111. A Monograph of the of the Family . By C. TATEREGAN, B.A., F.2.8.

Received October 13, read November 17,1903.

[PLATES1X.-XXI.] ri 1HE Loricariidae are a family belonging to the suborder Ostariophysi, and are found only in the rivers of South America, ranging from Panama and Trinidad or Porto Rico to Montevideo. The Ostariophysi resemble the Malacopterygii, the most primitive of‘ Teleostean Fishes, in having the air-bladder, if well developed, provided with a duct, the ventral fins abdominal in position, and a mesocoracoid element present in the pectoral arch ; they are, however, distinguished by the modification of the anterior vertebrae, which are usually fused, some of their lateral and superior elements forming a chain of ossicles (ossicles of Wcber) connecting the air-bladder with the auditory organ. The 0stariophysi are arranged by Boulenger in 6 families-Characinida, Gymno tidae, CyprinidE, Siluridae, A spredinidae, and Loricariidae, the last being equivalent to the Siluridae Hypostomatina of Gunther, with the exception of the genera Callichthys, Sisor, Erethistes, Exostorna, and Pseudechmeis, or to the Loricariidae and Argiidae of Eigenm ann. The Loricariidae share with the Siluridae most of the characters which serve to distinguish that family from the more generalised Characinib, the most important of which are the reduced maxillary, the absent symplectic and suboperculum, and the union of supraoccipital and parietals to form a single parieto-occipital bone ; but they present certain features of still greater specialisation-such as the absence of parapophyses, the sessile ribs, and the compressed caudal vertebrae-which warrant their separation as a distinct family. Most of the fishes of this family can be recognised by the armour of bony plates protecting the body, although the degenerate Argiinae are naked ; the inferior sucker- like mouth is also characteristic 1. It seems not improbable that the Loricariidze may have evolved ‘from the Siluridae in the neighbourhood of the Uoradina, which they resemble in many respects.

’ It appears that io nature these fishes fasten themselves to stones by means of the sucker-like mouth, whilet in captivity they have been observed to adhere to the bottom or sides of the vessel in which they are placed. Respiration seems then to he effected by taking in water through the gill-openings and expelling it again by the same passages in a reverse direction. VOL. XVII.-PART 111. No. l.-October, 1904. 2D 193 ME. C. TATE KEGAN OX THE

The present paper deals mainly with the material in the British Museum, including the types of the described by Gunther and Boulenger, but the author, during a visit to Paris, has been able to examine all the types of the species belonging to this family which have been described and figured by Castelnau and most of those of Cuvier and Valenciennes. In addition, some of the types of species described by Eigenmann have been received on loan, since the reading of this paper, from the Museum of Comparative Zoology, Cambridge, U.S.A., and descriptions of these are incorporated. The author wishes to express his gratitude to Mr. Boulenger for his ever-ready advice and help ; to Prof. Vaillant, through whose kindness he was permitted to examine the specimens in the Jardin des Plantes ; to Dr. Pellegrin for his great courtesy and personal attention during his visits to the Paris Museum; and to Dr. Garman, who has sent to the British Museum examples of several species previously unrepresented in its collection, and, in the case of those species which have been described from one or two specimens only, has lent the types. A tribute must be paid to Dr. Steindachner for the excellence and accuracy of his descriptions of so many fishes of this family, which have made it a comparatively easy matter to assign species described by him to their natural systematic position without having seen actual specimens of them.

The paper of Prof. and Mrs. Eigenmann on the “ South American Nematognathi ” has been of considerable assistance, especially proving of value as a ground-work and for purposes of reference. In the present paper 189 species are recognised as valid, 34 of which are described as new to science. Our knowledge of the geographical distribution of these species is so very incomplete that generalisations are of little value. Most of the genera seem to be represented in the principal river-systems of South America, the rivers south of the La Plata system excepted, and the species seem in many cases also to have a very wide distribution, as might be expected when it is considered that the Amazon system actually communicates with that of the Orinoco on the north, whilst the head-waters of the southern tributaries of the Amazon are in many cases only separated by a few miles from those of the rivers of the La Plata system. Whilst it is clear that the Magdalena, Orinoco, Amazon, and La Plata systems, and also the coast-rivers of Eastern , have each several characteristic species which do not extend into other river-systems, still in the present state of our knowledge it is difficult to say which are the species with their distribution thus restricted. Two important genera, Cilcetostomus and ilrges, occur only in the Andes of l’eru, Ecuador, Colombia, and Venezuela, and are characteristic of mountain-streams in which the other genera of this family me not found ; both have been recorded from considerable altitudes, and it would almost seem that each newly explored valley may be expected to furnish sGme new form. FISHES OF THE FAMILY LORICARIIDX. 193

In the following list, arranged in systematic order, a * indicates the locality from which a species has been recorded. The somewhat artificial divisions which are here established for convenience are :-(1) Western coast-streams of and Ecuador : (2) Rio Magdalena system (with Panama) ; (3) Venezuela and Guiana (Orinoco and Essequibo systems with Trinidad) ; (4) Upper Amazon (Amazon and its tributaries above its junction with the Yapura); (5) Middle and Lower Amazon; (6) Rio Paranahyba and Rio San Francisco and their tributaries, and smaller coast-streams in their neighbourhood; (7) Rio Parahyba, Rio Grande do Sul, and other coast- streams of S.E. Brazil; (8) Rio La Plata and. its tributaries.

-......

1 1. ~ 2. I Siibfam. 1. PLECo 8 TOMINX. I Gen. I. PLECOSTOMUSGronow. Suhgen. n. Pleeostomzcs. 1. P. guacari Lace!,...... I .. I .. 2. P. commersonii C. 4 V...... * *,* 3. P. punctatus C. 4 V...... 3k * 4. P. spinosissimus Sttlr...... I * 5. P. fests Blqr...... * 6. P. verres C. 4 V...... * .. * 7. P. carinatus Xtdr...... * 8. P. vaillanti Stdr...... 0 .. 9. P. emarginatus C. 4 V...... * * * * 10. P. alatus Casteln...... * * 11. P. cordova: Gthr...... 12. P. borellii Blgr...... 13. P. latirostris Rgn...... * 14. P. ternctzi Blyr...... ,...... 15. P. macrops Eiym...... * 16. P. garmani &n...... * 17. P. Yobini C. & P...... * ...... 18. P. une Stdr...... * 19. P. wuchereri cfthr...... * 20. P. luctkeni Stdr...... 21. P. auroguttatus Knev ...... *. a. .. * Subgen. b. Pogonopontn Rgn. 22. P. wertheimeri Stdr...... 23. P. pellegrini Ryn...... * 24. P. genibarbis c. 4 v...... * 0 Subgen. c. Rhinekpis Spix. 25. P. parahybm Xtdr...... 26. P. asper Spix ...... *. .. #f

Gen. 11. HEXIPSILICHTRYSEigm. 1. H. gotio Ltkn......

...... ~ .- - 194 MR. C. TATE REGAN ON THE

~. ~. ~ .... 1. 2. 1 3. 4. 5. 6. 7. 8. ! -- __ I- Gen. 111. ANCISTRUEKner. Subgen. a. Ancistmur. 1. A. annectens Rgn...... * 2. A. undecimalis Stdr...... * 3. A. etentacdatus S'iz ...... * 4. A. gibbiceps Kner ...... y * 5. A. lituratus Kner ...... * # 6. A. multiradiatus Elancock ...... * * * 6 a. A. multiradiatus, var. alternans Rp...... 7. A. punctatus Uthr...... 0 * * 8. A. medians Kner ...... * ...... I ...... 10. A. aspidolepis Gthr...... I .. x 11. A. brachvurus Kner ...... I ...... y 12. A. scaphhhynchus Kner ...... * * 13. A. oligospilus Uthr...... y 14. A. vittatus Stdr...... * * 15. A. bachi Blyr...... * 16. A. schomburgkii Gthr...... * 17. A. megacephalus (fthr...... * 18. A. platycephalus dlyr...... * Subgen. 6. Parancistrus Bllir. 19. A. niveatus Casteln...... * 20. A. punctatissimus Stdr...... * * 21. A. aurantiacus Casteln...... *

Subgen. C. Lasianeistrus Xgn. 22. A. heteracanthus Gthr...... * 23. A. pictus Casteln...... * 24. A. mystacinus Kner ...... * 25. A. guacharote C. 4 J? ...... * Subgen. d. Pseudancistrus Blkr. 26. A. setosus Blgr...... * 27. A. barbat,us G. V...... 1) 28. A. depressus Gthr...... * 29. A. guentheri Rgn...... 1) 30. A. angulicauda StcZr...... * 31. A. parahybre Eigm...... *...... *

Gen. IV. PANAQUE:Eigmn. 1. P. cochliodon Kner ...... * .. .. 2. P. dentex cfthr...... * 3. P. gibbosus Stdr...... * 4. P. nigrolineatus Ptrs...... _- * .. *

Gen. v. CEBTOsTOMU6 Tschudi. 1. C. loborhynchus Tschudi ...... * 2. C. rnaculatus Rgn...... * 3. C. marcapatae Bgn...... * 4. C. brevis Rgn...... * -. . ~__ . .FISHES OF THE FAMILY LORICARIIDE. 195

-.- ~- I I 3. ~ 4. 1 5. 6. 1 7. I .-I. - Gen. V. CHBTOSTOXU~(con.).

Gen. VI. XEINOCARARgn. 1. X. latifrons Qthr...... 2. X. gymnorhynchus Kner ...... 3. X. dolichoptera Kner ....., ...... 4. X. temminckii C. 4 V. , ...... 5. X. hoplogenys Gthr...... 6. X. cirrhosa C. 4 7...... , ...... * .. 9. X. brevipinnis Xgn...... 10. X. montanaRgn......

Gen. VII. PBEUDACANTRICCSBlkr. 1. P. serratus C. cj 7...... 2. P. spinosus Casteln...... 3. P. fordii Qthr...... 4. P. hystrix C. V......

Gen. VIII. AOANTHICU~Spix. 1. A. hystrix Spb ......

Subfam. 2. H P P o P T o P o M A TI B B, *' Gen. IX. HYPOPTOPOMAGthr. 1. H. thoracatum Bthr...... a. .. .. 2. H. carinaturn Stdr. * ...... af 3. H. guentheri Blgr...... 4. H. joberti Yuill...... * 5. H. steindachneri Blgr...... 6. H. gulare Cope ...... ,. af

1. 0. a5nis S&...... 2. 0. vittatus Rgn...... j 3. 0. vestitus Cope ...... : i 196 MB. C. TATE RECAPS ON THE

Gen. X. OTOCINCLUS(con.). 4. 0. flexilis Cope ...... * 5. 0. maculicauda Stdr...... *

~ 1 I Subfam. 3. LORICARIINB. 1 ,I I Gen. XI. LORICARTAL. . .I I I Subgen. a. Zhineloricaria Blkr. I 1. L. filamentosa Stdr. * I 2. L. nigricauda Rgn...... , .. .. I .. *I ...... ~ 3. L. parva Blgr...... I.. .. * 4. L. phoxocephala .. ... * .... ~ .. .. * 5. L. lauceolata Gthr. .. *' 6. L. jubata Blgr...... 7. L. uracantha Kner 4 Stdr...... * .. 8. L. magdalense Stdr...... * I. I 9. L. cadeae Hensel...... * 10. L. teEeanaStdr...... 11. L. catamarccnsis Berg ...... I .. * 12. L. konopickyi Stdv...... 13. L. steindachnori Rgn...... * * 14. L. lima hizer ...... , ...... * 15. L. strigilata Hensel ...... / ...... * 16. L. microlepidogaster Rgn...... ,I ...... * 17. L. latirostris SZgr...... i ...... ill

Subgen. b. Pseudoloiicnria Blkr. I 18. L. Isviuscula C. 4 P...... 1'3. L. punctata Rgn...... I ...... * Subgen. c. Loricariiclithys Blkr. I I I 20. L. maculata BZ...... *.( .. .. * * I 21. L. typue Blkr...... / .. .. I * * * .. .. * 22. L. acuta C. 4 V...... j .. .. I 6. * * 23. L. nudirostris Kner ...... I .. .. * 24. L. labialis Blgr...... , ...... * 25. L. spixii Stdr...... I.. I., ...... * 26. L. anus C. 4 P...... j ...... * * I Subgen. d. Lwicnrin. 27. 1,. evansii Blgr...... I ...... * 28. L. nudiventris C. 4 V...... I .. .. *. .. .. * 89. L. macrope Bgn...... (I ...... * 30. L. cataphracta L...... *i ' .. * * * 31. 2. cariuata Castetn...... l ...... * .Ep .. .. * 3% L. simillima Ryn...... * 33. L. apeltogaster BZy-...... * I , " I I FISHES OF THE FAMILY LORICSRIIDB. 197

~~ I I I I 2. 3. 8. I - .... -

Gen. XI. LORICARIA(con.). Subgen. d. Loricaria (con.). 34. L. macrodon Kner ...... *; 35. L. vetula C. 4 V...... *; 36. 1,. variegata Stclr...... 8 37. L. macromystax Gthr...... 38. I,. lamina Gthr...... 39. L. laticeps Hgn...... * 40. L. platycephala Ziner ...... *

Gen. XII. HBYIODONTICHTHYSBlkr. 1. H. depressus Kner ...... 2. H. acipenserinus Kner ...... I I Gen. XIII. OXYLORICARIABlkr. 1. 0. loricariiformis Xtdr...... 2. 0. platystoma Gthr...... 8 3. 0. brevirostris Eiynt...... 4. 0. guentheri Rp...... 5. 0. robusta Xp...... * 6. 0. lyra Rgn...... 7. 0. rostrataspix ...... 8. 0. barbata Kno- ...... * 9. 0. panamensis Eip, ...... u 10. 0. frenata Blgr...... 8

Gen. SIV. FARLOWELLAEigm. 1. F. gladius BJgr...... 2. F. gracilis Rp...... * 3. F. oxyrhynchus Kner ...... 4. F. knerii Stdr...... 5. F. acus h'iier ...... u 6. amaeonurn Gthr. *. .. " F...... 1. F. gladiolus Gthr...... a. ..

I Subfam. 4. N E o P L E c o s TO M IN E. I' I Gen. XV. NEOPLECOSTOXCSEigm. I 1. N. granosus C. 4 V...... *

Subfam. 5. ARQIIXB. Gen. XVI. ARGESC. &- V. 1. A. homodon Rgn...... u 2. A. boulengeri Ryn...... 3. A. guentheri Bkjr...... u 4. A. cyclopum Humboldt ...... 0 198 MR. C. TATE REGAN ON THE

I I I!1. j 2. I 3. ~ 4. 5. 6. 7. 0. -- __ I ‘ Gen. XVI. ARQEB(con.). 5. A. whymperi Blgr...... rc 6. A. cigenmanni Rgn...... 0 7. A. vaillanti Rgn...... i .. 9. A. brachycephalus Gtlir...... 10. A. chotae Rgn...... 11. A. rnarrnorat,us Rgti ...... If 0 12. A. longifilis Stdr...... * 13. A. sabalo C. 4 V...... 0 14. A. taczanowskii Blyr. *.I *...... y 15. A. prenadilla C. 4 V...... 16. A. fissidens Rqn...... * .. .y 19. A. simonsii Rp...... *’

Gen. XVII. ASTROBLEPWHumholdt. 1. A. grixalvs Humb...... rc - --

Before passing to the systematic portion of this work it may be as well to summarise the more important sexual differences and the changes which take place during growth. Sexual Daferences. The differences between the sexes in certain genera are very remarkable. In the forms with the body protected by bony scutes it is often the case that the little spines which cover these and the plates of the head are stronger in the males than in the females. This feature is greatly exaggerated in Plecostomus spinosissimus and I-’. festcle, in whicl: the males have the head and body covered with quite strong spines. In other cases specialisation has resulted in the great development of spines or bristles in certain areas only. Thus in setosus and its allies, and in many species of Oxyloricaria, Farlowella, and Loricarin, the sides of the head in the male are margined with bristles, which are either absent or considerably shorter in the female. In Loricaria Zanceolata and its allies similar bristles occur in addition on the supra- occipital, the nuchal scutes, and the spine of the pectoral fin. In many species of Loricnrin which have the sides of the head bristly in the males, the head is also some- what broader and the snout more obtuse in specimens of that sex, a feature which is perhaps more strongly marked in L. Zatirostris than in any other species (see Pi. XVI.). In most species of Xenocara the naked margin of the snout is much wider in males tliau in females, and whilst the latter have a more or less distinct marginal series of FISHES OF THE FAMILY LORICARIIDB. 199 tentacles, the former have in addition a Y-shaped group of tentacles, the limbs of the Y starting in front of the nostrils and running forward to meet in the middle line 011 the upper surface of the snout (see P1. XIV.). In Loricaria mutn niid related forms the males are distinguished by the great development of the lower lip, which covers the entire under surface of the head behind the mouth. Steindachner, in describing L. spixii, states that this is due to the nursing- habits of the males, which are said to carry the eggs, during development, between the lower lip and the head. However, I)r. Siebenrock has kindly informed me that this statemerit was based only on the structure of the lip, which seems adapted for such a purpose, and that none of' the specirnens in the Vienna Museum show eggs in tliis position. In the Argiina the males are readily distinguishable by their elongate anal papilla. Changes duxing Growth. As in most Teleostean Fishes, the young have the head and eye proportionately larger than those of the adult, and as a general rule the head is relatively less broad, the snout shorter, and the interorbital space rather narrower in young specimens. In the P1ecostomin;e the lower surfwe of the head and abdomen is naked in all very young specimens, and may remain so throughout life in some species, whilst in others a more or less complete covering of small granular scales may be acquired.. In those species of Loricaria which have the abdomen covered in the adult with small granular plates or scales, the latter are acquired during growth in a similar manner. In the Plecostominze the dorsal fin is both higher and longer in the young than in the adult, the difference being very considerable in some species-e. g., in Plecostomics eitcarginatits the length of the base of the dorsal is equal to its distance from the adipose fin in small specimens, c\ hcwas in large ones it is contained 14 times in that distance. In the Loricariinze the dorsal fin appears to become relatively higher during growth. Throughout the family the pectoral fin becomes relatively longer during growth, so that in the same species it may barely reach the base of the ventral in a small example and yet extend nearly to tlle posterior extremity in a large one. There is no difference in the length of the pectoral in individuals of different sexes if they are of the same size. The length of the barbel is dependent on the size of the specimen ; thus in Arges bnrclycephctlus small speciniens (80-110 mm.) have the barbel extending beyond the gill-opening, whilst in larger ones (200-220 mm.) the barbel extends only 9 of the distance from Its baee to the gill-opening. Considerable changes of colour and marking may take place during the growth of certain species. Iii many Loricariina the young have 4-6 broad dark cross-bands on the back, which are absent in the adult. In the Plecostominze, in species such as Plecostonzus guacari, P. verres, P. puncfatus, &c., which have the head, body, and fins covered with dark !spots, these are much fewer and larger on the head and body VOL. XVII.-PART III. No. 2.-October, 1904. 2E 200 MR. C. TATE REGAN ON THE

in the young, whilst the fins have a few cross-bars, which later on break up into spots, and these spots, like those on the head and body, seem to continually increase in number with increased size of the . Measurements. In descriptions of species the total length means the distance from the tip of the snout to the base of the caudal fin ; but where the total length (in millimetres) of the largest specimen described is given, the length of the middle rays of the caudal fin is included. In the armoured forms the length of the head is measured to the posterior margin of the temporal plate, but in the naked-bodied Argiinae to the edge of the bony operculum. In some cases it bas been found useful to indicate the size of the mouth by giving the proportion of the length of the mandibular ramus-i. e. of that box-like tooth-bearing portion of the mandible which borders the mouth-to the interorbital width. In counting the scutes, that one which is posterior to the clavicle is reckoned as the first, whilst the movable plates covering the bases of the caudal rays are not included. The length of the base of the dorsal fin is measured from the base of the first developed ray to the base of the last, the membrane behind the last ray being excluded. The sign f prefixed to the name of' a species indicates that the author has not examined actual specimens, but has based his description on the published RCCOUl1ts.

Family LORICARIIDB. Ostariophysi with parietals and supraoccipital united to form a single parieto-occipi tal bone, without suboperculum, symplectic, or metapterygoid ; palatine articulating posteriorly with the praefrontal and anteriorly with the small rod-like maxillary, which bears a barbel ; praemaxillaries not protractile, forming the entire upper border of the mouth. Caudal vertebrae compressed, with expanded and compressed neural and lmmal spines ; praecaudal vertebrae without parapophyses ; coalesced anterio