jeiCjvxjw_c^\^HE^ HARVARD UNIVERSITY LIBRARY OF THE MUSEUM OF COMPARATIVE ZOOLOGY VA'V^-AvN^, \^Ha. fiDemoirs of tbe fiDuseum of Comparative Zoology AT HARVARD COLLEGE Vol. L. No. 2 A COMPARATIVE STUDY OF THE HERPETOLOGICAL FAUNAE OF THE ULUGURU AND USAMBARA MOUNTAINS, TANGANYIKA TERRITORY WITH DESCRIPTIONS OF NEW SPECIES BY T. BARBOUR and A. LOVERIDGE WITH FOUR PLATES CAMBRIDGE, U.S.A. IPrinteC for tlje Museum December, 1928 fiDemoirs of tbe fiDuseum ot Comparative Zoology AT HARVARD COLLEGE Vol. L. No. 2 A COMPARATIVE STUDY OF THE HERPETOLOGICAL FAUNAE OF THE ULUGURU AND USAMBARA MOUNTAINS, TANGANYIKA TERRITORY WITH DESCRIPTIONS OF NEW SPECIES BY T. BARBOUR and A. LOVERIDGE WITH FOUR PLATES CAMBRIDGE, U.S.A. IPrinteD for tFje /IDuseum December, 1928 ^ A COMPARATIVE STUDY OF THE HERPETOLOGICAL FAUNAE OF THE ULUGURU AND USAMBARA MOUNTAINS, TANGANYIKA TERRITORY ^YITH DESCRIPTIONS OF NEW SPECIES By T. Bakboue and A. Loveridge The material discussed in the following pages was collected between September and December, 1926, by the junior author while on an expedition in the interest of the Museum of Comparative Zoology. Both authors have collaborated in a critical examination of the material. When, in the field notes on habits, the singular pronoun is used, it will be understood that it is the collector who is recording. A selection of the specimens obtained has been presented to the Musemn of the University of Michigan, and paratypes of fifteen of the new forms have been sent to the British Museum. CONTENTS Introduction 87 List of Collecting Stations 96 List of Species Collected Tortoises 104 Snakes 104 Lizards 138 Chameleons 170 Caecilians 178 Toads and Frogs 184 Bibliography 263 INTRODUCTION The object of the expedition was prhnarily to discover to what extent the little- known fauna of the Uluguru Mountain rain-forest was related to that of the better-known Usambara Mountains. The two ranges are separated from one another by at least a hundred and twenty miles of hot, low-lying country, while the Nguru Mountains form a small connecting-link between. 88 BARBOUR AND LOVERIDGE We consider that the results show a remarkable affinity in the forms inliabit- ing these ranges, which argues that deforestation accompanied by a lessened rainfall has occurred in the intervening area. Both ranges contain many forms common to the West African forest but this is more noticeably the case with the Usambara Mountains. The Uluguru Mountains on the other hand are shown to support forms (Chrysochloris sp. n. and Scolecomorphus spp. n.) wliich link them with the Nyika Plateau of Nyasaland. The easiest way to demonstrate this relationship is by giving a list of all the known reptiles and amphibians from these mountains. Where numbers are given they represent the number of specimens collected during the 1926 expedition now reported upon. A second table is given showing savannah forms which occur on the lower slopes of the Uluguru Mountain, based on a collection made near Government House, Morogoro, in 1918 by the junior author. Such species are marked with an 'M.' They are almost all widespread forms which do not occur in the rain-forest. It might be well to take this opportunity to refer to the many records of rare, or West Coast, species credited to Tanga, a coastal port lying between Mombasa and Dar es Salaam and only about fifty miles from Amani and the Usambara Range. Many of O. Neumann's specimens were labelled 'Tanga am Usambara.' To-day the Usambara Mountains are in the political province of Tanga. In Ger- man times Sigi (an hour's walk from Amani) was linked with Tanga by railway and it was customary for residents at the coast to spend their vacations, or recuper- ate, in the mountains; in fact at Mlalo near .Ambangula there is a sanitarium. It appears probable that many species purporting to come from Tanga in reality are from the mountains, otherwise it is difficult to understand why species that can live on the coastal plain at Tanga have not spread to Mombasa or Dar es Salaam. Among the minor objectives of the trip we ' had set out to secure topotypic material of the six snakes, four lizards, nine chameleons, and nine amphibia which had been described from the Usambara range. We were so far successful that twenty-five of the twenty-eight were collected, each in its own type locality. The third expectation of the trip was to find new species in the little-known range of the Uluguru Mountains; not only was this hope fulfilled in the case of mammals and birds, but also among the reptiles and amphibia of which the undermentioned genera, species or races are now described for the first time. ' The junior author was accompanied by his wife. AFRICAN HERPETOLOGY 89 Snakes. Typhlops uluguriuensis sp. n. Prosymna ornaiissima sp. n. Aparallactus uluguruensis sp. n. Lizards. Hemidactylus persimilis sp. n. Agania mossambica montana sub. sp. n. Agama colonorum iisambarae sub. sp. n. Scelote^ uluguruensis sp. n. Caecilians. Scolecomorphus uluguruensis sp. n. Scolecmnorphus attenuatus sp. n. Boulengerula uluguruensis sp. n. Frogs. Hyperolius mariae sp. n. Megalixalus uluguruensis sp. n. Leptopelis parkeri sp. n. Lepiopelis uluguruensis sp. n. Breviceps usambaricus sp. n. Hoplophryne iiluguruensis gen. et sp. n. Hoplophryne rogersi sp. n. Parhoplophryne usambaricii,s gen. et sp. n. The Following Changes in Nomenclatuke or in East African Records are Proposed Typhlops gierrai Mocquard becomes Typhlops punclatus gierrai (Mocquard). Glaucoma conjuncta Angel (not of Jan) = Leptotyphlops distanti (Boulenger). Leplodira lornieri Werner becomes CrotaphopelHs hotamboeia tornieri (Werner). Calamelaps polylepis Bocage = Calamelaps unicolor Reinhardt. Elapops modestus Angel (not of Gtinther) = Aparallactus concolor (Fischer). Elapsoidea nigra Giinther = Elapsoidea guentheri Bocage. Gymnodactylus africanus Werner becomes Paragonatodes africanus (Werner). Hemidactylus mabouia (part) Loveridge (not de Jonncs) is Hemidactylus sp. n. Agama colonorum (part) Loveridge (not of Daudin) is Agama mossambica mossambica Peters. Agama cyanogaster Angel (not of RUppell) = Agama atricollis Smith. Mabuiopsis jeanneli Angel = Mabuia irregularis Lonnberg. Ablepharus massaiensis Angel = Ablepharus wahlbergii (Smith). Chamaeleo tenius excubitor Barbour becomes Chamaeleo fischeri excubilor (Barbour). Bdellophis vittatus Boulenger becomes Scolecomorphus vittatus (Boulenger). Scolecomorphus kirki (?) of Loveridge (not of Boulenger) is Scolecomorphus sp. n. Boulengerula boulengeri Loveridge (not of Tornier) is Boulengerula sp. n. Rana aberdariensis Angel = Rana tiidti Boulenger. Rana merumontana Lonnberg becomes Rana fasciala merumonlana Lonnberg. RarM merumontana Loveridge (not of Lonnberg) = Rana galamensis Dumeril and Bibron. MaUgania bufonina Boettger is recognised as young of Rana adspersa (Dumeril and Bibron). Phrynohatrachus boulengeri De Witte = Phrynobatrachus acridoides Cope. Hylarthroleptis janenschi Ahl = Phrynobatrachus aC7-idoides Cope. Arthroleptis variabilis Matschie = Arthroleptis slenodactylus Pfeffer. Arthroleptis whytii Boulenger = Arthroleptis slenodactylus Pfeffer. Arthroleptis lonnbergi Nieden = Arthroleptis slenodactylus Pfeffer. Arthroleptis niethneri Ahl = Arthroleptis slenodactylus Pfeffer. Arthroleptis brei'ipes Ahl = Arthroleptis slenodactylus Pfeffer. Arthroleptis wahlbergii Procter and Loveridge (not of Smith) = juo. Arthroleptis slenodactylus Pfeffer. Hylamhates vermiculatus Boulenger becomes Leptopelis vermiculatus (Boulenger). Hylambaies johnstoni Boulenger becomes Leptopelis johnstoni (Boulenger). Hylambates argenteus Pfeffer = Hylambaies maculatm Dumeril. Breviceps verrucosus Tornier and Nieden (not of Rapp) is Breviceps sp. n. 90 BARBOUR AND LOVERIDGE Material Collected The four months collecting resulted in 4,039 specimens comprised as follows: SNAKES Four hundred and six snakes, referable to thirty-two species, were collected. Nieden (1910) lists seventeen species as occurring at Amani; all of these were taken except Dendraspis angusticeps. It need not be supposed that Amani is exceptionally rich in serpent life. A great many species have been recorded from there as a result of the extensive clearing of forest for plantations; moreover, many of the natives engaged on cleaning the plantations are Wanyimwezi whose attitude to snakes is fearless. Owing to the interest shown by Mr. F. N. Rogers, the Custodian of .^\inani Institute in 1926, his men brought me any snakes they came across during their day's work, so that over a hundred snakes, represent- ing twenty-one species, were obtained there in three weeks. LIZARDS Two hundred and sixty-four lizards, referable to twenty-five species, were collected. Nieden (1910) lists seven species as occurring at Amani; all of these with the exception of Mabuya ?naculilabris maculilabris were taken. Fifteen species were collected in the Usambara and sixteen in the Uluguru. Lizards are certainly scarce in these mountains. CHAMELEONS Four hundred and twenty chameleons, referable to eight forms, were col- lected. Chameleons are much more abundant in the Usambara range than they are in the Uluguru INIountains. C.AECILIANS Two hundred and fortj'-six caecilians, referable to five species, were collected. Only two species were previously known from Tanganyika Territory, both hav- ing been described from the Usambara Mountains. The three new species come from the Uluguru range. TOADS AND FROGS Two thousand, seven hundred and three tailless batrachians, referable to fifty species, were collected. Nieden (1910) hsts fifteen species as occurring at Amani; all except three were taken there, and seven others not listed by Nieden were also found at Amani. AFRICAN HERPETOLOGY 91 I. FAUNA OF, OR ON EDGE OF, IL\IN-FOREST BETWEEN 3,000 AND 8,000 FT. (Species marked with an * occur at lower levels also) bNAKES Uluguru Mtns. Usambara Mtns. References Typhlops uluguruensis sp. n 4 — Typhlops kleebergi — X a Typhlops punctatus gierrai and intermediates 1 11 b *Natrix olivaceus 30 30 *Boaedon lineatus 4 17 c Lycophidion meleagris 2 5 *Lycophidion capense 7 5 Lycophidioii semicinctum — X d *Mehelya capensis — 3 *Mehelya nyassae — X e *Chhrophis neglectus 15 14 Chlorophis macrops — 22 *Philothamnus semivariegalus 1 9 f *Prosymna amhigua M X g Prosymna ornatissima sp. n 4 — Homalosoma lutrix — X h *Dasypeltis scaber 1 1 Geodipsas vauerocegae 7 11 Geodipsas procterae 2 — Crotaphopeltis hotambocia tornieri 25 20 i *Thelotornis kirllandii 2 13 *Dispholidus typus — 7 j *Calamelaps wiicolor 1 3 Chilorinophis butleri — X k Aparallactus werneri 3 46 Aparalladus uluguruensis sp. n 3 7 Elapsoidea guentheri (including nigra) 5 42 Naja nielanoleuca — X 1 *Causus defillipii 2 — Bills gahonica — 4 Atheris ceratophorus — 2 *Atractaspis rostrata 4 — Lizards Paragonatodes africanus 1 21 *Hemidaclylus viabouia 2 4 *Hemidaciylus persimilis sp. n 1 — Lygodactylus fischeri 1 18 m *Lygodactylus grotei~. 1 1 *Agam.a mossamhica montana sub. sp. n 17 23 Agama colonorum usambarica sub sp. n — 14 *Varanus niloticus seen 1 Laceria jacksoni — X n Holaspis guentheri ? seen 2 *Gerrhosaurus major M X o *Mabuya comorensis 3 28 *Mabuya varia varia 11 6 *Mabuya striata 1 2 Siaphos kilimensis 9 4 Scelotes eggeli — 21 Scelotes uluguruensis sp. n 13 — Scelotes tetradactyla 1 — jsambara Mtns. References 3 80 p 295 6 X q 1 2 r 1 X s 92 BARBOUR AND LOVER IDGE Chameleons uiuguru Mtns. *Chamaeleo dilepis dilepis 2 Chamaeleo fischeri matschiei X Chamacleo fischeri muUiluberculatus — *Chamaeleo deremensis 2 Chamaeleo melleri M Chamaeleo spinosus — Chamaeleo tenuis X Rhamphnleon brevicaudatus 12 *Rhampholeon kersteni — Caecilians Scolecomorphus villains 6 5 Scolecomorphus uliiguntensis sp. n 132 — Scolecomorphus attenuatus sp. n 2 — Boulengerula boulengeri — 58 Boulengerula idugurvensis sp. n 43 — Toads and Frogs *Xenopus laevis X 4 Bufo brauni 29 35 JSu/o micranolis 1 — Neclophryntrides tornieri 134 15 Nectophrynoides vivipara 166 — Rana nulti 53 Rana fasciala merumonlana — *Phrynobatrachus krefftii — *Phr!/nobatrachus ogoensis ? 157 Arthroleplides marliensseni 24 * Arlkrole ptis stenodactylus 37 Arlhroleplis adolfi-friederici 32 Arthroleptis sehubolzi 5 Arlhroleplis xcnodaclylus 22 Arlhroleplis minulus — Phrynopsis usarnbarae ■ — Chiromanlis rufescens — Hyperolius concolor 1 Hyperolius mariae sp. n ; — *Hyperolius punclicidaliis X Hyperolius argus 69 *Megalixalus loveridgii 22 Megalixalus uluguruensis sp. n 16 Leplopelis aubryi — Leplopelis rufiis 20 Leplopelis idugiiruensis sp. n 22 — Leplopelis parkcri sp. n 1 — Leplopelis vermiculalus — 1 Leplopelis johnsloni 4 — Z3 Callulina krcffli 24 26 Spelaeophryne melhneri S — Breviceps uluguruensis 55 — Brcficeps rugosus 43 — Breviceps usamharicus sp. n — 105 zi Hoplophryne uluguruensis sp. n 42 — Hoplophryne rogersi sp. n — ■ 4 Parhoplophryne usambaricus sp. n — 1 136 u 106 156 3 11 197 153 19 X V X w X X 70 y 106 165 z 2 11 Zl X Z2 AFRICAN HERPETOLOGY 93 KEY TO REFERENCES OF TABLE I a. Werner's description in 1004 of the type from Usambara. b. Matschie's record of T. rschri^hti and Xieden and Sternfeld's records of T. piinctatus are prob- ably punclaliis typica or intermediates. c. Matschie's record of Boodon capensis from Derema. d. Sternfeld's record of Usambara. e. Werner's type of Gonionotophis degrijsi from Usambara. f. Matschie's record of P. neglectus at Derema and Sternfeld's of P. thomensis at Nguelo. g. Sternfeld's record from Usambara. h. Sternfeld's record from .\jnani. i. First recorded from the Usambara by Werner as Leptodira rufescens. ]. Commonly reported by natives as occurring at Nyange in the Uluguru Mountains. k. Sternfeld's record of Amani. 1. Sternfeld's record of Usambara. m. Matschie's type of L. conradti in 1892, Werner's 1895 record of L. thomensis and probably Nieden's 1910 of L. capensis. n. Xieden's record of Usambara. o. Werner's type of G. bergi. p. Tornier's record of Ukami. q. Tornier's record of Usambara. r. Tornier's record of Ukami. s. Tornier's record of ICerogwe at foot of Usambara Mountains. t. Loveridge's record of Bagilo, Uluguru Mountains. u. Nieden's record of Ukami and Amani Rana delalandii (= angolensis). V. Ahl's type from the Usambara. w. Tornier's record of Lewa in Usambara. X. Matschie has recorded H. concolor from Derema; perhaps these are H. mariae. y. Matschie has recorded H. cinctiventris from Derema; the determination is a little doubtful. z. Ukami and Usambara records of M. fornasinii and M. leptosomus are referable to this species, zi. Nieden has examined all old German East Africa records of aubnji and concludes that they are all riifus except those from the Usambara. Z2. Nieden's (1915) records from Amaniand Derema. z3. Reported by Lonnberg from Mombo at foot of the Usambara Range. 24. Nieden's records of B. verrucosus from Magrotto in Usambara. 94 BARBOUR AND LOVERIDGE II. SAVANNAH OR OPEN FOREST FAUNA OCCURRING UP TO 3,000 FT. (For species common to rain-forest also see those in Tabic I marked with an *) Q *Morogoro Amani Reforenres to oNAKES etc. etc. Aniani column Typhlops mandensis M — Typhlops punclalus pundatus — X a Typhlops mucntso M X b Typhlops lumbridformis — X c Typhlops unitaeniatus — X d Lcplotyphlops dislanii M — Python scbae M X e Tarbophis semiannulatus M ? f Crotaphopeltis hotamboeia hotamboeia 1 1 g Amplorhinus noiolacnia M — Rhamphiophis oxyrhynchus M X h Psammophis sibilans M X i Psammophis subtaenialus M X j Naja nigricollis M X k Dcndraspis angusticeps M X 1 Bitis arieians M X m Lizards Heinidadylus squamulatus M — Hemidactyhis ruspolii M — Hanidactylus brooki M X n Lygodaclylus picturaius picturatus 2 X o Zonurus tropidostcrnuni M — Lalastia johnstoni M — Gastropholis mitata ? X p Eranias spekii spekii M 1 Gerrhosaurus flamgularis flamgidaris M — Mabuya maculilabris maculilabris 1 X q Mabuya megalura — 1 Riopa sundevallii M 1 Ablcpharus wahlbergii M — Melanoseps ater longicmida 1 X r Toads and Frogs Xenopus mulleri 7 X s Bufo regularis 3 3 Rana oxyrhynchus M 1 Hyperolius fulvotrittalus 11 17 Hyperolius microps — • 2 Brenccps mossaynbicus 3 X t Phrynomcrus bifasciala ? X u Hemisus mamoratus M — * An ' M ' in the Morogoro column implies the species has been collected there by Loveridge. AFRICAN HERPETOLOGY 95 IvEY TO REFERENCES OF TABLE II J- k. 1. m. n. o. P- q- r. s. t. u. Nieden and Sternfeld's records, see 'b' of Table I. Sternfeld's record from Tanga. It may be remarked that the colouring of T. p. gierrai is almost the same as T. mucruso. Werner's type of T. kleebcrgi from Usambara is treated as specifically distinct (see Table I) so that it is doubtful if this species has actually been taken in the mountains. Sternfeld's record from Bukuri in Usambara. Werner also. Occurs at Sigi below Amani. ? Werner's record of Dipsas ohtusa Gmelin in 1895. Werner'.s record of Lcptodira rufescens is C. h. iornieri. Sternfeld's record of Tanga. Psammophis regularis Sternfeld has since been recorded by Sternfeld from Amani; if he is correct in this, undoubtedly regularis, which was described from the Cameroons, is merely a fortuitous variation of sihilans. Under any circumstances it is difficult to imagine a Psammophis occurring at Amani; perhaps it came from Sigi. Sternfeld's record of Usambara. Two in the collection of the Amani Institute are without locaUty but the species occurs at Sigi. Nieden's record of Amani. Sternfeld's record of Usambara. There is a specimen in Amani Institute from Sigi. Tornier's record from Derema. Tornier's record of Majamboni in Usambara. Werner's type of G. prasina. . Nieden's record of Amani. Tornier's type from Kerogwe at the foot of the Usambara Mountains. Undoubtedly the species occurs at Morogoro. Nieden has recorded it from Usambara. Tornier's records from Magila and Rubugwe in Usambara. Undoubtedly the species occurs at Morogoro. Nieden has recorded it from Usambara. SUMMARY Snakes. 34 species known from the Uluguru and 39 from the Usambara Range; of these 26 are common to both. Liz.^RDS. 27 species occur in the Uluguru and 23 in the Usambara Range; of these 18 are com- mon to both. Ch.^meleons. 6 forms are known from the Uluguru and 9 from the Usambara Range; 6 are com- mon to both. Caecilians. 4 forms occur in the Uluguru and 2 in the Usambara Range; only one is common to both. Frogs and Toads. 33 species are known from the Uluguru and 32 from the Usambara Range; about 20 are common to both. 96 BARBOUR AND LOVERIDGE ITINERARY AND LIST OF COLLECTING STATIONS Bagilo. Camped on a little knoll near the first huts of Bagilo 'Village' on the Morogoro-Bagilo path. Very little was collected in the vicinity, most of such work being done in the rain-forest a mile or more from Bagilo on the track to Morogoro. The forest floor was covered with leaf mould and proved a favour- ite haunt of Breviceps and Bufo, while guerezas and squirrels lived in the tree- tops. These lichen-laden trees are so tall that they must be seen before their stupendous size can be appreciated; their big limbs support masses of ferns, while lianas and other tropical epiphytes depend from them like ropes; these are much used by the squirrels. A broad river flows through the forest in the valley bottom, becoming a rushing, rock-studded torrent in the open. Here Arthroleptides occurs in suitable spots. The river is fed by tributary streams so choked with fallen trees and smothered in vegetation that following their courses was difficult. Along their banks grew the wild banana plants beloved by Callulina and Leptopelis. Above the camp was another much drier patch of forest that was unpro- ductive except for birds. Across to the west, in damp forest intersected by many moist ravines, Nectophrynoides and Arthroleptis were procured. Curiously enough it was the only patch of forest in the vicinity where one could count on obtaining these genera ; elsewhere their occurrence was fortui- tous. Here also, in open glades not far distant from a magnificent, forest-girt waterfall, Crotaphopeltis and Scelotes were discovered beneath fallen logs. Nyange. The camp site was one of necessity rather than of choice, for I was conducted to it one rainy morning by the chief, who had already caused huts to be erected for my native personnel. It was an ideal camp site but, being situated in the valley, and to a great extent surrounded by the maize plan- tations of the villagers, it involved long tramps daily to the forests which capped the surrounding hills. Mammal collecting was disappointing in the extreme, but birds were abundant in the forested hills to the east. It rained nearly every day that I was at Nyange and in consequence the natives were kept busy weeding their plots. In many places these clearings abutted on the forest, or were even on ground quite recently covered by forest, of which smouldering logs and stumps were all that remained. As the parents weeded, the children gathered the grass and uprooted vegetation into heaps which were left scattered about on the somewhat bare ground, for the maize was little more than a foot high in most of the cultivated areas. AFRICAN HERPETOLOGY 97 The Nyange children took four days to appreciate the advantage of a market for snakes, but when they at last understood that fifty cents (= 12 cents U. S. A. currency) would be paid for each one, they lost no more time but brought in nearly a hundred in a week. I shall never forget the Nyange kiddies bringing in their catches in old whisky bottles with maize-cob stop- pers, or their peals of laughter as I shook out the snakes, or as they went racing down the path to turn over more heaps or logs ; for it was in such situations that they found them. I have to thank the Nyange boys for the only three new snakes taken during the trip, a Typhlops, Prosymna, and Aparallactus. Agamas occurred on the rocks along the banks of the stream; the only really abundant lizard was Mabuya varia varia. The reptile fauna of the val- ley reflected the conditions, for it was such as might be found at Morogoro at the foot of the mountains. Caecilians were also obtained in the gardens, but salientia, with the single exception of Phrynohatrachus ogoensis, were very scarce. A great effort had to be made to secure even the small series of Spelaeophryne methneri, while those of the new genus Hoplophrytie were not obtained in the valley but in a moist ravine on a mountain top many miles away. Mkarazi. a broad shallow river frequented by cormorants and Varanus niloticus flows in the valley. On either side of it is a remnant of forest and another patch occurs on the top of a hill to the east of the road. This forest is very dry and there are no guerezas in it. Blue monkey, lemur, Peter's elephant- shrew, Nandi cat, and hyrax, however, are to be found, and I imagine that be- fore the deforestation of the surrounding country took place this was prob- ably rain-forest. Trumpeter hornbills and guinea-fowl were collected but the avifauna was essentially similar to that at Morogoro. Scelotes tetradactyla occurs alongside siaphos kilimensis. The new Boulengerula uluguruensis, and Spelaeophryne methneri, were also found in a damp spot beneath a large log; apart from these creatures the fauna was uninteresting. Mkangazi. a forced halting place on the uphill climb to Nyingwa. The coun- try is largely under cultivation with the exception of one or two small clumps of trees; there is no trace of forest, the massive rounded mountains being covered with dry stubble or native potato plants. I arrived at 1 p.m., and the rain descending at 3 p.m. continued till far into the night. Left at daybreak. Two days before passing the place on my return journey I sent word to the chief that I would halt for two hours and buy anything that the small boys would bring in except certain species which I enumerated. Rhynchocyon 98 BARBOUR AND LOVERIDGE petersi, Sarothrura rufa elizabethae, and Bufo micranotis were the only rarities so obtained. Tawa. a low-lying village in hot cultivated country. Halted for two hours to give the porters a rest and allow stragglers to catch up. Some children brought a few Megalixalus and Hyperolius. With this exception, though warned in ad- vance two days before, the people had caught nothing. While lunch was being prepared I did some collecting in the vicinity. Nyingwa. For some reason best known to himself the petty chief at Nyange pictured Nyingwa as a naturalist's Eldorado 'and it was only a day's march.' The chief was old but he must have been very active on the day that he accomplished it and he certainly did not carry a porter's load. It took us three days, for it involved a climb of 5,000 feet and the going for the last two miles was very hard on the porters. The foot of the mountain is well cultivated and contains many small banana plantations in which great numbers of Hyperolius argzis occur. As we climbed up, first Jumbe Saidi and then his father Jmnbe Chibi urged us to camp, but I insisted on being conducted up an hour's walk beyond the last huts, till we found a suitable clearing on the edge of the forest. There was a great deal of bamboo growing in the forest, and as a token of altitude we had a few giant groundsel below the camp — these groundsel were ten or twelve feet high ! The camp site was mainly selected on account of the croaking or calling of frogs in the vicinity. It proved rich collecting for amphibia but not a snake or a lizard was seen during the whole week ; some lizard eggs were the only sign of the existence of reptiles at this altitude though some open spaces seemed ideal for Mabuya. Why Nyingwa should be lacking in reptiles when they occur at greater heights on other mountains, remains to be explained. Abbott's Duiker dwells in the forest but is rare; the Mountain Duiker is common, but wary through much hunting. The forest is rich in bird-life and several species which were unknown to me were secured. Noteworthy among the amphibians was the occurrence of two species of Scolecomorphus as well as Nectophrynoides and Breviceps. These were the dominant forms, the last two being associated with the bamboos. ViTURi. WTien at Nyange, news was brought to me of Anomalurus at a place called Vituri some three hours north of our camp. The path wound up one mountain and down another through forest all the way. It was exceedingly AFRICAN HERPETOLOGY 99 difficult walking — or climbing — for the porters, but, travelling unliampered with a single attendant gun boy, I accomplished the journey in about two hours. The scattered huts of Vituri mark the northern limits of the rain- forest and, as one emerges from the latter, there is a perfect view across the shimmering plains to Nguru Mountain, which forms a Unk between the Ulu- guru and Usambara ranges. Vituri forest was comparatively dry at the time of our visit, and collecting was disappointing till I found a deep and moist ravine with wild bananas growing on its slopes. In these bananas were obtained Megalixalus, Hj-pero- lius, Nectophryne and a fine series of LeptopeUs, two of the three species being new. The snake and lizard fauna differed but little from that of Morogoro ex- cept that it contained a few more mountain forms and lacked a good manj' plains species. Siaphos kilimensis, a new Scelotes and Melanoseps ater longi- cauda were found. Below the camp the mountains sloped away very abruptly to the plains below. Half an hour after leaving camp one entered upon a reddish soil like that of Kilosa together with its typical flora, maiombo bush, and the old familiar fauna of common things associated with it. TtJMVi. This marked the foot of the mountains and was about the same altitude as Morogoro. It seemed strange to see the larger Agama mossambica 7nossam- bica on the tree-trunks in place of Agama m. montana subsp. n., Eremias spekii running about the paths once more and Mabuya striata plentiful again. No halt was made, but half-a-dozen species of lizards were seen in a mile, while in the mountains scarcely as many could be found in a day, even when they were searched for. Dar es Salaam. While we awaited the arrival of Mrs. Loveridge from England, collecting was carried on in a swamp to the south of the town, and very near the ocean. Here was found the very rare Hyperolius microps discovered by Livingstone in Rovuma Bay farther south. The other amphibia and reptiles were typical of the coastal plains. Amani. This place is world-famous for its Agricultural Research Institute and Arboretum. The Institute is at an altitude of 3,000 feet, yet, as it is only fifty miles from the sea as the crow flies, it appears to be as cool as 5,000 feet in the Uluguru Mountains. Climatically the different levels are rather con- fusing. A week before our arrival eight inches of rain had fallen in twenty- 100 BARBOUR AND LOVERIDGE four hours, being the heaviest deluge recorded since 1905. Its effects were shown in the landslides and washaways which blocked the road. At the foot of the mountain (Sigi, possibly 1,000 feet) palms have been assembled from every tropic clime, and as one ascends, the place is planted with acres of various trees from each of the five continents. Every species is represented by well-labelled groves, not single trees, and the winding drives are kept in beautiful condition. As one ascends to Mt. Bomoli (above the Institute) fir and pine forests replace the blue gums, while heather covers the open spaces about the rocky outcrops. The house which we occupied at Amani was in the grounds of the arbore- tum, yet close to the virgin forest; it was along the well-made paths of the latter that most of our collecting was done. The forest defies description; the trees are of immense height, and consequently arboreal animals — guerezas, blue monkeys, squirrels — find safe refuge and are niunerous. Of squirrels three genera were collected in one day. Forest birds are everywhere, perhaps the most notable, either for their beauty or their calls, being Turacus fischeri and Heterotrogon vittatum viltatuvi. Tree-planted hillsides, free from undergrowth but strewn with dead leaves beneath which small Arthroleptis find a moist retreat, rotting logs with sod- den interiors, marshy bottoms, and rank sedges in some ravines, swift streams flowing through well-kept pastures in others — these are but a few of the great variety of habitats which Amani has to offer small creatures. The streams emerge from the wooded valleys but arboreal frogs are scarce — or were so at the time of our visit. Hylanibates veriniculatus was obtained only after an infinity of trouble. Callulina kreffti was not uncommon but the only really abundant species were Breviceps usambaricus sp. n., Phrynobatrachus krejjtii in a certain stream in the forest, and Rana nutti where the same stream flowed through cleared lands. The arboreal lizards were the more interesting, Paragonatodes africanus occurring in the forest, and Holaspis guentheri on the outskirts, but very rare. Mabuya comorensis is the common skink of the forest paths. Chameleons of seven species were collected, and twenty-one species of snakes also, of which twenty were secured in nineteen days. Most of these were caught by natives engaged in cleaning the plantations. Arboreal, terrestrial, and burrowing vipers {Athens ceratophorus, Bitis gabonica, Atradaspis rostrata) were present but the only really common venomous snake was Elapsoidea guentheri. AFRICAN HERPETOLOGY 101 Derema. Very near to .\mani, the boundary being a small stream. Derema is somewhat lower. A large swamp, caused by the damming back of a stream flowing down the valley to the saw mill, is bordered by sedges in which Mega- lixalus and several species of Hj^jerolius find a refuge. Forest is present on the surrounding hills which have not been cleared for coffee planting, and sup- ports a fauna similar to that of Amani. Chameleons are abundant on the coffee trees, where they perform useful work in capturing insects. MisALAE. Little collecting was done at this spot. We arrived at 11.30 p.m. in pouring rain and would have left at dawn but for the difficulty of obtaining porters. The camping place was on land long since cleared of forest and now devoted to native gardens. Outcrops of rocks provided basking places for Mabuya varia varia; and Agama mossambica montana occurred on the trunks of the few trees left standing in the gardens. KizERUi. A day was spent here, as the previous day's march had been in driving rain most of the way and only the clothes we stood up in were not soaked. The forest here is magnificent but the natives are accomplished poachers, with game fences and pits scattered through the bush, so that animals are not very abundant. We devoted our attention to amphibia, there being little sunshine to tempt reptiles from their retreats. A Bitis gabonica was brought in and the chief assured me that B. arietans occurs on the hillsides some dis- tance below. From bush and grass country in the vicinity we also received a tree viper {A . ceratophorus) . The most interesting lizards found were Para- gonatodes africanus and Siaphos kilimensis; while in the few wild, and more numerous domestic bananas, Nectophrynoides, Callulina, and Hyperolius fulvovittatus occurred. Conditions reminded one of those at Nyange in the Uluguru range. KizARA. We halted for breakfast in the well-populated valley, which seemed very hot after the deliciously cool rain-forest through which we had come while crossing the mountains separating Kizerui from Kizara. Owing to the steep- ness of these peaks, the chief at Kizerui had furnished us with double the num- ber of porters requested. The Kizara chief and his men, informed a week be- forehand of our arrival, met us in the mountains a mile and a half above the village. After the usual greetings a hundred of them formed into a singing, drum-beating procession and so escorted us to the rest-camp, passing several bands of hand-clapping women on the way. To offset this royal welcome they 102 BARBOUR AND LOVERIDGE had only a boomslang and a couple of chameleons to show us; so, the day- being young, we breakfasted at the mission and pushed on and up to the mountains again. Mt. Lutindi. This was selected as being one of the highest spots in the Eastern Usambaras. It is at the very edge of the broad, hot valley which divides the eastern from the western ranges. Collecting was disappointing, as the forest was very dry and the boulder-strewn, wind-swept summit of the mountain was covered with short gi'ass and heather-like vegetation wherever the shallowness of the soil permitted its survival. On two aspects the mountain side is precipitous, with a sheer drop of a thousand feet to the valley below. The fauna is that of Amani but poorly represented by a few examples of a few species. The only amphibian of outstanding interest taken was Lepto- pelis aubryi and its young. BuMBULi. Situated in the Western Usambaras, Binnbuli is separated from Mt. Lutindi in the Eastern by a broad sandy valley inhabited by the savannah fauna of widely spread species. The Western Usambaras differed from the Eastern in the greater extent to which deforestation had been carried. Much of the country through which we passed on the way here was clothed with short grass on which the flocks and herds of the villagers were grazing. The spot at which we stayed at Bumbuli was in a warm valley just below a mountain whose lower slopes were planted with bananas. The summit was clothed with what should have been rain-forest but at the time of our visit was very dry, yet remnants of a rain-forest fauna were still present ; among the birds triunpeter hornbills and plantain-eaters were conspicuous. Paragon- atodes africanus, Siaphos kilimensis, Bufo brauni, and numerous other rain- forest species occurred among the reptiles and batrachians. Phillipshof. Camp was made on the site which had been selected by the Field Museum Expedition a couple of months previously : rolling downs of grazing land, marshy swamps and slow-flowing streams in the bottoms, scattered patches of rain-forest on the uplands, with vast stretches of forest nearby on either side of the Malindi road. Mammals and birds were numerous in this forest, but reptiles, with the exception of chameleons, were remarkably scarce; in fact I caught only one snake {Chlorophis neglectus) during the fortnight we spent at Phillipshof. It is the type locality of Chamaeleo fischeri vmUihiber- culatus, of which a long series was secured in the hope of discovering the range AFRICAN HERPETOLOGY 103 of variation and exactly how many races occurred in the Usambara Mountains. Some interesting observations on the egg-laying habits of Lygodactyhis fischeri were made. The amphibia presented little of interest apart from the finding of Lonnberg's Rana merumontana, thus establishing another record demonstrat- ing the homogeneity of the fauna of the East African mountains. APPROXIMATE ALTITUDES OF COLLECTING LOCALITIES* Uldguru Mountains Bagilo 5,000 ft. Mkangazi 3,000 Mkarazi 1,500 Nyange 2,500 Njdngwa 7,500 Simbini 3,000 Tawa 1,000 Vituri 2,000 Usambara Mountains Amani 3,000 ft. Bumbuli 2,000 Kizara 3,000 Kizerui 4,000 Lushoto 5,000 Misalae 3,000 Mt. Lutindi 4,000 Phillipshof 5,500 Sigi 1,000 Soni 2,000 * We should like to thank Mr. H. P. Rowe, Deputy Director of Tanganyika Surveys, for assistance in estimating some of these altitudes. Also Dr. Sandground for identifying the helminthological material, and the artists. Miss H. M. Saunders, Miss O. Otis and Mr. J. Henry Blake, for their excellent drawings. 104 BARBOUR AND LOVERIDGE SYSTEMATIC LIST OF SPECIES COLLECTED TESTUDINIDAE Pelusios nigricans castaneus (Schweigger) Emys caslanea Schweigger, 1814, 'Prodr. Chelon.,' p. 45. A small water tortoise was brought to me by two boys, who said that they had picked it up on the road about ten miles south of Nyange, Uluguru Mountains. I could obtain no other information as to the occurrence of tortoises in these mountains beyond Salimu's ' statement that they are found from time to time but are nowhere common. The word kikui is employed by the Wakami for all species of tortoises. TYPHLOPIDAE Typhlops uluguruensis sp. n. 4 (M. C. Z. 230S0-3) Nyange, Uluguru Mtns., 6-11. x. 26. Type. No. 23080. Museum of Comparative Zoology. Sex cf. From Nyange, Uluguru Mountains, Tanganyika Territory. Collected by A. Loveridge, Octo- ber 6th, 1926. Paratypes. Nos. 23081, 23082, 23083. Affinities. Apparently most nearly related to Typhlops kleebergi Werner and T. gracilis Sternfeld, though in some respects not unlike T. pallidus (Cope). Werner described T. kleebergi from the Usambara Mtns. in 1904. Sternfeld omitted it from his list of the snakes of German East Africa published in 1910. In 1915 Boulenger placed it in the synonymy of T. lumbriciformis (Peters). Unless inaccurately described, however, it is a valid species related to T. gracilis and differing from T. lumbriciformis in lacking a subocular and in that the diameter of its body is contained in its length 36 instead of 60 times. The relation of these three mountain-inhabiting East African snakes can best be shown by the following table: T. gracilis T. uhiguruensis T. kleebergi T. lumbriciformis Scale-rows 22 20 18 18 Body-diameter con- tained 80 times in the 50 36 GO length Nasal divided from Nasal divided from Nasal semidivided ? 2nd labial to rostral 1st labial to rostral Subocular absent absent absent Large subocular In a letter recently received from Dr. Werner he informs us that he cannot agree with Dr. Boulenger's action in placing T. kleebergi in the synonymy of T. lumbriciformis. ' Salimu bin Asumani, an Mkami collector frequently referred to in these pages. AFRICAN HERPETOLOGY 105 Diagnosis. Differs from all East African members of the genus, except gracilis, in lacking a praeocular and subocular. It differs from gracilis in its rounded in- stead of sharply horizontal-edged snout, 20 scale-rows instead of 22, and less slender proportions, for the diameter of its body goes into the total length 50 in- stead of 80 times. Description. Snout prominent, rounded, with scarcely inferior nostrils. Rostral very large, extending backwards to an imaginary line connecting the commisures of the mouth; eye indistinguishable (in one of the paratypes a little pink pigmentation indicates its position); nasal divided, the suture extending from the second labial through the nostril to the rostral ; no praeocular or subocu- lar; ocular scarcely half the size of the huge nasal, in contact with the third and fourth upper labials; a small postocular; four upper labials ; three lower labials. Diameter of body contained 50 times in the total length (48 to 51 in the paratype series); tail longer than broad (in paratype No. 23081 it is as long as broad), sharply pointed but not terminating in a spine; 20 scales round mid-body. Coloration. Uniformly flesh-pink in life; colourless or straw-coloured in alcohol. Measurements.^ Type Length of head and body, 235 mm. Length of tail, 5 mm. Diameter at mid-body, 4.5 mm. Paratypes Total lengths, 245, 240, 230 mm. Tail lengths, 3, 4, 4 mm. Diameters at mid-body, 5, 5, 4.5 mm. Diet. Only one snake was opened up for examination. The stomach contained many termites and the intestine was full of undigested heads. Habitat. Two, at least, of the series were taken under the rotting grass roof of a collapsed hut which had been built close to the edge of the rain-forest. Such huts are always teeming with termites. Typhlops braminus (Daudin) Eryx braniimts Daudin, 1803, Hist. Nat. Rept., 7, p. 279. 1 (M. C. Z. 2400S) Dar es Salaam, 4. xi. 26. Variation. Normal number of twenty mid-body scale rows. Coloration. Plumbeous black, not brown. Enemies. Recovered from the stomach of a burrowing viper (Atractaspis rostrata) in the Botanical Gardens. The head to mid-body is in a fine state of preservation ; the latter half of the body was entirely digested away. ' All measurements are in millimeters. 106 BARBOUR AND LOVERIDGE Typhlops punctatus punctatus (Leach) Acontias punctatus Leach, 1819, in Bowdich, ' Miss. Ashantee,' p. 493. 1 (M. C. Z. 23092) Dar es Salaam, 9. xi. 26. In a previous paper ^ the junior author drew attention to two other specimens from this locaUty with which this example agrees in colouring and other peculiari- ties. It has 32 scale-rows instead of 30, which is maximum for punctatus, and the body diameter is included in the length 38 times instead of 30. Its length is 300 mm., and the mid-body diameter 8 mm. More material may demonstrate an East Coast race with slenderer body. In some respects it resembles T. iettensis; but the position of its eye, which is clearly beneath the ocular scale, is unlike tettensis, mossambicus or trilobus as figured by Peters. Typhlops intermediate between punctatus and gierrai 1 (M. C. Z. 23093) Amani, Usambara Mtns., xii. 1926. This very interesting snake is coloured like those specimens of gierrai marked (iv) below, and it is undoubtedly the same species ; yet it lacks the small scale between the praeocular and supralabials, and its body diameter is included 30 times in the total length. It has 28 mid-body scale-rows. In the collection of the Amani Institute are three similar examples from the Monga path which I pro- visionally referred to T. pimctatus. Under the circumstances, we consider that the close relationship between gierrai and punctatus is best expressed by making gierrai a race of the widely distributed punctatus. It is apparently a perfectly valid mountain form. Typhlops punctatus gierrai (Mocquard) Typhlo-ps gierrai Mocquard, 1897, Bull. Mus. Paris, p. 122 (Tanga). 1 (M. C. Z. 23084) Bagilo, Uluguru Mtns., 27. Lx. 26. 6 (M. C. Z. 23085-90) Amani, Usambara Mtns., 20-27. .>d. 26. 1 (M. C. Z. 23091) Mt. Lutindi, Usambara Mtns., 10. xii. 26. 3 (M. C. Z. 23473) Mlalo near Ambangula, Usambara Mtns., xii. 26. Nyoka midomo miwili (Kikami and Kiswahili); mkonko (Ksumbara). Variation. This series is of considerable interest as gierrai has hitherto only been known from the holotype. Our specimens all agree in the possession of a small scale resting on the second and third upper labials. This, together with the formula of body diameter into length (31 to 50 times in gierrai; 24 to 30 times in punctatus), is the only characteristic which serves to distinguish it from punctatus. Possibly the eye is more distinct in punctatus; it appears so from the series at our disposal. ' Loveridge, 1923, Proc. Zocil. Soc. Lond., p. 872. AFRICAN HERPETOLOGY 107 While eight specimens agree with the type in having 28 scale-rows, three of the Amani snakes have only 26. The body diameter of the type was contained 50 times in its total length, while in the present series it ranges from 31 to 43 times. There seems to be little or no sexual variation in the breadth of the tail as compared with its length ; perhaps it will be found that a ratio can be established by dividing the length of tail into the body length. We suggest that in males it goes from 55 to 60 times and in females 75 to 79 times, but, being unable to sex all the specimens with precision, we hesitate to say that this is definitely the case. The reasons for considering gierrai a subspecies of punctatus are given above. Colouring. Four distinct types of coloration occur. These enable us to trace the gradation between the extremes. (i) Bagilo. Above, uniformly blue-black except for a minute light spot on the base of each scale. Below, uniformly white in life, yellowish in alcohol, (ii) Amani. Above, like (i) except that the blue-grey spot on each scale is large, giving the snake a somewhat striped appearance. This colouring is identical with that of the specimen of punctatus from Dar es Salaam re- corded above, also with an example of punctatus from Liberia. Below, uniformly white, (iii) Amani. In one young snake the light blue-grey spot eliminates the blue-black colouring of the scale except from the lateral edges. This results in the snake having ten well-defined, dark dorsal and lateral lines. Below, uniformly white, (iv) Amani. A very large snake. Above, pale grey, each scale with lateral black edges, excepting some which are involved in large irregular patches of black. The black scales which go to form the patches have each a light speck at tlieir base which is only visible with a lens. Below, uniformly white. This specimen agrees with the type of T. gierrai and closely matches examples of T. punctatus from the Cameroon which we consider referable to the colour variety represented by the 'species' congestus Dumeril and Bibron. (v) Mt. Lutindi. A still larger individual about to slough. Uniformly opaque white, but with the colouring of (iv) faintly discernible on close examination; more readily seen in alcohol than when alive. Measurements. The Bagilo snake, measuring 390 (385 -|- 5) mm. in length and 9 mm. in diameter, was slightly larger than the Mt. Lutindi example, which was 108 BARBOUR AND LOVERIDGE 380 (375 + 5) mm., but had a greater body diameter, viz. 12 mm. Both are smaller than the 469 mm. type. Habitat. My wife called my attention to the Lutindi snake, which was lying in a neatly moulded cavity beneath a large rotten log in the forest. It had heavy deposits of fat and was evidently resting there until such time as it could slough. j\Iost of the Amani specimens were taken inside damp and very rotten logs on Mt. Bomoli, but a few were taken beneath the logs. The Bagilo reptile was turned up by a native who was engaged in hoeing over his garden which abutted on the rain-forest. Folklore. The Wasumbara natives say that the mkonko is the ' cow ' of the soldier ants {Dorylus nigricans subsp.) which are very abundant in these moun- tains. Wlien 1 asked what they meant by this they replied that sometimes when one encountered these fierce marauders on the march, one might see a Typhlops wriggling along with the column of ants protected by the warrior ants and in no way molested by any of them. I found the story very widespread but mainly based on hearsay : only one or two natives claimed to have seen the phenomenon themselves. One European, however, told me that he had seen such a procession at Amani about three years before. I examined about fifty-four columns of siafu, as these ants are locally called, during the following five weeks without seeing a single Tjqjhlops. If the story is true, and I see no reason to doubt it, the reason for the strange association should provide an interesting subject for study, as these soldier ants are notorious for attacking almost every living thing which comes within reach of their jaws. The native's explanation, purely theoretical of course, is that the ants keep their ' cow ' until there is a food shortage, when they kill it. I took particular pains to ascertain that the creature referred to was a Typhlops and not a caecilian. Of course, the species of Typhlops the world over are prone to feed on termites or ants and their integument is modified so as to afford protection against insects' jaws or stings. Typhlops mucruso (Peters) Onychocephalus mucruso Peters, 1854, Monatsber. Akad. Berlin, p. 61. 1 (M. C. Z. 23093) Kilosa, ix. 26. We are indebted to Mr. N. C. Miller of the Game Department for this speci- men. Variation. In examining this snake the opportunity was taken to see if East African specimens can be separated into two races {mucruso typica and m. humbo), on the basis of the eye being beneath the suture between praeocular and ocular AFRICAN HERPETOLOGY 109 (humbo) or clearly in the ocular {itmcruso). Out of six Kilosa snakes in the col- lection of the Museum of Comparative Zoology, three have the eye in the ocular and three beneath the suture. Of four from Morogoro (three loaned by the United States National Museum, Nos. 62891-3) two are in the ocular and two beneath the suture, thus confirming Boulenger's opinion that the forms are not separable. Nor has its position anything to do with age as judged by length. Two of the characters used by Miss Procter ' to differentiate T. excentricus from T. mucruso cannot be considered of use, for the nasals meet behind the rostral in all four Morogoro snakes, are just in contact in four of the Kilosa snakes and are separated in the other two. The diameter is included in the length from 22 to 34 times in these ten snakes; 25 to 37 times has been the recognized range hitherto. The scale-rows are 30 to 36, which is within the range of the species. LEPTOTYPHLOPIDAE [Leptotyphlops distanti (Boulenger)] ^ Glaucoma distanti Boulenger, 1892, in Distant, 'Naturalist in the Transvaal,' p. 175, fig. Glauconia conjuncta Angel (not of Jan.), 1925, Reptiles et Batraciens, in 'Voyage de Ch. Alluaud et R. Jeanne! en Afrique Orientale (1911-1912),' p. 31. One of the two specimens was examined in Paris by the junior author through the courtesy of M. Angel. The diameter of body was contained in the total length 45 times, and the length of tail between 9 and 10 times. The colour of this speci- men was unusual, being brown above and paler beneath. The second and larger specimen was of the usual jet-black type. COLUBRID.\E Natrix olivaceus (Peters) Coronella olivacea Peters, 1854, Monatsber. Akad. Wiss. Berlin, p. 622. 1 (M. C. Z. 23094) Bagilo, Uluguru Mtns. 28. ix. 26. 24 (M. C. Z. 23096-23118) Nyange, Uluguru Mtns. 6-11. x. 26. 1 (M. C. Z. 23095) Mkangazi, Uluguru Mtns. 12. x. 26. 3 (M. C. Z. 23119-30-31) Nyingwa, Uluguru Mtns. 16-19. x. 26. 1 (M. C. Z. 23132) Vituri, Uluguru Mtns. 28. x. 26. 3 (M. C. Z. 2313.3-5) Dar es Salaam, 4. xi. 26. 29 (M. C. Z. 23136-23159) Amani, Usambara Mtns. xi-xii. 26. 1 (M. C. Z. 23160) BumbuU, Usambara Mtns. 7. xii. 26. Variation. This series of sixty-three snakes was collected in the hope that it would throw some light on the variation in the scale-rows. These specimens ' Procter, 1922; Ann. Mag. Nat. Hist., 9, 685. 2 When brackets are placed around a species and its author it denotes that the species in question was not collected during the expedition, but it has been found necessary to refer to it for some reason or other. no BARBOUR AND LOVER IDGE effectually show that there is something in the suggestion made by the junior author* that there may be a transcontinental reduction in their numbers from west to east. While for forty years 19 has been considered the invariable niunber for this species, in reality 17 is the more usual number in East Africa, at least for mountain snakes, and 15 occurs on Pemba Id. No fewer than 56 of these 63 snakes have 17 mid-body scale-rows or even less. In counting the rows on the neck, mid-body and praeanal region, one finds almost every possible variation: 1 specimen with 20, 17, 16 3 specimens with 19, 17, 15 4 a ' 19, 19, 17 1 " ' 19, 19, 16 1 (( ' 19, IS, 17 39 (( ' 19, 17, 17 8 u ' 19, 17, 16 U u 19, 16, 16 " u 18, 19, 17 u " 18, 17, 17 a It IS, 17, 16 u u 17, 17, 16 By taking the neck count just behind the parietals one may often find 21 rows, but as the scales are somewhat irregular so close to the head, the counts were made at a point about opposite the fifth ventral scute. The ventrals range from 130 to 140 with an average of 135; the anal is in- variably divided; subcaudals range from 63 to 87. For purposes of comparison nine specimens from Stanleyville and the Ituri Forest, Belgian Congo, in the collection of the American Museum of Natural History were counted, and all but one had 21, 19, 17 scale-rows; the one excep- tion had 19, 19, 17. Some had only one or two rows of 21 scales very close to the back of head. Four other specimens in the United States National Museum from Leopoldville, Belgian Congo, Kenya Colony, and Morogoro, Tanganyika Terri- tory, agreed in having 19-21, 19, 17 scale-rows. We wish to express our thanks to the curators concerned for the facilities afforded for examination of these specimens. Coloration. In life a Bagilo female had the ventrals coloured a bright mauve. In Nyange snakes the variation was astonishing, the lateral edges of the ventrals being red, olive, grey, or pale blue. In one specimen the ventrals were bright yellow with grey edges. Some of the young have a black band three scales wide behind, and impinging on, the posterior border of the parietals ; this band is fol- lowed by an incomplete yellow one two scales wide, interrupted on the vertebral line. Measurements. The largest male (Nyingwa) measured 430 (300 -t- 130) mm., and the largest female (Nyange) 488 (350 -|- 138); the smallest snake, a male (Nyange) taken on 7. x. 26., measured 193 (138 -|- 55). These mountain exam- ' Loveridge, 1925, Proc. Zool. Soc. Lond., p. 71. AFRICAN HERPETOLOGY 111 pies of N. olivaceus will probably be found to average considerably less than specimens from other parts of Africa. Breeding. A Nyange female had very slightly developed eggs measuring 5 nrni. in diameter on 8. x. 26. Another Nyange snake held two eggs measuring 12 X 6 mm. on 9. x. 26.; an Amani female, four eggs measuring 20 x 8 mm. on 19. xi. 26.; another, four eggs 16 x 7 mm. on 25. xi. 26.; and yet another, eggs measuring 21 x 8 nmi. on 29. xi. 26. At the same time the rest of the large series of females, with but few exceptions, held undeveloped ova. Diet. Stomach contents of Nyange snakes, (i) Arthroleptis stenodadylus. (ii) A. stenodadylus. (iii) Eight small frog's eggs, (iv) Three round, white eggs measuring 4.5 mm. in diameter, apparently frog's eggs, (v) An ^Amani snake held a caterpillar. Parasites. Rarely infected. Nematodes {Kalicephalus sp. and Ophidascaris sp.) were found in only one Amani female. Enemies. Thelotornis kirtla7idii was seen to eat a N. olivaceus which was shar- ing the same cage; many others disappeared. The junior author has previously drawn attention to the frequency with which the tails are missing in this species. No fewer than twenty-nine out of the present series of sixty-eight are in this con- dition. In one case he was responsible, for having picked up a Dar es Salaam snake from beside an irrigation ditch, he was holding it by the tail, preparatory to dropping it into a bag, when the snake wriggled very violently so that it broke away, leaving the tail in his grasp. As Grayia is another genus where the caudal appendage is frequently short, it seems probable that waterside species are more subject to attack by herons and waders than other snakes and frequently escape by active wriggling. In many localities they doubtless suffer from the bites of turtles and fishes. Habitat. Under bark or logs at the edge of the forest, not a forest species; under flat stones or rubbish in the native plantations. After heavy rain these snakes were sometimes found on the paths. They haunt the banks of streams and irrigation ditches and the edges of swamps. BoAEDON LiNEATus Dumeril et Bibron Boaedon lineatwn Dumeril et Bibron, 1854, 'Erpet. Gen.,' 7, p. 363. 1 (M. C. Z. 23161) Bagilo, Uluguru Mtns., 29. Lx. 26. 3 (M. C. Z. 23162-4) Nyange, Uluguru Mtns., 11. x. 26. 17 (M. C. Z. 23165-81) Amani, Usambara Mtns., xi-xii. 26. Variation. Mid-body scale-rows 25-29. Ventrals 190-220. Anals single. Subcaudals 48-67. The series is almost equally divided as to sex and it is to be 112 BARBOUR AND LOVERIDGE observed that the sexes may be invariably distinguished by the ventral and sub- caudal scale counts which, in the present series, are: Males Females Ventrals 190-201 209-220 Subcaudals 60-67 48-59 Labials 8, in one specimen 9, on one side of the head only; 3rd, 4th and 5th, or 4th and 5th, or 4th, 5th and 6th, labials enter the eye. In No. 23166 a small scale, almost like a third postocular, is wedged in between the 5th and 6th upper labials. Praeoculars very variable — 1 or 2 ; in specimen No. 23170 they are in contact with the frontal. Measurements. Largest male 668 (533 + 135) mm., and female 915 (805 -|- 110) mm. Both from Amani. Breeding. Adult female from Nyange held 16 eggs measuring 33 X 18 mm. on 11. X. 26. Another from Amani held 7 eggs measuring 35 X 16 mm. on 29. xi. 26. Diet. A small (298 mm.) Bagilo snake w^as much distended with a mouse (Leggada hello). Small House Snakes certainly seem to live well, for no fewer than seven of ten young ones brought in at Nyange were gorged. A rat (Rattus rattus alexandrinus) was recovered from the stomach of an Amani House Snake. Defence. When pressed, the cloacal glands of a small snake discharged a minute jet of clear but strong-smelling fluid. Parasites. Tapeworms {Ophiotaenia sp.) and roundworms were taken from one Nyange female. Nematodes (Kalicephalus sp.) in several others as well as an Amani snake. A great many larval mites under the caudal scales of a Nyange individual. [Lycophidion abyssinicum Boulenger] Lycophidium abyssinicum Boulenger, 1893, Cat. Sn. Brit. Mus., 1, p. 342 ('243'), PI. XXII, f. 1. The snake captured between Kilimanjaro and Taita and referred to this species by Mocquard (see Angel, 1925, Reptiles et Batraciens, in 'Voyage de Ch. Alluaud et R. Jeannel en Afrique Orientale [1911-1912],' p. 32) cannot now be located in the Paris Museum. The identification must remain extremely doubt- ful, as L. abyssinicum has never been recorded so far south before. Lycophidion meleagris Boulenger Lycophidium meleagris Boulenger, 1893, Cat. Sn. Brit. Mus., 1, p. 337, PI. XXI, f. 2. 1 (M. C. Z. 23182) Bagilo, Uluguru Mtns., 20. ix. 26. 1 (M. C. Z. 23183) Vituri, Uluguru Mtns., 9. x. 26. 4 (M. C. Z. 23184-7) Amani, Usambara Mtns., xi-xii. 26. 1 (M. C. Z. 23188) Bumbuli, Usambara Mtns., 16. xii. 26. AFRICAN HERPETOLOGY 113 Variation. All have the normal 15 scale-rows at mid-body. The ventrals range from 147 to 164; subcaudals 22-34, thus increasing the known range for both ventrals and subcaudals. Taken by sexes the subcaudals in females are 22-32, in males 30-34. The series is too short to draw definite conclusions. The only points of difference, other than scale-counts, between this species and L. capense, as given in the Catalogue of Snakes, are that, in L. meleagris, the a. Diameter of eye equals its distance from mouth, while in capense it is greater. b. Rostral nearly twice as broad, as against twice as broad in capense. c. Praefrontals a little broader than long, as against as long as broad or a little longer in capense. d. Frontal as long as broad in both species, but in capense it is also sometimes a little longer. Characters 'b,' 'c, ' 'd' break down in the series before us. Further reference to these relationships will be found under capense. Coloration. The tongue is white, tipped with grey. Measurements. The largest male (Vituri) measures 353 (308 + 45) mm., largest female (Bagilo) 340 (310 4- 30) mm. Smallest specimen, a female taken at Bmnbuli 16. xii. 26, measures 217 (195 + 22) mm. Breeding. Ova only slightly developed in one female from .'Vmani, 19. xi. 26. Diet. A Bagilo snake, being tremendously distended, was found on examina- tion to have swallowed a skink {Siaphos kilimensis) ; in an Amani specimen there was a snake's tail. Lycophidion capense (Smith) Lycodon capensis Smith, 1S31, S. Afr. Quart. Jour. (1), No. 5, p. 18. 1 (M. C. Z. 231S9) BagUo, Uluguru Mtns., 4. x. 26. 5 (M. C. Z. 23190-4) Nyange, Uluguru Mtns., 2-11. x. 26. 1 (M. C. Z. 23195) Vituri, Uluguru Mtns., 27. x. 26. 4 (M. C. Z. 23196-9) Amani, Usambara Mtns., 29. xi-30. xii. 26. 1 (M.,C. Z. 23200) Bumbuli, Usambara Mtns., 16. xii. 26. Variation. All have the normal 17 scale-rows at mid-body. Ventrals range from 195 to 215; subcaudals from 40 to 57, thus increasing the range for both ven- trals and subcaudals — indeed, in both ventrals and subcaudals the numbers are consistently high. Taken by sexes the subcaudals in females range from 40 to 47, in males 52-57. Compare these with those in the Catalogue of Snakes based on South African, Angola, and Congo specimens which give subcaudals for females as 30-39, for males 34-45, which makes it impossible to determine sex on this basis if the range for the species over its whole area of distribution be taken into account, though it is probably good for any one locahty. 114 BARBOUR AND LOVERIDGE The series agrees essentially with the characteristics of the species except that two big specimens are in accord with L. meleagris in the diameter of the eye, which only equals its distance from the mouth. In one of these from Amani, the horizontal diameter is much shorter than the vertical diameter. The character of eye-diameter in relation to distance from the mouth appears to be a good one except for very old snakes. The safest way, however, to distinguish between these two closely related species is by the scale-counts, viz.: L. meleagris. Scale-rows 15, Ventrals 147-165, Subcaudals 22-34. L. capense. Scale-rows 17, Ventrals 163-219, Subcaudals 24-57. It seems very strange that they should occur together in four out of the five localities, but 15 scale-rows is certainly associated in these mountains with a lower number of ventrals 147-164, while conversely 17 scale-rows goes with a higher number, 195-215. One needs a large series of Lycophidion from the type locality of jacksoni to throw more light on the relationships of the three species. Coloration. A Bagilo snake, taken on the site of my camp after I left, is uni- formly jet-black. Measurements. The largest male (Amani) measures 480 (400 -|- 80) mm., and the largest female (Amani) 023 (550 -|- 73) mm. The smallest (Vituri) snake, taken on 27. x. 26., measures 196 (170 -}- 26) mm. Breeding. A Nyange female held 8 eggs measuring 20 X 8 mm. on 2. x. 26. A Bagilo snake, 7 eggs only 10 mm. long on 4. x. 26. An Amani reptile, 7 eggs measuring 17 X 9 mm. on 29. xi. 26. Diet. At Nyange two snakes were taken with skinks (Mabuya striata) in their stomachs. At .Amani, one with a Mabuya comorensis. Mehelya capensis (Smith) Heterolepis capensis Smith, 1849, 'Illus. Zool. S. Afr.,' 3, PI. LV. 2 (M. C. Z. 23201-2) Amani, Usambara Mtns., xi-xii. 26. 1 (M. C. Z. 23204) Misalae, Usambara Mtns., 7. xii. 26. Variation. Internasals nearer seven-eighths than two-thirds the length of the praefrontals. In No. 23202 the frontal is as broad as long. In No. 23201 the frontal is shorter than its distance from the rostral. Nos. 23202 and 23204 are normal in having 1 prae- and 2 postoculars but No. 23201 has 2 prae- and 1 postocular. Through the courtesy of the Custodian of the .Amani Institute I was enabled to examine the three local specimens in that collection, all of which agreed in having only 1 postocular, a condition, I believe, that has never been reported in any South African examples of this snake. The scale variation of AFRICAN HERPETOLOGY 115 these six snakes is as follows. Scales on neck 21-22. Scales at mid-body 15. Ventrals 219-224. Anal single. Caudals 50-56. Upper labials 7, 3rd and 4th entering eye. Lower labials 7 or 8. Praeoculars 1 or 2. Postoculars 1 or 2. Coloration. Tongue pink. The vertebral line and ventrals are not yellow except in snakes about to slough ; in a freshly sloughed specimen they are pure china-white. Measurements. Largest male (Misalae) 1250 (1073 + 177) mm., and the largest female (Amani) 1119 (1070 + 149) mm. Breeding. The latter held six eggs (23203) just ready for laying; they meas- ured 42 X 30 mm. The snake was found by my wife in loose soil and vegetable debris at the base of a large tree in an open pasture, 19. xi. 26. Its old yellowish skin was in process of being shed and the creature was very docile, making no attempt to bite. An old egg-shell was found among the debris. Temper anient. The docile nature of this species was further exemplified by the male which I caught at 11 p.m. on the Misalae path while we were on the march in drizzling rain. It offered no resistance to capture and though frequently handled the following day never once attempted to bite. Chlorophis neglectus (Peters) Philothamnus neglectus Peters, 1866, Monatsber. Akad. Berlin, p. 890. 1 (M. C. Z. 23205) Bagilo, Uluguru Mtns., 19. ix. 26. 8 (M. C. Z. 23206-13) Nyange, Uluguru Mtns., 1-9, x. 26. 6 (M. C. Z. 23214-9) Vituri, Uluguru Mtns., 9-29. x. 26. 1 (M. C. Z. 23224) Dar es Salaam, Uluguru Mtns., 6. xi. 26. 2 (M. C. Z. 23220-1) Amani, Usambara Mtns., 29. xi. 26. 2 (M. C. Z. 23222-3) Bumbuli, Usambara Mtns., 14. xii. 26. 10 (M. C. Z. 23224-33) Mlalo nr. Hohenfriedeberg, xii. 26. Also taken at Phillipshof but not preserved. Ngole in Kikami, but the name is also applied to Dispholidtis typus and probably Philotha7tinus semivariegatus as well. Variation. Mid-body scale-rows 15. Ventrals 141-164. Subcaudals 81-119. No fewer than twenty-five of these thirty snakes have the normal 8 labials, fourth and fifth entering the eye; two have fourth and fifth on one side of the head, and fifth and sixth on the other where there are 9 labials; one had 9 labials, fifth and sixth entering eye on both sides of head; yet another, which has the normal number of 8 labials on both sides, has the fourth, fifth and sixth labials entering the eye on one side only. No. 23224 has a small supernumerary scale wedged between the labials on the left side of the head. 116 BARBOUR AND LOVERIDGE What is stated below as regards determination of sex in C. macrops by means of scale-counts applies equally to the present species. Coloration. Wlaile this snake is almost invariably uniformly green, at Nyange a number were brought in on which there was a series of paired black spots on the nape; in many these had coalesced. A female at Vituri was marked in the same way. Measurements. Largest specimen, probably a female, from Mlalo measures 896 (612 + 284) mm. No young examples were preserved. Habitat. Taken on bushes, one from Dar es Salaam on the thatch of a watch- er's hut in the rice swamps. Breeding. A Nyange female held six eggs measuring 20 X 7 mm. on 1. x. 26; a Vituri snake three eggs measuring 34 x 10 nam. Diet. A Bagilo snake was captured on the edge of the rain-forest with an Arthroleptis stenodactylus in its mouth. A Vituri snake had also swallowed one of these frogs. A Nyange snake had a frog, a Buprestid beetle and a grasshop- per's leg in its stomach. The frog was too digested to be definitely determined but was apparently an Arthroleptis. It is reasonable to suppose that the insects were liberated from the frog's stomach. In captivity at Dar es Salaam, three or four of these snakes died after making heavy meals on Rana mascarcniensis, the prey being nearly digested away. One was seen to capture and swallow a gecko (Lygodactylus p. picturatus). Enemies. At least one green snake was eaten by a Thelotornis kirtlandii which was sharing the same cage. Parasites. At Bagilo some cysts (Acanthocephala immature) were found encysted on the outer wall of the intestinal tract, also a mature 9 Acanthocephala. At Nyange several snakes affected with subdermal parasites of the same species were brought in and died within a few days; no nematodes were located in their alimentary canals. At Vituri other freshly captured snakes were suffering from these encysted worms. Chlorophis macrops (Boulenger) Oligolepis macrops Boulenger, 1896, Cat. Sn. Brit. Mus. 3, p. 644. 22 (M. C. Z. 2.3236-57) Amani, Usambara Mtns., xi-xii. 26. Nyoka tja niani in Ivisumbara. Variation. This species was based on a single young snake from the Usam- bara Mtns., and was made the tjqje of the genus Oligolepis by Boulenger. It was transferred to Chlorophis by Sternfeld after the examination of a dozen AFRICAN HERPETOLOGY 117 snakes from the region. The species is vaUd, though it was based on an aber- rant individual. The present series are topot>T3es. Undoubtedly its most important character is tlie 13 mid-body scale-rows; all other members of the genus have 15. Twenty-one snakes in the present series have 13, while one male (23236) has only 11; this same reptile is also unique in possessing an undivided anal shield. It is to be noted that Sternfeld had also a single specimen with a similar anal. Ventrals range from 136 to 169 (tj^je had 148). Subcaudals range from 82 to 122 (type had 75). In the type there are 9 upper labials, fifth and sixth entering the eye ; this is the case in only two of the present series, while four more have this condition on one side of the head with the more normal 8 labials, fourth and fifth entering the eye, on the other side of the head; no fewer than sixteen snakes have 8 labials, fourth and fifth entering the eye, which must therefore be regarded as the normal state. It is interesting to note that where 9 labials are present it is almost invariably due to the division of the third labial so that it automatically results in the fifth and sixth entering the eye. Sternfeld has already reported some of these variations, including cases where the fourth, fifth and sixth labials entered the eye. He has also drawn attention to the moderate keeling of the ventrals, all of which points Boulenger seems to have ignored when compiling his key to the Snakes of East Africa (Proc. Zool. Soc. Lond., 1915, p. 622). Sex cannot be definitely told by counting subcaudal scale-rows as they over- lap. In males 85-122 (average 96), females 82-87 (average 85). The sexes in the above series are equally divided and it would appear that all snakes with more than 90 subcaudals are males. Relationship. The amendments to the original description as enumerated above cause this snake to fall into the neglectus section of the key, and indeed macrops is so nearly related to that species that it can be separated only on the basis of its 20 maxillary teeth (20-27 in neglectus) and number of scale-rows at mid-body. Were it not that two specimens of neglecttis were also brought in at Amani one would be tempted to consider macrops as a subspecies of neglectus, being in the same relation to it as is Typhlops gierrai to T. punctatus and Crota- phopeltis tornieri to C. hotamboeia,i.e.,a mountain form with reduced scale-rows. It is still possible that the neglectus referred to may have been caught lower down the mountain, for in an hour one can readily ascend a thousand feet. Coloration in life. Female (No. 23238). Above, head green, body and tail olive, thirty-two transverse bands of varying distinctness on the body; these are 118 BARBOUR AND LOVERIDGE formed by light-yellow spots on scales bounded by black spots. Below, chin and throat white, rest of under surface yellowish, each ventral scale edged with black laterally, their free edges tinged with bluish-grey. A young female (No. 23241). Above, head olive, body and tail brown, forty- two transverse bands of varying distinctness on the body; these consist of pale- green bands, turning to Cambridge blue on the flanks, bounded by black lines. Below, chin and throat white, rest of lower surface light green. A very old female (No. 23239). Above, uniformly brown or olive-brown. Below bluish-white. Of yet others I have noted a female bright green. A juve- nile male, very young, yet dark olive colour. Two males, taken the same day, differed in the colour of their labials, one being bright green and the other olive- coloured. Several of the smaller snakes were indistinguishable in colouring from C. negledus. Measurements. Largest male 810 (560 -1-250) mm.; largest female 915 (655 -|- 260) mm. Smallest specimen, a male taken 25. xi. 26, measures 200 (130 + 70) mm. Habitat. In bushes on the outskirts of the forest, especially those bordering streams. Breeding. Five eggs measuring 29 x 11 mm. on 19. xi. 26. Six eggs measur- ing 36 X 15 mm. on 20. xi. 26. Fourteen eggs (9 on one side, 5 on the other) measuring 30 X 12 mm. on 22. xi. 26. Three eggs measuring 34 X 10 mm. on 24. xi. 26. Most of the other females held eggs less developed. Diet. Stomachs of all fifteen snakes taken in November were empty. Philothamnus semivariegatus (Smith) Dendraphis {Philothamnus) semivariegata Smith, 1840, 'Illus. Zool. S. Afr.,' 3, Pis. lix, Ix, and Ixiv, fig. 1. 1 (M. C. Z. 23259) Vituri, Uluguru Mtns., 28. x. 26. 1 (M. C. Z. 23260) Zigi, Foot of Usambara Mtns., 29. xi. 26. 8 (M. C. Z. 2.3261-8) Amani, Usambara Mtns., xi-xii. 26. Variation. Mid-body scale-rows 15. Ventrals 164-172. Subcaudals 125-153. Nine upper labials, fifth and sixth entering the eye except in No. 23263 where it is the fourth, fifth and sixth that enter. The temporals are almost too variable to describe; five specimens have the normal 2 -|- 2; in four others either the upper or lower of the first pair is divided. The number of times the tail is contained in the body length is 1.74 to 1.79 in males, 1.75 to 2.29 in females. Ventral and subcaudal scale-counts are equally useless as an aid to distinguishing the sex. AFRICAN HERPETOLOGY 119 Relationship. We imagine that the reasons Boulenger (1919, Rev. Zool. Africaine, vii, p. 23) referred Cameroon snakes to semivariegatus instead of to niiidns was because he considered that the characters on which these snakes were held to be distinct were not specific. Schmidt (1923, Bull. Am. Mus. Nat. Hist., 49, p. 78) reports on a series of fifteen nitidus from the Congo, which adds considerably to our knowledge of the scale ranges in that species. Without consulting all the extensive literature it may be said that P. nitidus has 163-175 ventrals, 125-153 subcaudals, 1+2 (rarely 1+1) temporals; P. semivariegatus, 149-207 ventrals, 112-159 sub- caudals, 2+2 (rarely 1 +1,1 + 2, or 2 + 1) temporals. The greater scale range of semivariegatus is quite probably due to the fact that much larger series of this more widely distributed snake have been examined. The temporals seem to be the only character which distinguishes the two species and we consider it probable that nitidus is entitled to rank only as a subspecies of semivariegatus. Measurements. Largest male measures 1125 (717 + 408) mm.; largest female 1012 (705 + 307) mm. Both from Amani. Correction. Through the courtesy of M. Angel the junior author was able to examine the snakes ' from Kibwezi, Kenya Colony and Bura, Teita, K. C, identified by M. Mocquard as Philothamnus nitidus Gunther and P. dorsalis Bocage, and has no hesitation in referring them both to P. semivariegatus. The specimen referred to dorsalis has a piece of the anterior temporal split off so that it might be considered as 2 + 1 or 1 + 1 + 1 ; apart from this it differs from dorsalis only in that the frontal is shorter, not longer, than its distance from the end of the snout. Ventrals 190; caudals 136 — tip missing. The specimen re- ferred to nitidus has the frontal equal to, not shorter than, the parietals. Ven- trals 196; caudals 142 — tip missing; temporals 1+2. If M. Mocquard took the trouble to count the ventrals there seems no reason why he should have re- ferred tliis snake to anything but semivariegatus. ' Angel, 1925, Reptiles et Batraciens, in 'Voyage de Ch. Alluaud et R. Jeannel en Afrique orientals (1911-1912),' pp. 32-33. 120 BARBOUE AND LOVERIDGE Prosymna ornatissima sp. n. Plate2, fig. 2 3 (M. C. Z. 23269-71) Nyange, Uluguru Mtns., 4, 6, 7. x. 26. 1 (M. C. Z. 23272) Vituri, Uluguru Mtns., 9. x. 26. Type. No. 23271. Museum of Comparative Zoology. Sex 9. From Nyange, Uluguru Mountains, Tanganyika Territory. Collected by A. Loveridge, Octo- ber 7th, 1926. Paratypes. Nos. 23269, 23270, 23272. Affinities. In a genus whose characters are so very variable it is little more than speculation to say that it seems nearest to P. ambigua Bocage. See discus- sion under that species. Diagnosis. Its remarkable black and scarlet colouring distinguishes it from all other members of the genus. Its subcaudal range is higher than any other East African species. Description. Rostral very large and broad, with sharp horizontal edge sepa- rated from the praefrontal; a single internasal and a single praefrontal; frontal large, more than half the width of the head, as long as the parietals; loreal wedge- shaped, longer than deep; 1 prae- and 1 or 2 postoculars (2 on the right side of head, 1 on left. In paratype 23269 there is no postocular on the left side as it has fused with the supraocular. In all the other paratypes there is 1 post- ocular) ; temporals 1 4- 2 (in No. 23270 the anterior temporal enters the eye be- low the postocular on the right side of the head) ; six upper labials, third and fourth entering the eye; one pair of well-developed chin-shields followed by three more or less enlarged scales. Scales smooth, in 15 rows. Ventrals 150 (127- 132 in & paratypes); anal entire; subcaudals 27 (35-40 in & paratypes). Coloration in life. Above, black, head scarlet, except for an arrow-shaped ex- tension of the black body-colouring which reaches to the frontal, and a vertical black streak from the supralabials to the eye (in No. 23270 these streaks are united by a black band which crosses the praefrontal and the anterior edge of the frontal) ; thirteen transverse, more or less complete, scarlet bands on the body, four alternating latero-dorsal blotches on tail (in the paratypes fourteen on body and four on tail) ; these scarlet bands are very irregular, often interrupted on the vertebral line or more or less broken and alternating. Below, throat pink, the rest of the undersurface black except where the lateral scarlet blotches impinge on the outer edges of the ventrals so as to tinge them with red. AFRICAN HERPETOLOGY 121 Measurements of Type 9 Paratype & d' & Length of head and body 252 mm. Length of head and body .117, 210, 242 mm. Length of tail 34 " Length of tail 26, 54, 55 " Diameter of body 9 " Diameter 5-8 " Breeding. No. 23269 is very young, the ventral scutes in the umbiUcal region being still unhealed when it was caught on 4. x. 26. Habitat. I was climbing up the lower slopes of Mt. Tongoni when my gun bearer, who was preceding me, called out that a snake was wriggling into the undergrowth on the edge of the cultivated plot at a spot where it was only thirty yards from many acres of rain-forest. A woman who was hoeing weeds in the plot remarked that she had just hoed up and killed this snake; presumably the first part of her story was more accurate than the second! The other three ex- amples were found under piles of weeds in similar native clearings near the rain- forest. [Prosymna ambigua Socage] Prosymna ambigmts Bocage, 1873, Jour. Sc. Lisb., 4, p. 218. (Duque de Braganea, Angola.) Prosymna bocagii Boulenger, 1897, Ann. Mag. N. H. (6) 19, p. 278, fig. (Zongo, Ubongi Rapids, Congo.) Prosymna vassei Mocquard, 1906, Bull. Mus. Paris, p. 250. (near Mozambique.) IProsymna transvaalensis Hewitt, 1910, Ann. Trans. Mus. 2, p. 73. (Tzaneen, Zoutpansberg District, Transvaal.) In critically examining the status of the species of Prosymna before describing P. ornatissima, a re-examination was made of the Lumbo example of P. ambigua recorded, together with two found bottled at Morogoro, by the junior author in 1923.' The results of this investigation as set forth below force us to the conclu- sion that both P. bocagii Blgr. and possibly P. transvaalerisis Hew. must be added to the synonymy of P. ambigua. Alternately it may be found that true Angolan ambigua have 17 scale-rows only, while South and East African agree with the Congo bocagii in having 15; in this event P. bocagii will stand. Lumbo is on the seashore about three miles from Mozambique, so that the Lumbo snake is almost topotypic with Mocquard's P. vassei which Boulenger has already placed in the synonymy of his P. bocagii. In our opinion the speci- men before us bridges over the differences between P. bocagii and P. ambigua and a careful perusal of Hewitt's remarks, one paragraph of which we quote be- low, only confirms us in the view that we are here dealing with a very variable and wide-ranging species. For convenience of comparison we tabulate the scale formulae as follows: ' Loveridge, 1923, Proc. Zool. Soc. Lond., p. 880. 122 BARBOUR AND LOVERIDGE ?CALE FORMULAE OF PROSYMNA TYPES Scale-rows Ventrals Subcaudals Supralabials Labials enter- ing eye Praefrontal Frontal Postoculars Pairs of chin- shields p. ambigua of BIgr. Cat. Sn. II. 248 15-17 135-152 19-34 6-7 3rd and 4th Not bordering eye P. nmhigua Bocage'e Type 17 149 19 6 3rd and 4th Not bordering eye P. hocagii Boulenger's Type 15 167 19 6 3rd and 4th Bordering eye Anterolateral angles reach eyes 1 p. vassei Mocquard's Type 15 151 17 6 2nd, 3rd, 4th Upper end of small prae- ocular in con- tact with prae- frontal Anterolateral angles reach eyes ? Specimen from Lumbo Mozambique M. C. Z. 18.203 15 148 20 6 3rd and 4th Not bordering eye Anterolateral angles not reaching eyes 2 p. trans- vanlensis Hewitt's Type 15 156 22 (26) 6 3rd and 4th Presumably not bordering eye Anterolateral angles not reaching eyes 1 On right side supra and post- ocular are fused 2 Hewitt's P. transvaalensis may prove to be a good race based on the two pairs of chin-shields and coloration, the latter approximating to the condition met with in P. meleagris of West Africa. Besides his type from Tzaneen, he had 'also three specimens from Medingen, Klein Letaba (Rev. W. Krause). The largest of these agrees with the tj^De, excepting in that the supraocular is not fused with the postocular on either side. The two other specimens, both juvenile, are aber- rant, in that one of them has two postoculars and two suboculars, which latter separate the eye from the upper labials, while in the other specimen labials two and three on the right side, and two, three and four on the left side enter the eye. Moreover in these young specimens, the rostral is not so angular as in the adults.' The last point is interesting in view of Boulenger stating that his P. hocagii differs from P. ambigua 'in the more prominent, slightly turned up snout.' His P. hocagii had a total length of 340 mm., while the biggest specimen of P. ambigua mentioned in the Catalogue of Snakes was only 225 mm. AFRICAN HERPETOLOGY 123 DASYPELTIDAE Dasypeltis scaber (Linnaeus) Coluber scaber Linnaeus, 176G, 'Syst. Nat.' 1, p. 384. 1 (i\I. C. Z. 23273) Nyange, Uluguru Mtns., 7. x. 26. 1 (M. C. Z. 23274) .\mani, Usambara Mtns., 30. xi. 26. Variation. Mid-body scale-rows 25. Ventrals 241-244. Subcaudals 82-87. Coloration. Both are of the pinkish-brown type with vertical lateral stripes; in the larger snake these are almost united on the dorsal line by three minute cream spots between each pair of stripes; this snake has also a brownish dorsal band along the whole length of its body and tail. Measurements. The larger (Amani) of the two females measures 785 (640 -f 145) mm. Breeding. Three eggs measuring 25 x 10 mm. were found in the Amani snake on 30. xi. 26. Diet. Fragments of egg-shell, apparently of a weaver bird's egg, having red spots on a white ground, were recovered from the stomach of the Nyange reptile; its stomach also contained a stiff mass of yolk. Ene^niesf A second egg-eater was taken at Nyange in a grass hut. It was temporarily placed in a vivarium where it was apparently eaten by a Bird snake (Thelotornis hirtlandii) . BOIGIDAE Geodipsas vauerocegae Tornier Geodipsas vauerocegae Tornier, 1902, Zool. Anz. 26, p. 703. 1 (M. C. Z. 2.327S) Bagilo, LHuguru Mtns., 28. ix. 26. 2 (M. C. Z. 23279) Nyange, Uluguru Mtns., 4 and 9. x. 26. 4 (M. C. Z. 23280-3) Vituri, Uluguru Mtns., 28. x. 26. 11, (M. C. Z. 23284-94) .'Vmani, Usambara Mtns., xi-xii. 26. Distribution. This species is here recorded for the first time as occurring in the Uluguru Mtns., being until now known only from its type locahty 'Usam- bara.' Variation. Mid-body scale-rows 17. Ventrals 124-133. Anal single. Sub- caudals 36-48. Praeoculars 2. Temporals 1 -f 2 (except No. 23290 where there are 2-1-2). Supralabials 7, third and fourth entering the eye. It will be seen that the range of variation in this species is extremely small but it should be noted that the frontal is often considerably longer than its distance from the end of the snout; the nasal is frequently not in contact with the second labial; three 124 BARBOUR AND LOVERIDGE or four lower labials are in contact with the anterior chin-shields, which are longer or shorter than the posterior in many specimens (in the type they were ' as long as the posterior'). In those specimens that were carefully sexed in the field, the ventral range in males is 124-128, in females 131-133; subcaudal scale-counts of the two sexes overlap entirely. The length of the tail is contained in the body length 4.26 to 5.1 times in males, 4.99 to 5.69 times in females. Coloration in life. Nyange 9 (No. 23279). Above, brown; each scale edged with darker; supralabials yellow, their upper portion dusky; a broad black band on the nape touches the parietals and is followed by eight pairs of very indistinct black spots. Below, uniformly yellow except for dusky specklings on all the outer edges of the ventrals except the free edge. Young Bagilo 9 (No. 23278). Above and below, uniformly black except the head scales anterior to the eyes, the labials and the gular scales, all of which are black with a white centre. In the Amani series some snakes are uniformly black beneath, others pure white with brown specklings. Measurements. Largest (Amani) male measures 332 (275 + 57) mm. ; largest (Nyange) female 370 (313 + 57) mm.; the smallest snake, a female taken at Amani on 25. xi. 26, measures 168 (135 + 33) mm. Breeding. In the Nyange female were 8 eggs measuring 6 mm. long on 4. x. 26. Undeveloped eggs in an Amani snake collected 25. xi. 26. Diet. An undeterminable frog was in one of the Amani snakes. Parasites. Large numbers of mites were found under the ventral scutes of a Vituri snake (No. 23280). Habitat. The Bagilo snake was taken under a branch in a clearing in the rain- forest; the Nyange snakes, under logs in a clearing occasioned by the fall of a huge tree. Geodipsas procterae Loveridge Geodipsas procterae Loveridge, 1922, Proc. Zool. Soc. London, p. 313. 1 (M. C. Z. 23275) BagUo, Uluguru Mtns., 28. ix. 26. 1 (M. C. Z. 23277) Nyange, Uluguru Mtns., 6. x. 26. Distribution. Hitherto only known from two males taken at Bagilo in 1921 and 1922; both the present snakes are females. Variation. Mid-body scale-rows 17. Ventrals 147. Subcaudals 34. Like the second male taken, these females differ from the type in having the third, fourth and fifth labials entering the eye ; the anterior chin-shields are a little longer than the posterior; both agree in having 2 prae- and 2 postoculars. AFRICAN HERPETOLOGY 125 Diagnosis. This species is still readily distinguished from G. vauerocegae by its single subcaudals, more numerous ventrals, labials and praeoculars, viz.: G. procterae. Ventrals 147-153. Subcaudals 34-49. Labials 8, 3rd, 4th, and 5th entering eye (rarely only 4th and 5th). Praeoculars 2. G. vauerocegae. Ventrals 124-133. Subcaudals 34-48. Labials 7, 3rd and 4th entering eye. Praeocular 1 (rarely 2). Meas2irements. The largest (Bagilo) female measures 505 (450 + 55) mm. The tip of the tail is wanting. Coloration in life. In these females there is no trace of the dorsal stripe of the type. The Bagilo adult is uniformly purpUsh iridescent black above; below, bluish-white with the exception of the chin and throat, which are pure white. It differs from both the males in the colouring of the ventral surface. Breeding. The oviducts of the Bagilo snake contained 13 eggs measuring 18 X 10 mm. on 28. ix. 26. Diet. In the stomach of the Nyange snake was one of the frogs (Hoplophryne iduguruensis sp. n.) of which several were taken in the banana plants against whose base the snake was coiled. Habitat. The Nyange snake, which I personally caught, was found among vegetable debris at the base of a wild banana (Musa uluguruensis Warb. or ensete Gmel.) growing in a ravine near the summit of Mt. Tongoni. Crotaphopeltis hotamboeia hotamboeia (Laurenti) Coronella hotamboeia Laurenti, 176S, 'Syn. Kept.,' p. 85. 1 (M. C. Z. 23295) Nyange, Uluguru Mtns., 7. x. 26. 1 (M. C. Z. 23296) Dar es Salaam, 4. xi. 26. 1 (M. C. Z. 23297) Mombo at foot of Usambara Range, 31. .xii. 26. Variation. These specimens are all normal ; that is to say, they agree in having a single praeocular in contact with the frontal. Scale-rows 19. Ventrals 155-168. Caudals 40. The Mombo snake has temporals 1+1 and only the 4th and 5th labials entering the eye in Heu of the 3rd, 4th, and 5th. Coloration. The Mombo snake is strikingly different from any other example of the species which I have seen. The whole of the upper head, including upper lip, black. Symphisial and first three lower labials on one side, first four on left side, dusky, rest of the lower labials and throat pure white; the black of the upper surface encroaches on the ventrals to the same extent that the dorsal colouring does in Natrix olivaceus, more so in certain areas where it almost ex- cludes the white. 126 BARBOUR AND LOVERIDGE Diet. Stomach of the Dar es Salaam snake revealed the feet of a frog with large disks ; his intestines were full of ant remains which may reasonably be sup- posed to have been liberated from the frog's stomach. Parasites. The Nyange female had a cyst upon its liver; the Dar es Salaam snake's stomach held a single nematode. Enemies. One sunny morning, while walking along the path from Dar es Salaam to Mogogoni, my attention was drawn to a Sharp-snouted Snake {Rham- phiophis oxyrhynchus) engaged in trying to swallow a \^Tiite-lipped Snake whose head and neck were already in its mouth. The White-lipped Snake had, howe^'er, inserted its tail into the entrance hole of an ant's nest which was situated at the base of a bush. With the leverage thus obtained it steadily drew itself down the hole until but three inches of its neck remained visible. To reach the combatants I had to crawl in under the bush on hands and knees, after which I squatted above them in my cramped quarters, planning how to seize the Sharp-snouted Snake, which was well protected by thorny sprays. Though my deliberation was only a matter of seconds, the snake disgorged its prey and darted away like lightning. Simultaneously the "\Miite-lipped Snake withdrew its head into the hole. We hunted for the Sharp-snouted Snake but failed to secure it amidst the tangle of brambles and rank grass. Returning to the ant-hole, I was not surprised to see the Wliite-lipped Snake emerging literally covered with ants which were biting him unmercifully. Crotaphopeltis hotamboeia tornieri (Werner) Leptodira tornieri Werner, 1876, Sitzber. Akad. Wien, 116, p. 1S7.1. 7 (M. C. Z. 2.3298-23.304) B.agilo, Uluguru Mtns., 23-29. ix. 26. 13 (M. C. Z. 23305-17) Nyange, Uluguru JMtns., 1-9. x. 20. 5 (M. C. Z. 2331.S-22) Vituri, Uluguru Mtns., S-2S. x. 26. 19 (M. C. Z. 23.323-34) Amani, U,sambara Mtns., xi-.xii. 26. 1 (M. C. Z. 23135) Mt. Lutindi, Usambara Mtns., 10. xii. 26. Variation. Mid-body scale-rows 17-19. Thirty-seven of the forty-six snakes have 17 scale-rows, one has 18, seven have 19. Ventrals range from 145 to 172; the average in Uluguru snakes is considerably higher than in those from the Usam- bara, viz. 163 and 152; nor is this a question of sex, for the average by sex in the Uluguru is 165 for males, 162 for females; in the Usambara 151 for males, 153 for females. Subcaudals range from 35 to 56, the average in males being 48 and in females 45 (Uluguru males 50, females 48; Usambara males 46, females 42). It may be definitely assumed that in this species neither the ventral nor subcaudal scale-counts are any guide to sex. Chin-shields range from 3 to 6, with an average AFRICAN HERPETOLOGY 127 of 3 (nearly 4 in Uluguru). Thirty-three of the forty-five snakes have 3 praeocu- lars; six have 2 praeoculars on one side and 1 on the other; six have 1 praeocular on both sides of the head, but in one specimen these are semidivided. Seventeen snakes agree with the type in having the praeocular in contact with the frontal ; four are in contact on one side only; twenty-three agree with hotamboeia in not having them in contact. Thirty-three snakes have 14-2 temporals; six have 1 -h 1 ; one has 1 -(- 3 ; the remaining five are irregular combinations with 14-2 on one side of the head and 1 -h 1 or 1 4- 3 on the other. Eight (very rarely 9) upper labials, ten lower labials; in twenty-six snakes the third, fourth and fifth labials enter the eye; in the remainder the fourth and fifth, or fourth, fifth and sixth, sometimes on one side only. Relations. As tornieri intergrades with hotamboeia in so many respects, it ap- pears to us desirable to emphasize this relationship by making it a subspecies of the older name, of which it forms a good mountain race characterized by 17 (rarely 18 or 19) scale-rows; usually, but by no means always, 3 praeoculars of which the uppermost may, or may not, be in contact with the frontal. Coloration in life. Males (Bagilo). Above, uniformly plumbeous with iri- descent sheen; in one specimen the head is noticeably olive. Below, mental and a pair of lower labials on either side of it, dusky brown; rest of chin, throat and anterior ventrals, whitish; in one individual the white persists along the centre of the ventrals ahnost to the anus, the borders of the ventrals being dusky; in the other the dusky borders encroach upon the white so as to obliterate it about two inches behind the head, leaving the rest of the undersurface uniformly plumbeous, as are the undersides of the tails in both snakes. Eye reddish-brown with vertical black pupil. Meas^{re77ients. The largest male (Bagilo) measures 530 (445 4-85) mm.; the largest female (Vituri) 471 (437 4- 34) mm.; the smallest snake, a male taken at Amani on 25. xi. 26., measures 203 (170 4- 33) mm. Breeding. Nine small eggs measuring 7x4 mm. in a 511 mm. female taken at Bagilo on 29. ix. 26. Three eggs measuring 26 x 7 mm. in a 390 mm. female at Nyange on 4. x. 26. ; six eggs 16x8 mm. at same place on 8. x. 26. Four eggs 18x8 nxm. in a female at Amani on 25. xi. 26. Diet. A frog-eater. At Bagilo Callulina kreffti, Breviceps nigosus and a green grub were recovered from stomachs. At Nyange three Arthroleptis stenodactijlus in one snake, B. rugosus in another. At Vituri A. stenodactylus and Leptopelis uluguruensis sp. n. At Amani Arthroleptis xenodactylus and C. kreffti. At Mt. Lutindi a C. kreffti. 128 BARBOUR AND LOVERIDGE Parasites. Nematodes (Kalicephalus sp. and immature Ascarids) are com- monly found in the stomach of this species and were preserved from all localities. Cestodes (Ophiotaenia crotaphopeltis sp. n.) were associated with the former in a Nyange male. Enemies. A Bird Snake (TheoUornis kirtlandii) was discovered with a large Tornier's Snake in its mouth. The Bird Snake held on doggedly, occasionally chewing with its poison fangs. The Tornier's Snake felt about with its tail for twigs and branches on which to gain a purchase. After eight minutes spent in this way the Bird Snake tried to swallow and was then observed to be in diffi- culties; the Tornier's Snake had hooked its teeth into the mucous membrane of the Bird Snake's mouth. I therefore separated them. The Tornier's Snake tried to make off but the Bird Snake, which had itself withdrawn, returned swiftly, and seizing its victim began to swallow it again. The Tornier's Snake being a large one, deglutition was unusually laboured. It took an hour, all but four min- utes, from the time when I first noticed that this Bird Snake had seized one of its companions in the vivarium. Habitat. The first specimens were found at Bagilo beneath a massive piece of tree- trunk, which was resting on a litter of twigs which raised it somewhat from the ground. It was the only dry piece of ground which I had seen during the day, for the whole rain-forest region was sodden after the heavy downpour of the pre- vious day. The log was in a fairly recent clearing occasioned by the fall, and sub- sequent cutting up, of a giant tree. Six snakes were taken at the spot, which was typical of the situations in which the large Amani series were captured. Temperament. When uncovered these snakes started to wriggle away and when gently seized made no attempt whatever to bite ; this attitude was typical of the whole large series, which displayed a gentleness in sharp distinction to L. h. hotamboeia, which is notorious for its irascibility. Thelotgrnis kirtlandii (Hallowell) Leptophis kirtlandii Hallowell, 1844, Proc. Acad. Nat. Sci. Philad., p. 02. 1 (M. C. Z. 2.33.36) Nyange, Uluguru Mtns., 4. x. 26. 1 (M. C. Z. 23337) Vituri, Uluguru Mtns., 27. x. 26. 11 (M. C. Z. 23338-23348) Ajnani, Usambara Mtns., xi-xii. 26. 2 (M. C. Z. 23349-23350) Mlalo nr. Ambangula, Usambara Mtns., xii. 26. Lvkvkuru (Kikami); sungahuni (lusumbara). Variation. Ventrals 154-170. Subcaudals 113-1G9. In males the ventral count averages slightly lower than in females and the subcaudals slightly higher, but much overlapping occurs. There is also overlapping in the relation of tail AFRICAN HERPETOLOGY 129 length to body length according to sexes. In addition to the variations recorded in the Catalogue of Snakes, iii, p. 185, only eight specimens have the normal fourth and fifth labials entering the eye; five have the third, fourth and fifth; while two combine these conditions on the two sides of their heads. In one snake with nine labials on the left side of its head, the seventh labial is divided into three scales thus: { + 1. A single snake has 2 praeoculars. All but two have normal temporals, viz., 1 + 2; the others have 1+3 and 2 + 2. Measurements. The largest male measures 1215 (765 + 450) mm., yet has the tip of its tail missing; the largest female 1212 (740 + 472) mm. Both from Mlalo. An Amani female whose tail has been chopped off measures 50 mm. longer in the body than even the fine Mlalo reptile. Breeding. A Nyange female held five eggs measuring 15 x 5 mm. on 4. x. 26. Diet. As already related the Bird Snake w^ll eat other snakes when in cap- tivity. The Nyange female, being temporarily placed in a vivarium containing other snakes, was observed to be swallowing a Matrix olivaceous about 10 a.m. the following day. On chloroforming the Bird Snake the same evening, I found the head already digested from the body of the Natrix and also that a Chlorophis neglectus had been taken a day or two before. Crotaphopeltis hotamboeia tornieri was also eaten by this species. Parasites. A fragment of a cestode (Ophiotaenia sp) was found in a Nyange snake. DisPHOLiDus TYPus (Smith) Bucephalus typus Smith, 1829, Zool. Jour., 4, p. 441. 1 (M. C. Z. 2.3351) Bulwa, Usambara Mtns., 23. xi. 26. 1 (M. C. Z. 2.3352) .Ajnani, Usambara Mtns., xii. 26. 1 (M. C. Z. 233.53) Kizara, Usambara IMtns., 9. xii. 26. 1 (M. C. Z. 23354) Bumbuli, Usambara Mtns., 16. xii. 26. 1 (M. C. Z. 23355) Sakkarani, Usambara iSItns., 18. xii. 26. 1 (M. C. Z. 23356) Lushoto, Usambara Mtns., 19. xii. 26. 1 (M. C. Z. 23357) Mlalo near Ambangula, Usambara Mtns., xii. 26. A half-grown specimen was also collected at Mlalo nr. Hohenfriedeberg and taken to London alive. All the above are adult. Ngole in Kikami and Kisumbara; in Ivikami, however, this name is also applied to Chlorophis neglectus. Variation. Mid-body scale-rows 19-21, the latter on the Lushoto snake only. Ventrals 173-191. Subcaudals 114-130. Supralabials 7, third and fourth enter- ing the eye. Temporals 1+2. Praeoculars 1 and postoculars 3, except in No. 23351 where on the right side they are 2 + 3, on the left 1 + 4, a most unusual variation. By sexes, the ventrals in males 173-175, in females 183-191 ; subcaudals in males 124-130, in females 114-124, but three are omitted on account of the 130 BARBOUR AND LOVERIDGE tips of their tails being missing. Schmidt's Congo series (1923, Bull. Am. Mus. Nat. Hist., xlix, pp. 114-116) show that the lower number of ventrals in males is a guide to sex though they overlap in the region of 183-186; subcaudals are use- less as a guide to sex. Coloration. All are bright livid green or grey-green, except the Amani snake, which is reddish, and the wholly black Mlalo specimen. Measurements. Largest male measures 1652 (1200+452) mm.; largest female 1775 (1320 + 455, tip of tail missing). Possibly both of these are records. Breeding. Eggs measuring 43 x 17 nmi. were found in the Mlalo snake ob- tained by my collector during the last few days of December. Diet. The Bulwa snake had bird's eggs in its stomach; the Lushoto Boom- slang had recently swallowed a Two-horned Chameleon (Chamaeleo fischeri muUituberculatus) . Temperament. I was told that these snakes are very common in the gardens of Lushoto. My attention was drawn to one in a tall privet-like hedge surround- ing the hotel. I gently took the reptile by the tail and so held it as it quietly strained to be free, until my wife fetched my snake stick. It made no attempt to bite when pulled out of the tree and pinned down. Calamelaps tjnicolor (Reinhardt) Calamaria unicolor Reinhardt, 1843, Danshe Vidensk. Salsk, Afh., 10, p. 236, PI. I, figs. 1-3 Calamelaps polylepis Socage, 1873, Jom. Sc. Lisb., 4, p. 216. 1 (M. C. Z. 23358) Nyange, Uluguru Mtns., 11. x. 26. 3 (M. C. Z. 233.59-62) Amani, Usambara Mtns., 22. xi-xii. 26. Variation. Mid-body scale-rows 17. Ventrals 164-171. Subcaudals 24-27. Except that they have fewer ventrals than West African examples, the present series of five males conforms to the description of C. unicolor as given in the Cata- logue of Snakes, iii, p. 245. Relationships. We have carefully compared the specimen of C. polylepis from Lumbo, Mozambique,' with the present series and find that it is identical with them in all respects except in having 19 scale-rows. Boulenger - states of C. polylepis that it is 'in every respect like C. unicolor except that the scales are in 21 rows.' Tornier ' has recorded specimens with 17, 19 and 21 rows from the Tanga-Usambara region, and elsewhere the junior author ■• has reported on a ' Loveridge, 1923, Proc. Zoiil. Soc. Lond., p. 889. 2 Boulenger, 1896, Cat. Sn. Brit. Mus., 3, p. 246. = Tornier, 1901, Zool. Jahrb. (Syst.), 14, p. 85. * Loveridge, 1923, Proe. Zool. Soc. Lond., p. 889. AFRICAN HERPETOLOGY 131 specimen, probably taken in the ]Morogoro-Uluguiu region, with 19 scale-rows. As the distribution in East Africa appears identical for sriakes with 17, 19 or 21 rows, we appear to be dealing with a single species and therefore refer C. polylepis to the synonymy of C. unicolor. Not having examined the type of Schmidt's C. niangarae from Niangara, Congo Beige, ^ we hesitate to place it in the synonymy of C. unicolor, where, to judge from the description alone, one would imagine that it should go, for it is apparently a male of that species. Coloration in life. Uniformly iridescent black, some individuals opaque grey, the latter presumably about to slough though showing no other indications of it. I had intended inviting attention to the extraordinary superficial likeness of this snake to the very venomous Atractaspis irregularis wliich occurs in the same localities, but in referring to the literature I find that Tornier has already placed Atractaspis hildebrandtii Peters in the synonymy of C. unicolor, and the fact that Boulenger referred a specimen of unicolor to hildebrandtii is eloquent proof of the similarity. Measurements. The largest of these five males measures 472 (430 + 42) mm. Diet. A Nyange snake had eaten a Lygosoma kilimensis. One Amani reptile had an Aparallactus werneri in its stomach, and another a Boulengerida houlengeri. Habitat. One was hoed up in a garden at Nyange; the Amani specimens were taken in the clearing of land. Aparallactus werneri Boulenger Aparallactus werneri Boulenger, 1895, Ann. Mag. Nat. Hist., (6) 16, p. 172. 1 (M. C. Z. 23371) Bagilo, Uluguru Mtns., 28. k. 26. 1 (M. C. Z. 23372) Vituri, Uluguru Mtns., 19. x. 26. 1 (M. C. Z. 23373) Mkarazi, Uluguru Mtns., 23. x. 26. 43 (.M. C. Z. 23374-406) Amani, Usambara Mtns., xi-xii. 26. 1 (M. C. Z. 23408) Ivizerui, Usambara Mtns., 8. xii. 26. 1 (M.-C. Z. 23409) Mt. Lutindi, Usambara Mtns., 10. xii. 26. 1 (M. C. Z. 23410) Bumbuli, Usambara Mtns., 14. xii. 26. Variation. The interesting point about this species is its remarkable con- stancy to type. In the whole series there are only three minor variations, viz., nasal enters eye on right side in No. 23393 ; there are 1 prae- and 1 postocular on right side of head and 1 prae- and 2 postoculars on left side of head in No. 23405; the symphisial is well in contact with the anterior chin-shields in No. 23406. Other- wise every specimen is true to type in the number of scale-rows, anal, labials, position of labials beneath eyes, temporals and symphisial. ' Schmidt, 1923, Bull. Am. Mus. Nat. Hist., 49, p. 117, fig. 12. 132 BARBOUR AND LOVERIDGE The ventrals range from 141 to 163. Subcaudals 32-45. In 16 males the ventrals range from 141 to 151. Subcaudals 35-45. In 33 females the ventrals range from 146 to 163. Subcaudals 32-42. Coloration in life, cf Bagilo. Above, head and neck (for seven scale-rows behind parietals) iridescent black, followed by a narrow band (two scale-rows wide) of yellow, rest of the body olive-brown, each scale edged all round with black; a yellow spot on the first, fourth, fifth and sixth upper labials. Below, bright yellow, slightly darker on tail. 9 Amani. Above, head and neck black, more or less separated by a narrow yellow collar, rest of upper surface bright green with a reticulation of fine black lines caused by each scale having black edges. Below, chin white, throat yellow- ish, rest of under surface very bright yellow. Before sloughing uniformly grey above and dirty white below. 27. xi. 26. Measurements. Largest male measures 333 (305 + 28, tip missing) mm.; largest female 354 (295 + 59) mm. Both Amani. Breeding. Seventeen females examined at Amani in November had large, well- developed, very elongated eggs. Seven had two eggs ; eight had three ; and two had four. Those with two eggs were usually rather small snakes, but the largest eggs, measuring 39 x 6 mm. on 25. xi. 26., were in one of these. Apparently females are more numerous than males if one can judge on a basis of 49 snakes of which only 16 were males. Diet. A Vituri and three Amani snakes had each eaten a centipede. Enemies. A Werner's Black-headed Snake was recovered from the stomach of a Calamelaps unicolor. Habitat. Taken beneath logs, bark and stones, both within and without the rain-forest. Some were unearthed by hoeing up grass and weeds on the outskirts of the forest. In fact its habitat is almost identical with that of the North Ameri- can Ring-necked Snakes (Diadophis spp.) which it so closely resembles in colour, markings and size. Apaballactus ■ulugurxjensis sp. n. 3 (M. C. Z. 23365-65) Nyange, Uluguru Mtns., 1-6. x. 26. 7 (M. C. Z. 23366-70) Amani, Usambara Mtns., xi-xii. 26. Type. No. 23363. Museum of Comparative Zoology. Sex d'. From Nyange, Uluguru Mountains, Tanganyika Territory. Collected by A. Loveridge, Octo- ber 1st, 1926. Paratypes. Nos. 23364, 23365, 23366-70 and two duplicates. AFRICAN HERPETOLOGY 133 Affinities. Both by its size and colour this would appear to be a southerly representative of A. concolor though structurally far more Hke A. guentheri Blgr., of which it is certainly no close relative. Diagnosis. Differs from concolor in that the visible portion of the rostral is only a third its distance from the frontal, the nasal being in contact with the praeocular, the second supralabial separated from the praefrontal, symphisial in contact with anterior chin-shields, and only three lower labials in contact with anterior chin- shields. It differs from guentheri in those characters which are emphasized above by itaUcs, also in the undivided (rarely divided) nasal, the anterior chin-shields being much longer than the posterior, and in the entirely different colouring, habit and size. Descriptioii. Diameter of eye greater than its distance from the oral margin. Rostral broader than deep, the portion visible from above equal to one-third its distance from the frontal; inter nasals shorter than the praefrontals; frontal once and a half as long as broad, much longer than its distance from the end of the snout, a little shorter than the parietals; nasal entire (rarely divided), in contact with the praeocular; one postocular; temporals 1+1; seven upper labials, third and fourth entering the eye, fifth in contact with the parietal; sym- phisial in contact with the anterior chin-shields, which are longer than the pos- terior and in contact with three lower labials (in two of the paratypes the division of an anterior lower labial on the right side only has taken place so that on this side there are four lower labials in contact with the anterior chin-shield). Scales in 15 rows. Ventrals 144 (137-159); anal entire; subcaudals 51 (45-58). Coloration. Plumbeous or iridescent black above ; paler beneath, particularly on throat and under surface of tail, which are almost white. Measurements of Type (f. Paratypes Total length 360 mm. The largest para type male measures 400 mm. Length of head and body 2S5 " (320+ 80) Length of tail 80 " Largest female 401 " Diameter at mid-body 10 " (325+76) Diet. Six centipedes of two species (Alipes grandidieri and another which is preserved) were recovered from their stomachs. In one Nyange snake, besides two centipedes there was a snail. 134 BARBOUR AND LOVERIDGE ELAPIDAE Elapsoidea guentheri Socage Elapsoidea guentheri Bocage, 1866, Jorn. Sci. Lisboa, 1, p. 70, PI. I, fig. 3. Elapsoidea nigra Gunther, 188S, Ann. & Mag. Nat. Hist. (6) 1, p. 322 (Ushambola, Zanzibar). 5 (M. C. Z. 23411-5) Nyange, Uluguru Mtns., 1-11. x. 26. 39 (M. C. Z. 23417-54) Amani, Usambara Mtns., xi-xii. 26. 3 (M. C. Z. 234.55-7) Bumbuli, Usambara Mtn.s., 16. xii. 26. Variation. Mid-body scale-rows 13. Ventrals 153-162. Anal single. Sub- caudals 13-23. Certain specimens show here and there one or more of the paired subcaudals fused into a single shield. Praeocular 1. Postoculars 2, except in No. 23413 where there is only a single shield. Temporals 1+2, except in Nos. 23424, 23427, and 23435, which have an extra small scale between the anterior temporal and the lower postocular on both sides. In No. 23441 this is the case on the right side only. Usually 4 lower labials in contact with the anterior chin-shields; only eight specimens in the series have 3 lower labials in contact and then usually on one side only. There is absolutely nothing in the key-character based on the comparative length of the internasals and praefrontals to separate E. nigra from E. guentheri; in the present series the internasals range from less than half to seven-eighths the length of the praefrontals. Coloration in life. Above, highly polished iridescent black, with or without transverse white bars. In some apparently entirely black adults the reptile will, on being disturbed, inflate its lungs and therebj^ suddenly bring into prominence the hitherto concealed white bars. Below, usually uniformly iridescent-black, paler, or even white, upon the throat ; very rarely there are a few white patches in the middle of the ventrals. Young specimens show a series of broad black bands (these are just a Uttle blacker than the ground colour) narrowly edged with white on both sides. E. guen- theri, as known to the junior author at Nairobi, is banded with pink and black or red and black in life, and it is difficult to believe that it is the same species as the series before us, yet colour seems to be the only ground on which the two reptiles can be distinguished. The species is obviously extremely variable in coloration. Furthermore, there is in the collection of the Amani Institute a snake col- lected at Amani, 20. v. 1907, which appeared to me to be of the red-banded type, though I may be mistaken. In 1910 Nieden fists E. guentheri as occurring at Amani, though whether his specimens were identical with what have heretofore been called E. nigra it is impossible to say. AFRICAN HERPETOLOGY 135 Measurements. Largest male measures 544 (507 + 37) mm. ; largest female 569 (530 + 39) rmn. Smallest snake, a male taken 22. xi. 26., measures 264 (245 + 19) mm. All from .Ajnani. Habitat. Found in clearing land or amongst heaps of rubbish within or with- out the rain-forest. I encountered several moving along the paths in the rain- forest between noon and 3 p.m. Temperament. By no means a vicious snake; if gently handled it makes no attempt to bite. Wholly black specimens with their small heads, scarcely larger than the body, are sufficiently like an Atractaspis viper in life so that they are not to be readily distinguished. Breeding. A Nyange female on 1. x. 26 had two enormously elongated eggs 40 X 10 mm. in the oviducts. Four Amani snakes examined on 29-30. xi. 26 had 2, 3, 4 and 4 eggs respectively, the largest of which measured only 27 X 10 mm. The two Bimibuh females taken 16. xii. 26. held 4 and 5 eggs; those of the latter series measured 37 x 12 mm. Diet. At Amani this species seems to subsist largely on the caecilian Boulen- gerula boulengeri as the following records show. Snake taken at 8 a.m. on 20. xi. 26. had a perfect undigested caecihan in its stomach. Another taken at 3 p.m. the same day held one with only the head digested, the skull being nicely cleaned. On 22nd another snake had the head and forepart of a caecihan in its stomach. On 30th three caecilians were recovered from the stomachs of two snakes, one having swallowed two. Parasites. A mass of tapeworms (Ophiotaenia elapsoideae sp. n. and Ophio- taenia near 0. monnigi) in the stomach of a Nyange snake examined 1. x. 26., and in another taken 11. x. 26; also abundant in an .Amani snake collected 23. xi. 26. Nematodes (Kalicephalus sp.) and cysts in stomach of an Amani snake on 25. xi. 26., and nematodes in another taken 29. xi. 26. Mites under ventrals of Amani snake No. 23445. VIPERIDAE Causus defilippii (Jan) Heterodon defilippii Jan, 1862, Arch. Zool. Anat. Phys. 2, p. 225. 2 (M. C. Z. 23458-9) Nyange, Uluguru -Mtns., 11. x. 26. Kinangananga in Kikami, a specific name. Variation. Except that one of these females has 2 -|- 2 temporals, in heu of 2-1-3, these specimens are normal. Mid-body scale-rows 17. Ventrals 118-120. Subcaudals 19. 136 BARBOUR AND LOVERIDGE BiTis ARiETANS (Merrem) Vipera {Echidna) arietans Merrem, 1820, 'Tent. Syst. Amphib.' p. 152. 1 (M. C. Z. 23463) Dar es Salaam, 8. xi. 2(i. Known to both Wakami and Wasumbara as moma, a name the latter also apply to the Gaboon Viper. It is interesting to note that this species, so common in the savannah countrj' of Africa, and even on the lower slopes of the Uluguru Mtns. at Morogoro, is said by the Wakami not to occur in the mountains. BiTis GABONicA (Dumeril et Bibron) Echidna gnhnnira Dumoril et Bibron, 1854, 'Erpet. Gen.,' 7, p. 1428, PI. LXXX, fig. b. 3 (M. C. Z. 23460-1) Amani, Usambara IMtn.s., xi.-xii. 26. 1 (M. C. Z. 23462) Iviserui, Usambara Mtns., 8. xii. 26. Moma in Kisumbara, not specific as also applied to the PulT .\dder. Distribution. This species does not occur in the Uluguru Mtns. The local chief at Kizerui assured me that Bitis arietans is also to be found near his village. As I was there only for one rather rainy day I was unable to verify this but it would be interesting to know if the two species occur in the same habitat. At Amani the Puff Adder is found a thousand feet lower down in the Sigi country. Variation. Mid-body scale-rows 28-44. Anal single. Subcaudals 20-33. Labials 13-15, all of which are well within the recognized limits of variation for this species. Measrirements. ^Vhen freshly killed the largest of these four females measured 49 inches or 1160 (1083 -I- 77) mm., which is very Uttle under the maximum re- corded by Boulenger. Diet. This large female had a Byatt's Squirrel (Aethiosciurus byatti) in its stomach; another Amani snake, a rat {Rattus sp.). The Kizerui reptile had also eaten a rat (Lophnromys aquihis aquilus). Parasites. A cestode (Ophioiaenia gabonica (Beddard)) was taken from the stomach of the large Amani viper. Atheris ceratophorus Werner .Mheris ceratophorus A\ erner, 1895, Verb. Zool.-bot. Ges. Wien, 45, p. 194, PI. \', fig. 1. 1 (M. C. Z. 23464) Amani, Usambara Mtns., xii. 26. 1 (M. C. Z. 23465) Kizerui, Usambara Mtns., 8. xii. 26. Variation. Both males; the type was a female. Rostral a little more than twice (not two and a half times) as long as broad; 8-9 (10-11 in type) scales AFRICAN HERPETOLOGY 137 across the interorbital region; 3-5 (3 in type) enlarged horn-like supraciliary scales. In the Kizerui snake there are only 9 upper labials and only 3 lower labials in contact with each chin-shield. Both have 21 (not 25 as in type) scale- rows. Ventrals 144-152 (142 in type). Subcaudals 41 -|- and 56 (Type 55). Coloration. While the Kiserui reptile agrees with the type in its markings and dark olive colour, the Amani snake is without the markings and is a paler, more yellowish, green. Measurements. Largest male (Kizerui) measures 422 (342 -|- 80) mm. Diet. An arboreal frog, apparently Hyperolius sp., was recovered from the stomach of the Kizerui snake. Habitat. The natives informed me that these snakes are found in grass or low bushes at a height of three or four feet from the ground. A European told me, to my surprise, that he had often seen horned Cerastes {sic) vipers in the grassy plains of the Kondoa-Irangi district. This lies within the known range of the species but is an unlikely habitat, if correct. Atractaspis rostrata Giinther .itractaspis rostrata Gunther, 1868, Ann. Mag. Nat. (4) 1, p. 429, PI. XIX, fig. 1. 3 (M. C. Z. 234U6-S) Nyange, Uluguru Mtns., 11. x. 26. 1 (M. C. Z. 234669) Mkarazi, Uluguru Mtns., 22. x. 26. 1 (M. C. Z. 234670) Dar es Salaam, 4. xi. 26. Variation. Mid-body scale-rows 21-23 (one Nyange snake and the JMkarazi specimen have 21, the others 23 rows). Ventrals 235-276 (the previous recog- nized range was 227-248). Subcaudals 21-26. In every other detail they agree precisely with the description. Coloration. Three distinct types of coloration are represented. The tongue is white. (1) Uniformly iridescent black above and below (Nyange). (2) Opaque or slaty grey above, lighter below. This snake is presumably about to slough (Nyange). (3) Brown above, pure white below (Dar es Salaam). Measurements. Largest male measures 581 (542 -|- 39) mm. ; largest female measures 559 (525 -|- 34) mm. Both from Nyange. Diet. The last-mentioned snake had in its stomach a blind snake (Typhlops hraminus), the head and foreparts of which were in good condition, the rest digested. Habitat. I found the Alkarazi snake at 9 a.m. beneath a very large flat boulder in a field of stubble; the Dar es Salaam reptile at 11 a.m. under a pile of leaves and sweepings in the Botanical Gardens in the centre of the town. 138 BARBOUR AND LOVERIDGE Venom. A small boy, about eight years of age, was bitten on the finger by the smallest snake, which was only 323 mm. long. TVlien I arrived at the place two days afterwards, he had a very swollen hand and wrist but it subsided on the third day. The only treatment he received was a bread poultice which I applied on the evening of the second day to ease his mind as much as anything. GEKKONIDAE Paragon ATODES africanus (Werner) Gximnodactyhis africanvs Werner, 1895, Verb. Zool. bot. Ges. Wien, 45. p. 190, PI. V. 1 (M. C. Z. 24011) Vituri, IHuguru Mtns., 27. x. 2fi. 13 (M. C. Z. 24012-21) Amani, Usambara Mtns., 20-29. xi. 26. 1 (.M. C. Z. 24022) Kizerui, Usambara Mtns., 8. xii. 29. 3 (M. C. Z. 24023-.5) Mt. Lutindi, Usambara Mtns., 10. xii. 26. 2 (M. C. Z. 24020-7) Bumbuli, Usambara Mtns., 15. xii. 26. Eggs (M. C. Z. 24028) Phillipshof, Usambara Mtns., 21. xU. 26. 2 (M. C. Z. 24029-30) Kwai, Usambara Mtns., 24. xii. 26. Variation. These agree with the type very closely; occasionally there are 8 or 9 supralabials (6-7 in tj'pe series) ; the three chin-shields bordering the mental behind appear to be a constant feature. Three of the five males have 10 praeanal pores, the fourth 11, the fifth (Vituri) 13. Relationship. We are indebted to Dr. G. K. Noble for making a comparison of one of the above topotypic specimens with the genotype of his new genus Paragonatodes. He writes: I find that the species is undoubtedly referable to Paragonatodes. A number of .small differences appear. Thus I find rudiments of a second epibranchial present. The pelvis, hyoid, shoulder girdle, and other structures come within the limits of Para- gonatodes. Coloration in life. Amani female. Above, blackish with greyish-green mark- ings; a chain of these down the back are rather like a series of ornamental A's; in some individuals these coalesce to form a pale-fawn vertebral line with paired stripes at right angles to it, formed from the feet of the A. In specimens taken on green tree-trunks the green increases and brightens to olive. Labials dusky with pale-yellow spots, many inconspicuous yellow specks on sides of head, neck and limbs. Below, throat white with three dusky A -shaped markings diminishing in size towards the centre; breast, belly, and underside of limbs yellowish-green or yellowish-white, under-surface of tail bright reddish-orange. In males the gular A -shaped markings appear to be a more constant feature; a series of them, one within the other, follows the outhne of the lower jaw. AFRICAN HERPETOLOGY 139 Measurements. Largest male (Amani) measures 101 (47 + 54) mm.; the largest female (Kwai) is 119 (54 + 65) mum. The smallest specimen, a female taken on 15. xii. 26 which has part of its tail missing, measures 24 mm. in length of head and body. Breeding. At Amani six eggs, each measuring 9x8 mm., were found in the rotten wood of fallen logs, the adults being collected at the same place and time, viz., 22. xi. 26. As in most geckos, only two eggs are produced at a time, the pair of eggs being visible in several of the females captured. These differ from those of the genus Lygodadylus, Hcmidaclyhis mabouia and many other species in not being stuck together. At Phillipshof, on 21. xii. 26, my attention was attracted to a hole in a bank in the forest by a snake's cast skin and a number of broken egg-shells about the entrance. After opening up the passage I raked out the loose soil at the terminus and found no fewer than thirty-one whole eggs of Para- gonatodes africanus. Diet. An ant in one and a cockroach in another. Distribution. This lizard is recorded for the first time from the Uluguru Altns. , where it is decidedly rare. It has been recorded from IVIt. Meru by Lonnberg and from j\It. Kenia by Tornier. It is evidently a species definitely associated with mountain rain-forest. Habitat. At Amani four adults were collected in the rotten wood of fallen logs where their eggs were found. On three occasions in as many days I saw them on the trunks of living trees but in each instance the tree had a decayed interior into which the gecko retired. They emerge about 5 p.m., sunset being at 6 P.M. In shady forest they may be seen on the trees at almost any hour. Four or more were seen one afternoon as I was walking through the forest from jVIisalae to Kizerui. One was on the brink of a large pool in the forest. On my approach, it took to the water without the shghtest hesitation and swam out to a piece of floating-bark. As I very quietly took another step in its direction, it jumped off its raft and struck out boldly across the six feet of water to the opposite bank, which it gained before I did. Had it not paused to rest on landing, I should not have caught it. The species squeaks loudly when captured. At Lutindi three were taken on tree-trunks or under logs. A very young one was captured on the whitewashed outer wall of one of the Mission buildings at Bumbuh. The second was taken on a rock. This was very interesting, for the forest at Bumbuh has been cut back to a mere cap along the crest of the hills ; it is very dry and no longer rain-forest, and it appears as if the gecko is attempting to adapt itself to other environments. At Lushoto much of the original forest 140 BARBOUR AND LOVERIDGE has been cleared away and acres planted with eucalyptus. A Paragonatodes crossed the path as I was walking through one of these extensive plantations, and when I stooped to intercept it, the gecko sprang or leaped to the trunk of one of the trees and made good its escape. Hemidactylus mabouia (Moreau de Jonnes) Plate 4, figs. 2 & 4. Gecko mahouia Moreau de Jonnes, 1818, Bull. Soc. Philom. Paris, p. 138. 1 (M. C. Z. 24031) Bagilo, Uluguru Mtns., 1.5. ix. 26. 1 (M. C. Z. 240.32) Nyange, Uluguru Mtns., 1. x. 26. 1 (M. C. Z. 24033) Bagamoyo, 11. xi. 26. 1 (M. C. Z. 24034) Tanga, 15. xi. 26. 3 (M. C. Z. 24035-7) Amani, Usambara Mtns., 25. xi. 26. Eggs (M. C. Z. 24038) Amani, Usambara Mtns., 26. xi. 26. 1 (M. C, Z. 24039) Soni, Usambara Mtns., 18. xii. 26 Also seen, or captured as snake food, at Nyingwa (near foot of mountain), Mkangazi, Vituri, Uar es Salaam, Muheza, Kerogwe, Kizerui, Mt. Lutindi (at very foot of mountain). Kimbulubxdu in Kikami, chungula in Kisumbara, mjusi kafiri in Kiswahili. Breeding. Paired eggs were found in the bark of a coconut palm at Dar es Salaam on 4. x. 26. ; they contained well-developed embryos. A female was shot from beside a pair of eggs stuck to a cliff-face at Amani, 26. xi. 26. These eggs measured 13 mm. in diameter. Diet. Big spider in stomach of a Bagilo female. A cockroach was disgorged by a gecko caught at Vituri. Parasites. Nematodes of the family Physalopteridae were taken in a Baga- moyo specimen. Habitat. Nearly all these geckos were taken on the trunks of large, usually hollow, trees. At Amani, where they are not as plentiful as one would expect, they also occur in the houses. Folklore. The Wakami say that if this lizard's faeces fall upon one's clothes, the stain will not come out. Enemies. Salimu reported having seen a Mountain Kestrel (Falco tinnunculus carlo) drop down from the air and pick a lizard off a rock. 1 sent him to shoot the bird and from its crop we recovered the remains of a Hemidactylus mahouia. Hemidactylus persimilis sp. n. Plate 4, figs. 1 & 3. Hemidactylus mabouia (part) Loveridge, 1923, Proc. Zool. Soc. Lond., p. 936, Mbala record; p. 937, paragraphs 1 and 7 only. 1 (M. C. Z. 24040) Bagilo, Uluguru Mtns., 14. i.x. 26. Eggs and 14 (M. C. Z. 24041-9) Dar es Salaam, 9. xi. 26. Type. No. 24041 Museum of Comparative Zoology. Sex d". Dar es Salaam, Tanganyika Territory. Collected by A. Loveridge, November 9th, 1926. AFRICAN HERPETOLOGY 141 Paratypes. Rest of the series mentioned above and those enumerated below. Affinities. Almost indistinguishable from immature specimens of Hemidacty- lus mabouia. Diagnosis. Differs from H. mabouia in having 5-6 lamellae under the median digit instead of 7-9. When fully grown and breeding it measures only 103 mm., while H. mabouia attains to 188 mm. The egg of this species measures 8 to 8.5 mm. in diameter, that of H. mabouia from 12 to 13 mm. (An earlier record by the junior author gives 10 to 12 mm. but the specimens are not available for checking and identifying the 10 mm. specimens.) Description. Characters as in H. mabouia with a tendency to have fewer scales everywhere. Thus supralabials 10 (in mabouia 10-14), infralabials 8-9 (in mabouia 9-10), praeanal pores 9-15 on either side (in mabouia 15-30), lamellae under median digits 5-6 (in mabouia 7-9). Specimens with regenerated tails often have a transversely enlarged median series of scales. The tubercles are more numerous than in East African specimens of H. mabouia but not more so than in certain West Indian examples which we have examined ; these tubercles also appear to be more sharply conical but this character again breaks down when a large series is examined. One gecko (M. C. Z. No. 1327) from Port au Prince, Haiti, has only six subdigital lamellae under the median digit. Coloration. As in H. mabouia and varying with the environment. Measurements of Type cf. Paratypes Length of head and body . 47 mm. The largest male (Bagilo) measures 104 mm. Length of tail regenerated 40 " ("i3 + .51) Length of hind limb 9 " Width of head 9 " Width of body 9 " Distribution. Occurs along the East African coast from Mombasa to Mozam- bique and in Tanganjaka Territory inland to Kilosa, which is also zoologically in the coastal belt. Its distribution will probably be found to coincide with that of the coconut palm and, like Phelsuma, it will doubtless have been introduced to the islands off the coast by dhows which are thatched with palm fronds. The undermentioned specimens in the collection of the Museum of Comparative Zoology have been examined and are named as paratypes. The whole series of H. mabouia from Africa, the West Indies, Central and South America have also been examined but, with the one exception cited, they agree in having 7 or more lamellae under the median digit, usually six pairs and one single teitninal lamella. 2 eggs (M. C. Z. 1S431) Morogoro, Tanganyika Territory, 11. iv. 18. 7 (M. C. Z. 18515-7) Kilosa, Tangan>-ika Territory, 1921-23. 2 (M. C. Z. 18524-5) Mbala, Tanganyika Territory, 26. ii. 23. 4 (M. C. Z. 22978-81) Morogoro, Tanganyika Territory, 1916-1917. 4 (M. C. Z. 22982-5) Lumbo, Portuguese East Africa, vii-viii. 18. 1 (M. C. Z. 22986) Frere Town, Kenya Colony, 31. xi. 22. 142 BARBOUR AND LOVERIDGE Hemidaciylus mabouia also occurs in all the above-mentioned localities and in no way affects the junior author's previous records except in the case of Mbala cited above. Breeding. About two-score eggs were collected among rotting palm-leaves and piles of palm-thatching near native villages on the outskirts of Dar es Salaam. It was the fact that there were no adult H. mabouia to be found among these leaves that caused me to wonder if the abundant 'half-grown H. mabouia' haunting these piles had not some connection with the eggs. After catching a dozen of them it was discovered that they were not half-grown but fully adult specimens, eggs being clearly visible through the semi-transparent abdominal walls of the females. These eggs agreed in size with those found among the palm-fronds but differed considerably from those found in houses or in fissures in the bark of big trees. Diet. A Tineid moth in the stomach of the Bagilo gecko. Habitat. Besides the commonest habitat already mentioned, at Bagilo I saw ten of these geckos basking in the company of three Striped Skinks along a crack in the face of a wall of rock. They were at a height of twelve feet from the ground and all were black as the cliff itself, which was covered with some dark lichen. Hemidactylus tropidolepis Mocquard Hemidaciylus tropidolepis Mocquard, 1888, Mem. Cent. Soc. Philom., p. 113. 2 (M. C. Z. 26649-50) Near Ifllindini, Kenya Colony, 3. v. 26. Relations. The material consists of a very young and an adult female of the same body length as Mocquard's male type. The adult was submitted to Mr. H. W. Parker for the favour of comparing it with types of related species in the British Museum. He writes: 'I have compared this specimen with, the types of various spp. as follows; it differs from (a) isolepis in the smaller dorsal scales which are somewhat keeled and in the presence of scattered enlarged tubercles on the occiput. (6) homoeolepis in the larger dorsals, keels on the dorsals and the larger scales on the snout. (c) Bunocnemis modestus in the slightly keeled and striated dorsals which are unequal in size. 'Boulenger writing of isolepis says " H. tropidolepis Mocquard . . . appears ... to approach Bunocnen).is modestus . . . but differs in the keeled dorsal scales." Your specimen seems to agree very well with Mocquard's description.' AFRICAN HERPETOLOGY 143 Nieden ' has united H. squamulatus Tornier of Tanganyika Territory with H. tropidolepis which was described from Somaliland. Variation. Three, instead of two, pairs of chin-shields, the third pair very small. The keeled scales on the back are apparently less numerous than in the type. Measurements. 9 total length 70 (37 + 33) mm. Young, 40 (23 + 17) mm. Diet. Six termites and a grasshopper's leg were recovered from the stomachs of these specimens. Habitat. Taken on a creeper-covered tree-stump in the middle of a native clearing. The stump was the centre of a rank growth of weeds which had to be cleared away before the geckos could be captured. LYGODACTYLtrs FiscHERi Boulenger Lygodactylus fischeri Boulenger, 1890, Proc. Zool. Soc. London, p. SO, PI. VIII, f. 1. 1 (M. C. Z. 24063) Vituri, Uluguru Mtns., 30. x. 26. 1 (M. C. Z. 24064) Mt. Bomoli, Amani, Usambara Mtns. 2. xii. 26. 8 and eggs (M. C. Z. 2406.5-73) Phillipshof, Western Usambara Mtns. 21-23. xii. 26. 9 (M. C. Z. 24074-8) Kwai, Western Usambara Mtns. 24. xii. 26. Relations. Muller ^ after examining a series of Cameroon specimens of L. fischeri Blgr, a species described from Sierra Leone, found their range of variation included L. conradti, which Matschie described from Derema, Usambara, in 1892. Nieden ^ agrees with this and adds L. fischeri scheffleri, which was described by Sternfeld from Kibwezi, Kenya Colony, to the synonymy. The examples from Mbunyi, Kenya Colony, and Bagilo, Uluguru Mountains, referred to scheffleri by the junior author ^ must be considered fischeri if the East African race is in- valid. None of the specimens taken outside the Usambara range approach the size of those captured in these mountains. If the types of scheffleri were full grown (24 and 27 nam. snout to vent) they were much smaller than the mountain geckos, which attain 42 mm. d", and 39 mm. $, snout to vent measurements at Phillipshof. Variation. The following variations of the Usambara series are compared with the type of L. conradti and the results fall within the range of variation of West Coast fischeri. ' Nieden, 1913, Mitt. Zool. Mus., Berlin, 7, p. 66. 2 Muller, 1910, Konig Bayr. Akad. Wiss. Munich, 24, p. 558. s Nieden, 1915, Mitt. Zool. Mus. Berlin, 7, pp. 67-68. ■" Loveridge, 1920, Proc. Zool. Soc. London, p. 136, and 1923, Proc. Zool. Soc. London, p. 940. 144 BARBOUR AND LOVERIDGE Supralabials 5-8 (four sides with 5, seven with (5, twenty-three with 7, three with 8; type was described with 8), sublabials 5-7 (two sides with 5, twenty-five with 6, nine with 7 ; type had 7) ; three specimens have the nostril between 3 (instead of 2) scales and the rostral and first labial. Tornier has stated that the male has 7 praeanal pores; this is the case with only two of the males in the present series, the remaining eight having 8 pores. The double row of transversely enlarged subcaudals is not always well defined ; on reproduced tails the scaling is very irregular; in one specimen there are narrow transversely dilated plates but usually the under surface is covered with scales of various sizes without definite arrange- ment. The tail of this gecko is large and swollen, differing in this respect, as well as in the readiness with which it is dropped, from those of other members of the genus. The reproduced half of the tail of the Vituri gecko has very irregularly enlarged plates, but the majority are single, not paired as in the intact basal portion. Coloration. The Usambara adults are practically the same as described by Matschie: those collected on banana plants at Phillipshof are yellowish-grey, those under logs and rocks at Kwai are blackish-grey. The males are bright yel- low or orange beneath (I omitted to take down exact color notes at the time), probably like the young Bomoli gecko of which I noted that it had the under side of the stomach and hind limbs bright yellow, and the under surface of the tail orange. Some adults have a very few scattered spots beneath. Two of the Kwai females have a broad, dark-edged, fawn stripe from the posterior border of the eye to the tail, where it merges with the stripe from the opposite flank; in this they agree with the Bomoli young one. Newly emerged geckos are spotted on throat and flanks exactly like the specimen from Vituri; above they are uniformly dark plumbeous. The coloration of the Vituri gecko agrees, in its spotted under surface, with the specimen recorded from Bagilo by the junior author. They show no yellow or orange below, nor did the IVIbunyi examples mentioned above. Measurements. The largest male measures 80 (39 -I- 51) mm., another with reproduced tail measures 42 mm. from snout to vent; the largest female 87 (39 + 48) mm. This is therefore 37 mm. longer than the type of conradti, which was presumably a female. It will be seen that in the matter of size this Ly{;o- dactijlus stands alone among East African members of the genus. On hatching, a young gecko measures 28 (15 + 13) mm. The Vituri male measures 59 (34 -|- 25) mm., but its tail is regenerated. Breeding. An egg, measuring 7x5 mm., was found beneath a log on Bomoli, 20. xi. 26. At Phillipshof I was walking up the road in the evening of 21. xii. 2(j. AFRICAN HERPETOLOGY 145 when I noticed about half-a-dozen pairs of eggs lying on the bare soil of the some- what broken-away bank at the side of the road. I speculated as to whether these had been laid in trees above and fallen, yet none were broken and none were to be seen except at the one restricted spot. Then I examined the roots, exposed by recent rain, of a large tuft of grass on the edge of the bank ; here to my aston- ishment T unearthed forty-five pairs of eggs, making a total of fifty-one clutches. They were all apparently fresh and recently laid ; how came fifty-one lizards to this spot to lay? Though I searched the vicinity, no geckos were to be found ; doubtless they were concealed in hollow tree-trunks. The nearest were collected on banana plants two miles away! Diet. Ants, and a beetle, cockroach, and spider were discernible among the stomach contents of the three geckos examined. Parasites. Red acarine parasites are of frequent occurrence in the anal region and about the praeanal pores. Habitat. I caught the eight Phillipshof geckos among the basal stalks of the leaves of wild bananas growing on an open hillside. They displayed great agility in springing from leaf to leaf. The Kwai series was taken by my native collector 'under logs and rocks at the edge of the forest.' The Vituri gecko was taken on a tree-stump at the forest edge. Lygodactylus grotei Sternfeld Lygodactylus grotei Sternfeld, 1911, Sitzber. Ges. Naturf. Freunde Berlin, p. 24.5. Eggs (M. C. Z. 240.56) Mkarazi, Uluguru Mtns., 23. x. 26. 1 (M. C. Z. 24055) Vituri, Uluguru Mtns., 3D. x. 26. 10 and eggs (M. C. Z. 24057-60) Dar es Salaam, 9. xi. 26. Eggs (iM. C. Z. 24061) Tanga, 15. xi. 26. 1 (M. C. Z. 24062) Burabuli, Usambara Mtns., 14. xii. 26. Coloration. There is no mention of the following points in the original de- scription. The throats of both the Dar es Salaam males are more or less dark, with a hght a -shaped or fl -shaped marking in the centre. A Hght vertebral line is frequently, but by no means always, present. Measurements. The largest male measures 108 f54 -f 54) mm. ; the largest female 65 (34 -|- 31) mm. Breeding. Two eggs were found inside the outer dried leaf of a banana stem, two more beneath a large stone in the same 'field'; that the second pair are the eggs of L. grotei is only by inference, as the situation seems improbable for L. p. picturatus. They contained embryos on the point of hatching. A series of eggs collected at Dar es Salaam on 9. xi. 20 were taken among piles of coconut palm fronds adjacent to banana plants on which the adults were 146 BARBOUR AND LOVERIDGE collected. One or two pairs of eggs were also found among the withered outer leaves of the bananas. Four pairs of eggs were taken at Tanga on 15. xi. 26. Some of these measured 7 X 5.5 mm. and are thus slightly larger than those of L. p. picturatus hitherto recorded. One pair hatched out the same day they were collected; the emerging geckos measured 25.5 (13.5 + 12) nun. Diet. I observed one gecko stalk, and spring upon, a cockroach which was on the main stem of a shrub occupied by the lizard. Habitat. Besides the Dar es Salaam geckos, the Vituri specimen was also taken on a banana stem. The Bumbuli female was basking in the rays of the setting sun low down on the stem of a shrub in the Mission grounds. It may be remarked that both places were well away from the forest and at no great altitude. Lygodactylus pictukatus picturatus (Peters) Hemidadylus picturatvs Peters, 1870, Monatsber. Akad. Wiss. Berlin, p. 455. 1 (M. C. Z. 24050) Nyange, Uluguru, Mtns. 1. x. 26. 1 (M. C. Z. 24051) near Mkarazi, Uluguru Mtns., 23. x. 26. 6 (M. C. Z. 24052) Dar es Salaam, 4. xi. 26. 1 (M. C. Z. 24053) Bagamoyo, 11. xi. 26. 2 (M. C. Z. 24054) Tanga, 15. xi. 26. Variation. The female from Mkarazi has 7 supralabials on both sides; hitherto 5-6 has been the usual number. Measurements. The Nyange male far surpasses in size any other example of this species which we have seen; his big head reminds one of old males of Agama atricollis or Lacerta ocellata. He measures 78 (41 + 37) mm. Enemies. One was captured and swallowed by a Green Snake (Chlorophis neglectus) at Dar es Salaam. Phelsuma laticauda (Boettger) Pachydactyliis laticauda Boettger, 1880, Zool. Anz. 3, p. 280. 1 (M. C. Z. 24079) Dar es Salaam, 4. xi. 26. 5 (M. C. Z. 24080-4) Bagamoyo, 11. xi. 26. Variation. These differ from the original description in the following points. Diameter of eye contained thrice in length from snout, taken from anterior border of the eye; the rostral has a slight median cleft, nostril pierced between the first labial and three scales (in one specimen 2 scales), occasionally the rostral appears to border it also; 8-11 upper labials; 8-10 lower labials; 7-10 chin-shields (in one lizard an apical chin-shield is formed by division from the rostral) ; 1-3 scales separate the supranasals; 9-15 femoral pores on each side. Tornier ' and the > Tomier, 1900, Zool. Jahrb. Syst., p. 588. AFRICAN HERPETOLOGY 147 junior author ' have previously commented on similar variations in Zanzibar examples. Measurements. Largest male measures 137 (65 + 72) mm.; the largest female 117 (54 + 63) mm. Diet. Ants and beetles in the stomachs of two specimens examined. Parasites. Red acarine parasites near the anus in one or two geckos. Habitat. They live in the tops of the palms and are well known to the climbers for coconuts. The Dar es Salaam specimen was taken in long grass, doubtless moving from one tree to another. AGAMIDAE Agama mossambica mossambica Peters Agama mossambica Peters, 1854, Monatsber. Akad. Wiss. Berlin, p. 616. 1 (M. C. Z. 24128) Tumvi, foot of Uluguru Mtns., 1. xi. 26. Konghore in Kikami. Coloration, and correctio?i. The junior author, having reexamined two of the Lumbo, Mozambique specimens (M. C. Z. 13620-1) attributed by him to colono- rum,- finds that they are juvenile mossambica. The colour note, and last two para- graphs, therefore, refer to mossambica and not colonorum. Distribution. It was interesting to note the complete absence of this common and wide-ranging form in the mountains, where it was replaced by a smaller race. Not one was seen during the seven weeks spent in the Uluguru Range, but an hour's walk down from the camp at Vituri (where the smaller race was common) brought me to Tumvi, where mossambica typica graced a great many of the tree- trunks. The male, collected for record, was taken on a baobab. Agama mossambica Montana subsp. nov. 5 (M. C. Z. 24086-90) near Bagilo, Uluguru Mtns., 24. ix. 26. 3 (M. C. Z. 24091-93) Nyange, Uluguru Mtns., 11. x. 26. 3 (M. C. Z. 24094-96) Mkuyuni, Uluguru Mtns., 18. x. 26. 1 (M. C. Z. 24097) Tawa, Uluguru Mtns., 20. x. 26. a (M. C. Z. 24098-102) Vituri, Uluguru Mtns., 27-30. x. 26. 15 (M. C. Z. 24103-118) Amani, Usambara Mtns., 19-30. xi. 26. 1 (M. C. Z. 24120) Ivizerui, Usambara Mtns., 8. xii. 26. 1 (M. C. Z. 24121) Mt. Lutindi, Usambara Mtns., 10. xii. 26. 6 (M. C. Z. 24122-7) Bumbuli, Usambara Mtns., 14. xii. 26. Also seen at Muheza and Derema. AU agamas are called konghore in Kikami, kokolwe in Kisumbara. ' Loveridge, 1925, Proc. Zool. Soc. Lond., p. 72. * Loveridge, 1920, Proc. Zool. Soe. Lond., p. 140. 148 BARBOUR AND LOVERIDGE Type. No. 24189. Museum of Comparative Zoology. Sex ^. From rocks in a native shamba one mile below Bagilo on the Bagilo-Nyange path, Uluguru Mountains, Tanganyika Territory. Collected by A. Loveridge September 24th, 1926. Paratypes. Thirty-nine specimens as listed above. Affinities. So closely related to typical A. mossambica that it can apparently be separated only on size and colour. All scale-counts appear to overlap. Diagnosis. The total length of Peters' male type of mossambica 330 mm. Average of three topotypes (M. C. Z. 1308) though half-tail of one is missing 300 Length of Tumvi specimen (M. C. Z. 2412S) listed above 300 Length of largest male moniana, the type from Bagilo 264 Average of the eighteen largest males from L'luguru and Usambara 245 Average of thirty-nine adults from Uluguru and LIsambara 229 It will thus be seen that adults of the two forms may readily be distinguished by the following key : Total length of adult over 300 mm M. m. mossambica Total length of adult under 280 mm M. m. montana The typical form inhabits the coastal belt up to 2,000 feet and the mountain race from 3,000 to 5,000; it is probable that intermediates will be found between these altitudes. In appearance those at Muheza, the lowest altitude at which any were seen, looked hke moniana; unfortunately none were collected. Two examples from Kilosa (M. C. Z. 18273, 18564) mentioned in an earlier paper are referable to the mountain form. Variation. Praeanal pores in eighteen males range from 8 to 12. Coloration of Type &. Above, olive-brown, most of top and sides of head pale blue ; an indistinct band of brick-red across throat from side to side of neck, where it is conspicuous; an irregular in width, but continuous, blue vertebral stripe broadens considerably on the base of the tail but is replaced by brown half-way along the tail ; traces of some greenish-blue on the upper arm and thighs, other- wise limbs brown. Below, anterior portion of throat rich ultramarine blue, grad- ually darkening to form a blackish patch reaching to gular band; rest of under parts, including limbs and tail, creamy-white, a slight pinkish tinge on the breast between the fore arms. Coloration of Paratypes. The appearance in life was so very varied and hand- some that notes were made of all five adult Bagilo agamas taken on the same day. cT. Above, as in type, but the blue much less distinct and showing many black scales scattered on the head, back and tail; the gular band of type inter- rupted below, leaving two large splashes of brick-red, one on each side of the AFRICAN HERPETOLOGY 149 neck. Below, jaws bright blue, the throat between them pinkish; the black patch of type is represented by a smaller pale-green patch; chest to abdomen shows more pink than in type — otherwise they are similar. (a) 9. Above, mainly golden brown; snout, canthus, supraocular region and around eyes are black; a band of greenish-blue across top of head in front of eyes; the whole of the back part and sides of head, including upper and lower labials, greenish-blue. Below, throat whitish washed with blue, of which there is a small patch in front of collar; beneath chest, belly, limbs and tail, dirty white. (6) 9. Above, blue on head but not so bright as in 'a'; no red on back, its place being taken by pale yellowish-brown. Beneath throat, slatey grey sparsely flecked with white. (c) 9. Above and below, identical with 'a' excepting the throat, which is white, handsomely vermiculated with dark blue and with a dark blue patch pos- teriorly. Three Vituri females when brought in were brown but on being chloroformed turned red, with blue heads. Measurements of Type. Length of head and body 96 mm. Breadth of head 18 mm. Length of tail 168 " Hind Hmb 58 " Length of head 2.5 " Longest (4th) toe 25 " Measiiremetits. The largest male (Bagilo type) measures 264 (96 -t- 168) mm. ; the largest female (Amani) 274 (93 + 181) mm. Average total length of eighteen adult males is 245 mm.; average total length of twenty-one females is 212 mm. The smallest specimen, taken at Amani on 29. xi. 26., measures 76 (30 + 46) mm. Breeding. As I ran round a rock, hoping to surprise some of these shy agamas, I suddenly came upon a red-backed female with her back arched in a most peculiar manner. Approaching her was a male (type in above colour description). She promptly disappeared on catching sight of me and I shot the male, having little doubt that I had broken in upon a courtship scene. Bagilo, 24. ix. 26. The female 'a' held 8 eggs containing traces of embryos; these measured 15 X 9 mm. In 'c' were 6 round developing eggs measuring 8 mm. in diameter. No developing ova at all in 'b.' 24. ix. 26. Three females brought in at Vituri on 27. x. 26 held 8, 8, and 10 eggs respec- tively; two lots were much enlarged, nearly ready for laying; the other lot was only half developed. Diet. The stomachs of all five Bagilo agamas were crammed with ants; in addition two individuals had each eaten a grasshopper. Many of the Amani lizards were examined and all seen had fed on ants. 150 BARBOUR AND LOVERIDGE Parasites. Red acarine parasites (Trombidid) insert themselves beneath the ventral scales, giving the latter an eruptive appearance; they were present on specimens from nearly all localities. Nematodes {Physaloptera amaniensis sp. n.) were taken from the stomach of an Amani agama. A grub-like creature was found in the intestines of a Bagilo lizard. Habitat. At Bagilo these agamas frequent rocks of moderate size which are scattered about the native gardens, and waste lands thickly overgrown with grass and bushes. If disturbed when basking, they almost invariably run to the farther side of the rock and down among the matted vegetation at its base. This habitat is shared by Mabuya varia which seemed to occur in about the ratio of three skinks to one agama. Two males from Nyange were shot a couple of miles from camp among the rocks fringing the stream in the valley bottom ; one was apparently drinking at the water's edge when disturbed. The species is scarce at Nyange. An adult male was shot at Tawa while basking, at a height of four feet from the ground, on the vertical trunk of a Javan silk-cotton tree. All the specimens obtained in the Usambaras, except one from Lutindi, were found on tree-trunks outside the rain-forest. A female at Lutindi was found beneath a log where she had quite possibly re- tired to lay her eggs. Agama colonorxjm usambarae sub. sp. n. Plate 2, fig. 1 14 (M. C. Z. 24129-39) Soni, near Lushoto, Usambara Mtns., IS. xi. 26 and 31. xii. 26. Kokolwe in Kisumbara. Type. No. 24129. Museum of Comparative Zoology. Adult & taken on rocks at Soni, close to the Half-way House on the Mombo-Lushoto Road, Western Usambara Mountains, Tanganyika Territory. Collected by A. Loveridge, No- vember 18th, 1926. Paratypes. Nos. 24130 to 24139. Six males, six females and a young one all from the same locality. 31. xii. 26. Also three males, two females and three young from Mnazi, northern side of Usambara ^lountains. (Field Aluseum of Natural History Nos. 12280-7.) Affinities. These agamas have been compared with thirty-two West African colonorum from Senegal, Liberia, Cameroon, Gaboon and Congo, from all of which they consistently differ in the AFRICAN HERPETOLOGY 151 (1) Much broader head of old Usambara males. (2) More depressed body, though not so depressed as in lionotus. (3) Median Une of the back smoother, having less of a keel than in typical colononim. (4) Dorsal scales only very slightly mucronate, not terminating in such long spines as they do in colonorum. (5) Crimson-lake colouring of the heads of males as against the brick-red of Liberian specimens as figured by Sir Harry Johnston in Liberia, Vol. II, coloured plate facing page 814. See also Discussion below. Discussion. Schmidt ' has pointed out that Lower Congo colonorum have a higher number of canthal and supraciliary scales than those from the West Coast. West Coast 6-8. Lower Congo 7-9. An examination of the largest series in the Museum of Comparative Zoology, viz., 11 Liberian, 14 Cameroon and 14 Usam- bara, results in an indi\'idual range for each of 6-9, with an individual average of 7; the Field Museimi specimens range from 9 to 10, with an average slightly above 9, so that this character is of no assistance. Schmidt adds that the average number of mid-body scale-rows also seem somewhat higher in Lower Congo colonorum than in forest specimens : 64 versus 70, with 68 in Sudanese specimens. It is to be noted that the highest number of scale-rows are on females on the West Coast, on males in Usambara, so the proportion of these in a series may result in a higher average. After arrang- ing the Museum of Comparative Zoology material geographically from West to East, one gets the following result: 3 Senegal cohnorum range: 58-64 with average of 60. 11 Liberian " ' 60-64 " " 62. 14 Cameroon " ' 72-84 " " " 75. 2 Gaboon ' 61-64 " " 62. 1 Congo " ' 65 " '' 65. 3 Blue NUe ' 72-«4 " " " 76. 22 Usambara " ' 70-80 " " " 75. (The last hne includes Field Museum material; without, it was 76.) He also refers to Stejneger, whose eastern colonorum has fewer spines in the nuchal crest (10-11 versus 12-15), a point not borne out by Schmidt's Congo series of 134 examples: Average of 8 adult Liberian specimens is 12 (range 11-17). Average of 10 adult Cameroon specimens is 15 (range 8-25). Average of 10 adult Usambara specimens is 11 (range 10-13). Average of 8 Field Museum Usambara specimens is 13 (range 7-15) ; which lends but sUght support to Stejneger 's suggestion. ' Schmidt, 1919, Bull. Am. Mus. Nat. Hist., 39, Art. II, p. 473. 152 BARBOUR AND LOVER IDGE A single Angolan specimen has 22. This individual has dorsal scales much like the Usambara agamas and its ventral scales are smooth like those of our Congo specimen and the Usambara series. On the other hand Senegal, Liberian, Cameroon (young excepted) and Gaboon colonorum have faintly keeled ventrals, of which there is not the slightest trace in the Usambara series. Three specimens collected by Dr. John C. Phillips at Rhino Camp agree with colonorum in not being depressed, but with Usambara specimens in having smooth belly scales and the keeling and spinosity of the dorsals. It might be pointed out that the character of the relative length of the longest spines in the region of the ear as compared with the diameter of the tympanum is of little use as so much variation occurs; similarly the elongate scale on the middle of the snout is but little help. Seven of the Usambara specimens have 1, nine have 2, and two have 3 such elongated scales. Description. Longest spines about ear are less than half the diameter of the tympanum in type but range from L5 to 3 mm. in length in the paratype series, while the diameter of the tympanum ranges from 3.5 to 6 mm. in the same series and is 4 mm. in the type. Mid-body scale-rows in type are 70 but range up to 80 in the nine paratype males, with an average of 76. (In the eight female paratypes the range is also 70 to 80, with an average of 74.) There are 12 femoral pores in the type but the whole series shows a range from 10 to 14, with an average of 12. Dorsal scale-rows decrease in size from vertebral line, where they are irreg- ular in size, to the sides, where they change abruptly to very large, mucronate, but keelless ventrals that lose their spines and decrease in size towards the median line of the belly. Coloration of Type &. Above, crimson-lake on head, slightly orange on nape; back, including fore and hind limbs, brilliant ultramarine except along the verte- bral line, where it is pale bluish-white; some hght spots on back (in some fully adult specimens these are lacking) ; basal portion of tail bright ultramarine, there- after banded white and ultramarine, the latter darkening towards the tip, where the bands are almost brown. Below, bright crimson-lake on chin and throat, with four more or less a -shaped, dark -grey lines following the outline of the lower jaw (in old specimens these hues may be indistinct) ; breast, including under side of fore limbs to wrist and from gular fold to an imaginary straight line between the axilla, ultramarine; rest of under surface, including soles of fore feet but excluding base of tail, white (in one paratype the colouring of breast and belly is reversed, the former being white and the latter ultramarine; in adults the whole of the imder surface, except for a median praeanal patch on the belly, is bright ultra- AFRICAN HERPETOLOGY 153 marine; the praeanal region, soles of fore and hind feet, under side of thighs and tail in these individuals is buff) ; base of tail bluish. Coloration of 9 . Above, pale coppery-brown on head and body, limbs and tail grey; pale-green spots edged with black on head and neck; well-defined trans- verse bars of sepia on back and tail, and spots of the same colour on limbs; a con- spicuous band of brick-red, more or less edged with black, above fore limb and extending half an inch along the side; a similar but not quite such a long band on flank in front of hind limb. Below, throat creamy with three dusky bands fol- lowing its outhne and some rather indistinct spots or blotches in the centre; rest of under-surface, including Umbs, pure white. Measurements. Type o". Snout to vent 115 mm. Tail 193 mm. The nine paratype males range from 110 to 135 mm. in length from snout to vent and from 175 to 220 mm. in length of uninjured tails. The greatest total length from snout to tip of tail is 350 (130 -|- 220) mm. The eight paratype females range from 79 to 105 mm. in length from snout to vent and from 135 to 175 mm. in length of uninjured tails. The greatest total length from snout to tip of tail is 280 (105 -|- 175) mm. A young agama taken on 21. xii. 36 measures 91 (36 + 55) mm. Breeding. Four of the six (M. C. Z.) females held large, well-developed eggs. Size of 9 (head to vent) lOo mm. No. of eggs 6, measuring 19 X 11 mm. 104 " 7 " 20 X 13 94 " 7 " 22 X 13 « 92 " 6 " 20 X 13 They were all taken on 31. xii. 26, rather indicating a definite breeding season. Diet. Unfortunately the native collector removed the stomachs of those he procured. The first pair obtained by myself held a great many ants' heads, several beetle elytra and a single grasshopper. I w as unable to distinguish any vegetable matter. [Agama atricollis Smith] Agama atricollis Smith, 1849, 'Illus. Zool. S. .\fr.,' 3, Appendix, p. 14. Monsieur F. Angel in a recent paper (1925, 'Voyage de Ch. Alluaud et R. Jeannel en Afrique Orientale [1911-1912],' p. 13), records seven agamas from Kisumu, Kenya Colony, as A. cyanogaster Riippell, a species usually regarded as ranging through Arabia and Abyssinia. The junior author, having previously collected A. atricollis at Kisumu, was anxious to examine these specimens, which was possible through the courtesy of Monsieur Angel. Boulenger's' key character of Nostril pierced on the canthus rostialis atricollis Nostril pierced below the canthus rostralis cyanogaster ' Boulenger, 1885, 'Cat. Liz. Brit. Mus.,' p. 337. 154 BARBOUR AND LOVERIDGE breaks down in practice and it was this that probably led ^^locquard, who made the identifications, astray. A juvenile example of cyanogaster from Harar, Abys- sinia, in the Museum of Comparative Zoology (No. 8064), has the nostril placed in precisely the same position as all our series of atricollis from various East African locaUties. Some atricollis specimens collected at Lake Tanganyika by Mr. Nutt (B. M. No. 76.5.14.10-13) were examined by the junior author when at the British Museum in February. These have the nostril pierced below the canthus, i.e., in the same scale and precisely the same position as in typical Abyssinian cyano- gaster. The nostril in atricollis is round, while in adult cyanogaster it is more often pear-shaped, the apex pointing backwards or upwards. The range of cyanogaster as given by Angel (undoubtedly based on Moc- quard's identifications, the alleged Mozambique example being in the Paris Museum) , as extending to Kenya Colony and Mozambique, can hardly be correct. VARANIDAE Vakanus niloticus (Linnaeus) Lacerta nilotica Linnaeus, 1766, Syst. Nat., ed. 12, p. 369. 1 (M. C. Z. 24140) Mogogoni Swamp near Dar es Salaam, 9. xi. 26. 1 (M. C. Z. 24141) Sigi near Amani, Usambara Mtns., 29. xi. 26. Nhengi (ICikami); kerigi (Kiswahili); mhvlu (Kisumbara). Measurements. The female from Sigi measured 4 feet 8J inches (600-850 mm.) in length. Breeding. She held forty-three eggs, each measuring 50 x 35 mm. (approxi- mately 2 X 1| inches). Diet. In her stomach were two large scarabs and a crab; the half-grown male had also eaten a crab. Parasites. The male was infected with numerous nematodes (Tanqna tiara) and a cestode (Duthiersia fimbriata) ; the female had no visible internal para- sites but carried many ticks (Aponomma exornatum). Distribution. This monitor Hzard occurs in the Uluguru Mountains as high as 6,000 feet (Bagilo), but is not common there, as none were seen anywhere; they are plentiful at Morogoro at the foot of the range. AFRICAN HERPETOLOGY 155 LACERTIDAE Eremias spekii spekii Giinther Eremias spekii Gunther, 1872, Ann. Mag. Nat. Hist. (4) 9, p. 381. 1 (M. C. Z. 22994) Mainland opposite Kilindini, Kenya Colony, 3. v. 26. 1 (M. C. Z. 24142) Tumvi, foot of Uluguru Mtns., 1. xi. 26. 1 (M. C. Z. 2414.3) Tanga, 15. xi. 26. 1 (M. C. Z. 24144) Foot of Mt. Lutindi, Usambara Mtns., 13. xii. 26. Distribntion. Apparently it does not occur in the mountains though abun- dant at the foot of both ranges. The Tumvi specimen was found fast asleep at 8 A.M. on the sunny Vituri-Morogoro Road. HoLASPis GTJENTHERi Gray Holaspis guentheri Gray, 1863, Proc. Zool. Soc. London, p. 153, PI. XX, fig. 1. 1 (M. C. Z. 24145) Amani, Usambara Mtns., 20. xi. 26. 1 (M. C. Z. 24146) Derema, Usambara Mtns., 29. xi. 26. Chungida in Kisumbara. Distribution. These specimens are topotypes of Werner's H. g. lacvis which cannot be considered vaHd. The distribution of this forest form is remarkable, as it occurs at Bukoba, Zomba, Nyasaland and Angola. I think it very prob- able that it will some day be found in the Uluguru Mountains. Habitat. It is extremely rare at Amani and impossible to obtain except by shooting. I shot one on a eucalyptus (introduced) tree near the top of Mt. Bomoli, and two others were seen on eucalyptus trees. The custodian of Amani informed me he had seen them only on eucalyptus trees but I found the Derema specimen sunning itself near the base of a giant tree on the forest-edge. Their depressed outline renders them inconspicuous and I quite agree with Schmidt that the curious tail processes are an aid to climbing. They may be compared with the caudal scales of the flying squirrel (Anomaliirus m-ientalis) which haunts the same forests. 156 BARBOUR AND LOVERIDGE SCINCIDAE AIabuya comorensis (Peters) Euprepes comorensis Peters, 1854, Monatsber. Akad. Wiss. Berlin, p. 019. Mabuia comorensis var. infralineata Boettger, 1913, in Voeltzkow, 'Reise in Ostafrica,' 3, p. 239. (Europa Id., Mozambique.) 2 (M. C. Z. 24147-8) Bagilo, Uluguru Mtns., 20-23. ix. 26. 1 (M. C. Z. 24149) Mkarazi, Uluguru Mtns., 22. x. 26. 1 (M. C. Z. 24150) Dar es Salaam, 4. xi. 26. 25 (M. C. Z. 24151-70) Amani, Usambara Mtns., xi. 26. 3 (M. C. Z. 24171-3) Burabuli, Usambara Mtns., 14. xii. 26. Ghondo in Kisumbara. Variation. Scale-rows 34-38; these were carefully checked and recounted. The supranasals may, or may not, be in contact behind the rostral, in exactly equal proportions; the anterior loreal is usually in contact with the first labial though it very often is not; the praefrontals also may, or may not, be in contact. The species has a pretty wide range of variation, as will be seen from the descrip- tion in the 'Catalogue of Lizards in the British Museum.' Boettger, when describing his colour variant infralineata, states that he found no stinictural difference between it and typical comorensis. Its claim to recog- nition lies in a series of fine black Hnes (formed by a speck on each scale) from snout to, and on, the base of the tail. In our series such lines are absent or pres- ent only on throat and tail. It is, perhaps, a matter of personal opinion whether insular colour variants should receive recognition. Coloration. Bagilo 9 . Above, nut-brown sparingly speckled with cream, sides darker brown with indications of a black band separating sides from back, sides also freckled with cream; rim around eye light yellow; throat greyish-white with brown speckling sharply divided from the under side of neck and body to anus, all of which is clear bright yellow. Mkarazi d. Almost identical with that of the brilliantly coloured Frere Town specimen described in Proc. Zool. Soc. Lond., 1923, p. 956, but in this present specimen the cream-coloured subocular stripe is broken up into spots and there is no blue on the throat, which is shiny white. Testes large. Measurements. Largest male measured 237 (91 -f- 146) mm., but one with a reproduced tail had a head and body length of 98 mm. ; this patriarchal individual measured 20 mm. across its head and 25 mm. across mid-body; with age the head develops out of proportion to the body. The largest female measured 258 (98 -f- 160) mm. but was much more slender than the male. Both were from Amani. AFRICAN HERPETOLOGY 157 Breeding. Eleven eggs measuring 14 X 8 mm. in Bagilo 9 taken on 20. ix. 26. ; undeveloped in another female killed three days later at same place. Nine, eleven and twelve eggs in three females collected at Amani on 24. xi. 26. Diet. Large cricket in Bagilo skink. Large locust and a large praying mantis and a mollusk in the Mkarazi specimen. Cockroach in Dar es Salaam reptile. Five Amani skinks examined had grasshoppers and beetles in their stomachs ; in addition one had a caterpillar. Parasites. Red acarine parasites in ears of Bagilo skink. Nematode in stomach of an Amani specimen taken 25. xi. 26. Enemies. One recovered from the stomach of a snake {Lycophidion capense) at Amani. Distribution. The Comoro Id. Skink is apparently no exception to the gen- eral rule that most scarce creatures are generally common in some locality. Its mainland records have been very few, for only single specimens are usually met with, but at Amani it is a tolerably common reptile. Habitat. One Bagilo specimen was shot while basking on the edge of a fissure in the face of a chff ; it had for company many Hemidactylus persimilis. The Mkarazi male was basking on a tree stump in the dry forest near the Mvua River. The Dar es Salaam female was beneath a pile of coconut husks close to the incinerator on the sea-front near the golf course. At Muheza one was seen on the trunk of a rubber-tree in a plantation. The rain-forest is apparently the ideal habitat, for at Amani it was common in the forest between 8 and 12 a.m. These skinks emerge from the fringe of swept-up leaves which border each path, to bask in the sunshine, but so quick are they in their movements that few but naturaUsts would see them ; a rustle of the leaves is usually the only sign of their presence. At other times, and indeed at all hours of the day, they may be ap- prehended beneath, or within, the fallen tree-trunks that are scattered about the arboretum. Fairly often they are seen on the trunks of growing trees but not so frequently as on the banks flanking the paths. A pair were seen on tree-trunks at Mt. Lutindi. Mabuya maculilabris (Gray) Euprepis maculilabris Gray, 1845, Cat. Lizards, p. U-i. (West .Africa.) Mabma bovlmigeri Sternfeld, 1911, Sitz. Ges. Naturf., Berlin, p. 248. (Malvonde Highlands, S. E. Tang. Terr.) Mabuia maculilabris major Sternfeld, 1912, Wiss. Ergeb. Deutsch Zent. Afrika Exp. 1907-08, 4, 2, p. 232. (Central African Lake Region.) Mabuia maculilabris kii'idjwiensis Sternfeld, 1912, I.e. p. 233. (Kwidjwi Is., Lake Kivu.) Mahuia maculilabris wauensis Sternfeld, 1912, I.e. p. 2.33. (Wau Is., Lake Kivu.) 158 BARBOUR AND LOVERIDGE Mabuia macidilabns schubutzi Sternfeld, 1912, I.e. p. 233. PI. VI. fig. 3. (Fort Beni and Kassenje, Bel- gian Congo.) Mabuia maculilahris graueri Sternfeld, 1912, I.e. p. 234. PI. V'l, fig. 4. (Congo belt.) Mabuia maculilabris rohibechi Sternfeld. 1912, I.e., p. 234. (Nguru Mtns., and Langenbcrg in East Cen- tral and South Western, T. T.) Mabuia albot'ieniatu Boettger, 1913, in \'oeltzkow, "Reise in Ostafrika," 3, p, 350. (Pemba Id.) 1 (M. C. Z. 24174) Vituri, Uluguru Mtns., 30. x. 26. Material. In addition to the above we have used for the purposes of this in- vestigation forty-two examples from fifteen localities ranging from Sao Thome Id. off the West Coast to Pemba Id. off the East Coast, including a series from the Central African Lake region just brought back by Dr. Joseph Bequaert. Variation. Schmidt ' has already referred M. m. major and M. m. bergeri Sternfeld (Dufile, Sudan) to the synonymy of M. maculilabris, and a careful perusal of Sternfeld's descriptions, which are brief and based almost entirely on colour variations, forces us to the conclusion that in the present state of our knowledge they are not entitled to recognition. M. boulengeri is stated by its describer to be ' very near maculilabris ' from which it differs principally in the high number of keels on the scales (7-9 with here and there 10 or 11), the smaller number of supraciliaries (4, the tj^je of maculilabris had 5), the greater body length (toes of the adpressed hind limb scarcely touching fingers of the fore limb), and the colouring. Boettger has proposed to differentiate Pemba Id. specimens on the basis of 4 supraciliaries and the presence of a distinct light lateral line. The lateral line can be matched by a specimen taken on board ship near Lulonga, Belgian Congo (M.C.Z. No. 24812) ; except for the Pemba examples this is the only one of our long series in which it occurs. Below we have placed in tabular form particulars of our comparative material. 'West Africa' was given by Gray as the type locality of maculilabris, but it will be observed that a Sao Thome lizard had from 7 to 1 1 keels, as also one Cameroon lizard ; A Cameroon skink had 4 supraciliaries, Schmidt has recorded from 3 to 6 in his Congo series of 143 specimens, there are two with 3 or 4 among our Congo specimens, another from S. E. Ruwenzori has 4. It is true that the majority of Tanganyika Territorj^ lizards have the lower number, three of the four skinks have 4, and one of the two Pemba Id. skinks has 4, the other two have 5. The longer body of Tanganyika skinks is certainly noticeable when compared with Central African skinks, but many Cameroon reptiles have the digits only just meeting. In coloration a Mkindo River, T. T., reptile agrees very closely with the type of boulengeri but has 5 supraciliaries and only 7 carinations on its scales, ' Schmidt, 1919, Bull. .Vm. Mus. Nat. Hist., 39, Art. II, pp. o25-531. AFRICAN HERPETOLOGY 159 while the toes of the adpressed hind Hmb reach to the wrist of the back-pressed fore Umb. It seems therefore that the single specimen of boulengeri on which the description was based had a combination of characters which do not usually occur in one individual, but are common characters of maculilabris. It appears that while M. maculilabris is a common skink in West Africa — we have recently received forty-one examples from one locality in the Cameroons — it is a scarce reptile in East Africa, for the junior author secured only fourteen in eight years. There are marked family hkenesses in specimens coming from any given locaUty, both in colour (thus, while those from the Central Lake region (major) are strongly spotted with light and dark, which is also the case with the Sao Thome reptile. East African skinks are bronze with very few spots but can be matched by Cameroon skinks) and in other characters (there is a noticeable decrease in mid-body scale-rows, and possibly supraciharies also, as one proceeds from West to East) , so that it appears as if maculilabris is undergoing evolutionary differentiation, but these variations have not progressed far enough, or become sufficiently standardized, to merit racial recognition. LOCALITIES AERANGED FROM EAST TO WEST AFRICA Number Greatest Longest Mid-body Xo. of of bead and body tail scale- supra- LocALiTY skinks lengths in mm. in mm. rows ciliaries Pemba Id 2 62 123 30 4-5 Tawa, Tanganyika Territory .... 1 76 115 30 4 Mkindo River, Tanganyika Terr. 1 76 184 30 5 Kipera, Tanganjika Territory ... 1 S2 168 30 4 Kilosa, Tanganyika Territory .... 1 85 203 30 4 Lulonga, Belgian Congo 1 73 112 30 7 Rutshuru, Belgian Congo 5 82 123 32 5 Mt. Mokia, s.e. Ruwenzori 2 85 174 30-.32 4-6 Mutea, White NUe 2 84 180 32-34 5 Bumba, Belgian Congo 6 85 153 32-34 4-6 Metet, Cameroon 4 85 103 34-36 .5-tj Lolodorf, Cameroon . 8 88 135 33-36 4-6 Sakbayeme, Cameroon 8 83 156 34r-Z6 5 Sao Thome Id 1 84 95 32 5 Keels on scales 5 7 7 7 7 3-4 5-8 o-i 5-7 5-11 5-7 7-11 Mabuya megalura (Peters) EuprepHs (Mabuia) megalurus Peters, 1878, Monatsber. Akad. Wiss. Berlin, p. 204, PI. II, fig. 4. 1 (M. C. Z. 24175) Dar es Salaam, 4. xi. 26. 1 (M. C. Z. 24176) Amani, Usambara Mtns., 30. .\i. 26. Coloration. The Amani female has nine distinct longitudinal lines from throat to vent, while the Dar es Salaam specimen is normal below, i.e., iridescent white. Measurements. The male is unusually large, 223 (56 + 167) mm. 160 BARBOUR AND LOVERIDGE Mabuya varia varia (Peters) Euprepes (Euprepis) rarius Peters, 1867, Monatsber. Akad. Wiss. Berlin, p. 20. Mahuiii varia var. Inngiloba Methuen & Hewitt, 1913, Ann. Transvaal Mus., 4, 1914, p. 142. (Nama- qualand.) 4 (IM. C. Z. 241S1) Bagilo, Uluguru Mtns., 23. ix. 26. li (M. C. Z. 24182) Nyange, Uluguru Mtns., 8. x. 26. 1 (M. C. Z. 24183) Misalae, Usambara Mtns., 7. xii. 20. 3 (M. C. Z. 24184) Bumbuli, Usambara Mtns., 14. xii. 26. 2 (M. C. Z. 2418.5) Kwai, Usambara Mtns., 24. xii. 20. Kigologondwa in Ivikami, Variation. All agree in having the anterior loreal in contact with the first labial; four have the frontonasal widely separated from the frontal (2 Bagilo, 1 Nyange, 1 Misalae) ; in all but the Misalae skink (where they are equal) the frontal is a little shorter than the frontoparietals and interparietal together; in all but a Nyange specimen the frontal is in contact with the second and third supraoculars (in the Nyange specimen on one side only it is in contact with the first, second, and third); frontoparietals distinct; parietals more or less in con- tact behind the interparietal except in two Nyange skinks. The subocular is usually ' feebly but distinctly narrowed inf eriorly ' ; the condition of ' much nar- rowed inferiorly ' is nicely demonstrated in this series by several which have, by division of the subocular, given rise to an extra labial at its expense. The varia- tion in the length of the hind limbs is well shown, for some adpressed hind limbs barely meet the digits of the fore limbs while in others they nearly reach the elbows. All have tricarinate dorsal scales. One has a reproduced bifid tail. We are fortunate in having received from Mr. Hewitt a cotype of the form longiloba characterized by unusually long and sharp ear lobules. Whereas most East African examples are ty]5ically short-lobed, the Namaqualand paratjqie can be matched in the length of its ear lobules by individuals selected from large series with mainly short or medium-sized lobules, for example, M.C.Z. No. 18668 from Mtah's, Mkalama subdistrict of Tanganyika Territory. Coloration. In the largest male taken, a Bagilo skink, the black and white dorsal markings are sharply contrasted ; while below, from just in front of the hind limbs to the tip of the tail, is suffused with salmon. An unusually handsome skink with enlarged testes. Measurements. This male measured 172 (67 + 105) mm. ; the largest female, also from Bagilo, measured 191 (71 -|- 120) mm. Breeding. At Bagilo a 9 with 5 eggs showing traces of embryos was taken on 11. ix. 26, another with 10 eggs on the 14th, and a third with 9 eggs on the 23rd; all the eggs were in the same stage of development. x\.t Nyange dui'ing the first AFRICAN HERPETOLOGY 161 fortnight of the following month all the females appeared to be pregnant; a young skink was taken on 11. x. 26 which measured 56 mm. over all. Many young were seen at Vituri on 25. x. 26, Misalae on 7. xii. 26, Phillipshof on 23. xii. 26 and Kwai on 24. xii. 26. A young one taken at the last locality on that date measured 60 mm. Diet. A Bagilo skink disgorged a stick insect and a red grasshopper on being caught; the following day a second lizard disgorged a grasshopper in the same way. In the middle of the road at Bumbuli was a skink so preoccupied with its meal that it allowed me to walk up and leisurely plant my snake-stick upon its back. Habitat. On rocks at Vituri and Bomoli. At Mogogoni it was fairly fre- quently seen upon the trunks of the coconut palms, yet at Dar es Salaam, which is only a few miles away, I have never seen varia on a tree, as there are few trunks unoccupied by M. striata. At Misalae quite a number were seen on the rocks beside the path at 2 p.m., in hot sunshine; they were so tame that they allowed one to approach within a yard before retiring. [Mabuya irregularis Lonnberg] Mabvia (striata ? var) irregularis sp. n.? Lonnberg, 1922, Ark. Zool. 14, p. 4. Mabuia [Mabiiiopsis) jeanneli Angel, 1923, Bull. Mus. Hist. Nat., Paris, p. 490.; 192.5, 'Voyage de Ch. Alluaud et R. Jeannel en Afrique Orientale (1911-1912),' Paris, pp. 21-23, PI. I, fig. iii. Relationship. It is a matter of personal opinion if the subgenus Mabiiiopsis proposed by Mons. Angel is to be recognized. In that the whole family of skinks show great variability in the arrangement and division of their head shields, we rather doubt if the proposed division is a natural one; while the group of scales on the head of M. irregularis is certainly most unusual it has so many other characters in common with M. striata and M. v. varia that it seems inadvisable to separate it from those species by placing it in a subgenus by itself. See figures of M. varia in Tornier (1900), p. 597, and M. amtilabris in Schmidt, 1919, Bull. Am. Mus. Nat. Hist., 39, Art. II, p. 552. Distribution.. To those interested in the mountain fauna of East Africa, Mons. Angel's new record of two specimens from Kinangop in the Aberdare Range at 10,204 ft. will come as a welcome addition to our knowledge. It was described from Soy near Mt. Elgon in 1921, then recorded from Mt. Kenia between 10,000 antl 11,600 feet in 1923, and Mons. Angel records the fifth and sixth known specimens in 1925. Variation. The junior author on examining Mons. Angel's specimens finds them in agreement with those from Mt. Kenia in the six points cited in Proc 162 BARBOUR AND LOVERIDGE Zool. Soc, 1923, p. 952, except that in the larger example the anterior loreal is in contact with the first labial on the right side but scarcely so on the left; it also disagrees with them but agrees with Liinnberg's type in that (i) the parietals are not in contact behind the interparietal; (ii) it has a single pair of enlarged nuchals. Mabuya striata (Peters) Tropidolepisma striatum Peters, 1844, Monatsber. Akad. Wiss. Berlin, p. 36. 1 (M. C. Z. 24177) Bagilo, Uluguru Mtns., 24. ix. 20. 1 (M. C. Z. 24178) Dar es Salaam, 4. xi. 20. 2 (M. C. Z. 24179) Amani, Usambara Mtns., 19. xi. 26. Also seen at Nyange, Mkangazi, Vituri, Tanga and Muheza. KiguTuguja (Klkami Kiluguru, Kikujao); mjud islam CIviswahili); ghondo ya nyurnba (Kisumbara). Though primarily intended for the Striped Skink, tliesc names are sometimes applied to M. v. varia, most natives not distinguishing between them. The Iviswahili name, hterally ' lizard of peace,' is applied ' because it lives with people without making any trouble ' (this is in reference to its dwelling in the thatch of so many huts). Breeding. Numerous ova within the Bagilo skink held embryos on 24. ix. 26; this was also the case with a skink from Nyange found in the stomach of a snake 2. X. 26. Enemies. A large 9 recovered from a Lycophidion capense at Nyange; an- other from the stomach of a kestrel (Cerchneis tinnunculus carlo) at Mkangazi. Habitat. One was seen running up a cliff-face at Vituri. At Dar es Salaam I disturbed a Striped Skink on the steep bank of a dyke bordering a stream four feet in width. The Uzard made for the field but I headed it off; it rushed back to the bank, down which it tumbled into the water, apparently swam across to the other side under water, put out its head and then disappeared entirely. This species seems to be on the increase in Dar es Salaam, scarcely a tree in some avenues being without one. At Amani they were common on the trees bordering the avenues but were never met with in the forest where M. comorensis occurred. RiOPA SUNDEVALLII (Smith) Eumices (Riopa) sunderaUii Smith, 1849, 'Illus. Zool. S. Afr.' 3, Appendix, p. 11. 1 (M. C. Z. 24199) Dar es Salaam, 4. xi. 26. 1 (M. C. Z. 24200) between Sigi and Amani, 25. xi. 20. Vvtavuta in Kikami; nyoka a ghomlo in Kisumbara. Distribution. Sundevall's Skink has not been recorded from Amani and it is doubtful whether it occurs there ; the specimen referred to above was taken under a log about 200 feet above Sigi ; doubtless with the deforestation of the ravines, which are being planted with bananas, this skink will spread up to Amani in time. In the highlands of Kenya it is found at 6,000 feet, as well as at the coast. AFRICAN HERPETOLOGY 163 SiAPHOS KiLiMENSis (Stejneger) Lygosama kilimensis Stejneger, 1891, Proc. U. S. Nat. Mus., 14, p. 405. 5 (M. C. Z. 24186-90) Nyange, Uluguru Mtns., 7. ix. 26. 4 (M. C. Z. 24191-94) Vituri, LHuguru Mtns., 9, 2.5, 28, .x. 26. 1 (M. C. Z. 24195) Kizerui, Usambara Mtns., 8. xii. 26. 2 (M. C. Z. 24196-7) Mt. Lutindi, Usambara Mtns., 10. xii. 26. 1 (M. C. Z. 2419S) Biimbuli, Usambara Mtns., 14. xii. 26. Variation. These thirteen specimens agree in minutest detail with the de- scription of the type from Kilimanjaro except that the frontal does not equal the frontoparietals in length but the frontoparietals plus the interparietal; the frontal can hardly be called short, and perhaps the ear-opening is not always 'much larger' than the nostril; there are sometimes 8 supraciliaries. Coloration. A breeding male taken at Nyange exhibits the following fugitive colours. Below, throat to vent bright yellow, subcaudal scales with a rufous tinge. Otherwise the colouring is as described by Stejneger. Measurements. The largest male measures 198 (73 -I- 125) mm.; the largest female 178 f67 -fill) mm. but another with reproduced tail is 77 mm. in length from snout to vent. All three from Nyange. Breeding. At Nyange the male, whose coloration is described above, had en- larged testes. The female held four white eggs measuring 11x6 mm. Diet. The contents of seven stomachs examined were as follows : (i) Cricket and spider, (ii) Spider, (iii) Spider, (iv) Large grasshopper and what was ap- parently a hairy beetle larva, (v) Three large hairless caterpillars, (vi) Cock- roach and woodlouse. (vii) Nil. Enemies. Was first found in the stomach of Lycophidion meleagris at Bagilo, 20. ix. 26. Distribution. First discovered on Kilimanjaro, then Mt. Kenia. Lonnberg has recorded it from Kibonoto, and Tornier from the rain-forest of 'Kagera und Kongo' (Tornier, 1900). It is here recorded from the Uluguru Mtns. for the first time. Habitat. The Bumbuli specimen was taken crossing a leaf-strewn path in a banana plantation; all the rest were taken beneath logs in, or on the very edge of, damp rain-forest. Ablepharus wahlbergii (Smith) Cryptoblephm-us ivahlbergii Smith, 1849, 'lllus. Zool. S. Apr.' 3, Appendix, p. 10. Ablepharus massaienxis Angel, 1924, Bull. Mus. Hist. Nat., Paris, p. 52. 4 (M. C. Z. 24201) Dar es Salaam, 4. xi. 20. 1 (M. C. Z. 24202) Bagamoyo, 11. xi. 26. 164 BARBOUR AND LOVERIDGE Variation. Angel has described from Nairobi under the name of ^. massaien- sis a skink which differs from A. wahlbergii only in having 26 instead of 24 scale- rows and 2 supraoculars instead of 3. To check his conclusions we selected for examination two Nairobi specimens (M. C. Z. 18691), which were found to have 28 and 26 scale-rows respectively ; also a series of fifty ^mhlbergii collected in an area of one hundred square yards at Lumbo, Mozambique, the first two specimens having 24 scale-rows, the third 26. As for the supraoculars, both Nairobi skinks have the normal 3, as have all specimens in the Lumbo series. It may be sup- posed therefore that Angel's massaiensis is abnormal in this respect. Habitat. The Dar es Salaam skinks were taken among dried leaves beneath a mango tree near the golf course. ScELOTES EGGELi Tomier Plate 4, fig. 5 Scelotes eggeli Tomier, 1902, Zool. Anz. p. 700. 21 (M. C. Z. 24213-24223) Phillipshof and Kwai, Usambara Mtns., xii. 26. Nyoka a ghondo in Kisumbara. Relation between Scelotes and Sepsina. Hewitt ' has already given very excel- lent reasons for uniting Herpetosaura (Syn. Herpetoseps) with Scelotes, and more recently ^ has discussed the difficulties of differentiating the genus Sepsina from Scelotes and points out that the character of the palatine bones separated on the median line of the palate in Sepsina (meeting in Scelotes) breaks down in the genus Scelotes ; an examination of the twelve species of that genus in the collec- tion of the Museum of Comparative Zoology corroborates this view; in several of them the point of separation occurs about the median line so that the decision is quite an arbitrary one; the same applies to half-a-dozen species usually accredited to the genus Sepsina, which are also represented in the Museum of Comparative Zoology. A second point in which the genera are supposed to differ is the absence of the nasal in Sepsina, its disappearance resulting in the nostril being between four shields — a condition found in at least one species of undoubted Scelotes affinities (»S. brevipes). Hewitt suggests that 'A more striking point of difference between the two genera is in the size of the interparietal scale, small in Sepsina but large and broader even than the frontal in Scelotes. Yet, Sepsina grammica Cope is said to have the interparietal large and nearly as wide as the frontal.' ' Hewitt, 1921, Ann. of Durban Museum, p. 3. -' Hewitt, 1027, Records of the .Vlbany Museum, p. 404. AFRICAN HERPETOLOGY 165 We have examined Cope's type CM. C. Z. No. 5787) and find that the inter- parietal is as small as in the other members of the genus. Sepsina weberi Roux from the South West Protectorate, however, has an interparietal much broader than the frontal, and its nostril is placed like those of the species of the genus Scelotes rather than those generally accredited to Sepsina. Tornier evidently referred eggeli to Scelotes on the grounds of the palatine bones meeting on the median line of the palate. He makes no mention of the rudimentary nasal ring which is almost, if not entirely, indistinguishable in some of the present series. On the basis of the width of its interparietal scale, however, S. eggeli should be referred to Sepsina ; but this would create an anomaly, for its very near ally in the Uluguru Altns. is an undoubted Scelotes, therefore we find ourselves unable to maintain Sepsina as a distinct genus. From external char- acters it appears probable that these two genera should be united and the char- acters referred to should be employed to define subgeneric groups. S. eggeli was based on two specimens from Kwai (where some of the present series were also obtained); Phillipshof is only about three miles from Kwai. This series adds considerably to our knowledge of the species and it seems ad- visable to redescribe it from the figured female (M. C. Z. No. 24213) and this series, incorporating Tornier's description so as to state clearly the range of variation. Description. Snout moderate, about twice the length of the eye; rostral twice as broad as deep, with slightly lunulate upper edge; seven to nine upper labials, only 5th enters eye, 5th or 6th the largest; nostril pierced in the upper posterior part of the rostral, apparently separated from the postnasal by a nar- row rim (rudiment of nasal), bounded above by the supranasal and below by the first labial; a large loreal, deeper than broad, broadly in contact with the fron- tonasal, which is broader than long; two praeoculars, a large upper and a small lower, the latter wedged between the third and fourth upper labials; frontonasal separated from the rostral by the supranasals which are in contact; frontal as long as its posterior breadth, longer than the snout, its posterior border slightly concave; the first supraocular pentrates into the frontal, four supraoculars; six or seven (five) supraciUaries ; interparietal broader than long; parietals in con- tact behind interparietal; two pairs of enlarged nuchals; scales on lower eyelid plainly visible ; ear opening distinct ; 22 to 26 mid-body scale-rows ; limbs penta- dactyle, very short, the adpressed fore limb covering only three-quarters of the distance to the ear, shorter than the distance between ear and tip of snout; hind limb equals the distance between ear and tip of snout. The thick tail joins the elongated body with very slight diminution in girth. 16G BARBOUR AND LOVERIDGE Coloration of figured 9 . Above, iridescent plumbeous, each scale with a hghter spot. Below, creamy-yellow on chin and throat, sharply divided from the rest of the under surface, which is bright salmon-pink; and while each scale on the throat has a large black basal spot, these are wanting down the centre of the body from fore to hind limbs though present on flanks and tail. Considerable variation occurs but most specimens differ from S. uluguruensis in the heavy spotting of their under surface ; in a few like the figured & these are almost confined to the sides and thus approach the immaculate appearance of S. uluguruensis. Some of the variations are as follows : (a) 9 , 93 +57 regenerated. Below, except on chin and throat, the centre of each scale is greenish-yellow and each has a larger or smaller dusky spot at its base ; this colouring obtains only for a little more than the basal half of the tail, the regenerated portion being uniformly plumbeous both above and below, though somewhat lighter below than on the corresponding upper surface. (b) (f , 96 + 102 mm. and 9, 83 + 103 mm. Creamy-yellow over whole of under surface except the end of the tail, which is white; each scale has a basal spot in these specimens. (c) 9 , 82 + 103 mm. Throat white, fore limbs to anus pale orange, tail bright salmon-pink; each scale with a basal spot. (d) cf, 72 +86 mm. Uniformly opaque white on whole of lower surface except for a basal spot on each scale. It wall thus be seen that the common salmon colouring of the tail is not an infallible guide to sexing males though so often present in that sex, nor is it adopted solely by breeding males, as it is present on some immature skinks. (e) Embryos. 32 + 33 mm. Above, uniformly plumbeous; light grey or almost white beneath. Still younger embryos are, of course, unpigmented. Measurements. The largest female measures 205 (101 + 104) mm. The largest male measures 206 (90 + 116) imn., but No. 24220, which has a regenerated tail, measures 102 mm. in length from snout to vent. The youngest independent skink taken measures 35 mm. from snout to vent. For measurements of embryos see below. Breeding. The figured 9 , taken on 21. xii. 26, held three embryos measuring 65 (32 + 33) mm.; other females taken at the same time held less-developed embryos. Diet. Contents of seven stomachs examined: (i) Beetle, (ii) Beetle, beetle larva, cockroach, (iii) Cockroach, (iv) Cockroach, (v) Grub, (vi) Spider, (vii) Spider. AFRICAN HERPETOLOGY 167 Parasites. Most intestines contain very minute nematodes, invisible to the naked eye; in one stomach was a larger species. Distribution. So far known only from Phillipshof and Kwai in the Western Usambaras. Habitat. All save one were taken beneath logs on recently cleared forest or pasture land. While seated on the side of the road, which was cut through the forest, I heard a 'plop' behind me and, looking round, saw one of these skinks, which had fallen over a five-foot bank at the other side of the road and was wriggling about among the dead leaves, none the worse for its fall. SCELOTES ULUGURUENSIS Sp. n. Plate 4, fig. 6 5 (M. C. Z. 24204-8) Bagilo, Uluguru Mtns., 25-28. ix. 26. 8 (M. C. Z. 24209-12) Vituri, IHuguru Mtns., 27-30. .\. 26. VvtaviUa in Ivikami. Type. No. 24204. Musemn of Comparative Zoology. An adult male taken beneath a log at edge of the rain-forest at Bagilo, Uluguru IMountains, Tan- ganyika Territorj\ Collected by A. Loveridge, September 25th, 1926. Paraty-pes. Nos. 24205 to 24212 and four duplicates. Diagnosis. Together with Scelotes eggeli this new skink is readily distinguish- able from the only other members of the genus occurring in East Africa by the presence of five instead of four digits on both fore and hind hmbs. The inter- parietal in this new species being in contact with the third and fourth supra- oculars at once distinguishes it from Scelotes eggeli. Its nearest relatives of the genus Scelotes (and Sepsina) appear to occur on Comoro Id. and jMadagascar. Description. Rostral very broad with lunulate upper edge, seven or eight upper labials, fifth, which is largest, below eye (occasionally 4th and 5th below eye) ; nostril pierced between rostral, supranasal, nasal and first labial ; a large postnasal broadly in contact with the frontonasal; two praeoculars present, a large upper and a very small lower; frontonasal separated from the rostral by the supranasals, which are in contact; four supraoculars, the first three in con- tact with the frontal, the thii'd and fourth in contact with the interparietal; six supraciliaries, of which the first is in contact with the postnasal and the fronto- nasal ; scales on lower eyehd plainly visible ; a pair of parietals in contact behind the interparietal; two (sometimes a third indicated) pairs of enlarged nuchals; ear-opening distinct; 24 mid-body scale-rows; limbs pentadactyle. Coloration of Type & in life. Above, snout to end of body transparent reddish- brown, plates on head edged with darker, a black spot on the apex of each scale; 168 BARBOUR AND LOVERIDGE these spots coalesce on fore arm so that the latter is largely black ; the black spots on the hind limbs are so large that they give the latter the appearance of being black; tail deep plumbeous black; first four upper labials black, rest dusky; sides of body and anterior aspects of limbs creamy-white, each with a black spot. Below, translucent white on throat, body and limbs ; the blood vessels of the throat, and internal organs of the body, may be seen through the scales, to which they impart red and orange tints; the under side of the tail is opaque white with double rows of dusky spots laterally. The colouring of the female paratj^pes is identical with that of the male ; some unsexed specimens show pinkish beneath the tail. Measurements. Type cf'. Snout to vent 75 mm. Tail 96 mm. Largest Bagilo 9 . Snout to vent 88 mm. Tail 70 mm. (Perhaps regenerated.) Breeding. The Bagilo female, taken on 25. ix. 26, held four ova measuring 10 X 7 mm.; a Vituri female, collected 27. ix. 26, held four measuring 11 x 8 mm. These latter showed a trace of embryos. I might here mention a strange find, though in no way suggesting that it has anything to do with this viviparous skink. At Nyingwa, about 7,000 feet, no sign of a live Hzard was seen during the week we were there, but on 16. x. 26 I found six Hzard eggs beneath a thick matting of moss in the rain-forest. I measured them and found they were 10x7 mm. ; the only other possible sugges- tion could be that they belong to S. kilimensis or possibly some chameleon. Diet. The stomach contents of five specimens examined were as follows: (i) Smooth-skinned lepidopterous larva, (ii) Coleopterous larva, (iii) Beetles. (iv) Beetles, (v) Beetles, woodlouse, spider. Habitat. All the examples taken were more or less concealed in the rich leaf- mould underneath logs; these were in, or on the edge of, the rain-forest. SCELOTES TETRADACTYLA (Peters) Sepsina (Rhinoscinciis) tetradactyla Peters, 1874, Monatsber. Akad. ^\'iss. Berlin, p. 374. 1 (M. C. Z. 24203) Mkarazi, Uluguru Mtns., 22. x. 26. Variation. 26 scale-rows instead of 24, 5 supraciliaries on one side, 6 on the other instead of 4. Coloration. The black lines following the longitudinal scale-rows, and the blue network on the tail, are only very faintly indicated. The colouring of the upper surface would be best described as iridescent steely-brown. Measurements. This male is larger than the type, and measures 133 (92 -|- 41) mm. ; possibly the tail is reproduced, as was that of the type according to Boulen- ger. The latter measured 120 (82 -|- 38) mm. AFRICAN HERPETOLOGY 169 Diet. A woodlouse was in the stomach of this skink. Habitat. Taken burrowing into the mould beneath a log on the roadside a hundred yards beyond the ford across the Mvua River on the Mkarazi-Tawa Road- Distribution. The type, collected by Hildebrandt in 1874, was described as from the ' Zanzibar Coast ' ; since then Nieden has recorded one collected by Bohm at Marungu and three collected by Brown at Milanji. In the British Museum are specimens from the Shire Highlands and Zomba, Nyasaland. Melanoseps ater longicauda Tornier Melanoseps ater var. longicauda Tornier, 1900, Zool. Jahrb. Syst., 13, p. 602. 1 (M. C. Z. 24235) Vituri, Uluguru Mtns., 27. x. 26. Variations and Relations. Owing to the end of the tail being missing it is with great misgivings we refer this skink to Tornier's subspecies. As it stands it is nearer to Gunther's M. ater from the Zambesi, agreeing with that reptile in having 22 mid-body scale-rows and not with longicauda in having 19. It has 144 scale-rows from chin to vent as against 118-120 in Tornier's two types from Masailand and Kerogwe' (latter locahty at foot of Usambara Mtns.). Our reason for employing this name is to emphasize its identity with one taken on the Mkata Plains (near the Station and River) in 1921. The points in which that skink differed from Tornier's t>T3es have already been mentioned. = Vituri is about fifty miles from Mkata Station as the crow flies, although the habitat con- ditions could scarcely be more different. It is probable that Tornier's name will not stand but it seems advisable to await the collection of more material before placing it in the synonymy. Color in life. Above, uniformly steely blue-black. Below, very sHghtly paler, except for the rostral, supranasals, mental, postmental and first lower labials, which are china-whitie. Measurements. Male, from snout to vent 124 mm.; end of tail missing. Habitat. Taken beneath a log in the rain-forest in a ravine, which is the first on the left as one enters the forest at Vituri on the path to Nyange. The spot was close to the base of a big rock over which a stream trickled, which must be a considerable waterfall in the rainy season so that the place must be con- stantly moistened with spray. For three days I made persistent efforts to secure another specimen in this ravine but without success, though many caecilians were turned up during the hunt. 1 'Karagwe am Panani.' The tii'st word would appear to be a misspelling for Kerogwe, which is on the Pangani. = Loveridge, 1923, Proc. Zool. Soc, Lond., p. 963. 170 BARBOUR AND LOVERIDGE CHAMAELEONTIDAE Chamaeleo dilepis dilepis Leach Chamaeleo dilepis Leach, 1S19, in Bowdich, 'Miss. Ashantee,' p. 493. » 2 (M. C. Z. 24246) Nyange, Ulugum Mtns., ix-x. 26. 2 (M. C. Z. 24248) Amani, Usambara Mtns., 26. xi. 26. 1 (M. C. Z. 24249) Sakkarani, Usambara Mtns., 18. xii. 26. Kangawingo (Kikami); luvi (Kisumbara). Not specific. Distrihution. Also found at Bagilo, and nine previously taken at Mkuyuni, Uluguru Mountains. It must not be supposed that because this series is small dilepis is less coinmon in the Uluguru Mountains, though this was the case in the Usambaras, where fischeri subsp. is the dominant chameleon. In the Uluguru dilepis is plentiful; about a score were dispatched alive to the London Zoological Gardens and the natives were then discouraged from bringing them in. Color in life. 9 . Amani. Above, bright yellow-green with darker green saddle- like markings and some yellow spots on the hind limbs, lumbar region and tail, a violet-brown spot behind each occipital flap and a broad streak of same on flank. Below, ventral crest, praeanal region and lower side of limbs white. Diet, (i) Leaf-Uke wings of some large orthopteran and the head of an ant (?) in the stomach of the Nyange adult, (ii) Grasshoppers in young Nyange chame- leon, (iii) Stomach and intestines of Amani 9 crammed with coleopterous and orthopterous remains. Folklore. The Wakami have the following story of this reptile. Once upon a time the chameleon challenged a lion to a race; they lined up for the start but as the lion sprang forward the kangaivingo grabbed its tail and clung on. The lion ran for all it was worth and the chameleon held on for dear life. On reaching the winning post the Uon sank down to rest, at which the kangawingo cried out, 'Look where you are lying or you will crush me.' The lion was aston- ished and dispirited on hearing the chameleon's voice and exclaimed, 'What! can you really outrun me?' The reptile replied, 'Of course I can and as a mat- ter of fact have been waiting for you to come in.' This story seems to us the African counterpart of our fable relating to the choosing of a king by the birds, when the eagle was outwitted by the wren which hid itself in the former's plumage. Parasites. A new species of nematode worm {Physaloptera ortleppi) as well as Strongylurus brevicaudata, were taken in the intestines of chameleons collected at Dodoma. AFRICAN HERPETOLOGY 171 Chamaeleo dilepis qtjilensts Bocage Chaviaeleo dilepis var. guiletisis Bocage, 1866, Jorn. Sci. Lisboa, 1, p. 59. 16 (M. C. Z. 24250-9) Ifakara, Mahenge, District, ix. 26. Thirty individuals in all were brought back by a native who was sent to the Iringa highlands to get some chameleons there. Young dilepis dilepis is rather hard to distinguish from this race, as the occipital lobes are not so developed in their young as in the adults. Chamaeleo fischeri subspp. One of the principal objects of the expedition was to collect sufficient topo- typic material to throw light on the strange distribution of the alleged races of this interesting chameleon. As a result, over four hundred examples were col- lected from the four type localities (and intermediate places) of the four races said to occur in the Eastern and Western Usambaras. As sixty of the chameleons were brought back alive and left at the Zoological Gardens, London, only three hundred and eighty-four were at our disposal for study; these may be summarized by locality as follows: Locality Is Type-LoraUty of C. f. matschici Werner && 99 52 28 2 4 12 .5-Y 148 63 43 14-3 Amani and Derema, Eastern Usambaras ] r f , li> M' H Mlalo near Ambangula, Western Usambaras C. f. wcrneri Nieden Bumbuli north of Ambangula, Western Usambaras Sakkarani north of Ambangula, Western Usambaras Phillipshof near Lushoto, Western Usambaras C. /. muUituberculatus Nieden Mlalo near Ilohenfriedeberg, Western Usambaras C! C. f. wenieri Nieden) Topography Nieden, in 1913, gives the distribution of the races oi fischeri as follows: C. f. fischeri Reichenow. Nguru Mtns. and Unguu Mtns. (Is not the latter a corruption of the former?) C. f. matschiei Werner. Amani, Magrotto, Nguelo, Derema in Eastern Usambara. C. f. vosseleri Nieden. Amani, Magrotto, Nguelo, Buloa near Tanga, Ukami and Usaramo. C. f. werneri Nieden. Mlalo and Ambangula in Western Usambara. C. /. multituberc^datus Nieden. Phillipshof near Wilhelmstal in Western Usambara. 172 BARBOUR AND LOVERIDGE The following notes on the type localities may help to elucidate matters. C. f. fischeri Reichenow. Chamaeleon fischeri Reichenow, 1887, Zool. Anz., 10, p. 371. The Nguru Mtns. lie about halfway between the Usambara and Uluguru Mtns., that is to say, from sixty to ninety miles from each. The female of C.f .fischeri is still unknown; in all probability when it is found it will be similar to that of C. /. matschiei and cause the latter to fall into the synonymy of the former. C. f. matschiei Werner and C. f. vosseleri Nieden. It will at once be observed that the ranges of these two races overlap consider- ably, and more strangely still, C.f. vosseleri was also described from Ukami (i.e. Uluguru Mtns.) and Usaramo (i.e. in the Dar es Salaam District) as far to the south of the Nguru Mtns. as are Amani, Nguelo etc. to the north. 'Buloa near Tanga' is, we imagine, Bulwa near Amani, in Tornier (1897). Eismann's speci- mens are recorded as from 'Buloa bei Tanga in Usambara.' C. f. werneri Nieden. This subspecies was described from 'Mlalo and Ambangula in West Usam- baras,' thus ruling out Mlalo near Amani in East Usambara Mtns. Ambangula lies five miles due north of Kerogwe station on the Tanga Railway. Mlalo, though rarely shown on maps, is a mile from Ambangula; obviously this is the Mlalo meant. The matter is complicated, however, since a male cotype of Nie- den's Ch. f. werneri has been received in exchange from the Berlin Museum and is labelled 'Mlalo bei Wilhelmstal.' 'WHien Wilhelmstal was captured by the Brit- ish it reverted to its native name of Lushoto. On arrival at Lushoto the junior author made very careful enquiries as to the whereabouts of this third Mlalo and found that the nearest known ^Nllalo was near Hohenfriedeberg, to reach which one passes through Phillipshof and so away to the north. To make assurance doubly sure a series of sixty specimens was obtained from this Mlalo also. C.f. nmltituberculatits Nieden. Phillipshof is a well-known resort a few miles out of Lushoto. Chamaeleo fischeri matschiei Werner Chamaeleon matschiei Werner, 1895, Verb. Zool. bot. Ges. Wien, p. 192, PI. V, fig. 2. Chamaeleo iornieri Werner, 1902, Zool. Jahrb. Syst. 15, p. 417, PI. XXIV (Mozambique) Chamaeleo fischeri vosseleri Nieden, 1913, Sitzber. Ges. Naturf. Freunde, p. 2-11. 80 (M. C. Z. 24260-24285) Amam and Derema, Usambara Mtns., xi. 26. Additional material. In addition to the cotype and topotypes mentioned above, others of three forms have been received by exchange from the Berlin Museum : AFRICAN HERPETOLOGY 173 No. 20192. a' Ch. f. fischeri ex 'Buloa' (? Bulwa). Eismann coll. 20193. & Ch. f. watschiei ex Amnm. Vos?eler coll. 20194. o" Ch. f. vosseleri ex Tanga. Reimer coll. A. cotypc. No. 20192is indistinguishable from 20194, which is not surprising, seeing that 'Buloa' is given as one of the tj^je localities of vosseleri. Variation. Males. These have dorsal spines on the anterior part of the back and usually, but not invariably, on the tail, which is sometimes perfectly smooth. When sorted by the presence or absence of spines on the tail it was observable that, in the main, the absence of spines is correlated with short horns and smaller size and is evidently a juvenile character; though one large well-developed male had a smooth tail, while four half-grown and two young specimens were in an intermediate condition. Other variations consist in the casque of the parietal region being plated or covered with small scales, or characterized by many in- termediate stages between these two e.xtremes. There is also astonishing varia- tion in the breadth and length of the posterior part of the casque, which may be rounded or sharply pointed. Females. No fewer than twenty of these twenty-eight females are hornless or exhibit only the faintest indication of horns in the form of a pair of angular swell- ings on the canthi. The remaining eight, which with two exceptions are much larger in size than the other twenty, have horns varying in length from 4 to 8 mm. (measured from anterior under side at base, to the tip) . Horns in the female are therefore obviously an age character. This is corroborated by the fact that the dorsal crest in the horned females is slightly more developed than in the horn- less, which are practically smooth. All agree in the non-spinous condition of the tail, which is usually absolutely smooth. Coloration. In life these chameleons have a soft velvety appearance unhke any other species I have seen. Measurements. The biggest male (24260) collected during the 1926 expedition measures 345 (130 + 195 -|- 20 mm. horn) mm., but is surpassed by No. 20193, which is 399 (147 +234 + 18 mm. horn) mm.; the biggest female (24273) measures 284 (112 + 165+7 mm. horn) mm.; the smallest female (24285) measures 113 (46 + 73) mm. In this specimen there is only a slight swelling to indicate the position of the horns. Breeding. Many of these females have very large ova or eggs. Diet. About a dozen were examined which held beetle remains; some were those of brilliant green weevils. In addition to the beetles other insects were recognizable in six of the chameleons. These were (i) Grasshopper, (ii) Grass- 174 BARBOUR AND LOVERIDGE hopper, (iii) Grasshopper and caterpillars, (iv) Orthoptera and a caterpillar. (v) A bug like a frog-hopper, (vi) Hemipteron. Parasites. Larval mites (Erythraeidae) were collected from these reptiles, and nematodes (Strongyluris gigas) were almost invariably present. Habitat. Several which I captured personally were ascending trees at the edge of the forest. ^Vlost of the others were taken on bushes. Chamaeleo fischeri multituberculatus Nieden' Chaniaeleon fischeri multituberculatus Nieden, 1913, Sitzber. Ges. Naturf. Freunde, Berlin, p. 247. Chamaeleon fischeri xcerneri Nieden, 1913, Sitzber. Ges. Naturf. Freunde, Berlin, p. 243. 2 (M. C. Z. 24286-87) Mlalo near ,\mbanguhi, Usambara Mtns., xii. 26. 4 (M. C. Z. 242SS-91) Bumbuli, U.sanibara Mtns., xii. 20. 18 (M. C. Z. 24292-308) Sakkarani, U.sambara Mtns., xii. 26. 211 (M. C. Z. 24309-360) Ihillipshof, Usambara Mtns., xii. 26. 60 (M. C. Z. 24361-72) Mlalo near Hohenfriedeberg, Usambara Mtns., xii. 26. Relationships. C. f. werneri Nieden, 1913, has page precedence over C. f. multituherculatus but is preoccupied on account of C. werneri Tornier, 1899, which is an entirely dUTerent reptile from Southern Tanganyika Territory. This is for- tunate, as the name multituberculatus is well chosen and at once defines the principal difference between the males of the Western and Eastern Usambara races. These reptiles also differ in that the females of muUituberculalus are in- variably horned. C. f. icerneri from Mlalo near Ambangula is probably somewhat of an intermediate. There are males in our Phillipshof series with similar crests to that which is figured and described for C. f. werneri. Variation. Males. A series of spines, usually uninterrupted from nape to the middle of the tail, but sometimes very low or interrupted in the sacral region. The teeth composing this crest may be continuous or with small interspaces between each two; they may also be flat (i.e., blade-like) or cone-shaped, the two types being very different in appearance. There is great variation in the horns, which are usually more or less parallel to one another, though they fre- quently meet at the points and sometimes, though rarelj^ diverge as in Ch. tavetensis (which, however, is quite distinct). The horns may project in the line of the body axis or tilt upwards; in a few specimens they are decidedly down- pointing. Some of the horns have grown in peculiar twisted fashion, doubtless as a result of injury received when young; one such pair lean well to the right. The parietal crest may extend right to the end of the casque or be but a small raised crest in the middle of that area. The parietal region on either side of the crest may ' Though Cli. f. werneri has page precedence over Ch.J. iniillUiibcrculaius, the latter name must be used, as Ch. werneri is preoccupied. AFRICAN HERPETOLOGY 175 be raised with a swollen appearance, which is a sign that the reptile is well nourished, or deeplj^ sunken, saucer-hke, the evidence of a starved condition. The posterior edge of the casque may be sharply defined, upstanding (famished) or sunken into the outline of the back (well fed) ; quite apart from this there is astonishing individual variation — in some the posterior edge of the casque ter- minates in a sharp point, in others it is broad and spiky, or again broadly rounded with scarcely noticeable projections. Females. One of the best features by which to differentiate matschiei and rmiltituberculatus is that every female in the Western Usambara series is horned, almost to the same extent as the males, while in the females from Eastern Usam- bara the majority are hornless. The two topotj-pic females of werneri have rostral processes 3 nun. long exactly like the typical female, but they can be matched by specimens in the PhilUpshof series which perhaps are not so advanced as is usual in females of multituherculatus of the same size. Coloration. The chameleons with a pair of white Unes on the belly are always females but the absence of such lines is not evidence that the reptile is a male. These Unes in males are usually a sign of youth, being present in young specimens, and they are generally wanting in adults although occasionally persisting in large individuals. Measurements. Largest male measures 319 (137 + 175 + 7 mm. worn horn) mm.; largest female 272 (112 + 148 + 12) mm.; youngest male 118 (51 + 65 + 2 mm. horn) mm.; youngest female 95 (42 + 53 + swelling not projecting beyond snout) mm. Breeding. One of the two Mlalo (near Ambangula) females had undeveloped ova, the other held small eggs of 7 mm. diameter on 31. xii. 26. Eight out of fifteen collected at Phillipshof on 20. xii. 26 showed developing eggs. These varied in number from ten to twenty-one, more or less, depending on the size of the mother. The development is best shown thus: 10 eggs measuring 17 X 9 mm. 16 eggs measuring 7X7 mm. 11 " " 14 X 7.5 16 " " 14 X S 12 " " 1.5 X S 17 " " 14 X 8 13 " " 13 X 8 21 " " 1.5 X 8 Diet. Chiefly coleoptera and orthoptera. Defence. The fact that the horns of all the biggest males are worn down to half the length of those of the younger males in their prime is surely evidence that these are used as fighting weapons, as is the case with C.jacksoni vauerescecae. Parasites. Larval red mites of the family Erj^thraeidae {Pterygosoma sp.) were numerous, and nematodes (Strongyluris hrevicaudata) are almost invariably 176 BARBOUR AND LOVERIDGE present. After the chameleons were chloroformed and soaked over night in water, they were transferred to formalin ; then round worms would appear, attempting to escape from the nostrils, mouth or anus. Enemies. In the garden of Lushoto Hotel, I shot a A\1iite-breasted Crow (Corvus albus) which was being fed by its parents on a chameleon. Fifty feet away I caught a Boomslang {Dispholidus typus) with a chameleon in its stomach. [Chamaeleo fischeri excubitor (Barbour)] Chamaeleo tenuia excubitor Barbour, 1911, Proc. Bio. Soc. Wash., 24, pp. 219-220. We take this opportunity of pointing out that this chameleon from Mt. Meru district of Kenya Colony represents a race oi fischeri in which the male is horn- less. The tjTJe is a male and not a female as was at first supposed. Chamaeleo deremensis Matschie Chamaeleon deremensis Matschie, 1S92, Sitzber. Ges. Naturf. Freunde Berhn, p. 103. 1 (M. C. Z. 24234) Nyange, Uluguru Mtns., 11. x. 26. 1 (M. C. Z. 24236) Vituri, LHuguru Mtns., 30. x. 26. 3 (M. C. Z. 24237-9) Amani, Usambara Mtns., 24. xi. 26. 2 (M. C. Z. 24240-1) Derema, Usambara Mtns., 30. xi. 26. 1 (M. C. Z. 24242) Bumbuh, Usambara Mtns., 14. xii. 26. Coloration. In life these are a beautiful dark-green with lighter green and mauve markings. Matschie does not exaggerate when he speaks of it as a mag- nificent species with its three long horns and sail-like dorsal crest. Measuremerits. The largest male measured 328 (165 -I- 140 + 23 mm. nasal horn); largest female 295 (158 + 137) mm. Both Derema, i.e., topotypes, yet exceeding the tj^ies in size. Breeding. The Derema and Bumbuli females have their o\'a in much the same stage of development, i.e., about 8 mm. diameter and spherical. An 83 mm. (snout to vent) Amani male is interesting as it shows the advanced age at which the horns first appear, for in this specimen none of them are over 2 mm. long. Diet, (i) Many hard-shelled beetles, including Macropoda tuberculifera or allied species, (ii) Beetles, grasshopper and caterpillar, (iii) Weevil and other beetles, two large green locustids, two green stick insects and a spider, (iv) Beetles of many species, chiefly weevils, orthopteran, mantid, snail, (v) Beetles of many species, a millipede and skink scales. The only skinks seen in the plantation where this chameleon was caught were Mabuya comorensis and M . v. varia. (vi) Many weevils, grasshopper, millipede. By reason of their diet consisting so AFRICAN HERPETOLOGY 177 largely of beetles and orthoptera these chameleons are of no small economic value to coffee planters and should be strictly preserved. Defence. The claws of a struggling deremensis tear one's skin (epithelium) very easily. The horns are doubtless used for fighting; the nasal horn of the Nyange male (which was only 4 mm. short of being the largest male) measured 23 nmi., while its occipital horns were 26 and 27 mm. respectively. Parasites. Many nematodes were found in the Derema and Bumbuli speci- mens. Distribution. Described from Derema and later recorded from Massailand and Usaramo (i.e., Dar es Salaam District) ; the present are the first records of its occurrence in the Uluguru Mountains. Habitat. At edge of forest in most localities. At Derema they occur in coffee plantations which within the last decade were forest land. Chamaeleo spinosxjs Matschie Chamaeleon spinosus Matschie, 1892, Sitzber. Ges. Naturf. Freunde Berlin, p. 105. 1 (M. C. Z. 24243) Amani, Usambara Mtns., 26. xi. 26. Measurements. A female which measures 66 (40 + 26) mm. and is therefore somewhat smaller than the type from Derema, which was 87 mm. Chamaeleo tenuis Matschie Chamaeleon tenuis Matschie, 1892, Sitzber. Ges. Naturf. Freunde Berlin, p. 106. 2 (M. C. Z. 24244-5) Amani, Usambara Mtns., 27. xi. 26. Measurements. The male measures 101 (48 + 53) mm., and the female 129 (62 + 67) mm. The male is far larger than the type from Derema, which was only 69 mm. They were taken together in long grass on Mt. Bomoli. Werner (1911) has recorded this species from Ukami (i.e., Uluguru). Rhampholeon brevicaudatus (Matschie) Chamaeleon {Brookesia) brevicaudatus Matschie, 1892, Sitzber. Ges. Naturf. Freunde Berlin, p. 107. 6 (M. C. Z. 24373-S) Bagilo, Uluguru Mtns., 13-27. ix. 26. 1 (M. C. Z. 24379) Nyange, Uluguru Mtns., 1. x. 26. 2 (M. C. Z. 24380-1) Mkangazi, Uluguru Mtns., 12. x. 26. 3 (M. C. Z. 24382-4) Vituri, IHuguru Mtns., 29. x. 26. 1 (M. C. Z. 24385) Amani, Usambara Mtns., 24. xi. 26. Lwivi (Ivikami); luvi (Ivisumbara). It will be observed that this is the only species of chameleon that the Wakami have a special name for, and it is very similar to that employed by the Wasumbara for all species of Chamaeleo and Rhampholeon. 178 BARBOUR AND LOVERIDGE Variations. In his description of the female type from Derema, Matschie's measurements show that the tail is one-sixth the total length from snout to tip of tail, or three and one-third the body length alone, yet in the text he says 'less than one-fifth the body length,' which is somewhat ambiguous. In the present series of four females the tail is included in the snout-to-vent measurements 4.08 to 4.83 times; in the nine males, 2.21 to 3.75 times, which gives one a ready means of sexing these creatures by the longer tail of the male. In this small series the average of both sexes is 3.63 times. Measurements. The largest male (Nyange) measures 94 (12 +22) mm.; the largest female (Bagilo), 92 (75 + 17) mm. Breeding. The largest female, taken at Bagilo 13. ix. 26, held four eggs; these measured approximately 6 mm. in diameter. At Vituri on 30. x. 26 two females were taken beneath large stones on the banks of a stream at the edge of the forest. I assumed at first that they had burrowed beneath the stones to lay their eggs, but while one had six enlarged eggs measuring 10 x 6 nun., the other was scarcely adult (61 mm. over all) and its ova were not developed. A female was taken at Mkangazi on 12. x. 26 with 4 eggs each 6 mm. in diam- eter. On the same day, while on the march, Sahmu secured the smallest specimen I have ever seen; it walked comfortably into the end of Ramazan's flute, which had a diameter of half an inch. There Salimu attempted to imprison it by plug- ging the aperture with some wool. Ramazan, not knowing it was there, removed the plug and played lustily on the flute as he brought up the tail end of the safari. The loss to science was learned only when he marched tootling into camp ! Salimu seized the flute, but the chameleon had departed. Diet. Stomach contents: (i) Spider, (ii) Hard-shelled beetle, (iii) Reddish grasshopper, (iv) Grasshopper, (v) Cricket. Three others examined were empty. CAECILIIDAE ScoLECOMORPHUs viTTATUs (Boulcnger) Bdellophis mtlata-'i Boulenger, 1895, Proc. Zool. Soc. I^nd., p. 412, PI. XXIV, fig. 4. 1 (M. C. Z. 12179) Bagilo, Uluguru Mtns., 29. ix. 26. 1 (M. C. Z. 12180) Nyange, Uluguru Mtns., 4. x. 26. 4 (M. C. Z. 12181-4) Vituri, Uluguru Mtns., 28-30. x. 26. 3 (M. C. Z. 12185-7) .\mani, Usambara ISItns. xi. 26. 1 (M. C. Z. 12188) Mt. Lutindi, Usambara Mtns. 10. xii. 20. 1 (M. C. Z. 12189) Phillipshof, Usambara Mtns. 24. xii. 26. Moiivi (Kikami); bhfn (Kiluguru); mango flvisumbara). AFRICAN HERPETOLOGY 179 Relations. The examination of the skull of a topotypic specimen of B. vittatus shows that the character of 'Eyes distinct' which separates the genus from Scolecomorphus, whose eyes are below the cranial bones, is a sign of youth, ossi- fication developing with age so that the eyes of adults are concealed. Variation. 122-148 annuli, previous known range being 125-148. The very strong palatal teeth, four to six in number, appear to have escaped the original describer. The eye, visible in the 160 mm. type, is indistinguishable in specimens of 200 mm. and over. The younger the specimen the sharper its snout. The position of the tentacles alters in relation to the apex of the lower jaw; in the young they lie well in front, while in the two largest specimens they are on either side of the apex. The tjrpe is in poor preservation, little more than skin and vertebrae, hence the 'much flattened' habit; though living specimens are de- pressed, the impression is not nearlj^ so marked as one might be led to suppose. Coloration. In life. No. 12179. Above, glossy purplish-black. Below, pink except under side of tail, which is pure white. — No. 12188. Above, purplish-black; below, mauve. When picked up, this specimen during its struggles exuded a sticky substance in considerable quantities, the mauve colouring coming off on my hands. Coloration. In alcoliol. With one exception all these specimens have the blue- black upper surface sharply defined from the yellowish under side. The exception is No. 12183, in which the pigmentation of the upper surface irregularly en- croaches on the lower surface. The transitional stages from the narrow blue-black dorsal band of the Amani topotypes to the complete blue-black upper surface of the Uluguru specimens are well shown in this series. Measurements. Lengths range from 140 to 300 mm., mid-body diameters from 5 to 9 mm., such diameters being contained in the lengths from 25 to 36.1 times. Diet. The Nyange specimen held a smooth-skinned green caterpillar. The stomachs of half-a-dozen other caecilians which were examined were found to contain a white floury- or cheese-like mass impossible of determination. Defence. See coloration. Parasites. Minute nematodes {Aplectana loveridgei sp. n.) present in Nyange and Vituri stomachs. Habitat. Concealed in leaf-mould beneath rotting logs in the damp rain- forest at Bagilo. The Nyange caeciUan was also beneath a log in a similarly moist situation but it was on the surface and wriggled awaj' as the log was over- turned. The large Lutindi specimen was beneath a log in very dry forest but only fifty yards from a stream. The PhiUipshof example was also in dry forest but close to a stream. 180 BARBOUR AND LOVERIDGE Distribution. There is in the British Museum a specimen, identified by Mr. G. A. Boulenger, from Mombasa, Kenya Colony, collected by Dr. S. L. Hinde in 1909. SCOLECOMOBPHUS ULUGURUENSIS Sp. n. (?) Scolecomorphtis kirkii Loveridge (not of Boulenger), 1925, Proc. Zool. Soc. p. 76.5. Small dried speci- men from Bagilo. 3 (M. C. Z. 12190-12192) Bagilo, Uluguru Mtns., Lx. 26. 129 (M. C. Z. 12193-12293) Nyingwa, Uluguru Mtns., x. 26. Type. No. 12193. Museum of Comparative Zoology. Sex c?'. From Nyingwa, Uluguru Mountains, Tanganyika Territory. Collected by A. Loveridge, October 19th, 1926. Paratypes. The Nyingwa specimens enumerated above. Diagnosis. The type of S. kirkii has been examined and compared with some of the Nyingwa specimens. It is intermediate between S. uluguruensis and the species about to be described, in its 152 circular folds and its diameter into total length, which is 38.57 (to 44) times; but above all it differs very much in colour, which is dark olive above and brownish olive below (in alcohol). Description. Habit like Hypogeophis, head very small, body round, thick and heavy, tail blunt. Snout prominent, obtusely acuminate, projecting far beyond the lower jaw. Teeth well developed, 16 upper, 12 lower and 6 strongly recurved palatine teeth. Eye hidden. Tentacle round, exsertile, situated in a horseshoe- shaped groove opening anteriorly, just behind an imaginary line connecting the nostril with the apex of the lower jaw, below and behind the nostril but much nearer the mouth than the nostril. 133 annuh (124 to 151 in count of forty Nyingwa paratypes; it appears probable that males range from 124 to 139 and females from 140 to 151 but some overlapping may occur); annuh on the nape very pronounced, giving an upward tilt to the head in adults (in the young this area is as smooth as in adult B. vittatus) ; after first 14 (14 to 20 or thereabouts) rows on nape the annuli are interrupted on the vertebral line to the end of the tail (but not interrupted on the last inch of body and tail in most paratypes); anal opening close to tip of tail, penis extended in type. Coloration. In alcohol. Above, dull blue-grey (sometimes glossy in males) inconspicuously merging into the somewhat more plumbeous grey of the lower surface; throat, and a similar or more extensive area in front of anal opening, white {in life this is bright flesh-pink). Measurements: Total length of type 215 mm. Diameter at mid-body 9 mm. Diameter at mid-body 9 '' Smallest specimen 140 " Largest specimen ( 9 ) 342 " Diameter at mid-body 6 " AFRICAN HERPETOLOGY 181 This species becomes proportionately more slender with age, though this does not appear at first sight and one would be apt to assume that the contrary was the case. It is best shown in the following table. The average number of times the diameter is included in the length of: 14 specimens over 140 but under 200 mm. in length is 24.1 times (the range being from 21.4 to 27.7 times) 96 specimens over 200 but under 300 mm. in length is 26.4 times (the range being from 22 to .33.1 times) 19 specimens over 300 but under 342 mm. in length is 29.9 times (the range being from 2o to 34.3 times) Breeding. Four embryos were present in one of the oviducts of a 305 mm. female; one such embryo (unstraightened) measured 13 mm., the eye was very, distinct. Diet. A cheese-like substance in many stomachs, probably the remains of macerated termites. Distribution. From 6,000 to 7,500 feet in the Uluguru Mtns. Habitat. Taken in sodden leaf-mould beneath rotting, moss-grown logs in the rain-forest, generally in clearings caused by the fall of trees. SCOLECOMORPHUS ATTENUATUS Sp. n. 2 (M. C. Z. 12194-5) Nyingwa, Uluguru Mtns., x. 26. Type. No. 12194. Museum of Comparative Zoology. Sex $. From Nyingwa, Uluguru Mountains, Tanganyika Territory. Collected by A. Loveridge, Octo- ber 15th, 1926. Paratype. No. 12195 collected October 19th, 1920. Diagnosis. Teeth apparently somewhat smaller than in S. uluguruensis. Snout more bluntly rounded. Habit much more slender, the diameter being con- tained in the total length 39 to 45 times as against 24 to 29 times. When the type was brought in with a number of S. uluguruensis I was immediately struck by its long slender habit and jet-black color. Description. Habit more like sis Duin6ril et Bibron, 1841, 'Erp6t. Gen.,' 8, p. 367. Rana merumontana Loveridge (not of Lonnberg), 1925, Proc. Zool. Soc, p. 774 (Kabare, Bukoba). 1 (M. C. Z. 12714) Bagamoyo, 11. xi. 26. This very distinctively coloured species has been compared with specimens from Zanzibar and Tanganyika Territory received from the Berlin Museum. With these it agrees perfectly, as also with the brief description. 200 BARBOUR AND LOVERIDGE Rana adspersa (Dum^ril et Bibron) Pyxicephahis adspersus Dumfiril et Bibron, 1841, 'Erpdt. G6n.,' 8, p. 444. Maltzania bufonia Boettger, 1881, Abh. Senck. Ges., 12, p. 418, PI. 1, figs. 3a-e. (Senegambia.) Ra7ia mallzanii Boulenger, 1882, Cat. Batr. Sal. Brit. Mus., p. 34. Rana adspersa Lonnberg, 1910, in Sjostedt, Ivilimanjaro-Meru Exp., I, part 4, p. 21. PI. 1, fig. 3a and 3b. 1 (M. C. Z. 12680) Dar es Salaam, 4. xi. 26. Variation. Boettger's Rana bufonina is said to differ from adspersa in the absence of longitudinal dorsal skin folds and ridges and in the larger tympanum. A comparison between Lonnberg's excellent figure of a not fully adult adspersa from Tanga, and Boettger's figures of his type of bufonina from Senegambia, will show how very similar in appearance are these two creatures. On the basis of the material in the M. C. Z. we are able to state with conviction that Boettger's smooth-skinned 87 mm. type was only a young R. adspersa. The material referred to is as follows: M. C. Z. No. 10369. Frere Town, Kenya Colony. Length 56 mm. M. C. Z. No. 10367. Ivilosa, Tanganyilva Terr. Length 96 mm. M. C. Z. No. 12686. Dar es Salaam, Tanganyika Terr. Length 98 mm. M. C. Z. No. 10368. Nyambita, Mwanza, Tanganyika Terr. Length 109 mm. The first three of these are dorsally smooth-skinned, showing a gradual devel- opment of lateral tubercles encroaching on the back. No. 10368 possesses a rough back similar to Lonnberg's figure, and comparable to the condition of the series of adult South African R. adspersa in the collection. Probably methods of pres- ervation have as much to do with a strongly tubercular appearance as has age. Diet. Stomach full of ants and sand. Enemies. It had more than a score of dead soldier ants {Dorylus nigricans subsp.) attached to its limbs and belly. Habitat. A half-grown specimen taken on the edge of a water hole close to Mogogoni Swamp. Arthroleptides martiensseni Nieden Arthroleplides martiensseni Nieden, 1910, Sitzber. Ges. Naturf. Freunde BerHn, p. 445. 23 (M. C. Z. 12817-21) Bagilo, Ukiguru Mtns., ix. 26. 1 (M. C. Z. 12822) Vituri, Ukiguru Mtns., x. 26. 2 (M. C. Z. 12823-4) Amani, Usambara Mtns., xi. 26. 1 (M. C. Z. 12825) Mt. Lutindi, U.sambara Mtns., 10. xii. 26. Variation. The interorbital space equals the upper eyelid in large specimens. The deep concavity of the loreal region is as pronounced in the young as in the adults; but while in the young the nostril is midway between the eyes and the end of the snout, in full-grown frogs the nostril is once and a half to twice as far AFRICAN HERPETOLOGY 201 from the eye as from the end of the snout. In very young specimens, and in one adult male, the tibio-tarsal articulation of the adpressed hind limb only reaches the end of the snout. The skin sometimes exhibits scattered flattened warts and in one presents quite a granular appearance. Coloration. In life. o". Bagilo. Above, dark olive-green, lighter on thighs, the whole freely sprinkled with white ; black blotches on back, and more or less well-defined bars on limbs. Beneath, grey, darker on throat, pahns and soles, Ughter on chest, abdomen and thighs. 9 . Sigi below Amani. This very reddish individual was taken in the river- bed on reddish soil. The larger o" from Amani was dark olive when found under a log on the hillside about thirty yards from a stream. It was put into a white linen bag to be taken back to camp; on removing it from the bag it was found to have changed to a pale leaf-green. Measurements. The largest male (Bagilo) measures 74 mm.; largest female (Vituri) 59 mm. Smallest frog (Bagilo) 17 mm. Breeding. The Vituri female was full of eggs ; only one other large female was taken. Most of the Bagilo specimens were very young and were taken between the 15th and 20th of September. Diet. The examination of eight stomachs gave the following results : (i) Grass- hopper, (ii) Wasp, (iii) Beetles, (iv) Beetles and fly pupa, (v) Long-snouted hemipteron lantern fly. (vi) Four neuropterous nymphs and a large spider, (vii) Big spider, (viii) About twenty acacia leaf petals such as might have been floating on the surface of a pool. Parasites. Worms of the family Physalopteridae were found in the stomach of an Amani frog. Phrynobatrachus krefftii Boulenger Phrynobatrachus krefftii Boulenger, 1909, Ann. Mag. Nat. Hist., (8) 4, p. 496. 100 (M. C. Z. 52736-60) Amani, Usambara Mtns., 25-30. .xi. 26. 46 (M. C. Z. 52761-75) Mt. Lutindi, Usambara Mtns., xii. 26. 3 (M. C. Z. .52776-78) BumbuU, Usambara Mtns., 12. xii. 26. 7 (M. C. Z. 52779-85) Phillipshof, Usambara Mtns., 21. xii. 26. Variation. Some of the above topotypes from Amani have been carefully compared with the female cotypes in the British Museum. The digital disks are very strongly developed in this Phrynobatrachus. In the types the tibio-tarsal articulation reaches to the end of the snout, or beyond. In the present series in some examples it only just reaches the eye, in others the snout or beyond, nor can this character be used for separating the sexes. There is, however, a remark- 202 BARBOUR AND LOVERIDGE ably interesting difference between the two, for in the males not only is there the ranid thickening of the thumb-pad, but the tubercles of the toes and metatarsal region have developed into sharp little spines. Curiously enough there are little spinose points on the backs of the males, while the dorsal skin of the females is perfectly smooth. These characters are probably of use only in distinguishing sexually mature males, as none of the small frogs seem to have them. Coloration. In life. &. Above, olive with brown markings, slightly greenish on head; a fawn band, pinkish at its extremities, unites the upper eyelids; it is sometimes backed with a darker one; indications of another band at right angles to this one commence on the snout; hind limbs faintly barred with black, some white specks on sides of head, fore limbs and flanks. Below, upper aspect of lower jaw white, with minute brown specks; under aspect, a sharply distinct black hne running right round jaw; throat bright crome, a dusky band across chest between fore limbs, breast and belly satiny-white with a few brown flecks, under surface of limbs yellowish-green with brown spots. Length 40 mm. 9 . The bright chrome-yellow throats are found only in the males ; in the females they are white or greenish, as also in young immature frogs. The amount of marbling on the lower surface varies very much. Some females have a light band along either flank and a light band beneath the eye. Measurements. The largest male measures 50 mm. ; largest females all 40 mm. (type also 40 mm.). Smallest frogs 10 mm., at which size they first lose their tails. Breeding. At Amani. After a heavy shower just before sunset on December 2nd, we heard the frogs calUng energetically from the bottom of a ravine not very far from the Institute. We hastened down as the light was fading, and located the frogs in a trickling stream. The males were calling with vocal sacs inflated and we were able to watch them at close quarters. Others were in embrace and seemed to favour spots where water flowed beneath rocks, under which they would retire at the last moment. Spawn was found which already held small tadpoles; other free-swimming tadpoles with their hind legs showing were also collected. From subsequent observations it would appear that at this time of year, at any rate, the frogs may collect for spawning after any heavy shower. Description of Tadpole. Length of body nearly twice the width; about one half the length of the tail. Nostrils slightly nearer tip of snout than eye. Eyes dorso-lateral. Distance between eyes somewhat greater than the width of the mouth. Spiracle on the left side, directed upward; much nearer the eye than to the posterior end of the body, visible from above and below. Anal tube horizon- tal, in the lower edge of the subcaudal fin, shghtly bent to the right. Tail slightly more than four times as long as deep; sharply pointed; upper and lower fins AFRICAN HERPETOLOGY 203 apparently rather narrow but poorly preserved. Buccal disk rather large and transverse. Upper lip slightly arched and edged with small horny teeth. Lower lip and sides fimbriated and with many blunt papillae. Two rows of fine horny teeth between upper lip and horny beak, which is wide, arched, and well developed, considerably overlapping the well-developed but much smaller lower beak. Pos- terior to the beak, three long parallel series of horny teeth. The mouth disk thus differs widely from Power's figure of P. natalensis (Trans. Roy. Soc. S. Afr. 14, 1927, p. 239, fig. 3) in having rows of denticles between the marginal row on the upper Hp and the beak and the three long parallel rows posterior to the beak. Color. Greyish above, whitish or transparent below. Tail finely speckled with dusky brown. Diet. The examination of ten stomachs gave the following results: (i) Beetles. (ii) Beetle, (iii) Beetle larva, (iv) Larvae and spider, (v) Spider and ants, (vi) Large black ant, moth, hemipteron. Cvii) Hemipteron. (viii) Hemipteron. (ix) Weevil, millipede, small crab, (x) Cockroach. Habitat. The first specimen, a male, was taken in a banana plant! Whether this is usual in the non-breeding season I cannot say; it might explain the well- developed disks, but not the spines on the feet, which are sexual. No others were found in bananas. Over fifty were taken in a stream which cuts the Com- coro road or path ; three were found in a marsh and many in the trickle of water in a ravine between the Institute and native village. Others were seen in a pool in the forest just below Derema house. Those in the slow-flowing broad stream were usually found sitting in sunlit patches on spits of sand or exposed rocks. \\Tien approached they remained still for longer than most frogs but were exceedingly clever at hiding when once they moved. Generally they dived into the clear water and hid beneath a leaf or remained motionless on the gravelly bottom, where their colouring was admirably concealing. At Misalae they were seen in a stagnant stream which was exposed to the full glare of the sun as it flowed through a native plot. At Kizerui they occurred in pools and streams in the depths of the forest. Phrynobatrachus acridoides (Cope) Staurois acridoides Cope, 1867, Jour. Acad. Nat. Sci. Philad., 6, p. 198 (Zanzibar). Phrynobatrachus boulengeri de \\'itte, 1919, Rev. Zool. Africaine, 6, ii, p. 225. (Beira and Coguna, P. E. A.); Procter, 1920, Proc. Zool. Soc. London, p. 413. Phrynobatrachus natalensis (part) Loveridge (not of Smith) 1925, Proc. Zool. Soc. London, pp. 779-781. Arthroleptis janenschi Ahl, 192.3, Sitzber. Ges. Naturf. Freunde, Berlin, p. 100. (Tendaguru, near Lindi, T. T.) 60 (M. C. Z. 12805-910) Dar es Salaam, 6 and 9. xi. 26. 18 (M. C. Z. 12811-815) Bagamoyo, 11. xi. 26. 204 BARBOUR AND LOVERIDGE Relations. Thanks to the kindness of Mr. H. W. Parker we have recently received a cotype of P. boulengeri from Beira, Portuguese East Africa, which we have compared with Cope's original types and with the Dar es Salaam speci- mens listed above and find identical beyond the slightest question. Zanzibar, the type locality of P. acridoides , is almost opposite Bagamoyo on the coast and not more than forty miles from Dar es Salaam. Nieden (1915) referred 71 specimens from Zanzibar, Bagamoyo and Dar es Salaam to acri- doides, an undoubtedly correct decision. A comparison of the description of P. acridoides, with the fuller descriptions of P. boulengeri and A . janenschi shows that they are all in absolute agreement with the characters given by Cope except that Ahl's janenschi is said to lack a conical papilla on the tongue, a feature that may well have been overlooked. Ahl's description is, indeed, a very excellent one of P. acridoides from Dar es Salaam. In 1924 Noble, ' on the basis of the variability of his huge series of Phryno- batrachus natalensis from the Congo and the published description of P. boulen- geri, placed the latter in the synonymy of the former. This course was followed by the junior author in his 1925 paper cited above. The receipt of the cotype of P. botdengeri confirmed the conclusion already arrived at that P. boailcngeri was a much smaller frog ; Dr. Noble fully concurs in the conclusion that it is distinct from P. natalensis. Distribution. East African specimens in the collection of the Museum of Comparative Zoology are from the following localities: Kenya Colony: Frere Town, near Mombasa. Zanzibar: Zanzibar (cotypes of Staurais acridoides Cope). Tanganyika Territory: Gonya, Bagamoyo, Dar es Salaam, Duthumi, Tulo, Morogoro, Ivilosa. Portuguese East Africa: Beira (cotype of P. boulengeri); Masiene, near Chai-chai, Limpopo River. Tendaguru is near Lindi on the littoral between Dar es Salaam and Beira. All the locahties recorded above are on the coastal plain, or have a typical coastal fauna and an altitude of less than 2,000 feet. Variation. The key character of the interorbital space, being narrower than the upper eyelid (acridoides) or a little broader (botdengeri) , is of little use, for the interorbital space in the present series equals the upper eyelid or is a little nar- rower or a little broader. The tympanum, which is said to be distinct in the type, is just distinguishable below the postocular fold; toes, said to be two-thirds webbed in the type, are almost fully webbed in the present series except the last two joints of the fourth toe; the tips of the fingers and toes, said to be dilated in the tjTie, are usually not dilated though sometimes slightly dilated; the tibio- tarsal articulation of the adpressed hind limb reaches the posterior border of the ' Noble, 1924, BuU. Am. Mus. Nat. Hist., 49, Art. ii, pp. 188-191. AFRICAN HERPETOLOGY 205 tympanum or the tip of the snout or any point between these extremes; skin folds from the posterior border of the eye to the fore arm are conspicuous, but the convergent, then divergent, plicae from the orbits to the scapular region are scarcely distinguishable. Coloration. A dorsal stripe, narrow or broad, present or absent. Juvenile specimens from a sandy water hole near Dar es Salaam were so extraordinarily like the grains of sand in colour that they could rarely be detected until they moved. Measurements. The largest specimen, a female from Bagamoyo, measured 28 mm. ; the smallest frog (Dar es Salaam) measured 11 mm. Breeding. Nearly all the Dar es Salaam frogs are very small but exhibit no trace of tails, yet their presence showed that the breeding season was during the ' big rains ' — March-May. Diet. Beetles. Habitat. They were found in the grass bordering the irrigation ditches in the CathoUc Mission plantation, also around a small water hole dug in a sandy hol- low near the sea and some three miles south of the town. When sufficiently alarmed the young frogs sprang into the water, but quickly struck out for the edge as if instinct warned them that the dangers of deep water might be greater than those of the land. PHR-iT>fOBATRACHUS ? OGOENSIS (Boulcnger) Arthroleplis ogoensis Boulenger, 1906 (for 1905) Ann. Mus. Stor. Nat. Geneva, (3) ii, p. 162, PI. 1, figs. 7 and 8. (Lambarene, Ogow6.) Loveridge, 1925, Proc. Zool. Soc. London, p. 78.3 (Bagilo). 86 (M. C. Z. 12786-790) Bagilo, Uluguru Mtns., ix. 26. 50 (M. C. Z. 12791-795) Nyange, Uluguru Mtns., x. 26. 17 (M. C. Z. 12796-800) Mkangazi, Uluguru Mtns., x. 26. 4 (M. C. Z. 12801-804) Nyingwa, Uluguru Mtns., 19. x. 26. Also seen at Mkarazi, Uluguru Mtns., T. T. Relations. The two Bagilo frogs referred to Arthroleplis ogoensis by the junior author in 1925 are specifically identical with the present series from the Uluguru Mountains. ]Mr. H. W. Parker has sent us a topotype of A. ogoensis which has been examined by Noble, who was working here on amphibian generic characters at the time. He finds that it possesses the bony sternum and omosternum of Phrynobatrachus as defined by Hewitt, so that hereafter the species should be known as Phrynobatrachus ogoensis f Boulenger). Arthroleplis rouzi Nieden (Buddu Forest, Uganda) has already been placed in the synonymy of P. ogoensis by the junior author. A careful perusal of the description of Phrynobatrachus kinangopensis Angel from Kinangop, Aberdare Mountains, Kenya Colony, shows that, like A . rouxi and the topotjqae of ogoensis 206 BARBOUR AND LOVERIDGE referred to above, it has the first finger shorter than the second ; apart from this there is no difference between the two species except that kinangopensis is said to have the toes three-quarters webbed, and ogoensis nearly one-half webbed; in both it extends as a fringe nearly to the end or to the end. Without topotypic material we hesitate to consider kinangopensis definitely as a synonym. There is, in fact, one doubt in referring Uluguru frogs to ogoensis. They agree in every respect when compared with the topotype except in colouring, for like rouxi and kinangopensis it possesses a vertical stripe, yet one of the tj^es (fig. 8) also lacks this stripe and is coloured above exactly like the Uluguru frogs. It seems strange, however, that in such a long series — over 150 specimens — not a single Uluguru frog should have a dorsal stripe. West African frogs appear to be more spotted beneath than those from the Uluguru, where the throat only is be- sprinkled or blotched. Variation. The sternum and omosternum are definitely bony. Digital disks are usually present though occasionally lacking; in the Nyingwa series, for exam- ple, both conditions are found, though the frogs are beyond question of one species. Minute spines on the back are more numerous in the males than in the females. Coloration in life. Mkangazi specimen. Above, back olive, limbs yellowish, the whole variegated with dusky marks, some of which form bars on the limbs. Below, throat and belly satiny-white, the former characteristically speckled or marbled with black as well as the flanks, the edge of the lower lip presenting a chequered appearance. Measurements. The largest specimen, a female from Nyange, measures 27 mm. None of the Bagilo series measure more than 25 mm. The smallest frog (Bagilo) measures 11 mm. and has a half-absorbed tail 5 mm. long. Breeding. Nyange frogs were calling day and night in a marsh close to my tent; this was during the first eleven days of October. Females from all localities were distended with ova. Young measuring 12 mm. in length were taken at Nyange and Mkangazi. Evidently this species spawns twice a year in these mountains, namely, during the greater and lesser rains. Possibly the spines referred to above are present only during the breeding season. Diet. Beetles. Parasites. Red subdermal parasites are so frequently present upon frogs of this species in the Uluguru Mountains that they might almost be employed as an aid to identification! Enemies. A large frog from Nyange has the right leg missing from well above the knee. AFRICAN HERPETOLOGY 207 Habitat. At Bagilo, in quiet pools and backwaters of a swift mountain torrent. At Nyange, in soggy, poorly drained land with small stagnant pools which would normally dty up but for their renewal by frequent showers, as at the time of my visit. Distribution. No sign of this species was seen in the Usambara Mountains, where it is replaced by the much larger P. krefftii. This is rather surprising, as the climatic conditions and enviroimient appear ideal and one can but suppose that it is the competition of its near relative that has prevented its obtaining a foothold. Aktheoleptis stenodactylus Pfeffer Arthrokptis stenodactylus Pfeffer, 1893 (1892). Ivihengo, Tanganyika Territory. Arthroleplis variabilis Matschie,' 1893. Buea and Barombi, Cameroon. Arlhroleptis whijtii Boulenger, 1897. Kondowe to Karonga, Nyasaland. Arlhroleptis lonnhergi Nieden, 1915. iSIombo, Usambara, Tanganyika Territory. Arlhroleptis methneri AM, 1924. Matumbi-Holen, Tanganjdka Territory. A rthroleptis brevipes Ahl, 1924, Bismarkburg, Togo. Arlhroleptis wahlbergii Procter (not of Smith) 1920, Proc. Zool. Soc, p. 414. Morogoro and Amani, Tang. Territory. Loveridge, 1925, Proc. Zool. Soc, p. 718. Uliya-Madazini Road, Morogoro District, Tang. Territory. 6 (M. C. Z. 13111-5) Bagilo, Uluguru Mtns., i.x. 26. 13 (M. C. Z. 13116-20) Nyange, Uluguru Mtns., .x. 26. 1 (M. C. Z. 13121) Mkangazi, Uluguru Mtns., 12. x. 26. 15 (M. C. Z. 13122-6) Nyingwa, Uluguru Mtns., x. 26. 2 (M. C. Z. 13127-8) Vituri, Uluguru Mtns., x. 26. 4 (M. C. Z. 13129-32) Dar es Salaam, 4. xi. 26. 9 (M. C. Z. 13133-7) Amani, Usambara Mtns., xi. 26. 1 (M. C. Z. 13138) Mt. Lutindi, Usambara Mtns., 10. xii. 26. 1 (M. C. Z. 13139) Bumbuli, Usambara Mtns., 12. xii. 26. Variation. In 1920, when reporting on the junior author's collection of am- phibia from Central Tanganyika Territory, Miss J. Procter referred specimens of this genus from the Uluguru and Usambara Mtns. to A. wahlbergii, A. steno- dactylus and A. whytii. In 1925, in a paper on a second collection, the junior author followed Nieden and others in placing whytii in the synonymy of steno- dactylus and recorded A. wahlbergii, A. stenodactylus and A. adolfi-friederici from various locahties in the iNIorogoro District. During the present expedition 276 frogs of this group were specially collected with a view to elucidating the relationships of these species. A. stenodactylus is evidently a frog with a wide range, both horizontal and vertical; on the pres- ent occasion it was taken at sea-level (Dar es Salaam) to 7,500 feet (Nyingwa). East African examples appear inseparable from Cameroon specimens in the col- lection of the Museum of Comparative Zoology. As a matter of fact A. steno- dactylus has also been reported from Spanish Guinea. ' Recorded by .4ngel in 1925 as occurring at Gazi, Ivenya Colony. 208 AFRICAN HERPETOLOGY ■5: C 1 1 o E 03 c 1^ M ■4^ cr m >> a _o c 03 J3 *j 1 rounded, as long as hori- zontal diameter of the eye c 03 D, 03 •s .s- 00 03 O 1? a 2- — 3 t- O 03 C (»»+-. oi O i 2 ^. S a 05 '■B -«&• b« *-• 03.», J3 O a ■i 'S. 1st longer than 2nd, 3rd one and one-half times as long as 2nd, 4th one-half as long as first -0 ■S t a +3 _B Si papilla-like, inner meta- tarsal tubercle large, shovel-shaped smooth above and below, sides covered with small warts 1 1 c Oi .— 1 1 to c o.S o ^ 03 C o 3 3 - = = = 3 ■g 'p. a s 1 a; 3 - 3 = a 'i 1 00 3 oe-3 ^§ sij 1.1 S 1 0 33 1 C3 b4 O -O M O C a> >, 0) M C 03 s 3 2 3 ■§ C « aj 2 « il 4) *^ J^ 3 ;3 o c3 CO &. a 3 3 03 -3-a 03 3 2 Si -S CO o o X (-• P. a 1" ^ 3 3 * c3 c *j 03 ■-I *i 3 33 0 & tn _>. 1 •« 1 ■a c, 0 1 1.2 £ 3 |g §2 > 2 « (s-i a _o cc 0 D. a * n 3 0 s 11 ^- 03 il .a o CO o 00 '2 a 2 OJ 03 §2 oi a 3 c3 030 1 1 S3 CO O c 1 c E 1 1 1 1 '■B 3 0. ■5. 03 ■a 3 1 0^ bC 3 _o i 3 03 '■*^ 0 a •*^ 3 ~S II CO en 1 1 ■3 03 a 3 1 1 00 GO c3 C c 03 H as 2e2 1 <0 a 1 1 1 ■3 0 3 0 "3 3 9 >> 3 c^ g u C 5" 0. =3 2 §§ cr a> 0 3 1 m c3 1 Si .^ B-a 3 fcltl M o 1 •s 1 3 c 03 X 3 O 3 72 3 *^ c 6 't3 o g a CO 3 2 S a a 3 3 0! C a >> 0) 3 0 H i 3 £ ts£ 3 CC <^ 0 s. c. (-1 ■§1 0 c 3» +j . 0 2 5 03 tn 1 "« 3 ©"a -3 *^ a 13 03 BARBOUR AND LOVER IDGE 209 It then transpired that the key characters supposed to differentiate several alleged species were common to examples from several localities where large series were collected. For example, the tibio-tarsal joint of the adpressed hind limb is said to reach the tympanum or posterior border of the eye (whytii), be- tween tympanum and eye (brevipes) , eye (stenodactylus, lonnbergi, methneri) ; in series it may only just reach the tympanum or as far as the middle of the eye, a sexual rather than a specific character, the shorter limbed frogs being females. The papilla on the tongue is absent in several half-grown frogs or is, at least, not to be detected, while it is present in other frogs from the same locality; in three out of the four Dar es Salaam specimens it seems to be lacking. The ab- sence of this papilla was said by Nieden to be the only distinguishing feature differentiating lonnbergi from stenodactylus, while formerly it was supposed to be absent in the type of that species. Sometimes this papilla, as it lies in the Uttle pit from which it springs, is flush with the level of the tongue and its prominence appears to depend not only on age but a good deal on the state of preservation of the specimen. The shovel-like metatarsal tubercle so prominent in adults is small, but quite distinct, in the young (cf. examples from Victoria Falls, Zambesi River, W. S. Brooks leg. ; curiously enough no very young specimens were met with during the present expedition) and serves to distinguish them at once from the young of A. adolfi-friederici, which has a rounded, more oblong tubercle. It is, therefore, advisable to redescribe the species on the basis of this series only, and it will then be found to embrace every variation cited in the de- scriptions of the five species which we relegate to the synonymy of stenodactylus. Description. A large species of stout habit, rather squat. Head moderate, broader than long, varying with sex; snout rounded, as long as the horizontal diameter of the eye; canthus rostrahs obtuse or distinct (sharper in the young); nostril midway between eye and end of snout or a little nearer the end of the snout than the eye (both conditions occurring in adults from the same locality) ; interorbital space equals the breadth of the upper eyeUd or is slightlj^ narrower or slightly broader; tympanmn distinct, one-third or two-thirds the diameter of the eye (its size develops with age as in Bujo) ; tongue with, rarely without, a conical papilla situated in a small pit on the front. Fingers slender, unequal, their tips obtuse, sUghtly swollen but not dilated, 1st and 2nd equal, or 1st sUghtly longer or sUghtly shorter than 2nd, 1st longer than 4th, 3rd a Uttle longer than 1st in females, once and a half to twice as long in males. Tips of toes sUghtly swollen, not dilated; toes without webs or with 210 BARBOUR AND LOVERIDGE the very slightest trace; a single inner metatarsal tubercle, large, shovel-shaped as in Rana delalandii, as long as, or slightly longer than, the inner toe; no outer metatarsal and no tarsal tubercle; subarticular tubercles on fingers and toes well developed, papilla-hke; the tibio-tarsal joint of the adpressed hind limb reaches the tympanum or as far as the eye. Skin smooth above, below smooth or subgranular, sides granulate or slightly warty; very occasionally a narrow vertebral ridge from snout to anus. Coloration. In life. An adult breeding female from Nyingwa is fairly tjqjical, though great range of variation occurs and the vertebral line is generally wanting. Above, pale brown with a yellow, thread-like, vertebral line commencing on praef rontal region and continuing past the anus to the lower surface ; it is crossed at right angles on the anus by a similar line extending across the hind limbs from metatarsus to metatarsus; the vertebral line is flanked on either side by dark brown, light-edged chains of markings typical of the genus; other light-edged, irregular, darker brown blotches towards the flanks ; a very black canthal band, light-edged above, passes from end of snout through nostril and eye over the tympanum, where it terminates on a level with the buccal border; lips brown but heavily blotched with white; limbs barred with brown and speckled with white. Below, white with some brown on throat; two noticeable, elongate, brown blotches between fore arms, and brown marblings on flanks. Pupil black surrounded by a narrow gold ring ; iris yellow, or golden, obscured by black specklings. Five of the Amani specimens have immaculate throats, whereas all Uluguru frogs have the throat marbled. All the Dar es Salaam frogs (females) are immaculate on the throat, while two of them have each a broad, light, vertebral band. Measurements. Largest males (Nyingwa, Amani, JVIt. Lutindi) measure 35 mm. ; largest females (Amani) 44 nmi., others (Nyange) 42 mm. in length. The smallest male (Nyingwa) was 24 mm. and smallest female (Dar es Salaam) 20 mm. Sexes. The sexes showed a marked disproportion, only fifteen males being taken as against thirty-seven females. The males may be distinguished readily by: (i) Much longer 3rd finger, which is one and a half to twice the length of the 1st. (ii) Smaller size, not exceeding 35 mm. in this series, (iii) Darkly pigmented throat or chin. (iv) Loose skin of vocal sac on throat. AFRICAN HERPETOLOGY 211 Breeding. Two females and a male were brought to me at Nyingwa on Octo- ber 18th, 1926, which had been taken in shallow burrows in the native gardens, together with their eggs. These numbered 33, 40, and 53 respectively, while a fourth lot numbered SO. In the last instance it may well be that two lots of eggs were lumped together, for the collector was uncommonly stupid and, disregarding instructions, had put the parent or parents in a receptacle with other frogs of the same species. Until a European verifies the taking of a male with the eggs it might be as well to accept with reserve the fact of a male guarding the eggs. When at Vituri a fortnight later, 1 noticed a movement in some recently turned earth in the vicinity of my tent. Supposing some Lygosoma or Scelotes to be moving it, I pounced upon the spot and captured an Arthroleptis steno- dactylus; upon turning her over the mass of eggs was easily seen through the transparent abdominal wall. She was placed in a tin of earth and immediately began to work herself backwards into the loose soil ; her behaviour struck me as being very different from that of an imprisoned Rana, which would usually explore its prison and make every effort to escape by leaping against the glass cover, whereas the Arthroleptis settled down immediately. The ovaries of Bagilo and Nyange frogs examined in September were develop^ ing well, indicating that breeding occurs from October to November in the Uluguru Mountains. Diet, (i) Large orthopteran, related to Centrophilus (^Ir. Nathan Banks in- forms us), cockroach and skipjack beetle, (ii) Grasshopper, cockroach and beet- les, (iii) Grasshopper, cockroach and cricket, (iv) Two grasshoppers and beetles, (v) Grasshopper and beetle, (vi) Grasshopper, (vii) Cockroach, spider, mollusk. (viii) Acridian, centipede, (ix) Centipede and beetles, (x) Geophihd, GO mm. in length. Parasites. Red subdermal organisms on Bagilo frogs. Enemies. Twice recovered from stomachs of Natrix olivaceus at Nyange. A Chlorophis neglectus, captured at Bagilo on 19. x. 26, held a frog in its mouth. Three were taken from the stomach of a single Crotaphopeltis hotamboeia tornieri at Nyange. Others were recovered which were doubtfully referable to this species or A. adolfi-friederici. Habitat. The largest of the Dar es Salaam frogs was taken in a water-holding bamboo stake in the Botanical Garden. Most of the other specimens were cap- tured beneath logs in the rain-forest, or brought in by natives who had hoed them up in their gardens. 212 BARBOUR AND LOVERIDGE Arthroleptis adolfi-friederici Nieden Anhroleplis adolfi-friederici Nieden, 1910, Sitzber. Ges. Naturf. Freunde Berlin, p. 440. 22 (M. C. Z. 13140-44) Bagilo, Uluguru iMtns., ix. 26. 2 (M. C. Z. 13145) Nyange, Uluguru Mtns., 11. x. 26. 2 (M. C. Z. 13146-47) Nyingwa, IHuguru Mtns., 6 and 18. x. 26. 6 (M. C. Z. 13148-52) Vituri, Uluguru Mtns., x. 26. 190 (M. C. Z. 13153-60) Amani, Usambara Mtns., xi. 26. . 5 (M. C. Z. 13161-65) Mt. Lutindi, Usambara Mtns., 10. xii. 26. 2 (M. C. Z. 13166-67) Phillipshof, Usambara Mtns., 24. xii. 26. Relations. Nieden compared this frog with A. whytii Boulenger ( = steno- dactylus) , to which it bears a striking colour Ukeness and general resemblance. Constant features in which they differ, as demonstrated by the present series, are the longer hind limbs of adolfi-friederici, reaching to between the eye and snout or, in the case of males, not infrequently beyond; the rounded, not shovel- shaped, metatarsal tubercle, which is shorter than the inner toe; the well-devel- oped disks of both fingers and toes (these strong disks app>ear in all ages and while, perhaps, usually rounded, they are very frequently sharply pointed as in A. xenodadylus Boulenger). Coloration. The coloration is very variable, more noticeably so in life, when it tends to resemble the tones of the enviroimient. Three half-grown Bagilo frogs and one Nyingwa frog exhibit a broad, white, median, dorsal stripe. In most specimens the middle of the belly is immaculate white but in some it is marbled like the throat and lower sides. At Amani the variation exhibited was astonishing. Where bamboos shade the path, and bestrew it with their long white leaves, the frogs are white, or whitish, but retain their black markings. Fifty yards away, where the dead leaves are rufous-stained through contact with the red soil of which the path is composed, the frogs, particularly the smaller ones, were bright reddish-brown. On entering the nearby forest where black soil predominated, frogs were less plentiful but very dark in colour. Measurements. Largest male (Bagilo) measures 32 mm., next largest (Phil- lipshof) 31 mm.; largest female (Amani) 42 mm., range of largest from every locaUty 38 to 42 mm. Smallest frogs at Amani, 8 mm.; at Bagilo, 12 mm. Sexes. An interesting sexual difference, presumably developed for the breed- ing season, appears in one Bagilo male (13143) and one Philhpshof male (13167), where the second and third fingers exhibit on their inner side a series of tooth- like granules as figured by Boulenger for A. poecilonotus (1906, Ann. & Mag. Nat. Hist., Ser. 7, Vol. 17, p. 320, fig. 1). While the disks of the Bagilo male are sharply pointed, those of the Philhpshof frog are roimded. AFRICAN HERPETOLOGY 213 Diet, (i) Big black ant. (ii) Two species of caterpillars, (iii) Cockroach, (iv) Cockroach, (v) Cockroach, (vi) Cockroach and big spider, (vii) Spider, (viii) Freshwater shrimp. Apparently the cockroaches, which are very common among the dead leaves frequented by the frogs, form one of their principal articles of diet. Enemies. One from the stomach of a Crotaphopeltis hotamboeia tornieri at Vituri on 27. x. 26. Habitat. The species seems to court sunlight at j^^mani, where the majority were captured in the mornings on leaf-strewn paths exposed to the full glare of the sun. See also remarks under coloration. Arthroleptis schubotzi Nieden Arthrolepiis schubotzi Nieden, 1910, Sitzber. Ges. Naturf. Freunde Berlin, p. 440. 5 (M. C. Z. 13168-72) Bagilo, Uluguru Mtns., 16. ix. 26. Variation. Described in 1910 from a single specimen taken in the Usumbura Mtns., of western Tanganyika Territory, it has since been recorded from Langen- burg and west of Mpapua. Without any authentic examples for comparison, though the series agrees perfectly with the somewhat meagre description, the identification should be received with reserve. If, however, they are correctly determined, the author appears to have overlooked what is perhaps the most distinctive feature of this small species, viz., the nimierous, almost spinose, tuber- cles on the tibia and foot. All agree with the type in that the tibio-tarsal articula- tion of the adpressed hind limb reaches the hinder part of the eye. The very shght dilations of the toes, and those of some of the fingers also, are more or less pointed in all five frogs. Coloration. No colour description was given by Nieden. The following was made on day of capture. Above, grey, a somewhat indistinct chain of markings from the snout along the back to lumbar region, a black blotch above anus; two very striking black marks, one on either side, just in front of hind limbs; Ups black, marbled with white ; fingers and toes barred black and white, limbs barred with black especially noticeable on the tibia, bright red on loins and posterior aspect of thighs. Below, white, mottled or marbled with grey everywhere, but least conspicuously on the belly. Measurements. The series measures 19, 19, 17, 17 and 16 mm. respectively; the type measures 21 mm. Diet. The stomach of the only specunen examined contained a very small grasshopper. 214 BARBOUR AND LOVERIDGE Parasites. Small red mites on the thighs of one or two of the frogs. Habitat. Among fallen leaves on a damp path in a very wet section of the rain-forest. Arthroleptis xenodactylus Boulenger Arthroleptis xenodactylus Boulenger, 1909, Ann. Mag. Nat. Hist., (8) 4, p. 490. 13 (M. C. Z. 13173-177) Bagilo, Uluguru Mtns., ix. 26. 3 (M. C. Z. 13178-179) Nyange, Uluguru Mtns., 6 and 7. x. 26. 6 (M. C. Z. 13180-185) Mkarazi, Uluguru Mtns., 22. x. 26. 152 (M. C. Z. 131S6-209) Amani, Usambara Mtns., xi. 26. 1 (M. C. Z. 13210) Phillipshof, Usambara Mtns., 27. xii. 20. Variation. The character on which so much stress has been laid, that of the sharply pointed digital disks, is not specific: though it is usually present, many frogs have only rounded disks. The snout from the nostril is invariably s/iorter than the eye, much shorter in very young frogs. In both sexes the nostril is somewhat nearer the end of the snout than the eye, not equidistant. The first finger is much shorter than the second (stated to be shorter in the original description) and should therefore cause the species to fall into another section of Noble's key (1924, Bull. Am. Mus. Nat. Hist., vol. 49, Art. ii, p. 199). The tibio-tarsal articulation of the adpressed hind Umb reaches between eye and snout (as in type) almost invariably, but in occasional specimens it just falls short of the eye. Coloration. In the field the junior author thought that he had two distinct species, so different are the sexes in appearance and colour. Some with flattened heads and very depressed snouts, which lack the typical chain of vertebral mark- ings characteristic of the genus, being brown above (in alcohol), appear to be all adult females. The description of a Bagilo frog as taken down in the field was as follows. Above, yellowish-brown marbled with darker and stippled with white, a large orange blotch on the back just above the anal orifice (this spot spreads to right and left on to the thighs) ; sides dark brown separated from the dorsal colouring by a light pinkish-white line commencing on the snout and passing over the nostril and eye along the side to the hind limb; thighs and inner aspect of tibia of the same orange shade as the anal patch. Below, white marbled with greenish-grey, soles of feet darker, almost black. Others were found in which, in addition to the large orange blotch, the typical vertebral chain of markings was also present, but the majority lacked the orange blotch and possessed the chain of dorsal markings. The adults of this type which were examined all proved to be males, and it appears to be the universal colouring AFRICAN HERPETOLOGY . 215 of the young also; presumably the females lose the markings as they become adult. Frogs of this type of coloration agree very well with the description of A. reichei Nieden except in the absence of papillae and in their size. A. reichei is evidently a much larger species, measuring 29 mm. in the type, while the type of .4. xenodactylus was only 17 nmi. The colour in life of a Nyange frog of this male type was: Above, greenish- grey, a chain of brownish vertebral markings typical of the genus from snout to vent; groins and tibia rich vermilion, rest of exposed parts of limbs greenish-grey marbled with darker, one or two indistinct bands on the tibia; jaws and sides of body white, marbled or speckled with brown. Below, whitish, finely speckled on throat ; under surface of loins and hind limbs vermilion to pink. Measurements. The following measurements were taken of the last-men- tioned individual : Snout to vent 13.5 mm. Length of tibia 10 mm. Length of head 6.75 " Length of foot 15 " Breadth of head 7 " Length of ith toe 6 " Diameter of orbit 2 " Length of 3rd finger 3 " Largest male measures 22 mm.; largest female 20 mm.; smallest frog 6 mm. (all •Amani). In the Uluguru series the largest male (Nyange) measured 21 mm., and female (Bagilo) 19 mm.; smallest frog (Nyange) 7 mm. Breeding. Many of the Amani frogs hold a few large ova, disproportionately large for so small a frog. Some of these frogs were found in bananas, and the eggs were plainly visible through the transparent skin of the belly. Diet. Ants were found in one, and numerous amphipod crustaceans in two others, but in general the stomach contents are too small to be readily identified. Habitat. Single specimens were found hiding under logs at Bagilo, Nyange and Mkarazi. The great bulk of the Amani series were taken hopping about among dry leaves in a shady grove of chinchona trees on Mt. BomoH. The leaves in this grove were very dry on the surface but in some spots were slightly damp underneath, yet the whole place was Uterally alive with frogs of the genus Arthro- leptis, of which four species were present. Others again were taken on paths in dry forest, and after a shower of rain some were found in long grass. Salimu secured nearly twenty fully grown adults in bananas ; this may have some signi- ficance, as the greater number of those taken beneath the chinchona trees were very young or haK grown- Enemies. One recovered from the stomach of a young Crotaphopeltis hotam- boeia tornieri. 216 BARBOUR AND LOVERIDGE Abthroleptis minutus Boulenger ArthrolepHs minutus Boulenger, 1895, Proc. Zool. Soc. London, p. 539, PI. XXX, fig. 4. 9 (M. C. Z. 12826-30) Amani, Usambara Mtns., xi. 26. 10 (M. C. Z. 12831-35) Phillipshof, Usambara Mtns., 24. xii. 26. Variation. These specimens answer perfectly to the description oi A. parvulus Boulenger from Angola. Since writing the foregoing some of these frogs have been submitted to Mr. H. W. Parker, who has kindly compared them with the types of minutus and parvulus; he says that they 'certainly agree better with parvulus than with minutus,' but points out that, without carefully studying the whole genus to appreciate the relative values, it is almost impossible to give a verdict as to the specific identity or otherwise of A. minutus and A. parvulus. He continues: 'They look different to me, minutus being larger and heavier, but I have found no definite character by which they can be distinguished. Before I could state definitely that they were conspecific I should like to see material from the inter- vening territories; all our specimens of A. parvulus (8) are from Angola, whereas all those of A. minutus (12) are from the eastern side of the Great Lakes. In pass- ing, the record of the latter species from Portuguese Guinea (Boulenger, 1906, Ann. Mus. Civ. Geneva, (3) 2, 42, p. 161, Noble, 1924, Bull. Am. Mus. Nat. Hist. 49, ii, p. 316) is an error; the specimen appears to belong to A. qutterosus Chabanaud.' Hyperolius spp. Owing to the unsatisfactory condition of this genus, its bad need of revision, the meagre descriptions of many of the species, and the difficulty of distinguishing species which have little but colour to separate them, though doubtless perfectly distinct, it is quite possible that some of the following examples of concolor, mariae and puncticulatu^ are wrongly assigned. A large series was collected in the hope that it might clear up some of the existing difficulties, but the reverse rather seems to be the case. Hyperolius concolor (Hallowell) Ixalns concolor Hallowell, 1844, Proc. Acad. Nat. Sci., Phila., 2, p. 60. 1 (M. C. Z. 13261) Mkarazi, Uluguru Mtns., 22. x. 26. Variation. Structurally in agreement with topotypic material, the tibio- tarsal articulation reaching the eye. Coloration. If correctly determined, this young 20 mm. frog differs from topotypic Liberian examples in being almost colourless when preserved ; in place AFRICAN HERPETOLOGY 217 of the dark backs of adult Liberian specimens there is only very minute stippling, nor does it show any trace of pink on thighs or feet. In this it is much more like Belgian Congo (det. Noble) and Bukoba (det. Loveridge) frogs in the M. C. Z. collection. From the under-mentioned Derema frogs it differs in having the thighs finely stippled ; those of the Derema frogs are colourless, but Liberian specimens have a narrow line of stippling. In life this frog was 'uniform pale satiny-green, ideally suited to the situation in which it was found — a banana stem.' Hyperolius mariae sp. n. Plate 3, Fig. 1 5 && (M. C. Z. 13262-66) Derema, Usambara Mtns., 30. xi. 26. 8 9 9 (M. C. Z. 13267-71) Derema, Usambara Mtns., 30. xi. 26. 57 juv (M. C. Z. 13272-76) Derema, Usambara Mtns., 30. xi. 26. Diagnosis. Nearest to concolor, from which it differs in the smaller size of the breeding adults and in the almost entire absence of any pigmentation on the thighs, which is probably a character of some importance as it is related to the amount of thigh exposed when the frog is at rest; pigmentation in the form of stippling is present on the thighs of our topotype Liberian specimens of concolor as well as on examples from the Congo and Tanganyika Territory. If correctly referred to the new species, the juvenile frogs as well as one male and two females show by their canthal markings (lost in maturity) relationship with puncticulatus and argus. Indeed, except in the matter of size and markings, it is difficult to distinguish the new species, though it has apparently a greater amount of webbing on the hind feet though the character is not over-reUable in a large series. In all probability the specimens from Derema referred to concolor by Nieden are identical with tlie form now described. Type. No. 13267, Museum of Comparative Zoology. An adult female taken in sedges near the mill dam at Derema, Usambara Mtns., Tanganyika Territory. Collected on November 30th, 1926, by Mrs. M. V. Loveridge, after whom it is named. Description. Head slightly longer than broad; snout rounded, slightly pro- jecting, longer than the orbital diameter (reckoning snout from anterior border of eye); canthus rostralis distinct but rounded; loreal region vertical, very slightly concave; interorbital space twice as broad as upper eyeUd (one and a half times in some female paratypes); transverse orbital diameter equals the 218 BARBOUR AND LOVERIDGE distance from the anterior border of the eye to the nostril, also the distance be- tween the nasal openings (or longer than the internasal distance in some female paratypes) ; tympanum hidden. Fingers and toes moderate, dilated at their tips, fingers about two-thirds webbed, the web extending to the base of the disk of the outer finger, to the last joint but one on both sides of the third finger, between the last joint but one and the distal joint of the second, to the last joint but one of the first; toes fully webbed, that is to say, to the bases, or almost to the base of the disks; the tibio-tarsal joint of the adpressed hind limb reaches the end of the snout in the type (usually the eye, or between eye and nostril in all female para- types). Skin smooth above and below except on the breast, belly, and thighs, where it is granular. One or two hardly distinguishable granules at the commis- sure of the mouth. Variation in cf paratypes. The interorbital space is from one and a half to twice as broad as the upper eyelid ; fingers only half webbed ; tibio-tarsal artic- ulation of the adpressed hind hmb reaches to the eye, or slightly beyond. The skin of the gular disk is markedly granular. Variation in young. It is highly probable that the young of more than one species of Hyperob'MS are included, and that some are young princticulatus, but it is doubtful if they could be satisfactorily separated. They appear to be some- what different to the young puncticulatus from Phillipshof in general, though in- dividuals are similar. Fingers and toes of these young are much less webbed than those of the adults, the fingers exhibit a trace of web, and the toes might be said to be two-thirds webbed. Colour of 9 type in life. Above, uniform greyish-white on head, back, tibia, and outside edge of foot ; a black speck on snout, nostrils ringed with black, edge of upper eyelid black, an indefinite broad cream-coloured stripe on side with a broad black one below it; thighs flesh-pink or blood-red. Below, lower lips tipped with blood-red, throat white, rest of the under surface blood-red. Colour of 9 paratypes in alcohol. The seven females exhibit a gradation from light to dark on the backs; two specimens (23 and 26 mm. long), which are very pale above and show no pink, have a light canthal band of the argus type, passing through the eye to some distance along the flank ; there is a very slight concentration of black pigment above and below this colourless line. The remain- ing six (25 to 28 mm. in length and which are breeding females) do not show any light canthal band, but its place is taken by a dark spot around the nostril and another spot behind the eye, which exhibit a tendency to develop with age until they almost meet to form a very black canthal streak. There is also an isolated AFRICAN HERPETOLOGY 219 broad black lateral band of a different character, being subdermal rather than superficial like the nasal spots. The colourless thighs and feet still show very- pinkish in alcohol a year after being collected. Colour of d" paratypes in alcohol. The smallest male (20 mm.), also pallid and without any pink on limbs, shows a light canthal band combined with heavy- pigmentation around the nostrils, the pigmentation extending along the flank. The four remaining males (23 to 26 mm.) are like the adult females in having a dark ring round the nostrils which may umte with the black postocular spot to form a rich black canthal band. No. 13265 alone has a corresponding black, or sepia, spot on each elbow and knee, and a pair of them above the anal opening. The throat of the juvenile male is immaculate white; those of the others heavily speckled with black on the gular disk and usually, but not invariably, on the lips as well. Colour of young. I made no field notes on the coloration of the young, but wTiting from memory, 1 think that in Uf e they were a semi-transparent yellowish- green, yet dusky because of frequent minute black stippling. In alcohol they are rather colourless except for the abundant speckling ; some show a light, dark- edged, canthal and lateral band of the argus type as well as a light dorsal stripe ; others are puncticulate in addition. Measurements. Type 9 . Snout to vent 28 mm. Type 9 . Diameter of orbit .... 3.5 mm. Length of head 10 " Length of tibia 15 " Length of snout 4 " Length of foot 21 " Breadth of head 9.5 " Length from snout to vent in the paratype females 28 to 23 mm. Length from snout to vent in the paratype males 26 to 20 " Length from snout to vent in the young 16 to 13 " Breeding. Half-a-dozen of these young, taken on November 30th, 1926, have traces of tails, while one (No. 13275) has a tail 12 mm. long though the frog itself measures only 15 mm. from snout to vent. Diet. Six stomachs examined, all of which were empty. Habitat. The first frog was seen in silhouette as it squatted on the blade of a sedge at the edge of the mill-dam pond ; thereafter many were found right down inside the central leaves and could be reached only by pulling these apart. Hyperolius puncticulatus (Pfeffer) liappia ptiHcticulata Pfeffer, 1893, Jahrb. Hamburg. Wiss. Anst., 10, p. 99. 15 (M. C. Z. 13277-285) Derema, Usambara Mtns., 30. xi. 26. 33 (M. C. Z. 13286-295) Bumbuli, Usambara Mtns., 14. xii. 26. 14 (M. C. Z. 13296-305) PhiUipshof, Usambara Mtns., 24. xii. 26. 44 juv (M. C. Z. 13306-310) Philhpshof, Usambara Mtns., 24. xii. 26. Lunkelewa in Kisumbara. 220 BARBOUR AND LOVERIDGE Relations. While the Derema frogs are undoubtedly specifically identical with the Morogoro examples described and figured by Miss J. Procter ' and those from Bumbuli seem to be variants of the same thing, the adult Phillipshof frogs appear to be somewhat intermediate with argus. Coloration. The Derema frogs are well spotted above, except six of the smaller specimens; the very broad lateral band, bordered by a thick black line above and below, is a constant feature in the markings of the adult females in alcohol. In life one Derema specimen was coloured as follows. Above, brownish orange with or without black spots ; a white line edged with yellow and broadly with black both above and below begins on snout, passes over (and the black line through) nostril and eye and terminates on flank. A corresponding white spot ringed with yellow and black on heel. Thighs and feet yellow, tinged with black. Below, uniformly yellow except hands and feet, which are flesh-pink. Another was brownish-yellow above; a yellow line black-edged above and below, over (and through) nostril and eye to middle of flank; thighs and feet blood-red. Below, orange-yellow. The Bumbuli series, all taken at one spot, exhibit the most bewildering amount of variation from a uniformly coloured frog whose only marking is a black ringed spot in front of the right eye (a rudiment of the canthal band which is present in most of the others), to individuals with all the exposed surfaces ver- miculated, or with dorsal stripe and lateral markings. Scarcely two are iden- tically marked. The following descriptions of two unusual types, as taken down in the field, are given in full. No. 13286. An adult female taken eight feet from the ground in full view on the leaf of a banana. Outer fingers certainly one-third webbed, toes two-thirds webbed and nearly to the end of the outer toe. Tibio-tarsal articulation just touch- es posterior border of the eye. — Colour in life. Above, pale yellow so heavily stippled with minute black specks as to present a greyish-green ground colour ; the Ughter areas (Ughter because almost free from stippling) are edged with black and take the form of spots more or less free or confluent ; a pair of black a -shaped lines follow the outline of the snout to the eye and may be more or less continued on flank or broken up into black-edged, light yellow spots (like the canthal bands and lateral stripes of H. argus) ; tibia and foot, and fingers also to a lesser degree, reddish-orange. Below, lemon-yellow; fingers, thighs and toes tinged with orange. No. 13295. An adult female taken ten feet from the ground on the upper surface of a banana leaf. The outer toe is webbed to the disk, otherwise exactly ' Procter, 1920, Proc. Zool. Soc. London, p. 415, fig. 2. AFRICAN HERPETOLOGY 221 similar to No. 13286 in the points cited above. — Colour in life. Above, pure white, a black line from the tip of the snout through the nostril and eye to two-thirds of the way along flank ; a a -shaped mark on the snout corresponds with the outer lines but terminates before it reaches the eyes; supraocular region black, and an interrupted line, parallel with the lower one on flank, eventually joins with it; a pair of black, parallel, dorsal Unes are connected anteriorly with the black supraocular lines by transverse ones but are open in the frontal region; these parallel lines unite just above the anus; at irregular intervals lateral transverse Hues unite the dorsal with the flank lines; all of these black lines are tinged with lemon-yellow at their edges; hands, thighs and feet are reddish-orange. Below, lemon-yellow; fingers, thighs and toes tinged with orange. The coloration of the young from Phillipshof is much the same as those of H. mariae from Derema, but the A -shaped mark and lateral lines are invariably present. Measufemeiits. The largest male (Phillipshof) measures 34 mm. (average of 14, nearly 29 mm.) ; the biggest female (Phillipshof) measures 39 mm. (average of 45, nearly 31 mm.) . Arranged according to sex and locality they are as follows : Range of 2 Derema males 25-29 mm., average 27 mm. " 13 Derema females 22-33 " 30 " 10 Bumbuli males 26-31 " 28 " 23 Bumbuli females 28-36 " 32 2 Phillipshof males 31-34 " 32 " 12 Phillipshof females 25-39 " 31 Breeding. The ovaries of ten Bumbuli frogs examined on 14. xii. 26 were, with one exception, in a pepper-and-salt stage; ovules were fairly well advanced in the exception mentioned. Whatever species the adult Phillipshof frogs are, it is almost certain that the young (mostly 14 mm. in length) belong to the same kind, as thorough search was made without revealing any other Hyperoli in the neighbourhood. These young were found in a swamp close to our camp. The vegetation, largely composed of floating tussocks of grass, was scattered everywhere upon the sudd. The only successful way to secure these frogs, discovered by IMrs. Loveridge, who captured most of them, was to part these tussocks, press them outwards and then wait. In a few moments one or two of these delicate, almost transparent, little crea- tures would come scrambling into view and clamber up one of the few grass- stems remaining vertical. Diet. Thirteen Bumbuli frogs were examined and ten of these were found to contain insects as follows: (i) Two beetles (Chrysomelidae and Cleridae det. N. Banks), (ii) Two Cockchafer beetles, (iii) Beetles, (iv) Beetles, (v) Beetles. 222 BARBOUR AND LOVERIDGE (vi) Beetle, (vii) Beetle and orthopteran. (viii) Many small, black, spinose orthopterous legs, (ix) Earwig, (x) Two grasshoppers and a hemiptei'on. Habitat. At Derema in the heart of sedges growing by a swamp. At Bumbuli these frogs were found between 3 and 6 p.m. sitting in hot sunshine on the upper or under surface of (cultivated) banana leaves, often in the gutter of the leaves; once one was taken on the outside of the unopened central tube, never one on the stem of the banana. We secured them by striking the leaf with a ten-foot wand, which caused the frog to take a flying leap into space and generally resulted in its landing on the ground. At Phillipshof in wild bananas on the forest edge; young (see above) in swamp half a mile from these bananas. Hyperolius ARGUS Peters Hypcrolius argus Peters, 1855, Arch. Naturg., part 1, p. 57. 68 (M. C. Z. 1.3322-33) Nyingwa, Uluguru Mtns., 15. x. 26. Eggs (M. C. Z. 13334) Nyingwa, Uluguru Mtns., 19. x. 26. 1 {M. C. Z. 13335) Vituri, Uluguru Mtns., 29. x. 26. Variation. Although this is undoubtedly what is generally called argus, the orbital diameter is only equal to the distance between the anterior border of the eye and the nostril. The snout is not, as Boulenger says (Cat. Batr. Sal. 1882, p. 122), 'as long as the orbital diameter.' Four out of five frogs agree with the statement that ' the adpressed hind limb reaches the eye' ; the chief exception appears to be in fully adult females, where it falls slightly short. In very young frogs as well as in adult males it usually reaches the eye; exceptions are three adult males where it passes, and two where it falls short of, the eye. Coloration. These frogs differ markedly from that shown on the coloured plate XXII of Peters 'Reise nach Mossambique, 1882,' in having no red on the thighs and no yellow spots on the centre of the back, such spots being arranged in a lateral line on either flank. The evolution of these spots was well shown in the Nyingwa series as described below. It is to be noted that Boulenger (loc. cit. page 122) says 'thighs not colored.' In life very young frogs up to 40 mm. in length are mainly white, above and below; the fingers and toes are orange; the back and exposed surfaces of the limbs are stippled with black specks which are denser in certain parts, thus form- ing lines; two of these lines, enclosing a more creamy-white region, run parallel from the tip of the snout to just above the anus, where they nearly meet the pair AFRICAN HERPETOLOGY 223 from the opposite side; within the dorsal area so enclosed is another line from the supraocular region (where it touches the innermost of the outer pair) to the area just above the junction of the hind limb (where again it touches the inner line of the outer pair) ; another pair of lines enclosing a creamy stripe commences on frontal region and continue to just above the anus. As the frog develops, the vertebral lines vanish and the lateral ones become definite black Unes enclosing a bright yellow stripe, which, however, does not reach to the anal region but ends on the flanks and, undergoing a shortening process, tends to end above fore arms in fully adult individuals ; during this disappearing process of the lateral lines, isolated yellow spots are sometimes left and parallel indistinct specks of dusky spots indicate exactly where the parallel lateral hnes should be. Apart from these markings the coloration of the adult frog above is rich orange, beneath bright lemon-yellow. Iris yellow speckled with black, pupil black and round. Measurements. The largest male measured 34 mm., the biggest female 40 mm., the smallest frog, apparently a female, 20 mm. Breeding. One 9 deposited a lot of eggs on the banana leaf in which it was brought to camp, October 19th, 1926. These eggs appear to be perfectly spherical and are two-thirds white and one-third jet black. Captive specimens deposited many eggs during the last fortnight of October on the outside of the tin employed as a water dish. Diet. Of sixteen specimens examined, ten held determinable remains; often fragments of more delicate insects were present in addition. The ten stomach contents were as follows: (i) Ant head and a grasshopper's limbs, (ii) Grass- hopper's legs, (iii) Legs of a long, slender orthopteran, apparently an Acri- dian; also a cockroach, (iv) Cockroach, (v) Cockroach and its egg-purse, (vi) Cockroach and spider, (vii) Cockroach and caterpillar, (viii) Caterpillar, (ix) Caterpillar, (x) Beetle. Habitat. These frogs were taken on domestic banana plants about 7,000 feet or somewhat lower. As a hundred were purchased from native children between October 15th and 19th they must be very plentiful. Doubtless more could have been obtained if remuneration had been continued. An attempt to ship thirty back to the Zoological Gardens failed. At Vituri they were also abundant ; one was preserved for record and no attempt made to secure others as there were more interesting things to hunt for. 224 BARBOUR AND LOVERIDGE HyPEROLIXJS FULVOVITTATUS Cope Hyperolius fulvovittatus Cope, 1860, Proc. Acad. Nat. Sci. Phila., p. 517. 2 (M. C. Z. 13336-7) Nyange, Uluguru Mtns., 1. x. 26. 9 (M. C. Z. 13338-46) Tawa, Uluguru Mtns., 20. x. 26. 10 (M. C. Z. 13347-56) Dar es Salaam, 4. xi. 26. 17 (M. C. Z. 13357-60) Kzerui, Usambara Mtns., S. xii. 26. Remarks. A selection of these specimens was carefully compared by the junior author with the examples from Morogoro and Duthumi in the British Museum (Procter, Proc. Zool. Soc. 1920, p. 417) and found to be specifically identical. Coloration in life. The colouring of the living frog was very different from that of alcoholic specimens as described by Cope and Miss Procter. The following notes were made in the field. Nyange Above, a satiny pale greenish-yellow, two darker bands along either side of the spine from just above anus but converging to a point between the eyes, the whole of the upper surface sprinkled with minute brown specks; another band on either side and still another on outer aspect of tibia. Below white. Tawa Above, silvery white tinged with green, a darker greenish stripe from nostril through eye to flank and an almost parallel dorsal pair terminating, but nut meeting, between eyes; exposed surfaces of thighs and foot darker than the rest; the whole of the upper surface covered with very minute black specks. Below, pure white without a trace of markings of any kind. Dar es Salaam. .Above, upper aspect of snout, back, arms and thighs silvery white, hmbs heavily mottled with fawn-brown almost to exclusion of the white; a brown mark between eyes which terminates in two 'arms' posteriorly which frequently nearly meet two similar 'arms' projecting forwards from a brown spot above anus; these lines are almost absent in the adults; a dark lateral line from nostril to hind hmb, where it connects with one of the forward-projecting posterior 'arms' already referred to. The different aspect presented by these frogs is caused by a breaking down in the centre of the lateral Unes; except for this feature, in alcohol they do not differ mark- edly from the Tawa frogs. KizERUi In frogs from this locality the dorsal hnes do meet between the eyes. Measurements. Largest of three males (Dar es Salaam) 22 mm., average 20 mm. Largest of thirty-five females (Nyange) 26 mm., average 20 mm. Smallest frog, a female from Dar es Salaam, 15 mm. Sex. The strange disproportion in the sexes (3 c? cf, 35 9 9 ) may probably be accounted for by the males having some other habitat. Breeding. Large ova are present in the Nyange (1. x. 26) and Dar es Salaam (4. xi. 26) frogs only. Diet. Stomachs of ten specimens containing identifiable material held the fol- lowing: (i) Beetle, (ii) Beetle, (iii) Water beetle, (iv) Scarab beetle, (v) Long- winged green beetle, (vi) Cercopid bug. (vii) Cercopid bug. (viii) Grass- hopper, (ix) Spider, (x) Acacia leaves, doubtless accidentally introduced; AFRICAN HERPETOLOGY 225 some were also present, with food, in the stomach of M. uluguruensis from Vituri. Habitat. Nyange, Tawa, Kizerui, and some of the Dar es Salaam specimens were personally taken by the junior author inside the outer leaves of bananas, clinging to the main stem. The exceptions from Dar es Salaam were beneath the bark of some poles standing in a swamp; others were between layers of palm thatch of dilapidated huts in the same swamp which had been built as shelters for watchers of the rice crops. Hyperolius flavoviridis Peters Hyperolius flavoviridis Peters, 1855, Arch. Naturg., 21, part 1, p. 57. 2 (M. C. Z. 13361-2) Dar es Salaam, 4. xi. 26. Variation. The fourth toe in the male is longer than shown in the coloured plate (Reise nach Mossambique, xxii, figs. 4 and 5) and the snout is more acumi- nate. The only other difference appears to be that the tympanum is invisible in this specimen (hidden but discernible in type) ; otherwise both frogs agree very closely with Peters' description except for some minor differences of coloration. Coloration in life. The green was a most vivid shade exactly corresponding to the colour of the new grass among which it was caught. Above, brilliant green, yellow on snout and limbs except on exposed svu-faces of latter in repose, which are green ; a somewhat indefinite brown Une connects the nostril with the upper eyelid over which it passes ; a fight yellow lateral line commences at the eye and terminates on the flank; armpit and groin Ught blue, a trace of blue on the lower jaw, throat and limb joints. Measurements. Length of male 29 mm. Length of female 19 mm. Peters gives the total length as 26 mm. Habitat. Both wgre taken in a patch of fresh-grown grass on the northern end of Mogogoni swamp. The larger frog was secured only after a long chase, as it was so active. Hyperolius microps Giinther Hyperolius microps Giinther, 1864, Proc. Zool. Soc. London, p. 311, PI. XXVII, fig. 3. 3 (M. C. Z. 13363-5) Dar es Salaam, 4 and 10. xi. 26. 2 (M. C. Z. 13366-7) Derema, Usambara Mtns., 30. xi. 26. Variation, Sexual. They agree with Noble's key * in that the tibio-tarsal articu- lation of the adpressed hind limb extends beyond eye. It is to be noted, however, ' Noble, 1924, Bull. Am. Mus. Nat. Hist. 49, Art. ii, p. 252. 226 BARBOUR AND LOVERIDGE that there is a strongly marked sexual dimorphism. Had it not been for the fact that a pair (13363 o' and 13365 9) were taken close together in the same thatch we should have hesitated to say they were specifically identical. In life, however, the similarities are more apparent. In the males (13363-64) the tibio-tarsal articulation reaches beyond, or far beyond, the end of the snout, while in the three females (13365-67) it reaches to the eye or nearly to the nostril. Giinther's type was a male with sharp canthus rostralis; in these males it is somewhat rolinded and much sharper in the females. In the males (including type) the tympanum is hidden, but quite visible in the females, a very interesting sexual difference. Apart from these trifling variations they agree well with the somewhat meagre descriptions. The fingers are less than one-third webbed, the toes fully three-quarters webbed. Coloration in life. Descriptions of three of these very beautiful little frogs were made in the field. Male (No. 13363). Above, yellowish-green; limbs creamy white; a yellow line commences on snout between the nostrils and passes over eye and along the sides, more dorsal than lateral, to terminate just in front of junction of hind limb; in the supraocular region this line is interrupted by a dusky patch; a very few black and white flecks on back may be distinguished with the aid of a lens. Below, transparent whitish but discoloured by internal organs, which are clearly visible; the throat, which is tinged with blue, has also some white flecks. Female (No. 13365). (Many eggs observable through abdominal skin.) Rich grass-green above; the Une, yellow in male, is silver in female and hardly dis- tinguishable in front of eye ; its position is well indicated by deep black spots which are also found on the lateral line ; the whole upper surface is speckled with minute black spots; limbs more greemsh-yellow than in cf. Below, greenish tinged with blue on the throat. Female (No. 13366). (Eggs minute, not discernible through skin.) Above, olive-green; a hght line, edged above and below with yellow, commences on snout, passes through and over nostril and eye, broadens on flank and then terminates abruptly in an obtuse point ; a white, yellow-edged spot on each heel ; thighs and tibia yellowish heavily speckled with oUve, as is the outer aspect of foot to a less extent. Below, yellowish. In alcohol all become pallid except for the distinct black naso-lateral line of specks and the fine dusky speckling of the females. AFRICAN HERPETOLOGY 227 Measurements. Largest male 21 mm., the same as tjrpe ; largest female 23 mm. Both Dar es Salaam. Derema females both 22 mm. Breeding. The Dar es Salaam and one Derema frog held many large ova; these extend forward almost to the axilla on both sides of the body. The second Derema female may have recently deposited eggs, as the ovules were small. Diet. Ants and flies in stomach of only specimen examined. Distribution. Described from Rovuma Bay, the southern boundary of Tan- ganyika Territory, later recorded from Angola. The present is the first record known to us of its occurrence north of the Rovuma River. This minute species is rare in collections. Habitat. At Dar es Salaam taken in the palm-leaf thatch of a collapsed and sodden hut in the jNIogogoni swamp. At Derema in the sedges bordering the swamp above the saw-mills. Megalixaltjs loveridgii Procter Megalixalus loveridgii Procter, 1920, Proc. Zool. Soc. London, p. 418. 5 (M. C. Z. 13369-73) Nyange, Ulugum Mtns., 4-8. x. 26. 4 (M. C. Z. 13374-77) Tawa, Ulugum Mtns., 20. x. 26. 3 (M. C. Z. 13378-80) Mkarazi, Uluguru Mtns., 22. x. 26. 10 (M. C. Z. 13381-85) Vituri, Ulugum Mtns., 29. x. 26. 9 (M. C. Z. 13540-44) Dar es Salaam, 4-9. xi. 26. 151 (M. C. Z. 13545-55) Derema, Usambara Mtns., xi. 26. 14 (M. C. Z. 13556-60) Mt. Lutindi, Usambara Mtns., 10. xii. 26. Distribution. This large series was collected in the hope of shedding some Hght on the very peculiar distribution attributed to members of the fornasinii- leptosomus-loveridgii group. M. loveridgii was described by Miss Procter from a single female collected at Morogoro; therefore, the Vituri series is nearest to being topotypic. Later the junior author secured a male CM. C. Z. 10439) at Ivilosa. Some hving examples from Derema, presented to the London Zoological Society, are referred by Miss Procter to loveridgii as is also an example from Portuguese East Africa (Report of addition to the Menagerie, Zool. Soc. Lond., 1927, p. 9). Thus its distribution on the East Coast is ahnost coextensive with that of fornasinii and it occurs both to the north and south of the type locaUty of that species. M. fornasinii. As far as Tanganyika Territory is concerned this species has been recorded from Zanzibar (Tornier, Boulenger, Miiller and Boettger), Dar es Salaam (Tornier), Kingani (Pfeffer, Tornier) , Unguu = Nguru? (Pfeffer, Torn- ier), Mhonda (Pfeffer, Tornier), Tanga (Tornier), and Ugogo. Lonnberg has re- ported it from Mombo at the foot of the Usambara Mountains. With the excep- 228 BARBOUR AND LOVERIDGE tion of Zanzibar, and Ugogo (now known as Dodoma District) these localities are well within an irregular square lying between Dar es Salaam, the Uluguru Mountains, the Usambara Mountains, and Tanga, or the area skirted by the present collecting trip. As M. loveridgii is a very distinct species it seems highly probable that these records should refer to the latter. If the Ugogo record is correct it would indicate an anomalous distribution comparable to that of leptosomus which, Noble has pointed out, occurs in rain-forest and savannah area. Boettger has recorded as M. fornasinii var unicolor an example from Pemba lacking the vertebral line which Noble has already placed in the synonymy of the former. M. leptosomus. A 'young one' has been recorded from Mavene in Usambara by Boettger, and other specimens from Zanzibar, Tanga and Undussuma by Tornier. It would be of great service if someone could reexamine these specimens with a view to checking these identifications. Nieden united M. fornasinii and leptosomus and recorded them from Pemba Island, Tanga, Bagamoyo, Ukami, Mohorro, Lindi, and Rugwe near Lake Rukwa. Conclusion. At present the range of M. loveridgii appears to be East Africa from the Usambara IMountains (Barbour and Loveridge) and Kilosa (Loveridge) to the Zambesi (Procter). In all probability many records of fornasinii and leptosomus should be referred to this species. Variation. The series shows remarkable uniformity in structural characters; none show the webbing of the hands as extensively as in the drawing of the type — one might rather say fingers half webbed. (In the Tawa series they are perhaps one-third to one-half webbed.) They are much more webbed than in examples of fornasinii from Cameroon and Gaboon ; as also in a series of Congo specimens loaned by Dr. G. K. Noble for comparison. One of the Nyange frogs was compared with the type M. loveridgii, with which it agreed in length and proportions; at the time the junior author noted that its feet were shghtly more webbed than in fornasinii. On the whole the toes are three-quarters webbed: to the base of the subter- minal phalanx of the 4th toe, and to the disk on one side of the 2nd, 3rd and 5th. The length of the third finger, which is uniformly twice the length of the first, does not appear to be a very helpful character in differentiating loveridgii from fornasinii, {ox the two Cameroon and Gahoon fornasinii available for comparison appear to be identical in this respect. The more spinose character of the minute tubercles in loveridgii, is the readiest means of differentiating the two species. AFRICAN HERPETOLOGY 229 While in the description Miss Procter states, 'Length from snout to vent 3 1 times length of head/ in the table of measurements the length from snout to vent is given as 36 mm., and the length of head as 11 mm., i.e., between 3j and 3 1 times, so that this character again is of Uttle use. Peters gives the length offornasinii as 35 mm., and the head length as 10 mm., so in size there is only a small difference. The Cameroon and Gaboon specimens show a head into body length of 2.5 and 3 times respectively. In the small series of five females from Nyange the variation is from 2.7 to 3.18 times. It would appear as if the head is contained in body length in loveridgn 2.70 to 3.27 times and in fornasinii 2.50 to 3.50, the latter figm-es being based on the measurements of Peters' type and of the Cameroon, and Gaboon frogs. The tibio-tarsal articulation exactly reaches the eye in every specimen, except in the case of two Derema females where it falls slightly short. The tympanum is generally indistinguishable in males, more often present in females. On examining twenty specimens of each sex it was found present in about 10 per cent of the males and nearly 50 per cent of the females. Coloration in alcohol. In no specimen of the present series does the vertebral band appear shaped as in the figure accompanying Miss Procter's description; there is a good deal of variation in its outline, but where present it more closely approximates to that shown in Bianconi's and Peters' plates oi fornasinii. In alcohol the ground colour is substantially as recorded by Miss Procter. As to the dorsal band an examination by locaUty shows the following. Nyange All uniformly purplish brown above, without vertebral marking. Tawa As above, but one with faint suggestion of the vertebral marking. Mkarazi As in Nyange specimens, but one paler. ViTURi Six as in Nyange frogs; four show the vertebral marking and are noticeably paler. Dar es Salaam. .All except three show very pronounced vertebral stripe; two of these exceptions show it faintly, only one is without. Mt. Lutindi Eleven without marking, three with vertebral stripe. Derema 58 males without marking, iO with vertebral stripe ; 31 females without marking, 26 with vertebral stripe. It will thus be seen that this marking is more often absent than present in frogs from this region. Coloration in life. It is probable that this frog changes its colour to a consider- able extent. Nyange 9 . Above, oHve-brown on all exposed surfaces; scattered white spots chiefly on sides and Umbs but also a few on back. Below, including hidden upper surface of thighs, lemon-yellow with the exception of the throat, chest, and belly, which are white. Nyange 9 . Above, opaque silvery white in a shield-shaped patch extending from snout to vent, and side to side; similar patches cover the thighs and are, like the back, studded with black specks. Both were taken in bananas. 230 BARBOUR AND LOVERIDGE Dab es Salaam. .Four silvery white examples, each with a fawn-brown vertebral line. Taken in holes in a post. Of another, taken in a banana close by, it was noted: back and tibia silvery white, a brown dorsal stripe terminating in point between eyes; flanks dusky, almost indicating a lateral stripe from snout to arms; below, pure white. Measurements. Largest male 38 mm., smallest 31, average of a hundred males 33.71; largest female (Mkarazi and Derema) 38 mm., smallest (Tawa and Derema) 35 mm., average of a hundred females 35 mm. Sex. In the first specimens collected the junior author was astonished at the disproportion of the sexes, viz. 2 males, 43 females. These were taken in bananas. At Derema, where the collecting was done in sedges, the balance was the opposite, viz. 98 males and 57 females. It is entirely a coincidence that the total should be 100 of each sex. The sexes may be readily told by the gular disk of the male, which is lacking in females. Breeding. In the Uluguru Mountains these frogs did not appear to be breed- ing. Several Vituri specimens examined held only minute ova; in those from Dar es Salaam they were much larger; while each of ten females examined at Derema on 30. xi. 26 held a mass of developing eggs about 2 nam. in diameter. Diet. Of the stomachs of ten specimens which were examined : (i) from Vituri held a large, hard lump almost the size of the stomach. This, when broken up, was found to be composed of frog's spawn, (ii) The largest meal had been made by a Derema frog which held a 20 mm. caterpillar, an 11 nma. grasshopper (head and abdomen measurement) and a 5 mm. spider, (iii) Spider, (iv) Spider and beetle, (v) Beetle, fvi) Two species of beetles, (vii) Coccinelid beetle, (viii) Ditto, same species, (ix) Beetle and the wings of a hymenopterous insect, almost certainly an ant. (x) Two small grasshoppers. Parasites. Minute worms. Habitat. The first examples from Nyange were taken by the junior author from between the outer skin and stem of wild banana plants growing in ravines at a high altitude. Others from Chogwe near Vituri (included in Vituri series), in domestic bananas. At Vituri, in wild bananas. At Dar es Salaam four were taken in the holes of a carpenter bee in the posts of an abandoned hut in Mogogoni swamp. The swamp was at one time planted with rice, hence the watcher's hut, long since abandoned. Over a hundred were taken in a single day in the sedges at Derema, usually between outer leaf and main stalk, but concealed from sight until the plant was stripped of the outer leaves. At Kizerui and Lutindi they were found in both domestic and wild bananas, in one instance in the central shoot so favoured by the bat {Pipistrellus nanus). AFRICAN HERPETOLOGY 231 Megalixalus ulitguruensis sp. n. Plate 3, fig. 2. 16 (M. C. Z. 13311-120) Vituri, Uluguru Mtns., 27-31. x. 26. 1 (M. C. Z. 13321) BumbuU, Usambara Mtns., 14. xii. 26. 1 (M. C. Z. 13368) Derema, Usambara Mtns., xii. 26. Type. No. 13311, Museum of Comparative Zoology. An adult female found sitting on a leaf of a fallen wild banana plant in a ravine at Vituri, Uluguru Mountains, Tanganyika Territory. Collected by A. Loveridge on October 27th, 1926. Paratypes. As listed above. Relatmis. This very Hyperolius-hke frog would have been placed in that genus had not the pupil been observed to be vertical in life. In the second speci- men taken it was round. Under chloroform the round pupil became a narrow vertical sht, but during preservation reverted to the round shape. Its nearest relative appears to be M. madagascariensis, which has the same white skin and canthal stripe but whose males have a gular disk. Description. Head slightly broader than long; snout obtusely roimded or slightly truncated, sHghtly projecting, longer than the orbital diameter (reckon- ing snout from the anterior border of eye) ; canthus rostrahs distinct but rounded; loreal region almost vertical, very sHghtly concave (in young rounded and not concave) ; interorbital space one and a half times as broad as upper eyeUd; trans- verse orbital diameter equals the distance from the anterior border of the eye to the nostril (slightly longer in some paratypes) , also the distance between the nasal openings; tympanum hidden (absolutely in every specimen). Fingers and toes moderate, dilated at their tips; fingers about one-third (certainly less than half) webbed; toes fully webbed, that is to say, to the bases of all the disks, though in the case of the fourth toe this is achieved only by a narrow border of web from the sub terminal joint (unfortunately this deUcate, narrow, semi-transparent webbing has not been reproduced well in the accompanying figure) ; the tibio- tarsal joint of the adpressed hind limb reaches the eye in the type and in all paratypes. Skin smooth above and below, except on the belly and thighs, where it is granular, hardly perceptibly so on the thighs; some scarcely distinguishable granules at the commissure of the mouth. Color oj 9 type in life. Above, head, back and exposed surfaces of fore and hind limbs enamel-white; a pair of black-speckled, reddish-brown flecks on an- terior part of back; another pair, more dorso-laterally situated, just in front of junction of hind limbs, and a streak of similar color on each tibia; commencing on snout a black-speckled, reddish-brown network line (the interspaces being 232 BARBOUR AND LOVERIDGE enamel-white) passes through nostril and eye and terminates on flank. It will be noticed that when the hind limbs are arranged in the normal attitude of rest, the tibial streaks form a continuation of the lateral network hnes; fingers, thighs and part of feet yellow. Below, transparent yellow except for a broad belt of enamel-white across chest. The whole Vituri series agreed with above description except for very minor differences. Some have a third pair of spots between eyes. Colour in Alcohol. These spots appear brown in alcohol, while the enamel- white entirely disappears, the frog being flesh-coloured with minute speckes, which may be so numerous dorsally as to give a distinctly brownish appearance to the upper surface. Colour of Bumbuli paratype in life. The Bumbuli frog, while structurally agreeing with the Vituri series and having the same appearance in alcohol, dif- fered considerably when alive. The colouring noted in the field was as follows. Above on head, back, fore arms, tibia and edge of feet, white with patches ap- pearing sUghtly rubbed and showing yellowish-green ; black specks are scattered over the whole of the upper surface, including upper arm and thigh, which are yellowish; a more or less transparent, slightly greenish band from snout through nostril and eye to flank shows the black stippling more clearly; fingers and toes clear lemon-yellow. Below, transparent white, more opaque on limbs than on belly, where internal organs show through; fingers and toes lemon-yellow. Measurements. Type 9 . Snout to vent 31 mm. Diameter of orbit ... 3 mm. Lengtii of head 9 " Length of tibia 14 " Length of snout 4.5" Length of foot 19.5" Breadth of head 10 " Largest specimen (13314) also a female 32 mm.; next four, after type, 29 mm. each. The half-dozen smallest frogs are 24-25 mm. and apparently males. Sex. It seems highly improbable that all sixteen frogs are females. If, as we suppose, the smaller are males (and one examined proved to be so) , then these males are without a gular disk. Breeding. The adult females, including the type, hold numerous well-devel- oped eggs, visible in life through the transparent abdominal skin. Diet. Besides much smaller, fragmentary insect remains, the following were discernible in ten stomachs examined, (i) Large muscoid dipteron. (ii) Many fruit fhes (Drosophila), one adult cercopid homopteron, and several nymphal ones, (iii) Large cercopid of a different species, (iv) Cercopid bug. (v) Elater beetles, (vi) Beetle (?Lathrididae). (vii) Two chrysomelid beetles, (viii) Several AFRICAN HERPETOLOGY 233 beetles with long elytra, (ix) Beetle, bug and grasshopper, (x) Earwig, ant and a small neuropteran which Mr. N. Banks considers is almost certainly Myr- melon sp. Habitat. All the Vituri specunens were taken on wild bananas in the rain- forest at an altitude a little over 2,000 feet; the type at noonday was fully ex- posed. The Bumbuli frog was found at 9 a.m. sitting on a wild banana leaf about nine feet from the ground. The plant was growing on the edge of the patch of remaining (dry) primary forest above the Mission, which is about 2,000 feet. Parasites. A nematode in one of the Vituri frogs. Leptopelis aubryi (A. Dimieril) Hyla aubryi A. Dumeril, 1856, Rev. Mag. Zool. (2), 8, p. 561. 11 (M. C. Z. 13.561-71) Mt. Lutindi, Usambara Mtns., 10. xii. 26. Variation. The tympana of the two adults are approxhnately three-quarters and seven-eighths of the eye diameter instead of 'half as given by Boulenger (Cat. Batr. Sal., 1882, p. 135) ; in other respects they agree with the description. In the nine young the tympanum is entirely wanting. Coloration in life. A very beautiful pale green. Measurements. The two adult females measure 58 and 52 mm. respectively; the young, 11 mm. Breeding. Neither of the adults was in breeding condition, the ovaries being shrunken and holding minute ova. Nine of the young were found one morning sitting on the upper surfaces of the leaves of a single banana plant within fifty feet of where the adults were taken. Diet, (i) A woodlouse, a 25 mm. reduviid bug, a 16 mm. cercopid bug, to- gether with the head and limbs of a large grasshopper, (ii) A brown cricket. Distribution. This Cameroon species has already been recorded by Nieden from Amani. Habitat. Both adults were taken concealed in the central shoots of domestic bananas growing wild in the forest on the lower slopes of Mt. Lutindi. Special search failed to reveal any others except the young. Leptopelis rufus Reichenow Leptopelis rufus Reichenow, 1874, ;Vrch. Naturg., 40, part 1, PL IX, Figs la and lb. 12 (M. C. Z. 13572-78) Bagilo, Uluguru Mtns., 9-30. ix. 26. 8 (M. C. Z. 13579-85) Vituri, Uluguru Mtns., 27-30. x. 26. Variation. The size of the tympanum in relation to the diameter of the eye varies a good deal ; it is more often half than three-fourths the orbital diameter. 234 BARBOUR AND LOVERIDGE Coloration in life. In colouring, East African rnjus differ from topotj^pic Cam- eroon specimens as already remarked (Loveridge, Proc. Zool. Soc, 1925, p. 787). The four following types were all taken at Bagilo. 9 . 70 mm. Rich oUve-green; a conspicuous, 11 mm. long, black, yellow-edged 'query' mark on head and nape; two or three more irregular blotches on back; four black cross-bars on tibia, and three or four, more or less distinct, on the hind feet; similar cross-bars on fore arms; disks of fingers and toes bright yellow; lower surfaces white but so heavily stippled with green as almost to obscure the white. Eyes red. Another adult $ , 67 mm. in length, has a broad red band commencing on the snout and passing over the eye and along the sides till it terminates just before the hind limbs; the anus is ringed round by a similar red band. This seems to be the most common type, as eight of our twelve females possess this lateral band. d". 53 mm. Slightly reddish-green without markings on crown or back. A black band, edged above with a bright-yellow line, passes from the snout through nostril and eye to disappear just in front of the hind limb, where it is broken up by yellowish-white vermiculations. One or two faint indications of possible black bars on the Umbs; a black circle, edged above with yellowish-white, around anus; upper lip more or less edged with yellow, with a streak nearly running up to the anterior border of the eye. Below, white, less stippled with green than in the 9. Another apparently immature d" has a broad black, yellow-edged bar con- necting the upper eyelids, while there is a similar yellow-edged, black, a -shaped marking on the back between the fore arms. Yet another presumably juvenile cf has a narrow bar like the last, but the A -shaped marking is replaced by an almost round black spot, and the whole of the back is vermiculated with dark lines. Measurements. The largest male (Vituri) measured 46 mm., range being 34 to 46 mm., with an average of 40 mm.; the sexing of the smallest being some- what doubtful. Largest females (Bagilo and Vituri) 70 mm., range 54 to 70, with an average of 63.5 mm. All adult and breeding. Sexes. At Bagilo one male and seven females were taken; at Vituri, seven males and five females. Breeding. All the Bagilo females held developing eggs (9-30. ix. 26) but those in the Vituri frogs are much larger (27-30. x. 26), some being as much as 4 nm:i. in diameter. AFRICAN HERPETOLOGY 235 Diet, (i) Beetle, (ii) Beetle, spider, grasshopper, (iii) Grasshopper legs, (iv) Many grasshopper remains, (v) Grasshopper, (vi) Grasshopper, (vii) Grass- hopper, (viii) Large, wingless, yellow grasshopper of a species common in the bananas, (ix) Orthopteran legs and a cockroach, (x) Two big cockroaches, (xi) Earwig. Habitat. Taken in the wild bananas growing in moist ravines in the rain- forest. Leptopelis uluguruensis sp. n. Plate 3, Fig. 3. 1 (M. C. Z. 13586), Nyange, Uluguru Mtns., 5. x. 26. 21 (M. C. Z. 13587-96), Vituri, Uluguru Mtns., 27-30. x. 26. Type. No. 13586, Museum of Comparative Zo5logy. An adult male taken on a wild banana plant at Nyange, Uluguru ^Mountains, Tanganyika Territory. Collected by A. Loveridge on October 5th, 1926. Paratypes. Twenty-one specimens from Vituri near Nyange as listed above. Relations. These frogs have the undivided omosternum and bony metaster- num as described by Noble for the genus Leptopelis. In keys it usually falls near rufus but is much squatter in habit, with shorter snout, shorter hind Umbs, and a less distinct, or entirely concealed, tympanum, in addition to the very different coloration. Compared with many species at the British Museum. Description. Vomerine teeth in two juxtaposed groups between the choanae. Head broader than long ; distance between the nostrils equals distance between a nostril and the anterior border of orbit, two-thirds the longitudinal diameter of the eye; canthus rostralis short and distinct, loreal region concave; tympanum visible on right side of head, concealed on left (indistinguishable in ten para- types), one-quarter the longitudinal diameter of the eye opening. Inner finger scarcely webbed, outer one, one-third webbed, disks well developed. Toes webbed to the disks on the inner side of second, third, and fifth ; only as a not very distinct seam to the disk of the fourth ; first toe half webbed. The extent of this webbing has not been reproduced so fully in the accompanying figure as in the original drawing. A moderate, slightly compressed, inner metatarsal tubercle. The tibio- tarsal articulation of the adpressed hind limb just falls short of the eye (reaches eye in many paratypes but never beyond). Skin shagreened above, very granu- lar beneath except on tibia and sole of foot. Colour of cf type in life. Above, rich green, upper eyelids brownish, their out- side edge yellow ; seven orange spots irregularly dispersed on back ; thighs trans- 236 BARBOUR AND LOVERIDGE parent yellowish, each with an orange spot; lower arms the same but so liberally besprinkled with niinute brown specks as to appear brownish. Below, throat blue, rest of the under surfaces white, more or less tinged with blue, particularly at axil and groin. Iris white, speckled with black; pupil black. Of a 9 paratype from Vituri it was noted : colour as in type but back greenish- brown ; a very few yellow spots on the flanks. Colour of cf type in alcohol. Above, head purplish-brown shading to fawn on back, the spots yellowish-green, limbs fawn, thighs almost colourless. Below, white. Measurements. Type cf . Snout to vent 42 mm. Diameter of orbit .... 6 mm. Length of head 15 " Length of tibia 19 " Length of snout 6 " Length of foot 30 " Breadth of head 17.5 " The type is probably the largest male taken; largest female 59 mm.; smallest frog 28 mm., taken 28. x. 26. Breeding. All the females examined had well-developing ovules about 2 mm. in diameter, 27-30. x. 2G. In life these eggs could be observed through the fairly transparent abdominal skin. Diet, (i) Spider, (ii) Spider, (iii) Earwig and grasshopper, (iv) Pallid acridian. An interesting feature of the stomachs of all six specimens examined was the presence of acacia leaf petals in each one; in two there was nothing else; one held over thirty. These petals are continually twinkling down from the trees into the bananas where the frogs live, and form large drifts in the leaves. At first the collector was inclined to suppose that they were accidentally taken up with food, but it seems possible that some frogs may be foolish enough to snap them up in mistake for insects as they aUght on the large flat leaves of the bananas. Enemies. One of these frogs was recovered from the stomach of a snake (Crotaphopeltis hotamboeia tornieri) at Vituri. Habitat. The type was found sitting in the gutter-like stalk of one of the im- mense top leaves of a wild banana. It made no attempt to escape other than to squat back into its retreat. The Vituri series were all taken in wild or domestic bananas. Leptopelis parkeri sp. n. Plate 4, Figs. 9 and 10. 1 (M. C. Z. 13597) Vituri, Uluguru Mtns., 27. x. 26. Type. No. 13597, Museum of Comparative Zoology. An adult female taken on a domestic banana plant at Vituri, Uluguru Mountains, Tanganyika Terri- tory. Collected by A. Loveridge, October 27th, 1926. AFRICAN HERPETOLOGY 237 Diagnosis. Near L. uluguruensis , from which it is distinguished by its well- separated vomerine teeth, its longer hind limbs, smooth dorsal skin and colour- ing. Undivided omosternum and bony metasternimi. Description. Vomerine teeth in two well-separated groups between the choanae. Head broader than long; distance between the nostrils slightly shorter than the distance between a nostril and the anterior border of orbit, rather more than two-thirds the longitudinal diameter of the eye; canthus rostralis distinct, loreal region concave; tympanum just distinguishable on both sides of head (possibly an individual character), one-third the longitudinal diameter of eye opening. In- ner finger barely webbed, outer one one-third webbed, disks well developed. Toes webbed to the disks on the inner side of the second, third, and fifth ; only as a not very distinct seam to the disk of the fourth; first toe half webbed. A small, thinly compressed, inner metatarsal tubercle (considerably different in appearance to that of L. uluguruensis). The tibio-tarsal articulation of the adpressed hind limb reaches the nostril. Skin smooth above, granular beneath except on tibia and sole of foot; the granulations of the throat noticeable only on close inspection. Colour of 9 type in alcohol. Above, fawn with very irregular stipples and dashes of dark rufous-brown on the back, besides several light spots; a line of stippling connecting the upper eyeUds; loreal region to tympanum and along flanks mottled with white and brown, as are also the fore limbs with four cross- bars on the fore arm formed by a concentration of the mottling; on the tibia also are four cross-bars indicated in a similar manner; the upper aspects of thighs and feet mottled like the flanks. Below, immaculate white except for a Uttle mottling on the extreme edges of the lower jaw. Unfortunately, though in the field a note was made as to this being an unde- scribed species, pressure of other work prevented colour notes being made as to its appearance in life, which was very different from that of L. uluguruensis. Measurements. Type 9 . Snout to vent 45 mm. Diameter of orbit .... 5 mm. Length of head \7 " Length of tibia 22 " Length of snout 7 " Length of foot 32 " Breadth of head 16 " Breeding. Ovules small, ovaries contracted. Diet. Remains of what are apparently orthopteran limbs were found in the intestine, together with numerous leaf petals like those found in L. iduguruensis. The stomach was empty except for two of these petals. Named after Mr. H. W. Parker of the British Museimi, who has so kindly as- sisted these investigations by making comparisons with tjrpes in his care, and in many other ways. 238 BARBOUR AND LOVERIDGE Leptopelis vermiculatus (Boulenger) Hylambates I'ermicidatus Boulenger, 1909, Ann. Mag. Nat. Hist., (8) 4, p. 497. Nieden, 1910, Sitzber. Ges. Naturf. Freunde, Berlin, p. 448. Nieden, 1915, Mitt. Zool. Mus. Berlin, vii, p. 370. 1 (M. C. Z. 13598) Amani, Usambara Mtns., xii. 26. Relations. Nieden has pointed out that Tornier's Plate V, fig. 1 (Tornier, 1897, Kriechthiere Deutsche-Ost-Af rikas) , of H. aubryi really represents vermic- ulatus and that the accompanying text refers to both frogs, under the impression that vermiculatus was the young of aubryi. Nieden himself considers it more closely related to L. rufus. Our examination of the unforked omosternum con- firms this view and places the species in the genus Leptopelis. Variations. While in London on the way back from Tanganyika Territory the junior author reexamined the 34 mm. type and closely compared it with the 46 mm. specimen listed above. Some points in the original description which had puzzled him in the field were resolved as follows. The quotations are from the original description. ' Snout rounded, as long as the eye,' means from the anterior border of the eye to the end of the snout. ' tympanum . . . two-thirds diameter of eye ' in the type was found almost to equal the diameter of eye in our specimen. 'fingers with a mere rudiment of web' is most misleading, for the develop- ment is the same in both specimens, whose fingers are at least one-third webbed; the least webbed is the third finger, with two free joints, apart from the disk, on its inner side. 'toes half-webbed.' In the type the webbing on the third and fifth actually does not reach to the end as in our specimen. 'inner metatarsal tubercle rather small, oval, feebly prominent.' This tubercle in the larger frog measures 2 mm. in length and may be said to be well developed. 'the tibio-tarsal articulation reaches just in front of the eye.' In ours it reaches only to the posterior border of the eye. The type, however, is somewhat macerated, which may account for this difference. Measurements. Total length 34 mm. Sex. &. Diet. Stomach empty; intestine held what were apparently remains of termites. Distribution. The type, which is still the only example in the British Museum, came from .:Vmani. Apart from these the only other examples known to us are AFRICAN HERPETOLOGY 239 recorded by Tornier and Nieden from .\mani, Derema, Bulwa (Buloa) and 'near Tanga or jNIagrotto.' Nieden also lists two young examples from Uhehe. If these latter are correctly determined L. verinicxilalus should occur in the Nguru and Uluguru Mountains, which form connecting Unks between the Usambara and the mountains of the Uhehe district. Habitat. No frog gave us more trouble to locate than this species. Krefft recorded it as found on bushes in the evening after rain. Search was made at Derema as well as at Amani but without result. On departing from ^Amani we left a native behind to endeavour to secure three topotypic species which we still lacked. I had pointed out to him the only remaining wild banana plant any- where near Amani (already examined by Salimu, who found only Megalixalus and Hyperolius in it). It was growing on the precipitous side of a ravine near the sunmnt of Mt. BomoU. Abedi was urged to reexamine this banana and the special offer of a shilUng (25 cents U. S. currency) was made if he got the right frog. Two days before we were due to sail from Tanga, the native, Abedi, arrived with his frog and an example oi Athens ceratophorus, one of the other two species we lacked ! Leptopelis johnstoni (Boulenger) Hylatnbates johnstoni Boulenger, 1897, Proc. Zool. Soc, London, p. 803, pi. xlvi, fig. 4. 4 (M. C. Z. 13599-602) BagUo, Uluguru Mtns., 14. ix. 26. Relations. These frogs have the unforked omosternum of the genus Lepto- pelis. Variation. The tibio-tarsal articulation, which reached the eye in the type and in three of the present series {&,