 8 13,7 .095 1.8 13.5-14,1 33 1 14,3 — — — 34 8 14. 8 .137 2.4 14.2-15.2 35 10 14. 6 .206 4.2 14.6-15.3 235 Length of Rostrum N Mean SE CV Range 1 11 13. 2 . 126 3 .0 12.6-14.0 2 13 13.0 .115 3.1 12.4-13. 8 3 5 12.7 .094 1.5 12.4-12.9 4 8 12. 8 . 163 3.4 12.2-13. 3 5 9 12. 4 .137 3.1 11.8-12. 9 6 6 13.3 . 046 0. 8 13.2-13. 5 7 8 13,1 .166 3.4 12.3-13. 8 8 8 12. 9 . 157 3.2 12.4-13. 5 9 10 12. 9 .134 3.1 11.8-13.3 10 11 12.5 .108 2.7 11.9-12.9 11 16 12.7 . 080 2,4 12.3-13. 5 12 7 12. 8 .178 3.4 12.3-13. 5 13 12 12.3 . 101 2.7 11.8-12. 8 14 7 11. 8 .120 2.5 11.5-12. 4 15 6 12. 1 . 040 0.7 12.0-12. 2 16 11 12.6 .082 2.1 12.1-13. 1 17 8 12. 3 . 079 1.7 12.1-12.7 18 17 12.3 .069 2.2 11,9-12.9 19 23 12.4 .066 2.5 11.8-13. 2 20 11 12.4 . 114 2.9 11.7-12.9 21 43 12. 5 . 051 2. 6 12.0-13.1 22 10 13.1 .145 3.3 12.5-13. 8 2 3A 15 12.9 .101 2.9 12.5-13.9 23B 13 12.8 . 102 2.8 12.2-13. 5 24 10 13.2 . 100 2.3 12.5-13. 6 2 5A 8 12, 5 .117 2.5 12.3-13.1 2 5B 10 12. 6 .126 3.0 12.1-13. 2 26 9 12.9 . 222 4.9 12.2-14. 0 2 7A 12 13,1 . 139 3.5 12.7-13. 6 27B 14 12.9 . 088 2,5 12.4-13. 6 28 8 13.3 . 170 3.4 12.7-14.1 29 8 12 . 126 2.6 12.2-13.1 30 12 12,6 .131 3.3 12.0-13. 6 31 10 12,3 .173 4.1 11.5-13.3 32 8 12.2 .120 2.4 11.9-12,7 33 1 13,1 — — -- 34 8 13. 2 .149 3.0 12.7-13.9 35 10 13 .9 .086 1.9 13.6-14. 5 236 Frontonasal Length N Mean SE CV Range 1 11 27.3 .177 2.1 16.6-18. 2 2 13 27.0 .178 2.3 26.2-28. 2 3 5 26.2 . 203 1.6 25 .6-26. 6 4 7 26 . 8 .303 2.8 25.9-28. 0 5 9 26. 2 . 202 2. 2 25.5-27. 2 6 6 27. 2 . 157 1.3 26.7-27. 5 7 8 27.3 .337 3.3 26 . 0-28. 7 8 8 26.7 . 163 1.6 26.1-27. 4 9 10 26 . 3 .186 2.4 24.9-27.1 10 11 25. 6 . 156 1.9 24.8-26.1 11 16 26.1 .148 2,2 25.2-27.1 12 7 26.4 .254 2.4 25.8-27. 5 13 12 26. 1 .126 1.6 25. 4-26. 6 14 7 25,6 .179 1.7 25.2-26. 4 15 6 25.5 .186 1.6 24 .9-26. 1 16 11 26 .0 . 155 1.9 25.3-26.7 17 8 25.7 .166 1.7 24.8-26. 2 18 17 25.6 .116 1.8 24.9-26. 5 19 23 26.0 .103 1.9 25.2-26.9 20 11 26 . 0 .190 2.3 24. 9-26 . 6 21 43 25.8 . 079 2.0 24. 5-27. 1 22 10 26 . 5 . 194 2.2 25.1-27.1 23A 15 26.6 . 125 1.8 25.8-27. 4 23B 13 26 . 4 . 173 2.3 25.2-27.1 24 10 26. 7 . 134 1.5 26 . 0-27. 3 2 5A 8 25 . 8 .187 1.9 2 5.2—26. 8 25B 10 26.2 . 196 2. 2 25.5-27.1 26 9 26.2 . 195 2.1 25.5-27. 0 21A 12 26.8 .138 1.7 25.9-27.5 2 7B 14 26. 4 .137 1.9 25.4-27.3 28 8 27.2 . 234 2.3 26.4-28. 3 29 8 26.1 . 206 2.1 25.1-26.7 30 12 26.4 . 158 1.8 25.2-27. 5 31 10 25. 8 . 195 2.7 24,9-27. 5 32 8 25.8 . 285 2.8 24.9-27. 0 33 1 26,3 -- — 34 8 27 , 4 . 172 1.7 26.7-28. 2 35 10 27 . 8 .165 1.8 27.3-28. 4 237 Depth of Cranium N Mean SE CV Range 1 11 14. 3 . 099 2.2 13.8-14. 8 2 13 14. 5 . 075 1.8 14.2-15.0 3 5 13. 7 .135 2.0 13.3-14.0 4 8 13. 8 .141 2.7 13.4-14. 4 5 9 13.8 . Ill 2.3 13 . 3-14. 2 6 6 14.3 .152 2.4 13.7-14. 7 7 8 14. 4 .220 4.0 13.3-15. 2 8 8 13. 8 . 137 2.6 13.1-14. 2 9 9 13.7 .167 3 . 2 ' 13.4-14. 6 10 11 13.5 .085 2.0 13 .1-14.0 11 16 13.8 .049 1.4 13.5-14.1 12 7 13. 7 . 227 4.0 13.2-14. 8 13 12 13. 7 . 084 2. 0 13.2-14. 2 14 7 13. 8 .062 1.1 13.7-14.1 15 6 13.6 .061 1.0 13.5-13. 9 16 11 13 . 8 .085 2.0 13.3-14. 2 17 8 13.7 . 144 2. 8 13. 2-14. 2 18 17 13. 6 . 059 1.7 13.2-14.0 19 23 13. 7 .078 2.7 13 . 0-14.3 20 11 13.7 . 084 1.9 13.4-14.1 21 43 13.7 . 038 1.8 13.3-14. 3 22 10 13. 3 .059 1.3 13.1-13. 7 23A 15 13. 5 . 090 2.5 12.9-14. 2 23B 13 13 . 6 . 093 2.4 13.2-14.1 24 10 13 . 8 .086 1.9 13.4-14. 2 2 5A 8 13,1 .121 2.4 12.7-13.7 25B 10 13 . 4 .090 2.0 13.0-13.9 26 9 13. 5 . 128 2.1 12.8-14.0 2 7A 12 13. 6 .043 1.1 13.4-13. 9 27B 14 13. 8 , 087 2.3 13.1-14. 3 28 8 13,6 .098 1.9 13,0-13. 8 29 8 13,4 .118 2,3 12.9-13.9 30 12 13. 8 .101 2.1 13,2-14. 4 31 10 13.4 . 085 1.9 13,1-13. 8 32 7 13.6 , 103 2.2 13.2-13. 9 33 1 13,5 -- — — 34 8 14,8 ,117 2.1 14,3-15.3 35 10 14.3 .113 2.5 13,9-14. 9 238 Length of Posterior Palatine Foramina N Mean SE CV Range 1 11 4. 2 . 073 5.5 3.7-4.5 2 13 3.7 . 074 6.9 3.4-4.1 3 5 3.4 . 201 11. 9 3 . 0-3 . 8 4 8 3.6 .123 9.2 2.9-4.0 5 9 3.5 .064 5.2 3.3-3.8 6 6 3 .9 .126 7 . 3 3.5-4.2 7 10 3,8 . 102 8. 0 3.2-4.2 8 8 3.7 . 099 7.1 3.3-4.0 9 10 3. 6 .052 3. 6 3.4-3.8 10 11 3.7 . 040 3 . 4 3.6-4.0 11 16 3.7 . 065 6 . 8 3.3-4.2 12 7 3. 8 .078 7.0 3.4-4.0 13 12 3,8 .081 7.1 3.6-4.5 14 7 3.8 . 028 1,8 3.7-3.9 15 6 3.9 .125 7 . 1 3.4-4. 2 16 11 3. 6 . 079 6.9 3.2-4.0 17 8 3.8 .084 5,9 3.4-4.1 18 17 3.7 .052 5,7 3.2-4.0 19 23 3.6 .043 5.6 3.3-4.0 20 11 3. 5 .097 8.7 3.1-4.0 21 43 3.8 . 042 7.3 3,1-4.5 22 10 3.7 .078 6 . 3 3,3-4. 1 23A 15 3.5 .056 5.9 3.1-3.9 23B 13 3.6 . 085 8 . 0 3.2-4.1 24 10 3.8 .097 7.6 3.5-4,3 25A 8 3.6 ,101 7.3 3.3-4. 2 25B 10 3. 6 .061 5.0 3.5-4.0 26 9 3.7 . 053 4,1 3.4-3.8 21K 12 3.7 , 085 7.5 3.2-4,1 27B 14 3.7 . 060 5.8 3,3-4.1 28 8 3.6 .121 8,8 3.2-4.2 29 8 3 , 6 .060 4.4 3.4-3.8 30 12 3 , 6 .076 8.4 3.0-4.2 31 8 3.8 . 098 6. 9 3.4-4.2 32 8 3 , 8 , 087 6.1 3.4-4.1 33 1 3.2 — __ — 34 8 4.1 , 070 4,5 3.7-4.3 35 10 3.9 .083 6.7 3.6-4.3 239 Alveolar Length of the First Molar N Mean SE CV Range 1 11 2.8 . 038 4.4 2.6-2.9 2 13 3 . 1 .051 5.8 2.8-3.3 3 5 2.5 .096 7.8 2. 2-2.7 4 8 2. 5 . 070 7 . 3 2.2-2.8 5 9 2.6 .035 3. 8 2.4-2.7 6 6 2,8 .066 5.2 2. 6-3.0 7 10 2.7 .069 7. 7 2.4-3. 0 8 8 2.8 . 052 5.0 2.6-3.0 9 10 2.8 . 055 3. 8 2.6-2.9 10 11 2.7 .034 4.0 2.5-2.8 11 16 2.7 . 039 5.7 2.4-3 . 0 12 7 2.8 .023 2.1 2.7-2.9 13 12 2. 8 . 047 5.6 2.6-3.1 14 7 2,7 . 064 5.8 2.5-3.0 15 6 2.7 . 078 6.4 2.5-3.0 16 11 2.7 .043 4.9 2.5-3. 0 17 8 2.7 .051 5.0 2.6-3 . 0 18 17 2. 8 .025 3. 6 2.6-3.0 19 23 2.7 .020 3.4 2.6-2.9 20 11 2.8 .030 3.4 2.6-2.9 21 43 2.6 .021 5.2 2.3-2.9 22 10 2.7 . 033 3.7 2.5-2.8 23A 15 2.7 . 035 4.7 2.5-2,9 23B 13 2.8 . 035 4.4 2.6-3.0 24 10 2.8 .025 2.6 2.7-2.9 25A 8 2.8 . 039 3.7 2,7-3.0 2 5B 10 2.8 .048 5.0 2.7-3.1 26 9 2.7 .067 2.4 2.6-2.8 27A 12 2.8 . 029 3.4 2.7-3.0 27B 14 2.7 . 041 5.4 2.4-3.0 28 8 2.7 . 081 8.1 2.4-3.1 29 8 2.7 . 032 3.1 2.6-2.8 30 12 2,8 .033 3 . 6 2.6-3.0 31 10 2.7 .049 4,9 2.5-2.9 32 8 2.7 .035 3 . 2 2.6-2.8 33 1 2.7 — — — 34 8 3.0 .052 4.7 2.8-3.2 35 10 2.8 .018 1.3 2.7-2.8 240 Length of Maxillary Cheek Teeth N Mean SE CV Range 1 11 5.5 .063 3.7 5.2-5.8 2 13 5.6 .077 4.8 5.3-6.0 3 4 5.1 .082 3.2 4.9-5.3 4 8 5.1 .083 4.3 4.7-5.3 5 9 5.1 .031 1.7 3.0-5.3 6 6 5.6 .083 3.3 5.4-5.8 7 10 5.5 .052 2.8 5.2-5.7 8 8 5.3 .090 4.5 5.1-5.8 9 10 5.4 .076 3.2 5.2-5.8 10 11 5.2 .054 3.3 5.0-5.5 11 16 5.2 .068 5.0 4.9-5.7 12 7 5.4 .070 3.2 5.1-5.6 13 12 5.3 .061 3.8 5.0-5.6 14 7 5.2 .099 4.7 4.9-5.6 15 6 5.3 .118 5.0 5.1-5.8 16 11 5.3 .049 2.9 5.2-5.7 17 8 5.2 .105 5.3 4.9-5.8 18 17 5.4 .037 2.7 5.2-5.8 19 23 5.4 .042 3.7 5,0-5.9 20 11 5.4 .047 2.7 5.1-5.6 21 43 5.2 .026 3.2 4.8-5.5 22 10 5.2 .054 3.1 5.0-5.5 23A 15 5.] .051 3.7 5.0-5.5 23B 13 5.3 .071 4.7 4.9-5.6 24 10 5.3 .042 2.4 5.1-5.5 25A 8 5.4 .045 2.2 5.2-5.5 25B 10 5.4 .095 5.3 5.0-5.9 26 9 5.3 .051 2.7 5.1-5.5 27A 12 5.5 .047 2.8 5.2-5.6 27B 14 5.3 .059 4.0 4.9-5.7 28 8 5.1 .057 2.9 5.0-5.4 29 8 5.2 .083 4.2 4.9-5.5 30 12 5.3 .086 3.8 5.0-5.7 31 10 5.2 .045 2.4 5.1-5.5 32 8 5.2 .047 2.3 5.0-5.3 33 1 5.3 34 8 5.7 .070 3.2 5.5-6.0 35 10 5.4 .037 2.1 5.3-5.6 REFERENCES Alexander, B. 1907. From the Niger to the Nile. Longmans, Green, and Co., 2:xi + 620p. Allen, G. M. 1931. Type specimens of mammals in the Museum of Comparative Zoology. Bull. Mus. Comp. Zoo1., 71:229-289. 1939. A checklist of African mammals. Bull. Mus. Comp. Zool., 83:1-763. Anderson, J. 1902. Zoology of Egypt; mammals. Hugh Rees, Ltd., London, xvii + 374p. Baron, J, C., B. Hubert, P. Lambin, and J. M. Fine. 1974. Serological differentiation of two species of Taterillus (Rodentia: Gerbillidae) from Senegal: T. gracilis (Thomas, 1892) and T. pygargus (Cuvier, 1832). Comp. Biochem. Physiol., 47A:441-446. Bellier, L., and J-C. Gautun. 1967. Note sur les Lemniscomys de Cote d'lvoire. Rev. Zool. Bot. Afr., 75:282-287. Booth, A. H. 1960. Small mammals of West Africa. Longmans, Green, and Co., Ltd., London, xi + 68p. Chapin, J. P, 1954. Gazetteer for the birds of the Belgian Congo. Bull. Amer. Mus, Nat. Hist., 75: 638-738, Choate, J. R. 1970. Systematics and zoogeography of middle American shrews of the genus Cryptotis. Univ. Kansas Publ., Mus. Nat. Hist., 19:195-317. Cockrum, E. L, 1962. Laboratory and field manual for introduction to mammalogy. The Ronald Press Co., New York, v + 116p. Corbet, G. B., and D. W. Yalden. .1972. Recent records of mammals (other than bats) from Ethiopia. Bull. Br, Mus, Nat. Hist. (Zool.), 22:211-252. 241 242 Cuvier, F. 1825. Des dents des mammiferes, considerees comme caracteres zoologiques. F. G. Levrault, Strasbourg, lv + 258p. . 1838. Memoire sur les gerboises et les gerbilles. Trans. Zool. Soc. London, 2:131-148. Cuvier, G. 1833. Catalogue des preparations faites sur les animaux de la classe des mammiferes. Nouv. Ann. Mus. d'Hist, Nat., 2:427-459. Davis, D. H. S. 1966a. Introduction to the colloquium. In, Proceedings of the colloquium on African Rodents (Brussels-Tervuren, 1964). Ann. Mus. Roy. Afr. Centr., in-8°, Zool., 144:3-5. . 1966b. Contribution to the revision of the genus Tatera in Africa. In, Proceedings of the colloquium on African Rodents (Brussels-Tervuren, 1964), Ann. Mus. Roy. Afr, Centr., in-8°, Zool., 144:49-74. Davis, D. H. S., and X. Misonne. 1964. Gazetteer of collecting localities of African rodents. Mus. Roy. Afr. Centr., Doc. Zool., 7:vii + lOOp. De Beaux, 0. 1922. Mammiferi Abissini e Somali, Atti. Soc. Ital. Sci, Nat. Mus. Civ. Stor. Nat,, Milano, 61:2-34. Dekeyser, P. L. 1955. Les mammiferes de l'Afrique Noire Francaise, Inst. Franc. Afr. Noire Init. Afr., 1:1-426. Delany, M, J. 19 64, A study of the ecology and breeding of small mammals in Uganda. Proc. Zool. Soc. London, 142:347-370. Delany, M. J,, and B. R, Neal. 1966. A review of the Muridae (order Rodentia) of Uganda. Bull. Br, Mus. Nat. Hist. (Zool.), 13:295-355. De Saeger, H. 1954. Exploration du Pare National de la Garamba. Inst. Pares Nat. Congo Beige, Fasc. I, 1:1-106, 243 De Witte, G.-F. 1933. Reptiles recoltes au Congo Beige par le Dr. H. Schouteden et par M. G.-F. de Witte. Ann. Mus. Congo Beige Tervuren (Belgique) C. Zool., Ser. 1, Fasc. 2, 2:53-100. Dixon, W. J. (Ed.). 1973. BMD biomedical computer programs. Univ. Calif. Press, Berkeley, vii + 773p. Ellerman, J. R. 1941. The families and genera of living rodents. Br. Mus. Nat. Hist,, London, 2:xii + 690p. Gautun, J-C., and F. Petter. 1973. Liste preliminaire des rongeurs d'une savane centrale de Cote d'lvoire: Foro-Foro. Mammalia, 36:634-637. Genest, H., and F. Petter. 1973. Les Taterillus de Republique Centrafricaine (Rongeurs, Gerbillides). Mammalia, 37:66-75. Genoways, H. H. 197 3. Systematics and evolutionary relationships of spiny mice, genus Liomys. Spec. Publ. The Mus., Texas Tech Univ., No. 5, 368p. Genoways, H. H., and J. R. Choate. 1972. A multivariate analysis of systematic relationships among populations of the short-tailed shrew (genus Blarina) in Nebraska. Syst. Zool., 21:106-116. Genoways, H. H., and J. K. Jones, Jr. 1971. Systematics of southern banner-tailed kangaroo rats' of the Dipodomys phillipsii group. J. Mamm., 28:174-179. Good, A. I. 1947. Les rongeurs du Cameroun. Bull. Soc. etud. Cameroun, 17/18:5-20. Hatt, R, T. 1934, Fourteen hitherto unrecognized African rodents. Amer. Mus. Novitates, No. 708:1-15. . 1940. Lagomorpha and Rodentia other than Sciuridae, Anomaluridae and Idiuridae collected by the American Museum Congo expedition. Bull. Amer. Mus, Nat. Hist,, 76:457-604, Hopkins, B. 1965. Forest and savanna. Heinemann, Ibaden, Nigeria and London, lOOp. Hubbard, C, A. 1973. Observations on the life histories and behaviour of some small rodents from Tanzania., Zool. Afr., 7:419-449, 244 Hubert, B. 1973. Ecologie des populations de rongeurs de Bandia (Senegal) en zone Sahelo-Soudanienne. Unpublished These Doct. 3e cycle, l'Univ. Paris VI, v + 176p. Hubert, B., and J. C. Baron. 1973. Determination of Taterillus (Rodentia, Gerbillidae) from Senegal by serum electrophoresis. Anim. Blood Groups Biochem. Genet., 4:51-54. Jeannin, A. 1936. Les mammiferes sauvages du Cameroun. Encyl. Biol., Paul Lechevalier, Paris, 16:iii + 2 50p. Jolicoeur, P., and J. E. Mosimann. 1960. Size and shape in the painted turtle, A principal component analysis. Growth, 24:339-354, Jullien, R., and F. Petter. 1969. La faune du gisement d'Akjoujt (Mauritanie). Bull. Mus. Nat. Hist. Nat,, 2nd ser., 41:129 0-1291. Kruskal, J. B. 1964a. Multidimensional scaling by optimizing goodness of fit for a«,pjanmet= i-c hypothesis. Psychometrika, 29:1-27. . 1964b. Nonmetric multidimensional scaling: A numerical method. Psychometrika, 29:115-129. Malbrant, R. 1936. Faune du Centre Africain Francais (Mammiferes et Oiseaux). Encyl. Biol., Paul Lechevalier, Paris, viii + 430p. Matthey, R. 1969. Chromosomes de Gerbillinae. Genres Tatera et Taterillus. Mammalia, 33:522-528. Matthey, R., and M. Jotterand. 1972. L'analyse du caryotype permet de reconnaitre deux especes cryptiques confondues sous le nom de Taterillus gracilis Th. (Rongeurs-Gerbillidae). Mammalia, 36:193-209. Matthey, R., and F. Petter. 1970. Etude cytogenetique et taxonomique de 4 0 Tatera et Taterillus provenant de Haute-Volta et de Republique Centrafricaine (Rongeurs, Gerbillidae). Mammalia, 34:585-597. Mayr, E, 1969. Principles of systematic zoology. McGraw- Hill Book Co., New York, xi + 428p, 245 Mayr, E. 197 0. Populations, species, and evolution. The Belknap Press, Harvard Univ. Press, Cambridge, Mass. , xv + 4 53p. Meester, J. 1963. A systematic revision of the shrew genus Crocidura in Southern Africa, Transvaal Mus., Mem. No. 13, South Africa, Pretoria, vii + 127p. Moreau, R. E,, G. H. E. Hopkins, and R. W. Hayman. 194 6. The type-localities of some African mammals. Proc. Zool. Soc. London, 115:387-447. Ogilvie-Grant, W. R. 1913. On a collection of birds from southern Abyssinia presented to the British Museum by Mr. W. N. McMillan. Part I. Passeres. Ibis, Ser. 10, 1:550-641. Petter, F. 1952. Note sur un type de F. Cuvier: Gerbillus pygargus, "la Gerbille du Senegal." Mammalia, 16:37-39. . 1970. Capture de d'un Taterillus en Mauritanie. Mammalia, 34(2) : 325-326. . 1972, Museum National d'Histoire Naturelle, Paris, personal communication. . 1974. Tendances evolutives dans le genre Gerbillus (Rongeurs, Gerbillides). Mammalia, 37: 631-636. , 1975. Subfamily Gerbillinae, Part 6.3, In, J. Meester and H. W. Setzer (Eds.), The mammals of Africa, an Identification manual. Smithsonian Inst, Press, Wash, D, C,, 14p, (in press). Petter, F,, A, Poulet, B, Hubert, and F, Adam, 1972, Contribution a 1'.etude des Taterillus du Senegal, T, pygargus (F, Cuvier, 1832) et T, gracilis (Thomas, 1892) (Rongeurs, Gerbillides), Mammalia, 36:210-213. Poulet, A. R. 1972a, Ecologie des populations de Taterillus pygargus (Cuvier) (Rongeurs, Gerbillides) du Sahel Senegalais. Unpublished These Doct, de 3e cycle, Faculte des Sciences, Univ, de Paris VI, v + 87p. . 1972b. Recherches ecologiques sur une savane Sahelienne du Ferlo septentrional, Senegal: Les mammiferes. La Terre et la Vie, 26:440-472. 246 Poulet, A. R. 1973. Caracteristiques spatiales de Taterillus pygargus dans le Sahel Senegalais. Mammalia, 36:579-606. . 1974. Recherches ecologiques sur une savane Sahelienne du Ferlo septentrional, Senegal: Quelques effets de la secheresse sur le peuplement mammalien. La Terre et la Vie, 28:124-130. Pousargues, E. D. 1896. Etude sur les mammiferes du Congo Francais. Ann. Sci. Nat. Zool. et Paleo., Paris, 3:129-416. Power, D. M,, and J. R. Tamsitt. J973. Variation in Phyllostomus discolor (Chiroptera: Phyllostomatidae). Canad, J. Zool., 51:461-468. Prim, R, C, 1957. Shortest connection networks and some generalizations. Bell Syst. Tech, J., 36:1389-1407. Proceedings of the Colloquium on African Rodents (Brussels- Tervuren, 1964). 1966. Ann, Mus, Roy. Afr. Centr,, in-8°, 2ool„, 144:xiii + 192p. Ranck, G. L. 1968. The rodents of Libya Taxonomy, ecology and zoogeographical relationships. U. S. Nat, Mus. Bull. 275, vii, 264p., 8 PI. Ridgway, R. 1912. Color standards and color nomenclature, Publ. by the author, iv + 43p. + 53 PI, Robbins, C, B, 1972. Author's personal files, unpublished data, , 1973, Nongeographic variation in Taterillus gracilis (Thomas) (Rodentia: Cricetidae). J. Mamm., 54:222-238, . 1974a, Systematic status and karyotypic relationships of the genus Taterillus (Rodentia: Cricetidae) from Kenya, Mammalia, 37:642-645, , 1974b, Comments on the taxonomy of the West African Taterillus (Rodentia: Cricetidae) with the description of a new species, Proc, Biol, Soc. Wash., 87:395-403, Robbins, L. W. 1973-1975. Field Technician, Africa Mammal Project, Smithsonian Institution, Washington, D, C., personal communication. 247 Rohlf, F. J. 1972. An empirical comparison of three ordination techniques in numerical taxonomy. Syst. Zool., 21:271-280. Rosevear, D. R. 1953. Checklist and atlas of Nigerian mammals with a foreward on vegetation. F. Howard Doulton and Co., Ltd., London, 23p, . 1965. The bats of West Africa. Trustees of the Br. Mus. Nat. Hist., London, 418p. . 1966. Reflections on the collections of the brothers Fox. In, Proceedings of the colloquium of African rodents. Ann. Mus. Roy. Afr. Centr., in-8°, Zool., 144:75-87. . 1969. The rodents of West Africa. Br. Mus. Nat. Hist., London, xii + 604p. Ruppell, E. 1830. Atlas zu der reise im nordlichen Afrika. Zoologie. Frankfurt am Mein, Gedrucht und in Comm. bei Heinr. Ludn. Bronner, p. 7 3-7 8. Saint-Leger, J. 1937. Mammals collected by the Lake Rudolf Rift Valley expedition 1934. Ann. Mag. Nat. Hist., Ser. 10, 19:524-531. Schouteden, H. 1936. Contribution a la faune ornithologique du nort-est du Congo Beige. Ann. Congo Beige, C. Zool,, Ser. 4, Fasc. 2, 1:41-156. . 1943. Catalogue des mammiferes du Congo Beige et du Ruanda-Urundi. Rev. Zool. Bot, Afr., 37: 102-125. , 1944. Les Mammiferes du Congo Beige et du Ruanda-Urundi II. -Ungulata (2), Rodentia. Ann. Mus. Congo Beige, C. Zool., Ser. 2, Fasc. 3, 2:333- 576. . 1948. Faune du Congo Beige et du Ruanda- Urundi I. Mammiferes. Ann. Mus. Congo Beige, Sci. Zool., Ser, 8, l:viii + 331p, Seal, H. L, 1964, Multivariate statistical analysis for biologists, John Wiley and Sons Inc., New York, xi + 207pt 248 Setzer, H. W. 1956. Mammals of the Anglo-Egyptian Sudan. Proc. U. S. Nat. Mus., 106:447-587. . 1968. Directions for preserving mammals for museum study. Smithsonian Inst., Wash. D. C., Information Leaflet 380, 19p. Simpson, G. G. 1945. The principles of classification and a classification of mammals. Bull. Amer. Mus. Nat. Hist., 85:xv- + 350p. Simpson, G. G., A. Roe, and R. C. Lewontin, 1960. Quantitative zoology. Harcourt, Brace and World, Inc., New York, 440p. Sneath, P. H. A., and R. R. Sokal. 1973. Numerical taxonomy, W. H. Freeman and Co., San Francisco, xv + 57 3p, Sokal, R. R., and F, J. Rohlf. 1969. Biometry: the principles and practice of statistics in biological research. W. H. Freeman and Co., San Francisco, xiii + 776p. Sokal, R. R,, and P. H. Sneath. 1963. Principles of numerical taxonomy. W. H. Freeman and Co., San Francisco, xvi + 359p. Stoll, N. R., R. Ph. Dollfus, J. Forest, N. D. Riley, C. W. Sabrosky, C. W. Wright, and R. V. Melville (Eds.). 1961. International code of zoological nomenclature adopted by the XV International Congress of Zoology. Int'l, Trust for Zool. Nom., London, xvii + 176p. Thomas, 0, 1892, Descriptions of three new gerbilles in the British Museum collection. Ann, Mag. Nat. Hist,, Ser, 6, 9:76"79, , 1906. New mammals collected in north-east Africa by Mr. Zaphiro, and presented to the British Museum by W. N. McMillan, Esq. Ann, Mag. Nat. Hist., Ser, 7, 18:300-306. , 1910. Notes on African rodents, I. A new genus of gerbils, Ann, Mag. Nat. Hist., Ser. 8, 6:221- 224, 1911. On mammals collected by the Rev. G, T. Fox in northern Nigeria, Ann. Mag, Nat, Iiist., Ser, 8, 7:457-463, 249 Thomas, 0., and M. A. C. Hinton. 1920. Captain Angus Buchanan's Air expedition. I. On a series of small mammals from Kano. Novitates Zool., 27:315-320. Thomas, 0., and M. A. C. Hinton. 1923. On the mammals obtained in Darfur by the Lynes-Lowe expedition. Proc. Zool, Soc. London, 18:247-271. Thomas, O., and R. C. Wroughton. 1907. New mammals from Lake Chad and the Congo, mostly from the collection made during the Alexander-Gosling expedition. Ann, Mag. Nat, Hist., Ser, 7, 19:370-387. Toschi, A. 1946. Two rodents from the Masai reserve. J. East Afr. Nat. Hist. Soc., 18:142-147. Tranier, M., B. Hubert, and F. Petter. 1974. Taterillus de l'ouest de Tchad et du nord du Cameroun (Rongeurs, Gerbillides). Mammalia, 37:637-641. Vaden, Ralph, 1969. The University of Arizona, personal communication, Wassif, K. , R. G, Lufty, and S. Wassif, 1969. Morpho- logical, cytological and taxonomic studies of the rodent genera Gerbillus and Dipodillus from Egypt, Proc. Egypt. Acad. Sci,, 22:77-96. Wettstein, O, 1916. Neue Gerbillinae aus Nordostafrika. Sitzung der matematisch-naturwissenschaftlichen klasse vom 25, Mai 1916. Anz. K. Akad. Wiss. Wien., 53:151-154, . 1917. Wissenschaftliche ergebnisse der mit unterstutzung der kaiserlichen Adademie der Wissen schaf ten in Wien aus der erbschaft treitl Von Ff Werner unternommenen zoologischen expedition nach dem Anglo-Agyptischen Sudan. (Kordofan) 1914. II. Bearbeitung der auf der expedition gesammelten vogel und saugetiere, Denk, Kais, Akad. Wiss. Wien Math, -Nat,, 94:139p, Wroughton, R. C. 1906. Notes on the genus Tatera, with descriptions of new species. Ann. Mag. Nat. Hist., Ser. 7, 17:474-499. 1907. List of a collection of small mammals made by Mr. A, L. Butler in the Bahr-el-Ghazal. Ann, Mag, Nat. Hist,, Ser. 7, 20:500-504. 250 Wroughton, R. C. 1910. African gerbils of the genera Tatera and Taterillus, Ann. Mag, Nat. Hist., Ser. 8, 6:291-294. . 1911. List of a collection of mammals made by Mr. A. L. Butler on the Upper Nile. Ann. Mag. Nat, Hist., Ser. 8, 8:458-462.