M. m m t: m f ; M. 1^

m %

^KK)^\^

nwwrw ^. % $1 ■m' f

>r'^*^ij»r\

%,l

B M .1-

p 'W ''

m. :K' t

|t^. *

fc J=: I' li i^ *' W

^'^S,f^..^^

:-^*>;

#'. m-

..ji'Si- ,«?'

^a .j^;

■Jte-

;.>»:■...&' tr h fh^

1

-•■

J '

■P' >:

M

[h,

.fe^ 1

^'

.1

*'' w' i-

%' t"j^'

«( ti-

ll^ f' ¥' *■

,fe; I' i^

,*^l> t * «■ »: ri i^

^^ US

HARVARD UNIVERSITY

i

^

Library of the

Museum of

Comparative Zoology

BULLETIN

OP THE

MUSEUM OF COMPAKATIVE ZOOLOGY

AT

HARVARD COLLEGE, IN CAMBRIDGE

VOL. 119

CAMBRIDGE, MASS., U. S. A. 1958-1959

TiiK Cosmos Pkkss, Inc. CAMHHincK. Mass., U.S.A.

1^

CONTENTS

PAGE

No. 1. Revision of Five Aerioan Snake Genera. By

Ai-tliur TjOvoridpT. July, 1!)5(S .... 1

Xo. '2. Fi)V\i New Ra.iids ekom the Gulf of Mexico. By Henry l>. Bifjclow and William G. Schroeder. July, l!)."i8 1!)!)

No. 3. The General Histology and Topographic jMicro- anatomy of Attstijalorius glabrattts. By Gliia- Tung Pan. (18 plates). July, 1958 . ' . . 285

No. 4. Stttdtes on the Ant Fattna of Melanesia TIT. Kiiytidoponera in Western Melanesia and the Moluccas. IV. The Trihe Ponehini. By E. 0. Wilson. August, 1958 .801

No. 5. A Ne\v Species of Ghelid Turtle, Phrynops (Batrachemys) datilt, from Golombia. By Rainer Zangerl and l^'rcd Modem. (2 plates). August, 1958 ;17;]

No. G. Taractes asper and tiii: Systematic Relation- SHU's OK THE Steinegertidai-: and Trachybery- ciDAE. By Giles \V. Mead and G. E. Maul. (1 plate). Geto))er, 1958 ;W1

No. 7. Additions to the Pleistocene Mammalian Fauna FROM Melboitrne, FLORIDA, liy Glaytou E. Ray. November, 1958 4P.)

No. 8. Stitdies on the MoRPHOLO(iY and Function of the Skull in the Boidae (Serpentes). Part 1. Granial Differences between Python sebae and Epicrates cenchris. By T. II. Frazzetta. Jan- uary, 1959 . . . " 451

Xo. 9. The Genus Tetragnatha (Aranbae, Argiopidae) IN Michigan. By Arthur M. Ghickering. Februai-v, 1959 . . ' . . . . . . ' . 473

Bulletin of the Museum oi Comparative Zoology

AT HARVAED COLLEGE Vol. 119, No. 1

REVISION OF FIVE AFRICAN SNAKE GENERA

Bv Arthur Loveridge

CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM

July, 1958

Publications Issued by or in Connection

WITH THE

MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE

Bulletin (octavo) 1863 The current volume is Vol. 119.

Breviora (octavo) 1952 No. 89 is current.

Memoirs (quarto) 1864-1938 Publication was terminated with Vol. 55.

JoHNSONiA (quarto) 1941 -A publication of the Department of MoUusks. Vol. 3, no. 35 is current.

Occasional Papers of the Department of Mollusks (octavo) 1945 Vol. 2, no. 21 is current.

Proceedings of the New England Zoological Club (octavo) 1899-1948 Published in connection with the Museum. Publication terminated with Vol. 24.

The continuing publications are issued at irregular intervals in num- bers which may be purchased separately. Prices and lists may be obtained on application to the Director of the Museum of Comparative Zoology, Cambridge 38, Massachusetts.

Of the Peters ' ' Check List of Birds of the World, ' ' volumes 1-3 are out of print ; volumes 4 and 6 may be obtained from the Harvard Uni- versity Press; volumes 5 and 7 are sold by the Museum, and future volumes will be published under Museum auspices.

Bulletin of the Museum oi Comparative Zoology

AT H A E V A R D C O L Tj E G E Vol. 119, No. 1

REVISION OF FIVE AFRICAN SNAKE GENERA

By Arthur Loveridge

CAMBEIDGE, MASS., U. S. A.

printed for the museum July, 1958

No. 1 Revision of Five African Snake Genera By Arthur Loveridge

CONTENTS

Page

Tntvodiu'fioii iind Acknowledgments 1

(lenus Lyeodonoinorphu!^ Fitzinger (key on page 7) 6

Genus Kaivicitcrcs Loveridge (key on page 26) 25

(ienus P]iilflt]iaiu7iu.s Smith (key on page 52) 49

(ienus Frosymna Gray (key on page 129) 127

Genus Chilorliinopliis Werner (key on page 169) 168

Bil)liography 179

INTRODUCTION AND ACKNOWLEDGMENTS

This paper represents the author's last attempt at revis- ing snake genera. The five dealt with here were selected because of the contused status of their component species, and this re- (fuired elucidating in connection Avith the report on Nyasaland reptiles. The work was undertaken about 1950. The typescript was then laid aside with the intention of borrowing certain snakes on which (juestionable records were based, also in the faint hope of completing the African COLUBRIDAE, to which family these five genera belong. Eleven other colubrid genera have been dealt with already in three contrilnitions to this Bulletin published in 1939, 1940 and 1944.

The procedure and format of the current contribution is essen- tially that of its predecessors, i.e. an attempt to synopsize all pertinent information appertaining to every species assigned to the genera under revicAv. In the present instance the period covered is from 1880 to 1956, though the last year is possibly incomplete.

However, in the case of Lijcodonomorphus Fitzinger, with which I synonymize (Hypliolycus Giinther, I have treated the

4 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY

little-known^ hicolor of Lake Tanganyika in greater detail than usual. This is due to the collaboration of C. J. P. lonides, Esq., who generously procured more than 100 specimens of this aquatic reptile for study. The astonishing variation displayed by this fine series is published here for the first time. One might add that Mr. lonides also submitted comparably large quantities of at least one species for each of the four remaining genera.

Others in Africa who have aided this investigation by sending in material recently, are Messrs. D. C. Broadley (Southern Rho- desia), D. Vesey-FitzGerald (Northern Rhodesia), M. C. Lesage (Ghana) and J. S. Owen (Sudan).

During, and subsequent to, 1950 my esteemed colleague Dr. R. F. Laurent of the Musee Royal du Congo Beige has added several interesting species to the genera Lycodononiorphus and Prosymna. AVith the latter genus, Laurent's extensive AYest African material enabled him to clarify the situation respecting the races of ambigua occurring in the Belgian Congo. Though time and the accumulation of more material may demonstrate that two of these forms are untenable, I have accepted his ar- rangement except in one instance. Fortunately for me Laurent's researches prevented my merging a. hocagii with a. ainbigiia, something I was preparing to do because of the confusing mis- identifications of hocagii in the literature, as maj^ be seen from its synonymy. In acknowledgement of this, and as a tribute to Laurent's contributions to African herpetology, I take pleasure in naming the only novelty described in this paper :

Prosymna meleagris laurenti subsp. nov.

Circumstances have made it (juite impossible for me to borrow paratypes of Laurent's new forms as I should have done if not so pressed for time. In other respects I have endeavoured to bring these revisions up to date as of 1956. At the time {ca. 1950) I first did them 1 borrowed much material or pestered herpetological correspondents for information respecting speci- mens in their care. For either one or the other I am indebted to the undermentioned colleagues and gladly avail myself of this opportunity of expressing my thanks.

1 This was written before seeing Laurent's 19.56 paper listing 133 specimens I

LOVERIDGE: AFRICAN SNAKE GENERA 5

J. R. Bailey (Duke University) ; J. C. Battersby (British Museum) ; C. M. Bogert (American Museum of Natural His- tory) ; G. S. Cansdale ( Lynda! e Avenue, London) ; B. S. Chau- han (Zoological Survey of India) ; D. M. Cochran (United States National Museum) ; R. Conant (Zoological Society of Philadelphia) ; J. Eiselt (Naturhistorische Museum, Vienna) ; V. F. FitzSimons (Transvaal Museum) ; J. Guibe (Paris Mu- seum) ; N. Ilartweg (Michigan University Museum) ; A. Holm (Zoological Museum, Uppsala) ; R. F. Laurent (Musee du Congo Beige) ; R. Mertens (Senckenberg Museum) ; M. G. Netting (Carnegie Museum, Pittsburgh) ; H. W. Parker (British Mu- seum) ; J. A. Peters (Brown University, Providence) ; C. R. S. Pitman (Chelsea, London) ; M. Poll (Musee du Congo Beige, Tervueren) ; C. H. Pope (when at Chicago Nat. Hist. Mus.) ; H. Rendahl (Naturhistoriska Riksmuseet) ; W. Rose (South African Museum, Capetown) ; K. P. Schmidt (Chicago Natural History Museum) ; B. Shreve (Museum of Comparative Zoology) ; R. H. Smithers (National Museum, Bulawayo) ; 0. van Straelen (Li- stitut Royal de Belgique) ; P. E. Vanzolini (Departmento de Zoologia, Sao Paulo) ; H. Wermuth (Zoologisches Museum, Ber- lin) ; 0. Wettstein (Osterreichische Akademie, Vienna) ; E. E. Williams (Museum of Comparative Zoology) ; G. F. de Witte (Listitut Royal de Belgique).

When necessity arose to refer to this borrowed, or other, ma- terial, the institution where it may be found is indicated by the following abbreviations :

A.M.N.H., American Museum of Natural History, New York; B.M., British Museum (Natural History), London; CM., Car- negie Museum, Pittsburgh; C.N.H.M., Chicago Natural History Museum. Chicago; I, lonides field number; I.R.B., Institut Royal de Belgique; M.C.Z., Museum of Comparative Zoology, Cam- bridge; P.M., Museum National d'Histoire naturelle, Paris; R.G.M.C, Musee Royal du Congo Beige, Tervueren; S.M.F., Senckenbergische Naturforschende Gesellschaft, Frankfurt, a.M. ; S.R.M., Southern Rhodesia Museum, Bulawayo; T.M., Transvaal Museum, Pretoria; U.S.N.M., United States National Museum, Washington.

Localities. An asterisk (*) in front of a place name implies that a specimen or specimens from said locality is either in the collection of the Museum of Comparative Zoology or has been

6 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

examined for this revision as in the ease of types studied at the British Museum in 1951. For one species only {Lycodono- morphus r. rufulus) the asterisk is placed after the locality to indicate that the material from these places is in the Transvaal Museum and was examined for me by Dr. V. F. FitzSimons with his customary kindness.

Bihliography. Owing to the spate of semi-popular books on snakes that have appeared in recent years, it is becoming in- creasingly difficult to know what should, or should not, be in- cluded. Where a date is followed by a letter of the alphabet it indicates that during the year cited the author in question pub- lished more than one paper on African herpetology. The letter has chronological significance in a more comprehensive bibliog- raphy of African Herpetology (1880-1953) which it is hoped may be published in the not too-distant future.

Family COLUBRIDAE

Genus LyCODONOMORPHUS Fitzinger

1843. Lycodunoinuipltus Fitzinger, Syst. Eept., p. 27. Type by original

designation: Coronella rufula Schlegel =: Coluber rufulus Lichten-

stein. 1848. Lycodontomorphiis Agassiz, Nomen. Zool. Index Universalis, p. 628.

Emend, for Lycodoiiomorphus Fitzinger. 1863. NeusteropMs Giinther, Proc. Zool. Soc. London, p. 16, footnote. Type

by monotypy : Natrix laevissima Giinther. 18931j. Ablahophi.s Boulenger, Cat. Snakes Brit. Mus., 1. p. 318. Type by

monotypy : Coluber rufulus Lichtenstein. 1894b. Glyphohjcus Guntlier, Proe. Zool. Soc. London, p. 629. Type l)y

monotypy : G. bicolor Giinther. 1924b. Xerophidion. Werner, Sitzb. Akad. Wiss. Wien, 133, Abt. 1, p. 53.

Type by monotypy: A. hypsirhinoides Werner = Glyphohjcus

bicolor Giinther.

Definition. Maxillary teeth 18-25," small, sube(iual ; mandibular teeth longest anteriorly. Head slightly distinct from neck; eye

2 III siviuK lS-19 fur bicolor, possibly Cott ('1935, p. 965) failed to allow for a mlssinfe' tooth ; there are clearly 19 in a ? sluill (M.C.Z. 54091) and appear to have l)eeii 'JO in another (M.C.Z. ;;i>07t!). Giinther said "about 21" for the type, while l{oulenj;er (lS9(;d. p. <)15) merely stated that the dentition of Gli/pliolinus was similar to that of Laiiiinoiihix and Jiotliroli/cus. However, Laurent (19541), P 43) jiives 19-20 teeth for liotli his .s. snbtaciiiutii)^ and s. upcmhac.

LOVERIDGE : AFRICAN SNAKE GENERA <

moderate, with round or vertically subelliptic pupil; nostril directed upwards, in a semi-divided or divided nasal; a loreal; preoculars 1, rarely 2; lateral head shields separated by an in- conspicuous groove from the upper labials. Body cylindrical ; scales smooth, with or without apical pits, in 19-25 rows ; ventrals rounded; anal entire, rarely divided. Tail moderate, tapering; subcaudals paired.

Hemipenis of male not (hicolor and siibtaeniatus) or distally bifurcate (rufulus) ; sulcus spermaticus forked. Hypapophyses present posteriorly in the vertebral column.

Range. African lakes and rivers south of S.

Remarks. For earlier comments on the status of this genus and the number of pterygoid teeth, see Loveridge (1953e, p. 253) and Laurent (1954b, pp. 38-43). As Laurent's admirably described species and its race bridge the gap between hicolor and laevissi- mus in the west, as do r. mlanjensis and r. tvhytii between hicolor and rufulus in the east, there no longer remains any valid reason for retaining Glypholycus as a distinct genus. Bogert, on whose work (1940, pp. 18-19) I have depended in respect to dentition, hemipeneal and hypapophj-seal characters, concurs in this view.

It is true that in hicolor the 4th labial is the only one to normally enter the orbit, whereas in all the other species both 4th and 5th are normal, and the 4th only is very exceptional. Such a character, however, cannot be considered of generic sig- nificance.

Key to the Species

Our knowledge of the full range of ventral and subcaudal counts is so incomplete for certain forms, while in others the known counts display so much overlapping, that it appears next to impossible to devise a key that reflects relationships. However, if ranges are taken into account, and the table of statistical data that follows the key is consulted, the identification of specimens should be greatly simplified.

1. Midbody scale-rows 23, rarely 25 (25 present in only 2 of the 110 snakes examined) ; range: only Lake Tanganyika hicolor (p. 9)

Midbody scale-rows 21 to 23 (21 chiefly in ,J <5 , 23 in 9 $); range: western Angola and western Belgian Congo 2

Midbody scale-rows 19 to 21 ; range : eastern Belgian Congo and southern Nyasaland 3

8 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Midbody scale-rows 1'.), rarely 'Jl (21 present in only 2 of about 20 records of Inevi^sinuis) ; range: southern Tanganyika Territory; northern Nyasaland and Africa south of the Zambezi 4

2. Subcaudals 29-40 (9 9 29-34; c?40); ventrals 172-188 ($ 172; 9 9

17.')-188) ; range: southeastern Belgian Congo s. wpemhac (p. 12)

Subcaudals 41-58 (9 9 41-48; $i 52-58); ventrals 175-193 {$ i

175-180; 9 9 189-193); range: western Belgian Congo and western

Angola .s. suhtaeniaius (p. 13)

3. Throat of adult whitish to gray ; belly mostly black sparsely flecked with

cream; tail below entirely, or almost entirely, black (a young one under 230 mm. in total length, pure white; juveniles from 250-600 mm. gradually assume adult coloring, that is of 600-800 mm. speci- mens) ; range: montane streams of eastern Belgian Congo from north

end of Lake Tanganyika to region of Lake Upemba

r. Icleupi (p. 14)

Throat of adult to anus immaculate white ; tail from anus to near its tip white -with a dusky median line; range: montane streams of Zomba, Mlanje and Cholo Mountains, southern Nyasaland r. vtlanjensiit (p. 16)

4. Subcaudals less than 50 (but that of S $ unknown); range: southern

Tanganyika Territory (where thej- approach leleupi in color) to north- ern Nyasaland ; possibly also Charre, Mozambique

r. ivhytii (p. 17 )

Subcaudals more than 50 (54 to 81; those reported as having less than 54, which I have examined, all had regenerated tail tips); range: south of the Zambezi 5

5. Loreal well separated from first labial; ventrals 160 to 178; subcaudals

54 to 78 ; pupil subcircular, rarely round ; no light line from behind eye to above conier of mouth; upper labials immaculate; underside immaculate or, at most and that but rarely, a few dusky marks ; range: streams of Southern Rhodesia and Bechuanaland, south to

Natal, west to the Cape r. rufulus (p. 19)

Loreal in contact with first labial or, if separated, only by a granule; ventrals 172 to 183; subcaudals 58 to 81; pupil round; a more or less light line from behind eye to above corner of mouth; upper labials spotted ; underside from throat to end of tail with a dusky median line or occasionally a median series of dusky markings; range: Transvaal (1 record), south to Natal (1 record), east through Cape Province laevissimus (p. 23)

LO\^RIDGE : AFRICAN SNAKE GENERA

STATISTICAL DATA FOR THE SPECIES OF LYCODONOMORPHUS

(Abnormal variations are given in parentheses)

Sppcips

Midbody soale-rows

Malo vt'iitrals

Foinalo \('iitrals

Feinalo subcaudals

Malo subcaudals

bioolor

23 (-2.-,)

154-164

152-166

52-59^

63-71*

s. upeinbac

21-23

172

175-188

29-34

40

s. subtaeniatus

21-23

175-180

189-193

41-48

52-58

r. leleupi

19-21

165-171

164-174

52-61

65-67

r. mlanjensis

21

163-167

169

51

60-71

r. whytii

19

?

159-172

37-47

?

r. rufulus

19

162-17;*)

160-178

54-79

72-78

laevissim us

19(-21)

172-181

172-183

58-69

76-81

LYCODONOMORPHUS BicoLOR (Gimther)

1894b. Glypholycus bicnJor Giinther, Proe. Zool. Soc. London for 1893, p.

629, fig. 1: Shores of Lake Tanganyika.

1896d. Boulenger, p. 61.1.

1910a. Sternfeld, p. 15.

1915a. Boulenger, p. 201.

1915c. Boulenger, p. 619.

1924b. Loveridge, p. 4.

1929a. Werner, p. 50.

1933h. Loveridge, p. 232. 1933m. Witte, p. 86.

1937f . Loveridge, pp. 489, 496.

1941a. Uthmoller, p. 40 (as Glypholicm).

1947c. Laurent, p. 10.

1952. Witte, p. 18.

1953. Witte, p. 164, lig. 42. 1956. Laurent, p. 88, pi. ix, fig. 1.

1942b. Nerophidion liypsirliinoidis Werner, Sitzb. Akad. Wiss. Wien., 133,

Abt. 1, p. 54, fig. 7: No locality. 1928. Malcolm Smith, p. 496 (synonymizcs with bicolor).

3 50-59 flde Laurent (1954b, p. 43).

4 59-71 fide Laurent (1954b, p. 43).

10 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Common Name. Tanganyika White-bellied Water-Snake.

Description. Rostral l)roader than deep, well visible from above ; nostril valvular, directed upwards in a semidivided nasal that is sometimes completely divided, rarely entire' ; internasals narrow anteriorly, as long; as, or longer or shorter than, broad, shorter than the prefrontals ; prefrontals moderate ; frontal 1% (usually 11/2 ) to 2" times as long as broad, rarely' as long as its distance from the end of the snout, usually as long as its distance from the rostral, much shorter than the parietals; loreal longer than deep, in contact with, or separate from, the first labial, sometimes by the presence of a small scale*; eye small, pupil usually round, sometimes subelliptic" ; preocular 1 (on both sides of all 111 snakes) ; postoculars 2 (110 ex.), very rarely 1 (M.C.Z. 54936) ; temporals 1+ 2 (both sides of 103 ex.). rarely 1 +1 (on one side only of M.C.Z. 54886) or 1 + 3 (M.C.Z. 54916, and on one side only of 6 others) ; lateral head shields separated by a deep groove from the upper labials ; upper labials 7, 8 (eighth usually posterior to the gape) or 9 (left side of M.C.Z. 54879), the fourth, rarely fourth and fifth (M.C.Z. 54837, 54934, etc.) entering the orbit (67 ex.) or separated from it (at least on one side) by 1 or 2 minute sul)oculars, or by a suboeular and an extension of the lower postoeular (left side of M.C.Z. 54879), or by an extension of the preocular contacting the lower postoeular (right side of M.C.Z. 54857) ; lower labials 8 or 9 (ninth usually posterior to the gape), first 4 (first 3 only on left side of M.C.Z. 54861 and 54921 ; or first 5 in M.C.Z. 54935 and on one side only in M.C.Z. 54921 and 54942) in contact with the anterior sublinguals, which are much longer than the pos- terior. Midbody scale-rows'" 23-25 (23 in 109 snakes; 24 in 3, viz. M.C.Z. 54908 and 54940-1; 25 in 2, viz. M.C.Z. 54853 ( $ )

s Entire in M.C.Z. 54851, 54919, or on left side only of 54852.

6 Twice in M.C.Z. 54897.

" As long as its distance in M.C.Z. 26929, 54858, 54871.

8 Small scale given off by 1st labial in M.C.Z. 30076, 54869, etc. SmaU scale given oflf bv L'nd labial in M.C.Z. 54855, 54881. etc. Second labial transversely divided in M.C.Z. 54853, 54857, etc.

9 Subelliptic in M.C.Z. 30076, 54859, 54916, and on left of 54941.

19, given by Boulenger (1915c. p. 619), due to inclusion of tchytii ; 21. stated ill the original description, amended to 23 by Boulenger.

LOVERIDGE : AFRICAN SNAKE GENERA 11

and 54924 ( 9 )) ; ventrals 152-166^^ ( S S 154-164; 9 9 152- 166) ; anal entire; subcandals^" 52-71 ( 9 9 52-59; $ $ 63-71) pairs.

Color. Above, "'ray or plumbeous, rarely black, uniform except for the two or three outermost rows of scales which are white. Below, throat and belly white, uniform or occasionally some dusky flecking on the anterior third; tail white, sometimes uni- form in young but usually with a dusky longitudinal line along the median sutures of the subcaudals.

Size. Largest $ (M.C.Z. 54854), 563 (420 + 143) mm.; largest 9 (M.C.Z. 54902), 778 (615 -f 163) mm.; smallest $ (M.C.Z. 54856), 227 (175 + 52) mm.; smallest 9 (M.C.Z. 54852), 235 (188 + 47) mm. All from Kigoma, Tanganyika Territory.

Se.mial dimorphism. This is reflected by their respective tail lengths, tliat of 52 S S being included in total length 3.64 to 4.42 times, with an average of 3.89; in 49 9 9 the tail was included in total length 4.32 to 5.06, with an average of 4.66. For the purpose of these figures snakes with incomplete tails were discarded. The number of subcaudals appear to be an even surer guide to sex, the 9 9 having from 52 to 59, the S S from 63 to 71 pairs.

Compare with Laurent's (1956, p. 89) figures expressed as percentages with an average of .255 lor $ $ , .227 for 9 9 . None of his specimens attained the maximum size of those in the Museum of Comparative Zoology.

Breeding. A number of the 9 9 taken at Kigoma between April 4 and May 4, 1956, are gravid. They have not been critically examined for precise dates and the number of eggs produced at a time.

Diet. I am indebted to Dr. E. Trewavas for determining the species of thirteen fish that Mr. C. J. P. lonides removed from stomachs of these snakes, viz.

1 Boulengerochromis microlepis (Boulenger) from M.C.Z. 54892.

11 184. as given by Bouleuger (lS96d, p. 615) presumably a misprint. Battersby has recounted all British Museum bicolor ami finds them within the range given for M.C.Z. material.

1250 is the lowest given by de Witte (1952, p. 18), and Laurent (1956, p. 89), but the tail tips of all M.C.Z. snakes with less than 52 have been regenerated, though at times this is difficult to see. For dd, Laurent gives 59-71.

12 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

1 Haplotaxodoii Iriroii Poll from M.C.Z. 54947.

1 Lamprolorjus callipferns Boiilenger from M.C.Z. 54908.

2 Lamprologua savoryi elotujafiis Trewavas & Poll from M.C.Z.

54872 and 54922. 1 Lamprnlogus sp. from M.C.Z. 54916.

1 Lininochromis pfcffcri (Boulenger) from 1. 677-3.

4 Limnotilapia dardcnii (Boulenger) from M.C.Z. 54917, 54933, 54949.

2 Stolothrissa tanganicae Regan from I. 6773.

Witte (1952, p. 18) also mentions recovering a species of Limnothrissa. Except for a couple of cichlids, the stomach con- tents consisted of elupeids (Laurent :1956).

Habitat. C. J. P. lonides informs me that he captured these hicolor in the lake at night by the light of a pressure lamp, taking the snakes in a hand net about the size of a shrimping net. Laurent (1956) says they are to be found hiding beneath stones on the lakeshore in the vicinity of rivers. He quotes N. Leleup who captured several in the burrows of gerbils {Tat era sp.) where they presumably pass the hours of daylight.

Localities. Tanganyika Territory: *Karema, Mpanda ; Katabi near Bangwe ; *Kigoma; Kirando Bay; Lagosa Bay; *Sumbawa. Northern Rhodesia: *Mpulungu. Belgian Congo Tanganyika District : Kabimba Bay ; Mtoto Bay ; Pala Bay ; Uvira. Belgian Urundi: Makamba (but considered doubtful by the collector) ; Rumonge.

Range. Lake Tanganyika off the coasts of Tanganyika Terri- tory; Northern Rhodesia; Belgian Congo and Ruanda-ITrundi.

LYcoDONOAtORPiius suBTAEXiATUs UPEMBAE Laurent

IdSdni. Board on lineatns Witte (part: not Diuiieril & Bibron), p. 86. 1953. Witte, p. 168.

1954b. Lycodonomorphus suhtaeniatus upemhae Laurent, Museo do Dundo, No. 23, p. 41, figs. 5-8: Nyonga, Katanga, Belgian Congo.

Common Name. Eastern Congo White-bellied Water-Snake.

Description.^'^ Tnternasals shorter than the prefrontals; pre- frontals moderate ; frontal as long as, or shorter than, its distance from the end of the snout, much^'* shorter than the parietals;

13 Taken entirely from the original description, where Laurent tabulates the scale-counts tor the d and all 21 9?.

LOVERIDGE: AFRICAN SNAKE GENERA 13

preociilar 1 (41 sides), rarely 2 (3 sides); postoculars 2 (42 sides), rarely 1 (2 sides) ; temporals 1 + 2. Midbody scales with apical pits, "in 21-23 rows; ventrals 172-188 (S 172; $ 9 175- 188) ; anal entire; subcaudals 29-40 ( 5 9 29-34; $ 40).

Color. Above, coloring lighter than in s. suhtaeniatKs and the dark dorsal zone narrower; the two or three outermost rows of scales white.

Size. Only S (ex Nyonga), 454 (387 + (i7) mm.; largest 9 (ex Nyonga), 831 (739 + 92) mm.

Sexual dimorphism. The solitary $ and one 9 have 21 mid- body scale-rows, the remaining 20 9 9 have 23 rows. The $ has fewer ventrals and more subcaudals than any of the 9 9 ; also his tail is longer ( £ 157r ; 9 9 10 to 12% of total length).

Dentition. Maxillary teeth 19-20.

Localities. Belgian Congo: Kina-Mwena ; Nyonga near Lake Upemba.

Range. Southeastern Belgian Congo.

Lycodonomorpiius subtaeniatus subtaeniatus Laurent

1952b. Boaedon vlrgatus Laurent (not Hallowell), p. 199.

19.")4:b. Lycodonomorpiius iiuhtae7iiatu.<i futhtaeniatus Laurent, Museu do

Dundo, No. 23, p. 38, figs, l-i: Kcseki, near Kwamouth, Belgian

Congo.

Common Names. Western Congo White-bellied Water-Snake ; luhuchwe or lunoka at Dundo, where the same names are applied to Boaedo)i fuliginosus.

Description.^^ Rostral broader than deep, just visible from above ; nostril directed upwards in a semidivided nasal, inter- nasals shorter than the prefrontals ; prefrontals moderate ; frontal as long as, or slightly longer or shorter than, its distance from the end of the snout, shorter than the parietals; loreal much longer than deep, separated from the first labial ; preocular 1 ; postoculars 2 ; temporals 1 + 2 ; upper labials 8, fourth and fifth entering the orbit ; lower labials 8, first 4 in contact with the anterior sublinguals, which are subequal to the posterior. Mid- body scales with apical pits, in 21-23 rows (21 in 8 paratypes;

1*66-70% in s. iipeinbae as against 74-86% in s. subtaeniatus.

15 Based entirel.v on Laurput's admirable account of 5 cTd" and 8 ? 9 as I liavc seen no material of this rfcently doscrilKMl species. For furttier details see Laurent's remarks.

14 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

23 in 5) ; ventrals 175-193 ( $ $ 175-180; 9 $ 189-193) ; anal entire ; subcaudals 41-58 ($ 2 41-48; S $ 52-58).

Color. Above, dark olive gray ; flanks paler, usually bearing a longitudinal lateral band as dark as the dorsum but at times the two shades merge into one another; head with light lines analogous to those of Boaedon fuliginosus though not so clear and sometimes indistinct (for a detailed description of these markings consult original paper). Below, whitish, chin shield and first three lower labials marked with white, throat very slightly pig- mented; tail with a dusky longitudinal line along the median sutures of the subcaudals.

Size. Largest S (ex Keseki), 651 (529 + 122) mm.; largest 9 (ex Dundo), 1009 (872 + 137) mm.

Dentition. Maxillary teeth 19-20.

Hemipenes. Hemipenis not bifurcate, extending to the level of the tenth or eleventh subcaudal ; sulcus spermaticus forked near the sixth subcaudal; spines largest in the middle region, weakest at the base and towards the summit ; no calyces. Laurent invites attention to the close resemblance of this hemipenis to that of L. rufulus, still more so to that of " Glypholycus hicolor." He should be consulted for comments on its distinctness from Lam- prophis and Boaedon.

Sexual dimorphism. Laurent points out that 5 of the 8 9 9 had 23 midbody scale-rows, whereas all the $ $ had 21. Seem- ingly significant in this species, though not so in hicolor, for dimorphism has progressed to the point where S $ have fewer ventrals as well as more subcaudals; in addition, S tails are always longer {S S 22-24^0; 9 9 14-17%).

Diet. A skink {Mahitya sp.) was recovered from the stomach of one of these snakes. This is especially interesting in view of the fact that Ucolor is a fish-eater, while rufulus and its race appear to prey chiefly on amphibia.

Localities. Angola: Dundo. Belgian Congo: Kalina; Keseki near Kwamouth ; Kunungu, Lac Leopold II; Leopoldville ; Lo- mami; Luluabourg, Kasai.

Range. Western Angola to western Belgian Congo.

Lycodonomorphus rufulus leleupi (Laurent)

lltfiOli. Ablahophis whytei leleupi Laureut, Revue Zool. Bot. Afr., 43. p. 351: Kundelungu (Irsac biological station), 1750 metres, iioithwest Lake Tanganyika, Belgian Congo.

LOVERIDGE: AFRICAN SNAKE GENERA 15

1953. Witte, p. 165; figs. 43 a-c; col. pi. iii, fig. 3; pi. xvi, fig. 2. 1956. Li/codonomorphus wliytei leleupi Laurent, p. 86.

Common Name. Congo Dark-bellied Water-Snake.

Description.^^ Rostral broader than deep, just visible from above ; nostril directed upwards, between two nasals ; internasals as long as, or longer or shorter than, broad, as long as, or shorter than, the prefrontals ; prefrontals moderate ; frontal as long as, or longer than, its distance from the end of the snout, shorter than the parietals; loreal longer than deep, well (only barely in 2 snakes) separated from the first labial, enters orbit on left side of one specimen (No. 34-54) ; preocular 1, rarely 2 (on only 3 out of 40 sides) ; postoculars 2; temporals 1 + 2 (34 sides) or 1 + 3 (14 sides) ; upper labials 8, fourth and fifth enter- ing the orbit ; lower labials 8, first 4 or 5 (6 sides only) in contact with the anterior sublinguals (anterior and posterior fused on right side of No. 112-122), which are subequal to the posterior. Midbody scale-rows 19 to 21 (19 in only 7 of the 24 snakes known); ventrals 164-174^' {S S 165-171; $ 9 164-174); anal entire ; subcaudals (46-52 in type series, fide Laurent, but pos- sibly tips regenerated) 52-67 (9 9 52-61; S S 65-67).

Color. Above, dark olive brown to glossy black, uniform (in- cluding outermost scale-rows). Below, anteriorly whitish to gray; belly of a 230 mm. juvenile is wholly white below, but larger young show progressively encroaching dark patches until the adult condition of a usually black belly sparsely flecked with cream is attained. One subadult from Kabwe River displays a more or less median black streak on its belly, while its tail, like those of the adults, is mostly black.

Size. Largest S (LR.B. 41-70), 540 (417 -f 123) mm.; largest $ (LR.B. 61), 851 (690 + 161) mm. ; smallest $ (LR.B. 4205-06), 230 (182 + 48) mm.

Dentition. Maxillary teeth 24 ; mandibular 24-27 (Laurent : 1956).

Breeding. On April 3, ova were small in a Lusinga $ , but between June 14-18 five adults from the same locality held eggs ranging from 25 to 35 X 15 mm. Similarly were a gravid 9

16 Based on original description ; plus 20 specimens, from 3 localities, each of whicti lias been examined by me for every clniracter cited.

17 162-174 flde Witte (1953 p. 167).

16 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

taken at Kabwe between May 21 and 25, and another on June 1 at Pelenge.

Diet. A young frog (Hemisus sp.) in one; in others various fish {Barhus sp. and Kneria sp.) fide M. Poll (Laurent :1956).

Localities. Belgian Congo: Kabwe River; Kundelungu; Lu- singa; Mukelengia River; Pelenge River (for precise position of these places see Witte, 1953, p. 167 and end map).

Range. Montane streams from north end of Lake Tanganyika to region of Lake ITpemba, Belgian Congo.

Lycodonomokphus rufulus mlaxjensis Loveridge

1953e. Lycodonomorphus rufulus mlanjensis Loveridge, Bull. Mus. Comp.

Zool., 110. p. 253: Ruo River, Mlanje Mountain, Nyasaland. J953i. Loveridge, p. 476.

Common Names. Mlanje White-bellied Water-Snake (Eng- lish) ; chirumi (Manganja).

Description.^^ Rostral broader than deep, just visible from above ; nostril directed upwards, between two nasals ; internasals as long as, or longer than, broad, slightly shorter than the pre- frontals ; prefrontals large, squarish ; frontal as long as, or longer than, its distance from the end of the snout, shorter than the parietals; loreal longer than deep, well separated from the first labial ; preocular 1 ; postoculars 2 ; temporals 1 -f- 2 ; upper labials 8, fourth and fifth entering the orbit ; loAver labials 8, first 4 in contact with the anterior sublinguals, which are subequal to the posterior. Midbody scale-rows 21 ; ventrals 163-169 ( 6 $ 163- 167; 9 169) ; anal entire; subcaudals 51-71 ($ 51; $ $ 60-71).

Color. Above, glossy black; upper and lower lips white, the upper heavily infuscated, the lower only anteriorly so; the outermost scale-row and edges of the one next above it, whitish, more or less obscured with dusky infuscations. Below, faintly pinkish white ; throat to anus immaculate except for an occa- sional tiny fleck (on three ventrals of type) ; tail with a longi- tudinal line formed of dusky median spots.

Size. Largest 6 (M.C.Z.' 51050 :the type), 735 (550 -f 180) mm., only known 9 (B.M. 24.2.1.27), 780 (642 + 138) mm. The tail being 170r of lior total length as against 21-24% for the three males.

i"* Based on :', r^' ' (from Mlanje and Cliolo in .M.C'.Z.) and a $ (Zonilia in Britisli Mnscnin ).

LOVERIDGE : AFRICAN SNAKE GENERA 17

Remarks. L. r. mlanjensis is distinguished from rufulus and all its races except r. leleupi of the Congo, by possessing 21 mid- body scale-rows. From r. leleupi (and laevissimus, of which two individuals with 21 scale-rows have been reported), it is readily distinguished by its immaculate white throat and belly. Whether atldilional Nyasaland material will demonstrate that it should be merged with r. whytii remains to be seen.

Diet. A large ranid tadpole was present in the stomach of the type.

Parasites. Nematode worms (Kalicephalus sp. ; immature As- caridae; also larval Anisakinae, probably Clcoascaris sp.) were abundant in the type.

Temperament. Even Africans appear to recognize the docility of this species as will be seen from their attitude as described by Loveridge (1953e, p. 255).

Habitat. Apparently bide beneath stones in shallow streams.

Localities. Nyasaland: *Muyenda Stream, an affluent of Nswadzi River, Cholo Mountain ; *Ruo River, Mlanje Moun- tain ; *Zomba Mountain.

Range. Montane streams of southern Nyasaland.

Lycodonomorpiii's rufulus wiiytii (Boulenger)

]S97c. (ilypholycus whytii Boulenger, Proe. Zool. Soe. London, pp. 800, 802,

pi. xlvi, fig. 2: Fort Hill, Nyasaland.

1898. .Johnston, p. 3Gla.

ISlijc. Boulenger p. 619.

192'.);i. Werner, p. 50.

iy;i5. Ahlahoplii.s ■whytii Cott, p. 96,1.

1940. Bogert, p. 18.

1952. Ablaboj)liis rufulus wliytei louides, in Swynnerton, p. 10.

1953e. Lycndonomorphus rufulus wliytli Loveridge, pp. 255, 317.

19.'3i. Loveridge, p. 476.

IQHHe. Loveridge, p. 182.

Cumnion Name. Whyte's Water-Snake.

Description.^^ Rostral broader than deep, just visible from

19 Based on ? ? (from Bimgwe : Songca and Fort Hill) and Cott's Charre snake which is too dried and shrivelled to be of much good ; however its tail with ^" subcaudals does .sccoi to be intact. Cunha's two snakes from Massangulo, Portuguese N.vasaland, reported as Ahluhophis rujulufi, were actually lioacdon fulifjino.'iux.

18 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

above ; uostril directed upwards, between two nasals ; internasals as long as, or. longer than, broad, shorter than the prefrontals ; prefrontals moderate ; frontal as long as, or longer than, its distance from the end of the snout, shorter than the parietals ; loreal longer than deep, well separated from the first labial ; preoculars 1-2 ; postoculars 2 ; temporals 1 -f- 2 ; upper labials 8, fourth and fifth entering the orbit (or 9, with fifth and sixth entering on right side of Charre 2 ) lower labials 8, first 4 in contact with the anterior sublinguals, which are subequal to the posterior. Midbody scales without apical pits, in 19 rows; ven- trals 159-168 (all 5 2 ; 172 not 176 as given by Cott if Charre snake included) ; anal entire; subcaudals 46-47 (all 9 5 ; 37-47 if Charre snake included).

Color. Above, dark olive to slate gray, uniform, or the outer- most scale-row on either side pinkish or yellowish white. Below, j)iiikish or j^ellowish white, uniform (Fort Hill), or a few dusky l)atches down the middle of the venter (Rungwe), or a dusky median line commencing on the throat, rapidly broadening to occupy all but the lateral edges of the ventrals and subcaudals.

Size. Largest 9 (M.C.Z. 52637), 709 (590 + 119) mm. from Songea; Rungwe 9 (A.M.N.H. 38984), 625 (519 + 106) mm.; Fort Hill type 9 (Brit. Mils.), 565 (465 + 100) mm.

Dentition. Maxillary teeth apparently 24 on one side of dam- aged Rungwe snake {fide Bogert).

Sexual dimorphism. Tails of these 3 9 9 are from 15-17% of total lengths.

Diet. Hind feet of an amphibian in stomach of Rungwe snake (Bogert) ; liind limbs of a Rana ftiscigula in the Songwe speci- men.

Parasites. The entire digestive tract of the Songwe snake was i-iddled with worms. These have been identified by J. T. Lucker as a 9 Oxyuroidea besides both sexes of a KalicepJialus, probably K. micriirus. In the mesentery were two $ Dracunculus sp. and numerous encapsuled larvae of one of the Physalopteridae.

Localities. Tanganyika Territory: Rungwe Mountain; *Songea boma. Nyasaland: *Fort Hill. Mozambique: * Charre (if refer- able to this race).

Range. Southern Tanganyika Territory and northern Nyasa- land, south ( 1 down eastern side of Lake Nyasa) to the Zambezi at Charre, Mozambique.

LOVERIDGE : AFRICAN SNAKE GENERA

19

Lycodonomorphus rufulus rufulus (Lichtenstein)

1823. Coluber rufulus Lichtenstein, Verz. Dubl. Zool. Mus. Berlin, p. 105:

South Africa. 1826. Fitzinger, p. 57.

1831. Cormiella leucopilus A. Smith, S. African Quart. Jour. (1), No. 5,

p. 17: 1837. Coronella rufula Schlegel (part), p. 74, pi. ii, figs. 18-19. 1843. Traill, p. 136 (as in Schlegel, subcaudal counts in error). 1843. Lyoodonomnrphus rufula Fitzinger, p. 27. 1861. Fitzinger, p. 408.

1847. Lamprophis rufulus A. Smith, pi. Iviii. 1863a, Jan, p. 285. 1866. Jan, pi. iv, fig. 1. 1884a. Muller, p. 284.

1887b. Boettger (part), p. 161 (omit Clarkebury). 1887h. Boulenger, p. 177. 18S9. Boettger, pp. 290, 294. 1890. Miiller, p. 693. 1891a. Matschie, p. 609. 1892a. Boulenger, p. 176.

1854a. Ablahcs rufula Dumeril & Bibron, p. 308. 1858e. Giinther, p. 30. 1867a. Steindachner, p. 62. 1893b. Ahlabophis rufulus Boulenger, p. 318. 1898. Boettger, p. 35. 1898. Werner, p. 143.

1898. Sclater, p. 98. 1902. Lampe & Lindholm, p. 17. 1905h. Boulenger, p. 255. 1907J. Boulenger, p. 486. 1907c. Roux, p. 733. 1908b. Boulenger, p. 228. 1908. (iough (part), p. 21. 1908. Odhner, p. 5.

191011. Boulenger, p. 503. 1910a. Hewitt, p. 57. 19101). Sternfeld, p. 15. 1910a. Werner, p. 355. 1912. FitzSimons, F.W., pp. 82, 91.

1912. Hewitt, p. 267.

1913. Hewitt & Power, p. 161. 1916. Andersson, p. 40.

20 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY

19251). Flower, p. 966.

1929. Flower, p. 181.

1929. Rose, pp. 146, 157, figs. 91-94, 102.

1929a. Werner, p. 48.

1931. Power, pp. 43, 48.

1935. Power, p. 333.

1937e. Hewitt, p. 57, pi. xviii.

1940. Bogert, p. 18.

1946a. FitzSimons, V.F., p. 355.

1950. Fantham & Porter, pp. 601, 603.

1950. Eose, pp. 225, 269, 272, 320, figs. 140-143, 164.-°

Further citations of "rufulus''' will be found under laevissimus and Boaedon f. fuliginosns.

Common Names. Brown Water-Snake (Hewitt). It has also been called Olive-brown Water-Snake by F. W. FitzSimons, South African Rufous Snake by Flower, Black Water Snake by Rose.

Description. Rostral broader than deep, just visible from above ; nostril directed upwards, between two nasals ; internasals as long as, or longer than, l)road, much shorter than the pre- frontals; prefrontals moderate; frontal II/2 to 1% times as long as broad, as long as its distance from the end of the snout, shorter than the parietals; loreal longer than deep, well separated from the first labial ; preocular 1, rarely 2 ; post- oculars 2 ; temporals 1 -|- 2 ; upper labials 8, fourth and fifth entering the orbit ; lower labials 8, first four in contact with the anterior sublinguals, which are subequal to the posterior. Mid- bod}^