BULLETIN OF f*%u**a THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY VOL. 13 1965-1966 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) LONDON: 1967 DATES OF PUBLICATION OF THE PARTS No. i ...... 6 August 1965 No. 2 . ." . . 14 September 1965 No. 3 . . . . . 14 September 1965 No. 4 . . . . . -15 October 1965 No. 5 ...... i October 1965 No. 6 ...... i October 1965 No. 7 . . . . . .22 October 1965 No. 8 . . . . . .3 December 1965 No. 9 ...... 2 March 1966 PRINTED IN GREAT BRITAIN BY ADLARD & SON LIMITED BARTHOLOMEW PRESS, DORKING CONTENTS ZOOLOGY VOLUME 13 PAGE No. I. Primitive Cryptostigmatid mites from Rhododendron forests in the Nepal Himalaya. By J. G. SHEALS i No. 2. Japanese and other Ophiuroids from the collections of the Miinich Museum. By AILSA M. CLARK (PI. i) 37 No. 3. The freshwater Gastropod molluscs of West Cameroon. By C. A. WRIGHT (Pis. 1-3) 73 No. 4. Barbus (Pisces, Cyprinidae) of the Volta region. By A. J. & J. HOPSON 99 No. 5. Notes on the Cupuladriidae (Polyzoa, Anasca). By PATRICIA L. COOK (Pis. 1-3) 151 No. 6. Polyzoa from West Africa : the Cupuladriidae (Cheilostomata, Anasca). By PATRICIA L. COOK (Pis. 1-3) 189 No. 7. Descriptions of some Strongyles (Nematoda) from mammals in East Nepal : with records of other parasitic nematodes. By WILLIAM G. INGLIS & COLIN G. OGDEN 229 No. 8. Studies on the British Dermanyssidae (Acari : Mesostigmata), part i, external morphology. By G. OWEN EVANS & W. M. TILL 247 No. 9. A review of the Muridae (order Rodentia) of Uganda. By M. J. DELANY & B. R. NEAL 295 Index to Volume 13 357 PRIMITIVE CRYPTOSTIGMATID MITES FROM RHODODENDRON FORESTS IN THE NEPAL HIMALAYA J. G. SHEALS BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 13 No. i LONDON: 1965 PRIMITIVE CRYPTO STIGMATID MITES FROM RHODODENDRON FORESTS IN THE NEPAL HIMALAYA BY J. G. SHEALS British Museum (Natural History) Pp. 1-35 ; 62 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 13 No. i LONDON: 1965 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 13, No. i of the Zoological series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. Trustees of the British Museum (Natural History) 1965 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued August, 1965 Price Eighteen Shillings PRIMITIVE CRYPTO STIGMATID MITES FROM RHODODENDRON FORESTS IN THE NEPAL HIMALAYA By J. G. SHEALS INTRODUCTION THIS report forms the first of a series on the soil- and litter-dwelling Acari collected by the British Museum (Natural History) Expedition to East Nepal 1961-62. A full account of this expedition with details of the localities examined is given by Sheals & Inglis (1965) . In spite of the outstanding biogeographical interest of the Himalayan region as a complex transitional zone, very little attention has been paid to the soil and litter fauna of the area. In so far as the Acari are concerned, small collections were made in 1954 by the British Museum (Natural History) Expedition to West Nepal. Some of this material has been utilized in revisionary studies (Hyatt 1956, Evans & Hyatt 1958, 1960), but apart from these no account of the Himalayan free-living mites appears to have been published since the report on a collection of Oribatidae from the Sikkim Himalaya by Pearce (1906). The present paper is concerned with the taxonomy of the Oribatei Inferiores from Rhododendron forest soil and litter in four localities in East Nepal : a large forest of tree Rhododendrons (mainly R. arboreum) on the slopes of the Milke Danra ridge, 27 19' N 87 31' E, altitude 9,300-11,000'; an area of dwarf Rhododendron at Topke Gola, 27 38' N 87 35' E, altitude 12,500-13,000'; an area of tree and bush Rhododendron (mainly R. arboreum) at Selap, 27 38' N 87 49' E, altitude 10,100- 10,300' and thickets of bush Rhododendron (mainly R. arboreum) on the crest of the Singalila ridge near Sandakphu, 27 06' N 88 01' E, altitude 12,000'. The higher Cryptostigmata and other mites from these localities will be dealt with in later papers. In the descriptions the setal nomenclature used in the later works of Grand jean has been followed except in the case of the Nothroidea where it is convenient to follow the system used by Sellnick & Forsslund (1955). The type material is deposited in the British Museum (Natural History) . Family ACARONYCHIDAE Grandjean, 1932 Himalacarus chimalae 1 gen. et sp. nov. ADULT. The prodorsum (figs, i & 2) is approximately i6ofjt, long and 140(1, wide at the broadest point level with the sensilli. There is a prominent obliquely pro- jecting rostral protuberance (the naso of Grandjean, 1958) below which a bilobed eye such as Grandjean (op. cit.) has noted for other members of this family can easily be discerned. The rostral setae (ro) are stout, prominent, quite smooth and about 8o(jt in length while the lamellar setae (la) are barbed and about 95[z in length. The interlamellar setae (in) and the exobothridial setae (xi) are barbed and about I30|x in length, while the exobothridial setae (xs) are short (approx. 30(x long), smooth x From the Nepali chimal certain tree Rhododendrons J. G. SHEALS ps 2 FIG. i. Himalacarus chimalae gen. et sp. nov. Adult lateral. and rather blunt. The sensillus (bo), i85-i90[x long, it quite smooth and tapers gradually to a fine point. A rectangular design evidently formed by a thickening of the cuticle can be discerned between the interlamellar setae ; this is almost identical in form to the system of 'chitinous nervures' (false lamellae) described by Grand] can (1952) for Stomacarus tristani. The hysterosoma is about i75(J. in length and approximately 170^ wide at the broadest point level with setae e 2 . The 'asthenique zone' between the posterior border of the propodosoma and the median dorsal sclerite is very short. There are 17 pairs of setae in the gastronotal series. Setae GI, cp, d 2 , ei, 62, f 2 and hi are long, barbed, dark brown in colour and arise from tubercles, those of setae e t being particularly prominent. An internal thickening such as Grandjean (1952, I952a, 1957) has observed for Stomacarus, Archeonothms and Andacarus, respectively, can be seen below the tubercle of e\. Difficulty was experienced in securing specimens with completely undamaged setae but it would appear that the long barbed setae c lf c 2 , cp, d 2 and e 2 are about equal in length (i.e. about 170^) while the barbed setae f 2 and hj may be slightly longer than the total length of the idiosoma. Setae Ci and the PRIMITIVE CRYPTOSTIGMATID MITES FROM NEPAL short smooth c 2 lie together on a median dorsal sclerite, while setae d 2 and the short smooth di lie together on paired sclerites. The remaining gastronotal setae, with the exception of e 2 , are borne individually on small sclerites. Setae fj and those of the ps and h series are short, smooth and simple, with the exception of setae ps 2 which are shaped like spear-heads. Setae ps t and h 2 are rather blunt. h llL FIGS. 2-6. Himalacarus chimalae gen. et sp. nov. Fig. 2. Adult, dorsum. Fig. 3. Adult, ano-genital region. Fig. 4. Adult, infracapitulum, Fig. 5. Adult, palp. Fig. 6. Tritonymph, venter. 6 J. G. SHEALS The ano-genital region is shown in fig. 3. There are 3 setae in the anal series (AN) and 4 in the adanal (AD). All these setae lie individually on small sclerites. There are 9 pairs of genital setae and 3 pairs of aggenitals (AG). The 10 pairs of eugenital setae are all approximately similar in size and shape and it is suspected that all the adults collected were males. The coxisternal region is essentially similar to that of the tritonymph (fig. 6), the setal formula being (4-3-4-4). FIGS. 7-10. Himalacarus chimalae gen. et sp. nov. Adult, legs. Fig. 7. Leg I. Fig. 8. Leg II. Fig. 9. Leg III. Fig. 10. Leg IV. PRIMITIVE CRYPTOSTIGMATID MITES FROM NEPAL 7 The infracapitulum is shown in fig. 4. There are four oral lips and in comparison with other members of the family the lower lip is particularly large and easily seen. The upper lip is also prominent and bears on its under surface a series of minute denticles forming an inverted U-shaped pattern. The three pairs of adoral setae are simple, the anterior pair (or^ being shorter than the other two. The chelicerae have the characteristic 'ornithocephaline' outline common to other members of the family and carry two setae dorsally. Their fixed digits bear three prominent teeth and the movable digits only two ill-defined shallow projections. The five-segmented palp (fig. 5) has the setal formula (0-2-1-3-18). The solenidion of the palp-tarsus is borne on a prominence immediately distal to the lyriform fissure and of the 18 setae on this segment, 9 are eupathidia. The legs (figs. 7-10) are relatively stout and robust, legs I and IV being markedly longer than II and III. All the femora are divided. The setal formulae are : I (o-[4~6]- 5-6-31); II (i-[ 5 -5]-5-7-27) ; III ( 2 -[2-3]-3-6-26) ; IV (3-[3-3]-4-5-23), and the formulae for the solenidia are: 1(2-4-4) > H ( J 2 3) '> III (i-i-o) and IV (1-2-0). On tarsus I the famulus, which is slender and ciliated, lies on a small prominence between the lyriform fissure and the proximal solenidion. Of the 31 setae observed on this segment at least 9 can be distinguished as eupathidia. As in Acaronychus, only the pleural setae on tarsus II are eupathidial. On all four legs the ambulacrum has two well-developed lateral claws. A minute blunt central element can also be distinguished but this is evidently very much less developed than in other Acaronychid genera. TRITONYMPH. The number and disposition of dorsal setae are essentially similar to those of the adult. Ventrally (fig. 6) the anal and coxisternal regions are similar to those of the adult although in the figured specimen only 3 adanal setae were present on one side. There are 7 genital and 2 aggenital setae. The setal formulae for the legs are I (o-(4~6)-5-6-27), II (i-(3-5)-5~7-2i), III (2-(i-3)-3-6-2i) ; IV (3-(i-3)~4-5-i7). The formulae for the solenidia are the same as those for the adult. OTHER STAGES. Unknown. SYSTEMATIC POSITION The Palaeacaroidea, now comprising n genera, were first recognised as a group by Tragardh (1932) who considered they should constitute a separate sub-order of the Acari the Palaeacariformes. A few months later, however, Grandjean (1932) failed to confirm Tragardh's observation of the presence of mandibulary stigmata and concluded that the group should be incorporated with the Oribatei. In 1954, Grandjean considered the group should be regarded provisionally as a super-family the Palaeacaroidea comprising three distinct families, the Acaronychidae (=Archeonothridae), Palaeacaridae and Ctenacaridae. Himalacarus shows obvious affinites with the four genera, Acaronychus, Stomacams, 8 J. G. SHEALS Andacarus and Archeonothms which comprise the family Acaronychidae in so much as it possesses all the family attributes listed by Grandjean (1954), namely: 1. A 'naso' and its associated inferior protuberance. 2. A gastronotal series of 34 setae, the long dorsal setae being darkly pigmented. 3. No large asthenique zone. 4. A short hysterosoma without large mediodorsal and pygidial shields. 5. Normal genital papillae and unmodified anterior genital setae. 6. Four oral lips. 7. The anterior adoral setae being much shorter than the other two pairs. 8. A large rounded palp tarsus with 16-19 setae of which 9 are eupathidia. 9. A mandible with an 'ornithocephaline' outline. 10. A first tarsus with a large erect famulus implanted close to the lyriform fissure. Grandjean further recognised two sub-families, the Archeonothrinae for the 3 genera Archeonothrus , Stomacarus and Andacarus; and the Acaronychinae for the single genus Acaronychus. He separated the Archeonothrinae on the basis of their non-claviform famulus and normal female eugenital setae and on the presence of 2 aggenital setae, false lamellae and small internal apophyses at the bases of setae e 2 . The Acaronychinae were characterised by the absence of internal apophyses and false lamellae and by the presence of a dilated famulus, 3 aggenital setae and, in the female, swollen claw-like eugenital setae. However, Grandjean later (1957) observed that the female of Stomacarus macfarlani, the species which was subsequently made the type of Andacarus (Grandjean igsSa), had claw-like eugenital setae so that this feature had to be abandoned as a sub-familial character. Although having 3 aggenital setae, Himalacarus shows greater overall similarity to the Archeonothrine genera than to Acaronychus, thus the number of aggenital setae must also be eliminated from the list of sub-familial features, and, with the reduction of the number of diagnostic sub-familial attributes to two, it might be preferable to consider the family as a single relatively homogenous group. In table i, the condition of the genera in relation to the main characters which have been used in their classification is summarized. Since the main part of this account was prepared, Balogh and Csiszar (1963) have published a description of Stomacarus longicaudatus from South Argentina. Judging by the figure of the dorsum this species appears to be very close to H . chimalae and is probably congeneric with it. It can be noted that in S. longicaudatus the dorsal setae di and i\ are comparatively long and barbed whilst in H . chimalae these setae are shorter and smooth. MATERIAL. H. chimalae occurred only in the samples from the Milke Danra forest taken on 2nd December 1961. In all about 40 specimens were collected of which 15 were adult, apparently all males, the remainder being tritonymphs. The greater PRIMITIVE CRYPTOSTIGMATID MITES FROM NEPAL W Q othrus Himalacarus Very large Slender Present Present 3 ? Spearlike | 13 "bo w ^ '55 .2 3 (3 Tj CD O |H 1 CD "3 60 OT 3 Barbed s u, ^ -^j , , en .5 HOANO: O CD C < CD rrt Jj5 f? PC C en ^ rt CD ^ i-l c?) PM G en CD rt VH O N & ^. Presen I CD X! i M PQ PS V CD W s ' K tj CD H * ^ O ^ ^ *o be z d CD" ^ STATES Q JH -y O D hnT3 S 1 1s ^ J C/5 PL, Present Swollen Slender Absent G 'O TTpq & w H en CD < 'C ~ to "bo ^CD V G rrj SUTION OF IH 4-> | aj J d S y Tl CD g PC C w o |3 53 g^ CO ,_] C?) PM Present N Normal Slender Absent 'en O median ; 3 Smooth w s en H 2 CD M .1 CD g to ^ CD M S -I -3 H 13 K ^ ctf ^ V r3 en H ^s^ ^ fl" ^ ft 0) id CD o M u G c _o -r; c en IH ^ i i W X >^ "S *> f5 IH en tO 53 -Tn XJ CD en "p 03 CD ft 1 'ft u OS ^5 ^ H **< <1 ro do Cfi *tl cj' ro PQ (1 M a en ,0 CD ^ M ( O ifH IH r3 -4-* en IH "o CD c3 a CD J CD X5 ^ 13 en en 1 t) *O O en ^ -1 3 G ll P- < 'rt fig 8 - II > I2 [Possibly recognisable as a distinct form of digitula.] Diamphiodia digitula : Fell, 1962, 14. MATERIAL. Doderlein : Enoshima 22, i badly damaged specimen ; Enoshima 24, 3 specimens without discs; (PDoflein), Kachiyama, 5 specimens with discs, 6 without ; Doderlein : Enoshima, 4 specimens of forma leptopholida. The present material together with Matsumoto's observations and my drawings of the types of some of H. L. Clark's nominal species from Japan, made in 1953 in the U.S., suggest that Amphiura digitula tends to intergrade with Amphioplus ancistrotus (H. L. Clark). This is not so surprising as it may seem at first sight since the two genera are closely related and species of Amphiura with two distal oral papillae, such as A . digitula, need only the development of a single intermediate papilla to bridge the gap to Amphioplus. The species of both genera possess a first oral tentacle scale higher in the oral slit each side of the jaw, which is completely absent in Amphiodia, where the oral papillae are only numerically, not morphologically, intermediate between those of the type species of Amphioplus with four papillae and of Amphiura with two. H. L. Clark in 1911 referred both digitula and ancistrotus to Amphiodia but when Matsumoto studied the homologies of the oral papillae of Amphiurids he recognized that, although digitula has two distal oral papillae, these are based on the adoral shield rather than the oral plate and an oral tentacle scale (additional papilla in Matsumoto's terminology) is present ; accordingly the species should be referred to Amphiura. There are several comparable species of Amphiura, notably A. koreae, in which there are two distal oral papillae, similarly arising from the edge of the adoral shield (though the inner of the two may be based about the point of junction of the adoral shield and the oral plate) . Even if Fell's Diamphiodia proves to be sufficiently natural for recognition as a genus distinct from Amphiodia, there is no question that 50 A. M. CLARK digitula could be referred to it since the type species of Diamphiodia has oral papillae similar to Amphiodia. Matsumoto also referred Amphiodia ancistrota to Amphioplus because of its total of four (sometimes even five) oral papillae together with an oral tentacle scale. Both these species have rather unusual paired digits outside the radial shields, which prompted the name digitula ; also both have two tentacle scales of moderate size, somewhat similar arm plates and radial shields which are divergent and more or less narrow (the width is exaggerated in H. L. Clark's figure 69 (1911) of A. ancistrotus in comparison with the type specimen). The type of Amphiura digitula is distin- guished particularly by the spiniform shape of the inner one of the two distal oral papillae and by the diastema between this papilla and the infradental one (charac- teristic of the genus Amphiura as opposed to Amphioplus) also by the fine, almost granuliform, scaling of the disc ventrally and marginally, contrasting with the smoother and larger dorsal scales. In the type of Amphioplus ancistrotus there are four rounded oral papillae in continuous series and the ventral disc scales are smooth and not extremely fine. Between or close to one or other of these two extremes come two other Japanese amphiurids described by H. L. Clark in 1915 as the types of new species. One of these was referred to Amphiura and the other to Amphioplus (incidentally supporting Matsumoto's generic dispositions), so that Dr. Clark failed to compare them with either digitula or ancistrotus since he had left these both in Amphiodia in 1911. a FIG. 2. Amphiura digitula forma leptopholida H. L. Clark. Holotype of A. leptopholida, M.C.Z. no. 1365. (a) Dorsal and (b) ventral partial views of disc and oral area. The first of these nominal species (fig. 2), which he named Amphiura leptopholida, has oral papillae almost identical with those of the type of digitula, the inner of the two distal ones being spiniform ; also it has pairs of digits outside the radial shields. The JAPANESE AND OTHER OPHIUROIDS 51 only differences are that the ventral disc scales are not very small and granuliform but smooth and somewhat larger and the second from lowest of the four arm spines has a very slight terminal hook rather than a simple tip. The four specimens in the Munich collection from " Enoshima " differ from the complete specimens labelled "Enoshima 22 " or " 24 " and the rest from Kachiyama (all of which are immediately recognizable as digitula) in having the ventral disc scales smooth, not granuliform, and the second arm spine distinctly hooked on the segments beyond the base of the arm. Since the disc is easily lost in this species, judging from the proportion of incomplete specimens in the present collection, it is possible that the difference in disc scaling at least is correlated with regeneration. Three of the four Enoshima specimens exceed by 2 or 3 mm. the 6 mm. disc diameter of the types of both A. digitula and leptopholida, so the presence of five rather than four arm spines basally is only to be expected ; the fourth specimen, disc diameter 5 mm., has four spines basally. The differences being of such small magnitude, I believe that the types of digitula and Uptopholida are conspecific, nevertheless I think it worthwhile to distinguish specimens with smooth disc scales and hooked arm spines (if these characters prove to be consistently linked) as forma leptopholida, though such infrasubspecific taxa have no status in taxonomy, according to the Code of the International Commission. The second nominal species of H. L. Clark, 1915, is Amphioplus lobatod.es. Apart from minor differences in the proportions of the oral shields and lesser curvature of the second arm spine, there seems to me no reason why it should not be referred to the synonymy of A . ancistrotus. The Munich collection also includes three Amphiurids which, like Amphiura leptopholida , present an intermediate condition between A. digitula and Amphioplus ancistrotus. These are described under the heading of the latter species. Amphioplus ancistrotus (H. L. Clark) (Text-figs. 3 and 4) Amphiodia ancistrota H. L. Clark, 1911, 161-162, fig. 69. Amphioplus ancistrotus: Matsumoto, 1917, 171-172, fig. 43; Chang, 1948, 54-55, fig. n, pi. viii, fig. 7 ; Djakonov, 1949, 54, fig. 69 ; 1954, 6l >' Fell > I 9&2, 17. Amphioplus lobatodes H. L. Clark, 1915, 254-255, pi. vii, figs. 12, 13. MATERIAL. No details, presumably southern Japan, 2 specimens ; Haberer, Sagami Bay, 7. vii. 1904, 1 specimen of a distinct form ; Doderlein, Enoshima 24, 2 specimens of the same form. As mentioned under the heading of Amphiura digitula, I consider Amphioplus lobatodes to be a synonym of A . ancistrotus. The drawing of the type, which I made in 1954 when visiting the United States, shows no digits distal to the radial shields in the radius depicted. Unfortunately I omitted to notice whether this was also true of the other radii. H. L. Clark's brief description, largely comparative with A . lobatus, makes no mention of this feature. If digits are really absent, then it may be possible to distinguish lobatodes from ancistrotus but I doubt this. 52 A. M. CLARK In the two specimens with no detailed locality the oral papillae are all short and rounded. Although H. L. Clark's figure of the holotype shows the third (from inner- most) papilla as conical and somewhat elongated, I found on examination of the holotype that the three inner papillae are all rounded, though subequal, while the third one is not erect (fig. 3#) ; in fact they are just as described by Dr. Clark, accor- dingly I have no doubt that these two specimens are conspecific with the type. imm FIG. 3. Amphioplus ancistrotus (H. L. Clark), (a) Holotype of Amphiodia ancistrota H. L. Clark, U.S.N.M. no. 25601, oral papillae, (b) to (e) holotype of Amphioplus lobatodes H. L. Clark, M.C.Z. no. 1480 ; (b) dorsal and (c) ventral partial views of disc and oral areas, (d) segment with twenty-second dorsal arm plate, (e) arm spines of twelfth free segment. In (c) the ventral scaling is in reality barely distinct, especially proximally where it is very tenuous and broken in most interradii. The other three specimens in the Munich collection, however, have the third oral papilla, and to some extent the second also, erect and spiniform. Matsumoto's figures 436 and c both show these papillae as elongated and acute, but in Chang's drawing all the papillae are short and rounded as in the type of Amphioplus lobatodes (fig. 3b-e). (It should be noted that Chang's specimen has the disc naked ventrally so the identification may be incorrect, however the disc diameter was only 3 mm. and the small size might contribute to this deficiency.) In most other characters, namely the form of the arm plates, tentacle scales, hooked second arm spine and particularly the paired digits outside the radial shields, these three specimens agree with the first two mentioned but their oral papillae provide such a sharp contrast that I consider it worthwhile to designate this form with the two middle oral papillae erect and spini- form as forma anisopapilla of Amphiura ancistrotus. Although such infrasubspecific names are not recognized by the International Commission on Zoological Nomen- clature and have no status in taxonomy, I think they still have their uses for the sake of comparison. The shape of the oral shields, the extent of contact of the adorals and the proportions of the radial shields all appear to be variable in A . ancistrotus and it is interesting that some of the shapes exhibited approximate to those found in Amphiura digitula. (It may be noted here that H. L. Clark's figure of the holotype of Amphioplus ancistrotus shows the radial shields shorter than they really are and therefore relatively too wide.) There are two minor differences between the specimen of forma anisopapilla figured and the type of ancistrotus, namely in the former the disc scales are particularly indistinct and the dorsal arm plates appear thinner and JAPANESE AND OTHER OPHIUROIDS 53 more transparent. However, I doubt whether these differences are significant. The two specimens of unknown locality, which may be designated as forma ancis- trotus, have disc diameters of 9 mm. and 6 mm. In both of them the second from lowest arm spine is hooked but the larger specimen has five arm spines on the first four or five free segments, while the smaller one has no more than four basally. FIG. 4. (a) to (d) Amphioplus ancistrotus forma anisopapilla nov. Enoshima 24. (a) Dorsal and (b) ventral partial views of disc and oral area, (c) segment with fourteenth dorsal arm plate, (d) segment with twelfth ventral arm plate, (e) and (f) Amphioplus conductus Koehler, holotype, U.S.N.M. no. 41161, (e) a pair of radial shields with hyaline processes distal to them, (f) the second from lowest arm spine of a middle arm segment. In (a) the dorsal disc scales are very indistinct, also the dorsal arm plate is semi-transparent and the underlying ossicles are indicated below it. In (b) three of the third oral papillae shown are more or less fore-shortened, whereas the fourth has become appressed. The larger Enoshima specimen of forma anisopapilla (fig. ^a-c) has the disc 7 mm. in diameter. The radial shields are 1-3-1 -45 mm. long and 0-4 mm. in maximum width ; the proximal angle is acute. The smaller specimen from the same locality has the disc 3 mm. across and differs in having the oral shields longer and more acute proximally. The Sagami Bay specimen has the disc diameter just over 7 mm. and, unlike the other two, has the primary rosette just distinct ; its radial shields are smaller, with length : breadth 0-8 : 0-2 mm. and only the third oral papilla is spiniform. There are a few species of Amphioplus which similarly have the middle oral papillae spiniform, but all of them have only a single tentacle scale. They include A. dispar (Koehler), 1897, from the Indian Ocean, A. aciculatus from off the Congo and A. acutus from the Antarctic, both species of Mortensen, 1936 (Discovery Report), also A. gastracantha and notacantha (Liitken and Mortensen), 1899, from the East Pacific, though the two last-named differ further in having a few disc spinelets so that they may not be referable to Amphioplus at all but to Amphiacantha. Amphioplus ancistrotus is closely comparable with A. diacritus Murakami (1943, p. 225), the type of which has the disc diameter as much as 10 mm. The plates of 54 A. M. CLARK the primary rosette in A . diacritus are small but distinct, the radial shields appear to be longer and narrower than in ancistrotus and not so divergent, also the arm spines number only three even at this large size, the middle one the longest but evidently not hooked at the tip (though, as all the arms are said to be broken at the base, since specialized spines only develop further out on the arm, in most Amphiurid species that have them, it is possible that all the segments with hooks have been lost). A. diacritus is possibly more closely related to A. rhadinobrachius H. L. Clark, another Japanese species with only three arm spines. It is noteworthy that in Murakami's figure of A . diacritus there are paired processes distal to the radial shields though none are mentioned in his description. A few other species of Amphiurids from the western Pacific have similar processes. One such is Amphiacantha acanthina (H. L. Clark), also from Japan, which has distal angles to the dorsal arm plates like Amphioplus ancistrotus but differs in having only three arm spines as well as in the generic character of possessing spines on the disc. A second species is Amphioplus conductus Koehler from the Philippines, in the holotype of which I found paired hyaline bifurcating digits distal to the radial shields (fig. 4^) not observed by Koehler. The only significant differences I can see between conductus and ancistrotus are that, at a disc diameter of 6 mm., A. conductus already has six arm spines on some basal segments and the second of these is not hooked but bifurcated ; also the dorsal arm plates lack the marked distal angle. Amphioplus asterictus H. L. Clark (Text-fig. 5) Amphioplus asterictus H. L. Clark, 1915, 252, pi. vii, fig. 9-11. MATERIAL. Haberer, Sagami Bay, 7. vii. 1904, i specimen lacking the disc. a 3mm FIG. 5. Amphioplus asterictus H. L. Clark. Holotype, M.C.Z. no. 1486. (a) Disc and one arm base viewed dorsally, (b) two jaws and part of the disc viewed ventrally, (c) spines of two proximal segments. In (a) the primary disc scales are slightly hollowed, not thickened. JAPANESE AND OTHER OPHIUROIDS 55 Since H. L. Clark published only photographs of the type, I give here a drawing of it which I was able to make at the Museum of Comparative Zoology, thanks to the kindness of Dr. Elisabeth Deichmann. Amphioplus japonicus forma parvus (Matsumoto) Amphioplus relictus (part), Koehler, 1922, 180, pi. Ixxi, figs. 7, 8 ; Fell, 1962, 17. [Non A. relictus (Koehler), 1898]. Ophiophragmus japonicus var. parvus Matsumoto, 1941, 334-336, figs. 3, 4. MATERIAL. Doderlein: Tango, 40 fathoms (73 metres), I specimen; Tango 4, 7 specimens ; Kagoshima, 6 specimens ; no details, 7 specimens. The generic position of Ophiophragmus japonicus has been the subject of some controversy. Matsumoto (1915) included the species in Ophiophragmus because of the erect marginal scales, despite the fact that it has four oral papillae. Since the possession of only three papillae is partly diagnostic of Ophiophragmus, H. L. Clark (1918, p. 271) referred japonicus to Amphioplus noting that the erect " fence " of marginal scales is not homologous with the articulated marginal papillae or spinelets found in the other species of Ophiophragmus, including the type, 0. wurdemanni (Lyman), 1860 (of which no good figures existed up to that time, Koehler (1914) having confused the issue by publishing photographs of Amphiodia limbata under the name of Ophiophragmus " wundermanni "). Likewise H. L. Clark referred Ophiophragmus affinis Duncan, 1887, to Amphioplus [where the name became invalid as a homonym of A. affinis (Studer)], noting that it is probably identical with A. relictus (Koehler), 1898. Matsumoto had considered Duncan's affinis, which came from the Bay of Bengal, to be a close relation of japonicus together with Amphi- pholis andreae Liitken, 1872, from Java, Amphiura praestans Koehler, 1905, from the Flores Sea and Amphiodia periercta H. L. Clark, 1911, from Alaska and Oregon, U.S. [The last-named of these is a true Amphiodia but the others are currently referred to Amphioplus.'] In 1922 Koehler also referred affinis Duncan, as well as his own Amphiura relictus, to Amphioplus on the grounds of differences in the internal structure in comparison with Ophiophragmus wurdemanni. In 1941 Matsumoto noted that the additional oral papilla above the main series of superficial papillae (alternatively called the first oral tentacle scale) is absent in japonicus unlike the type species of Amphioplus, A. tumidus (Lyman). Accordingly he disagreed with H. L. Clark's transfer of the species to Amphioplus and retained it in Ophiophragmus. Having removed the three distal oral papillae of one series in a specimen from this collection, I can confirm that Matsumoto 's observation is correct. However, in the closely related species Amphioplus hastatus (Ljungman) the oral tentacle scale may not be visible in some specimens, even when the overlying papillae are displaced, though its development is clearly variable since in others a small scale or papilla can be distinguished. A. hastatus may also have spinous projections on the uppermost row of ventral scales similar to those occurring in japonicus, as Mortensen (1940, Echinoderms from the Iranian Gulf) has shown. A. M. CLARK w A j^ ^ M JJ3 'S ' ~ ^""^^^ ^*^^~^ *^"^^^ I O " ' O C-l O *O *O ^O O f^ l^ t^"* C *\ t*^ TJ* r^-* ^ ^o c^i c^ ro *o ^^ ^^ **^ ^^ ^O *^" '^h vo '"? to Tj--<1-Tt-'O T *' (r 5rOiO>O Nio ft, ^^3,666666666< 5OOOOOOOOOOOOOO OO -< w 1 N .5 ^3^" roO OO^IH^ -I" g .a vi vi g 5 ^ i 9 a o>O ioO vD r-- ^^ . <2 s M w N ^ a? 3. O O t^ O t^* O IH j M C C^C^I NNCI N S fl C< W * f S -si t-i c3 a &* PH ^ S ,fi ** -^- "S JAPANESE AND OTHER OPHIUROIDS 57 Accordingly, together with Fell (1962), I agree with H. L. Clark tha.t japonicus is better referred to Amphioplus, although I believe that, together with A. hastatus, depressus, relictus, andreae, laevis, praestans, megapomus, miyadii and some others, it may be possible to distinguish it as forming part of a subgenus marked off from Amphioplus tumidus, the type species of the genus, by the reduction of the oral ten- tacle scale coupled with linear arrangement of the oral papillae, contiguity of the radial shields and enlargement of the two tentacle scales of each arm pore. The present specimens are referred to Matsumoto's forma parvus, of A. japonicus because none of them have the row of enlarged outermost dorsal disc scales shown by Matsumoto in his figures of japonicus itself. I have not seen any specimens of the latter and am unable to assess whether or not the forma is worth retaining. The table given here shows some measurements of 21 specimens of A. japonicus forma parvus. It indicates that the relative size of the radial shields decreases as growth proceeds, from a maximum of just under half the disc radius at about 3 mm. disc diameter to a minimum of only a quarter when the diameter is over 8 mm. Since this ratio has been used to distinguish between related species belonging to this section of the genus Amphioplus it is clearly advisable that the total size of the disc should be taken into account when making such comparisons. When describing parvus in 1941, Matsumoto commented that it is allied to Amphio- plus megapomus H. L. Clark. Having examined the discless type specimen of megapomus I cannot agree with this. The dorsal arm plates of megapomus have a median distal peak, so appearing trilobed, and the arm spines are much more acute than in japonicus. In 1915, H. L. Clark noted that complete specimens of A. megapomus have relatively wide radial shields as in A. japonicus. Despite this, I believe that A. megapomus is more likely to prove conspecific with A. miyadii Murakami, 1943, also of Japanese origin, since it agrees in the oral structure and in the distinctive trilobed dorsal arm plates. Regeneration of the disc in those species of Amphioplus which are particularly liable to shed it, such as A. integer, results at first in abnormally short radial shields, though their relative length increases as growth proceeds. If the specimens with discs studied by Clark in 1915 were regenera- ting, the shields would probably be abnormally short. However, another possibility is that these intact specimens were not conspecific with the types of megapomus since Clark noted that their arm plates were not identical. A . miyadii is certainly more closely related to A . laevis (Lyman) and praestans (Koehler) with similarly elongated radial shields, than it is to A . japonicus. In contrast to the type of forma parvus, the majority of the present specimens have the six primary plates distinct, if not by larger size then by a slight hollowing of their surface. In the smallest specimens the primaries are partly contiguous with each other, having interstitial scales only at their corners (as in Koehler's photograph of the small type of A. hastatus (1927), Ark. Zool., 19, pi. 3, fig. 2) but in larger speci- mens the primaries are more or less widely separated. In 1922 Koehler referred sixty-four specimens from Kagoshima, Japan, as well as a number of Philippine specimens to Amphioplus relictus, noting that the relative width of the oral shields is variable and that some specimens (of which he figures one 5 8 A.M. CLARK from Kagoshima) have a distinct marginal row of erect disc scales. I have found that some of the specimens earlier (1905) identified as A. relictus by Koehler, at least those from " Siboga " stations 51 and 71 (Makassar and Molo Strait in the East Indies), have a distinct enlarged row of marginal disc scales contrasting with the uppermost row of smaller ventral scales, which are erect and project to form a scal- loped edge to the disc, as in some specimens of A. japonicus. Nevertheless, I think that Koehler's Kagoshima specimens are more likely to be referable to A . jpaonicus than relictus since the ventral view of one (1922, pi. 71, fig. 7) suggests that the distal edge of the oral shields is simply convex as in all the present specimens, without having a distinct, more or less constricted, distal lobe, as occurs even in Koehler's Manila specimen (pi. 71, fig. 6) where the shields are particularly wide in comparison with those of most Philippine and East Indian examples of the species, in which the distal lobe is usually as long as the proximal angle. [In the Manila specimen the shape is very like that found in one of the syntypes of A . andreae, seen by me at the Museum of Comparative Zoology, and it is possible that it should have been referred to andreae and not to relictus.'] As these remarks indicate, the distinctions between Amphioplus hastatus, depressus, relictus, andreae and japonicus are very subtle and in need of further consideration when a true assessment of variation and growth changes can be made together with re-examination of the types. Amphiodia craterodmeta H. L. Clark Amphiodia craterodmeta H. L. Clark, 1911, 155-157, fig. 65 ; Matsumoto, 1917, 182 ; Djakonov, 1938, 463 ; 1949, 54, fig. 72 ; 1954, 62. Diamphiodia craterodmeta : Fell, 1962, 14. MATERIAL. Domaschnew, nos. 53 and 55, 20. vi. 1900 and 27.vii.i9oo, 4 speci- mens ; Brashnikow, no. 6, July, 1899, 2 specimens ; Schmidt, nos. 9 and 10, Mauka, south-west Sakhalin, 46-47 sagenes (c. 100 metres), S.vi.igoi, 19 specimens. (All N. Japan Sea). Of the two specimens from Brashnikow's number 6, the larger has the disc diameter 8 mm. and the smaller 4-5 mm. The former has no distinct primary disc scales, the oral shields are very small with the distal lobe both sunken and constricted, also the infradental papillae are widely spaced and head the series of three oral papillae each side and the disc plates are all extremely well-defined. The smaller specimen from no. 6 differs in having the primaries much larger than the other scales. In comparison with Amphiodia fissa, this species differs in having narrower (but still mainly contiguous) radial shields, entire dorsal arm plates, smaller oral shields and larger adorals, the primary disc scales not so different from the other scales, four arm spines proximally and the ventral tentacle scales longer. Amphiodia fissa (Liitken) (Text-fig. 6) Amphipholis fissa Liitken, 1869, 12-13 (3O-3 1 )- JAPANESE AND OTHER OPHIUROIDS 59 Amphiodia rossica Djakonov, 1935, 4^5, figs. 7 and 8 ; 1938, 465-466, fig. 8 ; 1949, 55, fig. 71 ; 1954, 64, fig. 1 8. Diamphiodia fissa and rossica : Fell, 1962, 14. MATERIAL. Schmidt, nos. 3 and 5, Mauka, south-west Sakhalin, 20 and 15 sagenes (42 and 31 metres), 3.vi.i9Oi and 4.11.1901, 5 specimens ; no. 21, Posjet Bay, near the bay of Minanosok, 12 sagenes (26 metres), 6 . iv . 1900, 4 specimens ; no. 81, entrance to the bay of Tshogu-tschien-dogu, n specimens. a FIG. 6. Amphiodia fissa (Liitken). Specimen from Peter the Great Bay, M.C.Z. no. 4622, labelled as A . rossica Djakonov. (a) Dorsal and (b) ventral views of parts of the disc and oral area, (c) segment with fifteenth (split) dorsal arm plate, (d) lateral view of an arm base showing the individually projecting disc scales ; the middle arm spines fore-short- ened. NOMENCLATURE. Liitken described this species in Danish from material collected on the Amur coast (i.e. probably in the vicinity of Vladivostok), giving also a short latin diagnosis but no figures. The very unusual feature for an Amphiurid of the subdivided dorsal arm plates, which gave the species its name, taken in conjunction with the peculiar erect marginal scales of the disc, leaves me in no doubt that Amphiodia rossica Djakonov (also from the Amur region), which shares these charac- ters, was based on material of this same species. In his original description of Amphiodia rossica (1935) Djakonov made no mention of Amphipholis fissa but in 1954 he commented on the possible identity of the two ; however, with no published figure of the type of A. fissa in existence he evidently thought that sufficient doubt exists to justify the continued use of the name he had himself established. Since Liitken's name Amphipholis fissa has remained in oblivion for well over 50 years, I submitted to the International Commission on Zoological Nomenclature a petition to obtain the suppression of the name A. fissa as a nomen oblitum. This case 60 A. M. CLARK is suspended in view of the controversy regarding nomina oblita in general raised at the last colloquium on nomenclature. REMARKS. The smallest of the five specimens from Mauka has the disc diameter only 3'5 mm - an d the proximal tentacle scale on the lateral plate is reduced or absent on all but the basal pores, though the scale on the ventral arm plate is quite large. The dorsal arm plates are nearly all split and fairly regularly so, also the marginal disc scales are erect. In another small specimen with disc diameter 3 mm. the ventral scale is also absent on segments beyond the tenth and sometimes as far proximally as the fifth. The larger specimens have the disc diameter about 10 mm. Yet another small specimen, with disc diameter only 2-75 mm., has the dorsal arm plates fan-shaped and not split. It also has the plates of the primary rosette in direct contact with each other, whereas in specimens with disc diameter 5 mm. or more the primaries are all separated. The splitting of the dorsal arm plates seems to start with the distal plates and progress proximally, but even when the plates are not split the arms appear distinctly carinate. One specimen with disc diameter 4-5 mm. has the first five to seven plates entire ; it also has three arms regenerating, each with a soft, worm-like tip, not yet distinctly segmented. There are two tentacle scales only to about the twentieth segment, then one. Other variable characters include the erection of the marginal disc scales, which may be lacking, also the development of the distal lobe of the oral shields. Since Djakonov's figures are unnatural with regard to the mouth parts and not quite in agreement with the specimens I have seen, it seems worthwhile to give here a figure of an oral angle drawn from a specimen presented by Djakonov to the Museum of Comparative Zoology, Harvard. Family OPHIOTRICHIDAE Ophiothrix sp. ? eusteira H. L. Clark Ophiothrix eusteira H. L. Clark, 1911, 265-267, fig. 132 ; Matsumoto, 1917, 222-223, fig- 60. MATERIAL. Doderlein, y.xi.iSSi. Ill, 170 fathoms (311 metres), i specimen. The single specimen has the disc diameter 7-5 mm. The radial shields are large, c. 2-4 mm. long, and completely bare. The disc scales are clearly visible below the sparse covering of small, spaced, tapering stumps, though some have about three long thorns remaining on their tips. In the middle of the disc are a few short spines only, agreeing more with Matsumoto's figured specimen than with the holotype, which has the central spines very long. There is more resemblance to the specimen figured by Koehler (1922, pi. 45, fig. 6) under the name of Ophiothrix koreana, which I think would be better referred to 0. eusteira in view of my observations (below) about the types of 0. koreana. This specimen is only doubtfully identified as Ophiothrix eusteira because it already has as many as nine arm spines on the second free segment, whereas the holotype, with disc diameter 1-5 mm. greater, has only six or seven spines, according to H. L. Clark. JAPANESE AND OTHER OPHIUROIDS 61 Ophiothrix koreana Duncan (pi. I, figs. 3 and 4.) Ophiothrix koreana Duncan, 1879, 473-476, pi. xi, figs. 28-32 ; H. L. Clark, 1911, 257-262, figs. 127-128; Matsumoto, 1917, 220 ; Murakami, 1942, 20 ; 1943, 232-233 ; 1944, 267 ; Djakonov, 1954, 88. MATERIAL. Doderlein : Yogashima, 3 specimens; Yogashima, 2.xi.i88i, 3 specimens; Yogashima 2 & 3.xi.i88i, I specimen; Enoshima 18, 183 metres, 8.xi.i88i.II, i specimen ; Tango, 73 metres, 2 specimens. Duncan's figures of the syntypes of Ophiothrix koreana were not very good so I give here photographs of the upper surfaces of two of them. These show that, unlike the specimens figured under the name of koreana by H. L. Clark, which subsequent writers have taken as models of the species, their radial shields carry a number of thorny, mostly trifid, stumps at their distal ends as well as proximally, though the stumps appear more sparse on the shields than on the scales. The largest syntype, with disc diameter 8 mm., has about 30 stumps on each radial shield including four or five at the distal tip, while the smallest specimen, disc diameter 4-5 mm., has about 12 stumps on each shield. This smallest syntype, as well as one other out of the six, has some thorny disc spines in addition to the stumps but the remaining four have no spines and superficially resemble H. L. Clark's figure of Ophiothrix hylodes (1911, p. 263, fig. 130), which Matsumoto and subsequently Clark himself referred to the synonymy of Ophiothrix marenzelleri Koehler, 1904^. The type of hylodes has the disc stumps (or " stout, blunt, rough spines " in Dr. Clark's terminology) with irregular thorns along their lengths and at their tips rather than the trifid form found in the types of 0. koreana and also in the specimens identified as 0. marenzelleri by Koehler (1922). However, some of the specimens of the present collection also have " stumpy spines " of this kind, so the form of the stumps cannot be considered as diagnostic. H. L. Clark did not mention 0. marenzelleri in his work of 1911, but commented that 0. hylodes might be " only an extreme variety of koreana ". Mat- sumoto repeated this supposition and noted that " the sublittoral form (of marenzel- leri) approaches 0. koreana more closely than the littoral form ", this sublittoral form sometimes having a few long spines on the disc as well as the close covering of thorny stumps or tubercles. Matsumoto did not comment on the density of tubercles on the radial shields as being any less in the littoral form than in the sublittoral but Koehler (19040, p. 104) noted that in the types of 0. marenzelleri the radial shields are almost as closely covered with stumps as the rest of the disc. In the present collection I have identified as 0. marenzelleri specimens in which the stumps appear equally dense on the radial shields as on the rest of the disc and those with the radial shields distinctly barer I have named 0. koreana. However, so close are the types of 0. koreana to the form generally accepted as 0. marenzelleri that I think the latter name might well be considered a synonym. Judging from the variation of Ophiothrix fragilis in the north Atlantic it is quite possible that the Japanese specimens hitherto designated as koreana, marenzelleri, hylodes and possibly even eusteira H. L. Clark, 1911, all represent a single very variable species. Alter- natively the specimens with almost or completely bare radial shields such as those 62 A. M. CLARK figured under the name of 0. koreana by H. L. Clark in 1911 might be referred to 0. eusteira. A much larger collection than that available to me is necessary before a proper assessment of this problem can be made, so notorious is the genus Ophiothrix for its variability. Ophiothrix stabilis Koehler Ophiothrix stabilis Koehler, 19040, 84-86, figs. 46-49 ; Matsumoto, 1917, 224 ; Murakami, 1944, 268. Ophiothrix ciliaris : H. L. Clark, 1911, 257. [Non O. ciliaris (Lamarck), 1816.] MATERIAL. Schmidt, 1901, Nagasaki, 3 specimens. These three specimens were taken with one which I have referred to Ophiothrix marenzelleri. They are immediately distinguishable from it by the disc stumps which are shorter and better-termed tubercles than stumps ; they also have a coronet of usually five to seven divergent points. Some disc spines are also present as in the holotype 0. stabilis but there is a difference in the shape of the ventral arm plates. None of these have the distal edge markedly convex as Koehler describes it in the holotype (though his figure may not be reliable since it shows the ventral-most arm spines with their hooks directed inwards instead of outwards or proximally, when aligned perpendicularly, as in other species of Ophiothrix and in the three specimens of 0. stabilis in the present collection). Since Matsumoto used the convex or concave shape of the distal edge of the ventral arm plates as an important character in his key to the Japanese species of Ophiothrix, the identification of these three specimens as 0. stabilis invalidates this particular dichotomy of his key. This dichotomy was already untrue for 0. eusteira, included with stabilis as having convex edges to these plates, whereas only two proximal plates are so depicted in H. L. Clark's figure of eusteira and none at all in Matsumoto's own figure, where all the plates appear incurved in the middle of the distal side. Ophiothrix panchyendyta H. L. Clark Ophiothrix panchyendyta H. L. Clark, 1911, 264-265, fig. 131 ; Matsumoto, 1917, 219. MATERIAL. Okinose Bay, 600 metres, i specimen. The disc spines appear longer than shown in H. L. Clark's figure of the type. The disc diameter of this specimen is 10-5 mm. and the majority of the spines towards the centre of the disc are 1-7-2-0 mm. long. The length of the corresponding spines was not specified in the type. There are hardly any short stumps on the disc besides the spines. Close to the proximal interradial edge of each radial shield is a row of four to eight short, tapering thorny-tipped spines or stumps, but superficially the radial shields appear relatively large and very bare. There is no spine in the middle of the first dorsal arm plate, unlike the type of 0. panchyendyta, but the arm plates are similarly conspicuously granular in surface texture. Although H. L. Clark sus- pected that neither of these last two characters might prove to be diagnostic, the latter at least holds good for this specimen. JAPANESE AND OTHER OPHIUROIDS 63 Ophiopsammium rugosum Koehler Ophiopsammium rugosum Koehler, 1905, 116-117, pi- x "i n g s - 4~6 ; 1930, 197. MATERIAL. Sprater, Gulf of Siam, 1910/4500, clinging to a pennatulid, 10 speci- mens. The outlines of the radial shields are visible through the skin and granules. The disc spines are very variable in occurrence, even in different interradii of the same specimen ; there may be just two widely separated spines right on the edge of an interradius or about 10 conical spines more or less inset or some shorter conical tubercles (all these with truncated tips). One specimen has two to four very short, thick, pointed tubercles in each dorsal interradius level with the tangent to the distal ends of the radial shields, then at the very edge of the disc or slightly towards the ventral side there are two separate clusters each of about four short conical spines. Another specimen has about 35 very large low granules in the central part of the disc and central interradial areas, not projecting out of the skin. The colour of the disc is pinkish dorsally and dark-red ventrally. There are also patches of intense black spots dorsally of varying extent and conspicuousness in the different specimens. One has particularly large spots in asymmetrical areas of the disc and arms, so that one pair of radial shields and one adjacent shield are covered with big spots and one pair with only fine spots, while the rest are parti-coloured. On the arms the spotted patches extend over about four segments. The tube feet are dark red. These specimens appear to have been epizooic on the pennatulid, since they are well entwined within its branches. Possibly the peculiarities of the genus are correlated with this habit. Family OPHIODERMATIDAE Ophioconis permixta Koehler Ophioconis permixta Koehler, 1905, 14-15, pi. ii, figs. 4 & 7 (5 in caption). Ophiurodon permixtus : Matsumoto, 1915, 84 ; 1917, 315 ; Koehler, 1922, 352. MATERIAL. Xenia reef, Dar-es-Salaam, Tanganyika, 3 specimens. The use of the original generic name for this species results from a comparison between the three present specimens and two examples in the British Museum collections of Ophioconis forbesi (Heller), 1863, the type species of Ophioconis Liitken, 1869, one specimen from the Adriatic originating with Dr. Heller himself and the other from La Ciotat in the south of France. Unfortunately I have no material of Ophioconis grandisquama Koehler, 1904, the type species of Ophiurodon Matsumoto, 1915 but there are several examples in the British Museum collection of Ophioconis cupidum Koehler, which Matsumoto referred to Ophiurodon together with Ophioconis cincta Brock and 0. permixta Koehler, all of which I consider to be congeneric with Ophioconis forbesi. Matsumoto's restricted Ophioconis included only 0. forbesi and 0. brevispina Ludwig, 1880, both from the Mediterranean and neither known to Matsumoto at 64 A. M. CLARK first hand. He distinguished Ophioconis from Ophiurodon, Ophiuroconis , Ophiuro- chaeta, Ophiolimna and Ophiarachna by the arm spines, which he said are " very short, lying flat on the arm, hyaline ". In fact the dorsal arm spines of the first two free segments on both specimens of Ophioconis forbesi seen by me are equal to or even slightly longer than the corresponding dorsal arm plates and in the better-preserved (French) specimen (pi. I, figs, i & 2) most of the spines are erect, only those of the distal parts of the arms having become flattened against the arms in preservation. In Ophioconis brevispina and in 0. vivipara Mortensen, 1925, from Morocco, the arm spines are much shorter, only about half as long as the corresponding segments near the bases of the arms, judging from the figures, but it is 0. forbesi which is the type species and criterion of the genus. The three specimens of 0. permixta in the Munich collection are all small, with the disc diameter little more than 3 mm. but their arm spines have very similar propor- tions and alignment to those of 0. forbesi and, apart from the scattering of spinelets among the granules of the disc, there is little difference between them, certainly not enough to warrant a generic separation. Hertz (1927) has described a subspecies nueva of Ophiurodon grandisquama based on four specimens from Madeira. She maintains that these are very similar to a Japanese specimen of 0. grandisquamawhich she has studied. Certainly the arm spines appear from her photographs to be rela- tively longer than those of Ophioconis forbesi , since most of the spines of the proximal half of the arm exceed or at least equal the segments in length. Nevertheless I can see no difference of sufficient magnitude to justify generic separation. If Hertz is right in asserting that nueva is closely related to grandisquama then Ophiurodon must be considered as a synonym of Ophioconis. However, this needs confirmation from a study of Japanese specimens. Certainly I do not think that the presence of disc spinelets in permixta warrants a generic distinction from Ophioconis forbesi, par- ticularly as H. L. Clark (1938) has found the occurrence of such spinelets in the related species Ophioconis cincta Brock to be very variable. In eleven specimens from northern Australia he found that five had a distinct marginal fringe of disc spinelets, as in the type of cincta, while five others had these marginal spinelets poorly developed or absent and the last specimen had both marginal and dorsal spinelets, as in 0. permixta. This last observation casts doubt on the validity of permixta as distinct from cincta, but much better sampling of these rather inconspicuous ophiuroids is needed before a true appreciation of the specific limits can be reached. Ophiurodon cupidum should also be referred to Ophioconis ; having no disc spinelets it is the Indo-Pacific counterpart of the Mediterranean Ophioconis forbesi , the main difference being only of colour pattern. In the small specimens of 0. permixta the second tentacle scale is present for the first six to twelve segments, on one arm even to the seventeenth. This record from East Africa provides a considerable extension of range for Ophioconis permixta, which was previously known only from the East Indian area. Pectinura anchista H. L. Clark Pectinum anchista H. L. Clark, 1911, 23-25, fig. i ; Matsumoto, 1917, 322. JAPANESE AND OTHER OPHIUROIDS 65 MATERIAL. Haberer, no. 4168 (pt.), Sagami Bay, between Ito and Hatsushima Island, c. 150 metres, March 1903, 2 specimens. Matsumoto thinks that Pectinura anchista may be a synonym of P. cylindrica (Hutton), 1872, from New Zealand. On morphological grounds it seems to me to be equally likely that anchista is synonymous with P. aequalis (Lyman), 1880 and on zoogeographical grounds more probable since the " Challenger " collected aequalis off New Guinea and it has since been recorded by Koehler (1904 and 1922) from the vicinity of Celebes and from the Philippines, while Murakami (1944) records it from Yaeyama, Japan. Murakami did not state the size of his specimens but Lyman's type of P. aequalis has the disc diameter 25 mm. and Koehler 's Philippine specimens are equally large or larger still. The presence of 10 arm spines in Koehler's specimens, whereas these from Sagami Bay have only seven proximally, may be attributable to the much smaller size of the latter, the disc diameter being only u or 12 mm. The holotype of P. anchista also has seven arm spines ; its disc diameter is 14 mm. I think that a direct comparison between Japanese and East Indian specimens of similar size will show the two to be indistinguishable. Family OPHIURIDAE Aspidophiura uniumbonata Murakami Aspidophiura uniumbonata Murakami, 1942, 21-22, fig. 8. MATERIAL. No details, presumably southern Japan, i specimen. The disc diameter is 4-5 mm. and the longest stump of arm remaining attached is only 3-5 mm. long. There is a suggestion of a boss on the central disc plate as in the type of Aspidophiura uniumbonata. Two of the five primary radials are irregularly subdivided but the other three are regular and about equal in size to the radial shields. The oral shields are wider than in either A . watasei or A.forbesi but agree with those of the type of uniumbonata. The only difference from the type is that the uppermost arm spine of at least the first and second free segments is just longer than the segment and the middle arm spine too is relatively longer, almost equal to the segment in length. In the two other species the spines are shorter. A number of other species were included in the collection for which no particular comment was elicited. They are as follows : Southern Japan Astrodendrum sagamimtm (Doderlein) Haberer, no. 4118, Fukuura, Sagami Bay, c. 150 metres, 1-2 . iii . 1903, 2 specimens. 66 A. M. CLARK Ophiacantha pentagona Koehler Doflein, no. 320, Sagami Bay, 180 metres, 25 .x. 1904, 10 specimens ; Doderlein : Enoshima 14, i specimen; Enoshima 22, 2 specimens; 2.xi.i88i, IV, i specimen. Ophiopholis mirabilis (Duncan) Doflein: Yogashima (Misaki), 150 metres, 31. x. 1904, 4 specimens; Sagami Bay, towards Boshu, 120 metres, i.x.1904, i specimen; Haberer, Fukuura, Sagami Bay, March, 1903, i specimen ; (Haberer?), no. 4356, Sagami Bay 7 . vii . 1904, i specimen. Ophiopholis sp.? brachyactis H. L. Clark Doderlein, Enoshima 18, 8.xi.i88i, II, i specimen. Ophiactis pteropoma H. L. Clark Doderlein : Enoshima 25, 2 specimens ; 2.xi.i88i, III, i specimen. Amphipholis sobrina Matsumoto Doderlein, Enoshima 22, i specimen ; Tango, 40 fathoms (73 metres), i speci- men. Amphiacantha acanthina (H. L. Clark) Doflein, station 16, no. 22b, 3 specimens. Ophiothrix marenzelleri Koehler Schmidt, Nagasaki, 1901, i specimen ; Haberer, no. 4356, Sagami Bay, 7. vii. 1904, i specimen; Doflein, Yogashima (Misaki), 150 metres, 31. x. 1904, I specimen; (Doflein?), Kachiyama, i specimen; Doderlein, 2.xi.i88i, III, 2 specimens. Ophiomastix mixta Liitken Schmidt, Nagasaki, i.iii.igoi, 3 specimens. Ophiarachnella gorgonia (Miiller and Troschel) Schmidt, Nagasaki, February-March, 1901, i specimen. Ophioplocus japonicus H. L. Clark Schmidt, Nagasaki, i.iii.igoi, i specimen. Ophiura kinbergi (Ljungman) Schmidt, Nagasaki, i specimen; Doflein, Tzushi (?), 130 metres, n.xi.i9O4, 2 specimens; Doderlein, Tagawa (?), 7.xi.i88i, II, i specimen ; (Doderlein?), Tagawa, i specimen. JAPANESE AND OTHER OPHIUROIDS 67 Ophiozonetta projecta (Koehler) Doderlein, Yogashima, i specimen. Ophiozonella longispina (H. L. Clark) Haberer : no. 4168 (pt.), Sagami Bay, between Iso and Hatsushima Island, c. 150 metres, March, 1903, 20 specimens ; Fukuura, Sagami Bay, March, 1903, i specimen ; Doderlein : Enoshima 22, i specimen ; 2 .xi. 1881, III, i specimen. Stegophiura sladeni (Duncan) Haberer : no. 4356 (pt.), Sagami Bay, 7.vii.i904, i specimen ; no. 4168 (pt.), Sagami Bay, between Ito and Hatsushima Island, c. 150 metres, March, 1903, i specimen ; Sagami Bay, 1900, 2 specimens. Stegophiura vivipara Matsumoto Doflein, no. 582, Uraga Channel, Sagami Gulf, 150 metres, 22.x. 1904, i speci- men ; (Doderlein?), Yogashima, 2 and 3.xi.i88i, 4 specimens. Ophioleuce charischema (H. L. Clark) Doderlein : Enoshima 22, i specimen ; Yogashima 2 and 3.x.i88i, i specimen ; no details, i specimen. Amur-Sakhalin (Localities probably east of Vladivostok, unless Sakhalin included). Ophiacantha adiaphora H. L. Clark Schmidt, no. 64, 19. v. 1900, i specimen. Ophiacantha bidentata (Retzius) Schmidt : no. 57, i specimen ; no. 48, off Cape Povorotny, 230-196 metres ; 1 8 specimens. Ophiopholis aculeata (Linnaeus) Schmidt : no. 68, 24^.1900, i specimen ; no. 46, Strelok Channel, i specimen ; Brashnikow : no. 7, June 1899, I specimen ; no. 14, June, 1899, i specimen ; no. 49, August, 1899, 4^ specimens ; nos. 19 and 14, 1899/1901, n specimens ; Domaschnew, no. 59, July, 1900, i specimen. Amphiura lepidevaspis Djakonov Schmidt, no. 10 (pt.), Mauka, S.W. Sakhalin, 46-47 sagenes (c. 100 metres), 8 . vi . 1 901 , i specimen . Amphiodia craterodmeta H. L. Clark Domaschnew, no. 55, 20 . vi . 1900, 3 specimens ; no. 53, 27 . vii . 1900, i specimen ; Brashnikow, no. 6 (pt.), July, 1899, 2 specimens ; Schmidt, nos. 9 (pt.) and 10 (pt.), Mauka, S.W. Sakhalin, 46-47 sagenes (c. 100 metres), S.vi.igoi, 19 specimens. 68 A. M. CLARK Amphioplus macraspis (H. L. Clark) Schmidt, no. 81 (pt.), 1900, 2 discless specimens. Stegophiura nodosa (Lutken) Schmidt, no. 47, 95-100 (? sagenes), 9. v. 1900, I specimen. Ophiura leptoctenia H. L. Clark Schmidt : nos. 8, 9, Mauka, S.W. Sakhalin, 46-47 sagenes (c. 100 metres), S.vi.igoi, 5 specimens ; no. 48 (pt.) off Cape Povorotny, 230-196 metres, 1900, 9 specimens ; no. 58, same locality, 200-132 metres, 3 specimens. Ophiura quadrispina H. L. Clark Schmidt, no. 48 (pt.), off Cape Povorotny, 230-196 metres, 1900, i specimen. Ophiura sarsi Liitken Schmidt : nos. 9, 10 (pts.), Mauka, S.W. Sakhalin, 46-47 sagenes (c. 100 metres), S.vi.igoi, 16 specimens; no. n, "east facing Patrok Bay (?)," 22.iii.i9OO, i specimen ; no. 30 (locality illegible), 17.^.1900, I specimen; no. 48 (pt.), off Cape Povorotny, 230-196 metres, 1900, 2 specimens ; no. 69, between Askold and (?) Skriplev Islands, 55 sagenes (c. 115 metres), 24^.1900, 2 specimens: Brashnikow : no. 6 (pt.), July, 1899, 3 specimens ; no. 35, July, 1899, 2 speci- mens. Ophiura sarsi vadicola Djakonov Schmidt : no. 18, off Cape Povorotny, 3.^.1900, n specimens ; no. 24, Aniva Bay, S. Sakhalin, 13-14 sagenes (27-29 metres), 28.viii.i9Oi, i specimen ; no. 46, Strelok Channel, 48-42 (? sagenes), 8. v. 1900, i specimen; no. 69 (pt.), between Askold and (?) Skriplev Islands, 55 sagenes (c. 115 metres), 24^.1900, i specimen ; Brashnikow, no. 35, i6.vii.i899, 7 specimens. Ophiura maculata (Ludwig) Brashnikow : no. 15, June, 1899, 2 specimens ; no. 27, July, 1899, i specimen. Miscellaneous Localities Aster onyx longifissus Doderlein Albatross st. 2892 (Santa Barbara Channel, S. California) 35 st. 2979 (Anacapa Island, S. California) 30 st. 3198 (off central California) 23 st. 2891 (Point Conception, Oregon) 26 Aster onyx loveni Muller and Troschel Albatross st. 2923 (off San Diego, California) i st. 3787 (off Punta Gorda, N. California) i st. 5637 (Molucca Islands) i JAPANESE AND OTHER OPHIUROIDS 69 Asteroschema (or Ophiocreas) spp. (Doderlein, 1927, regards Ophiocreas as hardly distinct from Asteroschema even at the subgeneric level, but Mortensen and Fell since consider Ophiocreas to be a distinct genus. With- out the disc no distinction is possible) . Albatross st. 5621 (Molucca Islands) i arm (Doderlein recorded Asteroschema (Ophiocreas) gilolense from this station) . Albatross st. 5634 (Molucca Islands) i arm (Doderlein recorded A. (0.) ambonesicum from this station). Amphiura chiajei Forbes Villefranche. Dr. Neresheimer i Ophictais savignyi Miiller and Troschel Ceylon. Haeckel i Ophiactis simplex (Le Conte) Academy Bay, Santa Cruz, Galapagos Islands Foerster, 15. v. 1959 2 Ophiothrix sp. prob. fragilis (Abildgaard) 44 48' N : 13 45' E (northern Adriatic) 36 metres, 26. ix. 1895 i Ophiothrix angulata (Say) Kingston Harbour, Jamaica. Dr. Heitz, 30.xii.i9O3 c. 50 Macr ophiothrix hirsuta (Miiller and Troschel) Ras Mohammed, Red Sea. Dr. Hofer, 1892 i Ophiothela danae Verrill Thursday I. Regenerating 3 arms and half the disc i Ophiocoma scolopendrina (Lamarck) Red Sea, Schadwan I. (? spelling), southside. H. Haas, 2y.iii.i956 5 Mombasa, E. Africa, May, 1960, Papp 5 Washington Museum, no. 46985, Philippines, E. A. Mearns, 1912 2 (badly damaged) Ophioderma cinereum Miiller and Troschel New Port, Curacao. Dr. Hellmich, 24.1937 i 7 o A. M. CLARK Ophioderma longicauda (Retzius) Tunis. Pet. Muller, 13. ix. 1960 i Bathypectinura conspicua (Koehler) Albatross st. 5215 3 st. 5219 2 Ophionereis albomaculata E. A. Smith Academy Bay, Santa Cruz, Galapagos Islands, Foerster, 15 . v. 1959 5 OphiuroUpis gelida (Koehler) German Siidpolar-Expedition : 25 . vi . 1902 4 17. iv. 1902, 385 m. (one infested with sponge lophon) 9 12 . viii . 1902 (largest with lophon) 5 14. iv. 1902 2 REFERENCES CHANG, F.-Y. 1948. Echinoderms of Tsingtao. Contr. Inst. Zool. Acad. Peiping, 4, 33-104, 24 figs. 1 1 pis. CLARK, H. L. 1911. North Pacific Ophiurans in the collection of the United States National Museum. Bull. U.S. not. Mus., 75, 1-302, 144 figs. 1915. Catalogue of recent Ophiurans. Mem. Mus. comp. Zool. Haw., 25 (4), 165-376, 20 pis. 1918. Brittle-stars, new and old. Bull. Mus. comp. Zool. Harv., 62 (6), 265-338, 8 pis. 1938. Echinoderms from Australia. Mem. Mus. comp. Zool. Harv., 55, viii + 596, 63 figs. 28 pis. - 1946. The Echinoderm Fauna of Australia. Publ. Carneg. Instn., 566, 1-567. DJAKONOV, A. M. 1935. New Ophiurans of the genus Amphiodia from the Sea of Japan. Dokl. Akad. nauk. SSSR., n.s. 3, 379-380. - 1938. The Echinodermata of Siaukhu Bay (Japan Sea). Reports of the Japan Sea Hydrobiological Expedition of the Zoological Institute, Academy of Sciences of the U.S.S.R. in 1934, pt- T > 4 2 5-498, 10 figs. [In Russian]. 1949. Definitions of Echinoderms from far-eastern seas. Bull. Pacif. Inst. Fish. Oceanogr. 30, 1-130, 22 pis. [In Russian]. 1954. Ophiuroids of the Soviet Union. Tabl. anal. Faune URSS, 55, 1-136, 47 figs, i pi. [In Russian]. DODERLEIN, L. 1898. Ueber einige epizoisch-lebende Ophiuroidea. In Semon, Zoologische Forschungsreisen in Australien und dem Malayischen Archipel. Denkschr. med.-naturw. Ges. Jena, 8, 483-488, pi. xxxvii. DUNCAN, P. M. 1879. On some Ophiuroidea from the Korean Seas. /. Linn. Soc. (Zool.), 14, 445-482, pis. ix-xi. - 1887. On the Ophiuridae of the Mergui Archipelago, collected for the Trustees of the Indian Museum by Dr. John Anderson. /. Linn. Soc. (Zool.), 21, 85-106, figs. 28-40, pis. viii, ix & xi. FELL, H. B. 1960. Synoptic keys to the genera of Ophiuroidea. Zool. Publ. Viet. Univ. N.Z., no. 26, 1-44, 6 figs. 1962. A revision of the major genera of Amphiurid Ophiuroidea. Trans, roy. Soc. N.Z. (Zool.), 2, 1-26, i pi. JAPANESE AND OTHER OPHIUROIDS 71 HERTZ, M. 1927. Die Ophiuroiden der Deutschen Tief see-Expedition, i. Wiss. Ergebn. Valdivia 22 (3), 59-122, 7 figs. pis. vi-ix. KOEHLER, R. 1897. Echinodermes recueillis par 1' Investigator dans 1'Ocean Indien. i. Les Ophiures de Mer Profonde. Ann. Sci. Nat., (8), 4, 277-372. 1898. Echinodermes recueillis par 1'Investigator dans 1'Ocean Indien. 2. Les Ophiures littorales. Bull. sci. Fr. Belg., 31, 54-124, pis. ii-v. 1904. Ophiures de 1'Expedition du Siboga. i . Ophiures de mer profonde. Siboga Exped., 45a, 1-167, 36 pis. i9O4a. Ophiures nouvelles ou peu connues. Mem. Soc. Zool. Fr., 17, 54-119, 98 figs. 1905. Ophiures de 1'Expedition du Siboga. 2. Ophiures littorales. Siboga Exped., 45b, 1-142, 18 pis. 1914. A contribution to the study of Ophiurans of the United States National Museum. Bull. U.S. nat. Mus., 84, 1-173, 18 pis. 1922. Ophiurans of the Philippine Seas and adjacent waters. Bull. U.S. nat. Mus., 100 (3), 1-486, 103 pis. - 1930. Papers from Dr. Th. Mortensen's Pacific Expedition, 1914-16. LIV. Ophiures recueillis par le Docteur Th. Mortensen dans 1'Archipel Malais. Vidensk. Medd. dansk. naturh. Foren., 89, 1-295, 2O pl s - LJUNGMAN, A. 1866. Ophiuroidea viventia hue usque cognita enumerat. Ofvers. Vetensk- Akad. Fork. Stockh., 23, 303-336. LUTKEN, C. F. 1869. Additamenta ad Historiam Ophiuridarum. c. Beskrivende og kritiske Bidrag til Kundskab om Slangestjernerne. K. danske Vidensk. Selsk., 5 (8), 24-109, 3 figs. - 1872. Ophiuridarum novarum vel minus cognitarum descriptiones nonnullae. Overs, danske Vidensk. Selsk. Fork., 77, 75-158, 2 pis. LUTKEN, C. F. & MORTENSEN, T. 1 899. Reports on an exploration off the west coasts of Mexico, Central and South America, and off the Galapagos Islands. XXV. The Ophiuridae. Mem. Mus. comp. Zool. Harv., 23 (2), 97-208, 23 pis. LYMAN, T. 1879. Ophiuridae and Astrophytidae of the " Challenger " Expedition. II. Bull. Mus. comp. Zool. Harv., 6 (2), 17-83, 8 pis. 1882. Ophiuroidea. Rep. Sci. Res. Challenger, Zool. 5, 1-386, 46 pis. MARKTANNER-TURNERETSCHER, G. 1887. Beschreibung neuer Ophiuriden und Bemerkungen zu bekannten. Ann. naturh. (Mus.) Hofmus. Wien, 2, 291-316, pis. xii, xiii. MATSUMOTO, H. 1915. A new classification of Ophiuroidea : with descriptions of new genera and species. Proc. A cad. nat. Sci. Philad., 67, 43-92. 1917. A Monograph of Japanese Ophiuroidea, arranged according to a new classification. J. Coll. Sci. Tokyo, 38 (2), 1-408, 100 figs. 7 pis. 1941. Report of the biological survey of Mutsu Bay and vicinities. Sci. Rep. Tohuku Univ., (4) 16, 331-344, 10 figs. pis. xxi-xxiii. MURAKAMI, S. 1942. Ophiurans of Izu, Japan. /. Dep. Agric. Kyushu Univ., 7 (i), 1-36, 12 fififS. I943 . Report on the Ophiurans of Palao, Caroline Islands. Report on the Ophiurans of Yaeyama, Ryukyu. Ophiurans from some Gulfs and Bays of Nippon. /. Dep. Agric Kyushu Univ., 7 (4-6), 159-204, figs. 1-17 ; 205-222, figs. 1-2 ; 223-234, figs. 1-2. I944 . Report on the Ophiurans from off Ogasawara Islands and from off the Yaeyama group, Nippon. Note on the Ophiurans of Amakusa, Kyusyu. /. Dep. Agric. Kyushu Univ., 7 (7-8), 235-257, figs. 1-14 ; 259-280, figs. 1-5, pi. i. PLATE i FIGS, i, 2. Ophioconis forbesi (Heller). B.M. reg. no. 94. u . 19.2, from La Ciotat, S. of France. X3- FIGS. 3, 4. Two syntypes of Ophiothrix koreana Duncan, B.M. reg. no. 80.1.3.13, from the Korean Straits, 42 metres, both X3. Bull. EM. (N.H.) Zool. 13, 2 PLATE i PRINTED IN GREAT BRITAIN BY ADLARD & SON LIMITED BARTHOLOMEW PRESS, DORKING 3 5EF19 THE FRESHWATER GASTROPOD MOLLUSCS OF WEST CAMEROON C. A. WRIGHT BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 13 No. 3 LONDON: 1965 THE FRESHWATER GASTROPOD MOLLUSCS OF WEST CAMEROON BY C. A. WRIGHT British Museum (Natural History) Pp. 73-98 ; Plates 1-3 ; 21 Text-figs BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 13 No. 3 LONDON: 1965 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 13, No. 3 of the Zoological series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. Trustees of the British Museum (Natural History) 1965 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued September, 1965 Price Sixteen Shillings THE FRESHWATER GASTROPOD MOLLUSCS OF WEST CAMEROON By C. A. WRIGHT SCHISTOSOMIASIS, the disease of man caused by blood-flukes of the genus Schistosoma, is commonly associated with the savannah and semi-arid regions of Africa and the Middle East. In 1953 a focus of infection with S. haematobium was reported by Zahra in two crater lakes in the rain-forest area of West Cameroon. In 1957 Mandahl- Barth (1957!)) reported two species of Bulinus (one of them new) from one of these crater lake foci and, in reply to a request made to Dr. B. O. L. Duke of the Helminth- iasis Research Unit at Kumba, I received a preserved sample of bulinid snails from the second of the two infected lakes reported by Zahra. In order to obtain living material of both snails and schistosomes from this unusual focus and to investigate the possible wider distribution of the disease in West Cameroon a short visit was made to the territory in October and November 1963. This paper reports the mala- cological results of the expedition. I am greatly indebted to Dr. S. P. Tchoungui, Federal Minister of Health of the Republique Federale du Cameroun and to Colonel R. Aretas, Conseiller Technique at the Federal Ministry of Health for their encouragement and support. It is a pleasure also to acknowledge the facilities made available by Dr. G. G. Dibue, Director of Medical Services, West Cameroon and Dr. B. O. L. Duke, Director of the Helminthiasis Research Unit at Kumba. Special thanks are due to Mr. and Mrs. Peter Moore of the Helminthiasis Research Unit for their kindness and hospitality ; it is largely due to their help that a series of minor disasters were overcome and the objectives of the expedition were achieved. Both in the field and subsequently in the laboratory I was assisted by Mr. M. S. Bennett. This work was made possible by grant No. AI-O365O-03 from the U.S. Public Health Service. TOPOGRAPHY OF WEST CAMEROON The territory now designated West Cameroon was formerly the south-western part of the British Trusteeship territory lying between Nigeria and the French-adminis- tered Cameroun. In a number of scientific papers the same area has been referred to by different authors as West Cameroon, the Southern Cameroons and the North- western Cameroons. The region extends about 450 miles north-eastward from roughly 4N., 9E. on the coast of the Gulf of Guinea to about gN., I3E. The country is mountainous with dense rain-forest in the south-western part and open grassland in the hills to the north-east ; its geography and geology have been very fully des- cribed by Geze (1943). The mountains are partly tectonic in origin but the majority are volcanic, forming a chain of which the Gulf of Guinea islands (Fernando-Poo, Principe, Sao Thome and Annobon) are a south-western extension. Overlying the 3 7 6 C. A. WRIGHT pre-Cambrian bedrock of gneisses and gneissic granites are a sedimentary series of Cretaceous age which form a band running roughly parallel to the coast and there is another band in the Mamfe region. Overlying the Cretaceous deposits in the south- western part of the country is a thick sheet of basalt of volcanic origin. Geze quotes evidence suggesting that volcanic activity in Cameroon began in the Creta- ceous period but Reyment (1954) does not support this view and considers that the activity has been entirely post-Cretaceous. There are some sedimentary grits and sandstones of Tertiary age in the coastal area and Quaternary deposits are confined to the mangrove swamps which surround much of the coast. Volcanic activity resulting in extensive layers of tuffs in some areas has continued into recent times and the last eruption of Mount Cameroon occurred in 1954. The crater lakes which were one of the major objectives of the present expedition are considered by Geze to be the result of violent gaseous explosions of relatively recent origin. This opinion is based on the absence of igneous material associated with the craters, other than that of the older basalt sheet through which the eruptions occurred or demonstrably younger elements. These younger igneous elements in- clude the island in Lake Barombi Kotto (the remains of a small volcano which appeared after the lake was formed) and the basalt stream on the north-east side of Lake Barombi Mbo. Evidence of the relatively recent origin of the craters is provided by the generally steep slope of their sides but in some this incline is less abrupt, suggesting a considerable variation in their age. Typically the lakes are roughly circular in outline and the internal slopes of the craters are thickly forested with dense vegetation extending down to and overhanging the water's edge. Hydro- graphic information is available for only the three main lakes in Kumba division, Barombi Mbo, Barombi Kotto and Soden ; the first and last of these are steep-sided and deep but the volcanic island in Barombi Kotto has filled a good deal of the lake, making it relatively shallow and creating a more gently sloping shore, particularly around the island. On the north-west side of Barombi Mbo the entering stream has created a small, swampy delta. The extension of this silt deposition has formed a shelf extending out into the lake about one hundred yards. The water depth over this shelf probably does not exceed ten feet and there is considerable growth of aquatic vegetation. Access to Lake Soden is difficult and the only point which was visited had a very steep slope and exceedingly dense, overhanging marginal vegeta- tion. The fringing forest appears to be uninterrupted and it is not known if there are any shallow areas. Lake Ejaghem near Mamfe has less steep banks and forest clearing has resulted in some gaps in the marginal vegetation ; this lake and the three preceding ones have endemic fish populations which are utilized to varying extents by the local people. In Bamenda Division there are a number of lakes whose origin is uncertain but which have many of the characteristics of the craters in Kumba Division. Only three of these were visited, Lakes Bambuluwe, Bafeng and Wum ; the first two are small, steep-sided and with dense marginal vegetation while Lake Wum is open and has some areas with a gently sloping margin with aquatic vegetation. Fish have been introduced into Lake Wum and appear to be thriving but there is no natural fish population in either Bambuluwe or Bafeng and FRESHWATER GASTROPOD MOLLUSCS OF W. CAMEROON 77 information from inhabitants of the village at Oku indicated that there are no fish in Lake Oku either. HISTORICAL Cameroon has attracted a good deal of malacological attention as a result of its rich and interesting fauna of terrestrial species but, for a territory so well-endowed with rivers, streams and lakes, the records of freshwater species are poor, von Martens (1877) described a large West African collection made by Professor Bucholz and this included six species of fresh- and brackish-water prosobranchs from the coastal region, mostly in the neighbourhood of Victoria. In 1891 the same author reported on a collection made by Preuss in the area of the " Barombi-Station " and mentioned the two large species of freshwater prosobranchs still found in Lake Barombi Mbo and its tributary stream. d'Ailly's (1896) account based upon ex- cellently documented collections made by the Swedish surveyor Dusen included nine species of prosobranchs and de Rochebrune (1898) gave a list of species received from Sjostedt, one of Dusen's colleagues, and mentioned two of the prosobranchs present in d'Ailly's material. O. Boettger (1905) published a list of Cameroons species based on material received from several collectors but here again the only freshwater gastropods noted were prosobranchs. Two small collections made by Dyke were reported upon by Spence (1925 and 1928) and included three species of prosobranchs, two of which were described as new. The first basommatophoran recorded was a planorbid, found on the northern slopes of Mount Cameroon and described by C. R. Boettger as Australorbis camerunensis (1941). A Swiss expedition in north-eastern Cameroon collected two freshwater prosobranchs not included in any of the earlier accounts (Forcart, 1951) and the Danish expedition to the French Cameroons in 1949-50 brought back only terrestrial species (N0stvik, 1956). Man- dahl-Barth (1957^ described Bulinus camerunensis and recorded B. truncatus rohlfsi from specimens collected by Mr. P. J. Moore in Lake Barombi Kotto. These two species together with Boettger's planorbid are the only basommatophoran pul- monates recorded from the area. The present account adds seven more planorbids, one lymnaeid and one ancylid to the list as well as an additional prosobranch not previously recorded. Family PLANORBIDAE Biomphalaria camerunensis (Boettger) Australorbis camerunensis C. R. Boettger, 1941 : 121. Biomphalaria camerunensis camerunensis Mandahl-Barth, igsya : 1142. MATERIAL : Bambalang, N'dop Plain, Bamenda Division, about 70 specimens collected from shallow water with thick mud bottom and scarcely any vegetation, heavily shaded by thickets of low-growing palm trees. 6th November, 1963. Shell (PI. I, figs. 16-23) : yellowish-brown, discoidal, flattened above with the centre slightly depressed ; Whorls weakly angled beneath, umbilicus wide ; well- developed spiral lines give marked granular microsculpture even on later whorls ; ?8 C. A. WRIGHT larger shells irregularly rugose and patchily eroded. Nearly all shells up to 12-5 mm. diameter have apertural lamellae (PI. I, fig. 23). The largest specimen collected had a maximum diameter of 17-3 mm., umbilicus 7-2 mm. and height 5-4 mm. The mean ratio maximum diameter/umbilicus diameter for 30 specimens of all sizes is 2-67 and for the ratio maximum diameter /height the value is 2-75. Anatomy : male copulatory organ large, penis-sheath longer than preputium (Text-fig. 4), mean ratio penis sheath /preputium for a series of adult specimens is 1-42 (range 1-07-1-86), penis slightly shorter than the sheath ; prostate long with 30 or more primary diverticula, all with secondary and many with tertiary branches. Albumen gland relatively small, oviduct and muciparous gland (the translucent white proximal part of the uterus) short but the distal, yellow oothecal gland is long ; vagina long, without marked dilatation, receptaculum seminis club-shaped (Text-fig. 4), vesicle about equal in length to duct. The pericardia of all specimens dissected were packed with metacercariae of an unidentified trematode and the gonad and part of the digestive gland of others were found to be completely destroyed by rediae of probably the same species of parasite. Radula (Text-fig. 10) : teeth small, tendency towards arrow-head shape of some lateral mesocones but majority triangular, ecto- and endocones roughly equal ; six to eight tricuspid laterals in each half-row, about the same number of intermedi- ates, division of the endocone starts at about tooth 15, most marginal ectocones undivided but some have outer edge serrated. Biomphalaria camerunensis has so far been known only from the type series of shells collected at Mongonge on the northern slope of Cameroon Mountain. Mandahl- Barth (ig57a) described a sub-species B. c. manzadica from a few localities in the Lower Congo and, on shell characters my material is closer to this sub-species than the nominate form in that the umbilicus diameter is greater than the height. How- ever, the illustration of the type specimen of B. camerunensis shows that it differs from any of my specimens in that the aperture and terminal part of the body whorl are deflected downward, a character which gives a misleading increase in the height of the shell in very large specimens. All other members of the Biomphalaria sudanica species group have the umbilicus diameter greater than the shell height as in my material and it seems unwise to set B. camerunensis apart because this proportion is reversed in a single, large, possibly aberrant specimen. Anatomically the present material differs from all other members of the B. sudanica group (with the exception of B. sudanica rugosa Mandahl-Barth, 1960) in that the penis sheath is consistently longer than the preputium. The radula teeth conform to the general characters of the sudanica group. Three shells (B.M. (N.H.) coll. no. 1936-6-2-24-26) from the Batouri District of East Cameroon which compare well with my material were labelled " Planorbis salinarum " by Connolly in 1936 and there is no doubt that they bear a close resemblance to that species. However, the umbilical diameters and heights of the Cameroons shells are proportionately slightly higher than in the type series of B. salinarum. It is probable that snails reported from Yaounde in East Cameroon as B. sudanica by Gaud (1955) are also B. camerunensis. The presence of apertural lamellae does not appear to have been noted before in FRESHWATER GASTROPOD MOLLUSCS OF W. CAMEROON 79 members of the B. sudanica species group. Smith (1881) originally described the form tanganyicensis (now considered by Mandahl-Barth (i957a) to be a sub-species of B. sudanica) as a member of the genus Segmentina but there is no trace of lamellae in any of Smith's specimens. The lamellae found in B. camerunensis do not differ from those reported in B. pfeifferi and in various Puerto Rican species by Richards (1963) who suggested that they may serve as supporting structures for the delicate aperture during aestivation. The unusually shallow and shaded habitat from which this sample was collected is most probably subject to desiccation in the dry season. The specimens brought alive into the laboratory showed a persistent tendency to crawl out of the water. Compared with Sudanese and Kenyan strains of B. sudanica kept in this laboratory B. camerunensis is an exceptionally slow-growing species and it has so far proved refractory to infection with strains of Schistosoma mansoni from Egypt and Liberia. Biomphalaria pfeifferi (Krauss) Planorbis pfeifferi Krauss, 1848 : 83, pi. 5, fig. 7. MATERIAL : Lake Wum, Bamenda Division, on sparse aquatic vegetation in 2-3 feet of water. Six specimens, yth November, 1963. Shell (PL I, figs. 14-16) : light yellowish-brown, whorls without marked angulation, well-developed spiral lines intersect fine growth lines to give a strong granular microsculpture. The dimensions of the largest specimen in the sample were, maxi- mum diameter 6-4 mm., umbilical diameter 1-8 mm., height 2-8 mm. The mean ratio for the whole sample of maximum diameter/umbilical diameter was 3-66 and of maximum diameter /height was 2-14. Anatomy : the state of preservation of the sample was poor. The two largest specimens (6-4 and 5-5 mm. diameter) were dissected and found to be fully mature. The penis sheath is shorter than the preputium, the prostate has about 15 primary diverticula, most of them with secondary branches. In contrast to B. camerunensis the oviduct is strongly convoluted and the muciparous gland encloses more than half the uterus. The receptaculum is club-shaped and the vesicle is about equal in length to its duct. The pericardia of both specimens contained many trematode metacercariae. Radula : the teeth are larger than in B. camerunensis, there are seven tricuspid laterals in each half-row, their mesocones are broad and spatulate and the marginal ectocones are divided. Biomphalaria pfeifferi has an almost universal distribution in the Ethiopian region and its presence in Cameroon is to be expected. Anisus coretus (de Blainville) Le Coret, Coretus Adanson, 1757 : 7-10, pi. I, fig. 3. Planorbis coretus Adanson, de Blainville, 1826 : 230. non Planorbis coretus Adanson, Dautzenberg, 1890. Coretus adansonii Gray, 1850 : 119, pi. 309, fig. 4. Planorbis (Gyraulus) gibbonsi Nelson. Binder, 1957 : I2O > n - I 7- 3 8o C. A. WRIGHT \ A ( \ FIGS. 1-3. Anisus coretus. i, Whole genital system. 2, Tip of male copulatory organ. 3, Prostate. FIG. 4. Biomphalaria camerunensis. Male copulatory organs and receptacula. MATERIAL : Lake Barombi Kotto, Kumba Division. Ten specimens and laboratory-bred material. I5th November, 1963. Pamol Estate, Lobe, Kumba Division, in ditches, about 70 specimens. i5th November, 1963. Shell (PL II, figs. 4-6) : small, discoidal, flattened above, slightly concave beneath, whorls 3-3^, rounded without angulation, sutures relatively deep, aperture rounded. The shells are glossy with fine growth lines, many of those from Lobe have a very fine spiral sculpture on the underside giving a wavy appearance to the growth lines. The largest specimen seen was laboratory bred and measured 2-7 mm. maximum diameter. The mean dimensions of five adult specimens from Lobe are 2-4 mm. maximum diameter, 1-24 mm. umbilicus diameter and 0-68 mm. height. Anatomy (Text-fig, i) : the male copulatory organ is small with the penis sheath and preputium about equal in length and the preputium slightly greater in diameter FRESHWATER GASTROPOD MOLLUSCS OF W. CAMEROON 81 than the proximal dilatation of the sheath. The penis is usually a little longer than the sheath, there is a small claw-like stylet at its tip and the opening of the vas deferens is sub-terminal (Text-fig. 2) . The prostate (Text-fig. 3) consists of a glandular part of the male duct with a single diverticulum. Examination by phase-contrast microscopy shows that the wall of the sperm-duct consists of irregularly rounded cells, the glandular prostatic part and the diverticulum are finely and densely granular and the vas deferens appears to be composed of connective tissue with sparse oval cells with well-defined nuclei. The seminal vesicle on the hermaphrodite duct is coiled and has small projections on its surface. The vagina is short and proximally dilated, the uterus is long, narrow and straight and the albumen gland is relatively large. The receptaculum seminis is ovoid and slightly shorter than its slender duct. Radula (Text-fig. 16) : teeth small and few in number, there are about eight tricuspid laterals, a single intermediate in which the ectocone is divided and a single five-cusped marginal in each half-row. Anatomically this material does not differ significantly from Binder's (1958) description of Gyraulus gibbonsi from the Ivory Coast. The terminal stylet of the penis is similar in both forms and differs from the smooth cap-like structure of Anisus misellus (Morelet) from Angola (Wright, 1963). The single prostatic diverti- culum of the Ivory Coast material is very much longer than in specimens from the Cameroons but this is probably a variable character in view of the occurrence of occasional individuals with a single diverticulum in A . misellus, a species which usually has a more normal prostate with several diverticula. Binder shows the seminal vesicle to be convoluted but smooth in his specimens and in this respect the Cameroon form resembles A. misellus which also has small projections from the surface of the vesicle. Binder referred his specimens to Gyraulus gibbonsi on the basis of an illustration by Pilsbry & Bequaert (1927) . G. gibbonsi was described by Nelson (1878) from Zanzibar but his illustration does not agree with his description in that it shows a shell com- pletely flattened beneath and with a sharp basal angle while the description says that the aperture is rounded. The whereabouts of the type-specimen are unknown but Pilsbry & Bequaert examined material from Zanzibar and found that it agreed with specimens from the Congo with a rounded aperture. I have also examined topotype shell material and have found that it resembles the form from the Cameroons except that it lacks any trace of spiral sculpture and reaches a much greater size at the 3^ whorl stage than does the West African form. There is, however, a very striking resemblance between my specimens and the description and photographs of Adanson's (1757) Le Coret, Coretus, published by Fischer-Piette (1942). Fischer-Piette pointed out that Adanson's species had never been rediscovered and that Dautzenberg's (1890) re-description of it from the type-locality at Podor, Senegal, certainly did not agree with the original specimens recovered later. Gray's (1850) species Coretus adansonii was based only on Adanson's original description as was the earlier Planorbis coretus of de Blainville (1828). Adanson emphasized the small size of this snail and reported copulation between individuals just over 3 mm. in diameter. The relationships of A. coretus to other species are not yet readily determined. 82 C. A. WRIGHT A. misellus is obviously close and it is possible that some of the many species of small planorbids indifferently described by Germain (1907, 1909, 1911, 1917) from West Africa and the Lake Chad region are no more than local variations, de Azevedo et al, (1961) have recently described the anatomy of A. natalensis from Mozambique as having a single prostatic diverticulum, no penial stylet and a terminal opening to the vas deferens. These characters are in contrast to the normal prostate and sclerotized terminal part of the penis in A . natalensis from Ethiopia and Trans- vaal (Wright & Brown, 1962). Gyraulus costulatus (Krauss) Planorbis costulatus Krauss, 1848 : 83, pi. 5, fig. 8. MATERIAL : Lake Wum, Bamenda Division. On dead leaves near the lake margin. 3 specimens. 7th November, 1963. Shell (PL II, figs. 7-9) : small, discoidal, pale yellowish brown, strongly and regularly ribbed with a well-marked equatorial carination. The largest specimen was broken but probably had 3-3 1 whorls ; an individual with 3 whorls measured 3-2 mm. diameter, 0-8 mm. umbilical diameter and 1-08 mm. high. Anatomy : one specimen was aphallic but the anatomy of the other two corres- ponded well with that described for this species from Angola (Wright, 1963). The penis papilla noted by de Azevedo et al. (1961) was observed but the Cameroon specimens had the penis relatively shorter with the stylet not reaching to the end of the sheath. Radula (Text-fig. 18) : teeth small, 9-10 tricuspid laterals and about 6 five cusped marginals in each half-row. The cusps of the laterals are a little broader than those in Angolan specimens. This species was probably more abundant than the number of specimens collected suggests but the samples obtained from Lake Wum were, for various reasons, in- adequate. The largest of the three specimens was grossly infected with rediae of an unidentified trematode. This was probably the same species of parasite as the metacercariae found in the pericardia of the other two individuals and the Biom- phalaria pfeifferi and SegmentorUs angustus from the same locality. Segmentorbis angustus (Jickeli) Segmentina angusta Jickeli, 187/1 : 22O > pi- 7> fi- 2 4- Segmentorbis angustus (Jickeli), Mandahl-Barth, 1954 : 9^- MATERIAL : Lake Wum, Bamenda Division. On dead leaves and emergent grasses. About 50 specimens, also laboratory-bred material. 7th November, 1963. Shell (PI. II, figs. 1-3) : lenticular, flattened beneath, relatively sharp basal angle, translucent, pale yellowish-brown, wild specimens often thickly crusted with dark brown deposits, umbilicus narrow and very deep. There are usually two, rarely three sets of lamellae in the body whorl, the basal is well-developed, straight, about half the width of the whorl, the inner curves strongly over the parietal wall, the FRESHWATER GASTROPOD MOLLUSCS OF W. CAMEROON 83 outer is usually weak, often sub-divided and sometimes absent while the dorsal is rarely present (Text-figs. 5 and 6). The shell is smooth and shiny with a faint trace of spiral sculpture on the underside, visible only under high magnification. Anatomy (Text-fig. 7) : penis-sheath about two-thirds the length and a quarter the width of the preputium, with a single, large, thin-walled flagellum attached to the proximal end of the sheath. There is a separate prostatic duct from which arise about fifteen unbranched diverticula. The seminal vesicle is widely dilated and convoluted with small surface protruberances. There is a slight dorsal dilatation of the vagina, the uterus is long and straight and the receptaculum seminis is clavate, about equal in length to its duct. Several specimens had heavy infections of trema- tode metacercariae in their pericardia. Radula (Text-fig. 17) : the teeth are similar to those of Anisus coretus but they are smaller and more numerous. The first 7-9 laterals are tricuspid, the endocone is subdivided in the next six or seven and from about tooth fifteen outward the ectocone is also divided giving six to eight typical 5-6 cusped marginals. 5. angustus was originally described from Ethiopia and is known from many places in East and South Africa. A single small specimen was recorded from Kikondja in the Congo by Pilsbry & Bequaert (1927). Their illustration of the shell shows a greater number of basal lamellae (4) than are found in specimens of com- parable size in the present material and the outer lamellae are even more degenerate in the Congo specimen. Mandahl-Barth (1954) notes that there is usually only a single set of septa but occasionally up to five sets are seen in this species in Uganda. He also mentions specimens of Segmentorbis with even larger numbers of septa in which the other shell characters do not differ from normal S. angustus. The form of the receptaculum seminis in the Cameroon specimens differs from that described by Mandahl-Barth (1954) for Uganda material and by de Azevedo et al. (1961) for specimens from Mozambique. Both of these authors show the vesicle to be small and spherical with a diameter only about a quarter the length of the duct. Even young specimens from Lake Wum have an elongate-ovoid receptaculum about equal in length to its duct and this form is similar to that in Ethiopian specimens (Brown, 1965). Segmentorbis kanisaensis (Preston) Segmentina kanisaensis Preston, 1914 : 265, pi. 18, figs. 17-19. Segmentorbis (Carinorbis) kanisaensis ; Mandahl-Barth, 1954 : 98, fig. 45. MATERIAL : Lake Barombi Mbo, Kumba Division. One specimen. Shell : lenticular, flattened beneath, sharp basal angle ; translucent, yellowish- brown, fine growth lines and slight spiral microsculpture on the underside ; a single set of well-developed lamellae (Text-figs. 8 and 9), the basal about two-thirds the width of the whorl, the inner strong, curved, the outer long, oblique and the dorsal a pronounced dot-like structure. Dimensions of the single specimen were diameter 2 mm., umbilicus 0-5 mm., height 0-6 mm. Anatomy : the specimen was immature but dissection showed the characteristic lack of a flagellum on the male copulatory organ. C. A. WRIGHT FIGS. 5-7. Segmentorbis angustus. 5, Dorsal view of shell. 6, Ventral view of shell. 7, Whole genital system. FIGS. 8, 9. Segmentorbis kanisaensis. 8, Dorsal view of shell. 9, Ventral view of shell. Segmentorbis snails were relatively common on dead leaves in Lake Barombi Mbo near the mouth of the entering stream on the north-west shore. Unfortunately all FRESHWATER GASTROPOD MOLLUSCS OF W. CAMEROON 85 of the wild specimens were lost and the single individual reported here was later found in a tank in the laboratory containing leaf-litter from the lake. This species is common in low-lying areas of West Africa from the Gambia to Angola and is also found in East and South Africa. Bulinus rohlfsi (Clessin) Physa rohlfsi Clessin, 1886 : 349, pi. 49, fig. 7. Bulinus truncatus rohlfsi ; Mandahl-Barth, I957(b) : 29, pi. 20, fig. 18. MATERIAL : Lake Barombi Mbo, Kumba Division. About 100 specimens received from Dr. B. O. L. Duke, 8th January, 1957 ; 20 specimens and laboratory-bred material, I4th November, 1963. Lake Barombi Kotto, Kumba Division. About 40 specimens and laboratory-bred material, 3ist October and I5th November, 1963. Shell (PL I, figs. 7-10) : light yellowish-brown, spire short, no marked ribbing or other micro-sculpture ; columella more or less straight, columellar margin reflexed, closing the umbilicus. Adult specimens from Lake Barombi Kotto usually have the aperture margin reflexed giving to the shell a bell-mouth appearance ; this distortion occasionally occurs in younger individuals and, if followed by a period of normal growth, results in the formation of a marked ridge on the body whorl. Similar reflexion of the aperture margin has not been seen in wild specimens from Barombi Mbo but it occasionally occurs in laboratory colonies from the Mbo stock. The maximum size in the Kotto sample is length 11-2 mm., width 8-8 mm. and aperture length 8-9 mm. From Mbo the largest specimen is 7-7 mm. long, 5-1 mm. wide and has an aperture length of 5-1 mm. This disparity in maximum size between the two populations has been maintained in laboratory colonies bred from the stocks. The colony from Kotto is also remarkable for the number of distorted shell-forms which it produces. The mean ratio shell-length/aperture length for adult specimens of both populations is 1-2. Anatomy : all specimens dissected from both populations were aphallic ; in some there is a trace of sperm-duct and a vestigial prostate but in the majority there is no development of the male system. Specimens from Mbo have the uterine glands fully developed and apparently functional at shell-length 4-0 mm. while those from Kotto are less well-developed at 6 mm. The mantle in specimens from Kotto is pale grey with black spots and patches while in the Mbo samples the markings vary, as they do in the same species from Angola, from almost unmarked to strongly patterned. Radula (Text-fig. 13) : teeth smaller than in B. truncatus, 3-5 tricuspid laterals, all with slightly arrowhead-shaped mesocones flanked by fine interstitial cusps, 13-18 marginals in which the mesocone is not so clearly dominant as it is in the next species. Bulinus rohlfsi is the oldest name of a considerable number of species and varieties described from Lake Chad. Mandahl-Barth (i957b) treats rohlfsi as a sub-species of B. truncatus and there is no doubt about the close relationship of the two species. However, the sub-species concept has little real meaning in the Basommatophora 86 C. A. WRIGHT and I prefer to regard rohlfsi as a distinct species. The large number of names given to this species in Lake Chad is some indication of the great variability of its shell form. The distinctions observed here between the populations from the two Kumba lakes are further evidence of this polymorphism. In both lakes B. rohlfsi is acting as intermediate host for Schistosoma haematobium. Bulinus tropicus (Krauss) Physa tropica Krauss, 1848 : Bulinus (Bulinus) hemprichii depressus Haas, 1936 : 28, pi. I, fig. 15. Bulinus tropicus tropicus (in part) ; Mandahl-Barth, I95yb : 19, pi. 10, fig. 4. MATERIAL : Babungo, N'dop Plain, Bamenda Division. On aquatic plants in a slow-flowing stream about i mile east of the village. 20 specimens, also laboratory bred material. 6th November, 1963. Shell (PI. I, figs, ii and 12) : thin, translucent, yellowish-brown, spire flat, aper- ture wide, columella slightly twisted ; early whorls ribbed, body-whorl smooth with fine growth-lines. Dimensions of the largest specimen were length 8-1 mm., width 6-9 mm. and aperture length 7-4 mm. In the laboratory shell-lengths in excess of 12 mm. have been achieved by mature individuals. The mean ratio shell-length/aperture-length for a series of juveniles is 1-03, a clear indication of the flatness of the spire. However, laboratory-bred specimens develop a prominent spire even in the first generation. Anatomy : penis sheath about twice the length of the preputium, its proximal dilatation is a little wider than the distal part ; epiphallus short and without dilata- tion. The largest wild specimen (8-1 mm.) was protandrously mature with apparent- ly functional copulatory organ and prostate but poorly differentiated uterine glands. The remainder of the sample were all juvenile but in all there were at least traces of rudimentary male genitalia and no potentially aphallic individuals were seen. The mantle is light grey with sparse, well-defined black spots and patches. Radula (Text-fig. 12) : teeth larger than in B. rohlfsi, 6-7 tricuspid laterals with sub-division of the endocone occurring in the seventh or eighth tooth ; lateral mesocones slightly arrow-head shaped, usually flanked by lanceolate interstitial cusps, inner edge of the endocones frequently corrugated. In the 20-22 marginals the mesocones remain undivided and are very prominent. Wild specimens of this form closely resemble Bulinus depressus, described as a sub-species of " B. hemprichii " from a canal near Lake Bangweulu in Zambia. In an earlier paper (Wright, 1957) I suggested that B. depressus was no more than a juvenile of B. globosus but, in a private communication, Dr. Mandahl-Barth drew my attention to certain features of depressus which indicate that it is definitely not related to the B. africanus complex and he (i957b) included it in the synonomy of B. tropicus tropicus. Examination of the immature wild material from Babungo left me in some doubt as to whether it should be assigned to the tropicus or truncatus species groups. However, electrophoresis of the egg-proteins of the one living specimen brought back to London demonstrated without doubt that this form belongs to the FRESHWATER GASTROPOD MOLLUSCS OF W. CAMEROON 87 B. tropicus complex (Wright & Ross, in press). This record is probably near to the extreme north-west limits of the range of this species group for I am unable to agree with Mandahl-Barth on the inclusion of B. guernei from West Africa in the tropicus group ; I have examined B. guernei from three of the four localities mentioned by Mandahl-Barth (ig^jb) and I have found a high proportion of aphallic individuals and arrowhead shaped lateral mesocones, Smithers (1956) has also reported wild- caught specimens infected with Schistosoma haematobium. These three characters all indicate an affinity with the truncatus group and no other species with tropicus characteristics has been reported from West Africa. It is possible that the B. strigosus reported by Gaud (1955) from Bangui is B. tropicus. Bulinus forskali (Ehrenberg) Isidora Forskalii Ehrenberg, 1831 : 20. MATERIAL : Ditches and streams on Pamol Estate, Lobe, Kumba Division. 36 specimens. 25th October, 1963. Shell (PI. I, figs. 5 and 6) : strongly turretted with well-marked shoulders and ribs on all whorls, distinct small spines on the shoulder of the third whorl where the ribs intersect the angle. Mean dimensions in millimeters of 14 adults (maxima in brackets) were length 5-4 (7-2), width 2-7 (3-2) and aperture length 2-7 (3-0). Anatomy : male copulatory organ small and slender, penis sheath /preputium ratio about 3/2. Radula (Text-fig. 15) : similar to that described for this species from Angola but few malformations present in this population ; 5 or 6 tricuspid laterals and up to eighteen marginals. B. forskali is probably the most widely distributed and variable species of fresh- water snail in Africa and its presence in Cameroon was to be expected. It was recorded by Zahra (1953) from the stream flowing into the lake past Barombi Mbo village but it was not found there during the present visit despite repeated examination of the area. Bulinus camerunensis Mandahl-Barth Bulinus camerunensis Mandahl-Barth, iQ57b : 31, pi. 22, fig. 21. MATERIAL : Lake Barombi Kotto, Kumba Division. About 150 specimens, also laboratory-bred material. 3ist October and i5th November, 1963. Shell (PL I, figs. 1-4) : high-spired, translucent, yellowish-white, most specimens show some quite marked ribbing and in a few there is a weak shoulder on the second and third whorls ; about half the shells examined show irregular spiral lines giving patches of reticulate sculpture. The mean dimensions in millimeters (maxima in brackets) for a sample of 50 wild adults are length 5-4 (6-9), breadth 2-9 (3-5) and aper- ture length 3-2 (4-0). Larger specimens, up to 11-5 mm. high have been bred in the laboratory. The mean value and range of the ratio shell-length/aperture-length is 1-65 (1-47-1-94). In only three out of the 50 wild snails measured was the breadth 88 C. A. WRIGHT of the shell greater than the aperture length. This is in contrast to the dimensions given in the original description. Anatomy : male copulatory organ small and slender, sheath usually longer than the preputium and its proximal dilatation about equal in diameter to the distal part. The relative lengths of the two parts of the copulatory organ vary widely, the mean ratio of sheath-length to preputium is 1-43 but it ranges from about i-o-2-o. The prostate is relatively small in all of the fully adult specimens examined. Mantle markings in wild specimens are irregular and somewhat diffuse black patches on a grey ground but laboratory-bred individuals have the mantle almost uniformly black. Radula (Text-fig. 14) : teeth similar to those of B. forskali but slightly smaller and nearly all of the lateral mesocones are flanked by fine interstitial cusps. There are five or six laterals as in B. forskali but the number of marginals is greater (up to twenty-four) in B. camerunensis. In his original description of this species Mandahl-Barth mentioned the possibility that B. camerunensis was no more than a local form of B. forskali but that it differed from that species by its smaller radula teeth. The contrast in shell-characters between B. camerunensis and the form of B. forskali cc-llected a few miles away at Lobe is well-marked and the present species shows greater similarity to B. beccarii (Paladilhe) from Aden (Wright, I963a) than to any other member of the forskali group. In the hope of finding further foci of this species six other crater lakes were visited but without success. Repeated attempts to infect B. camerunensis with the West Cameroon strain of Schistosoma haematobium have failed. Family LYMNAEIDAE Lymnaea natalensis Krauss Limnaeus natalensis Krauss, 1848 : 85, pi. 5, fig. 15. MATERIAL : Babungo, N'dop Plain, Bamenda Division. On aquatic vegetation in a slow-flowing stream about I mile east of the village. 12 specimens. 6th Novem- ber, 1963. Lake Wum, Bamenda Division. 6 specimens, yth November, 1963. Shell (PI. I, fig. 13) : thin, shiny, elongate-ovoid, spire short and sharply pointed ; the only ornament is fine growth lines. Mean dimensions in millimeters (maxima in brackets) of four adult specimens from Babungo are length 8-4 (9'i), breadth 4-6 (5-6) and aperture length 6-4 (7-0). Anatomy : penis-sheath approximately twice the length of the preputium, receptaculum seminis duct long with the receptaculum lying on the left side of the body. Two of the four adult specimens from Babungo had the accessory genital glands completely destroyed by an unidentified larval trematode but the other two had the characteristic distal dilatation of the prostate. Radula (Text-fig. 19) : eight to ten tricuspid laterals in which the endocones and mesocones are fused together for over half their length ; fifteen to twenty marginals, transition from the laterals occurring by sub-division of the ectocone which almost FRESHWATER GASTROPOD MOLLUSCS OF W. CAMEROON FIGS. 10-21. Radula Teeth. 10. Biomphalaria camerunensis. n. Biomphalaria pfeifferi. 12. Bulinus tropicus. 13. Bulinus vohlfsi. 14. Bulinus camerunensis. 15. Bulinus forskali. 16. Anisus coretus. 17. Segmentorbis angustus. 18. Gyraulus costulatus. 19. Lymnaea natalensis. 20. Lanistes libycus. 21. Potadoma freethi . go C. A. WRIGHT disappears in the outer teeth which have only four or five cusps derived from the endo- and mesocones. L. natalensis is widely distributed throughout Africa. The only feature of interest in the present material is the relatively small size at which maturity is reached. Family ANCYLIDAE FERRISSIA sp. MATERIAL : Ditches on Pamol Estate, Lobe, Kumba Division. 5 specimens. i5th November, 1963. Shell (PI. II, figs. 10 and n) : small, rectangularly ovoid, light yellow, apex striate, displaced to the right, smoothly rounded. The margin of all five specimens was of soft periostracum causing considerable variation in shell shape. The mean dimensions in millimeters (maxima in brackets) of four entire specimens are length 178 (1-92), width 1-2 (1-2), height 0-57 (0-64). No preserved specimens were available for anatomical study. Small ancylids were also found in Lake Wum and at Babungo, both in Bamenda Division but the material from both localities was lost. The small amount of material and lack of anatomical data do not permit an assessment of the affinities of this species but the proportions of the shell are similar to those of F. eburnensis Binder, 1957, from the Ivory Coast. Population variations in the outline of the shell from uniformly oval to rectangularly ovoid were noted by Binder. Family AMPULLARIIDAE Lanistes libycus (Morelet) Ampullaria libyca Morelet, 1848 : 28, pi. 3, fig. 9. Ampullaria bernardiana Morelet, 1860 : 190. MATERIAL : Lake Barombi Mbo, Kumba Division. On stones and sandy bottom near canoe beach on north-east side of the lake. 40 specimens. 24th October, 1963. Ditches on Pamol Estate, Lobe, Kumba Division. 12 specimens. I5th Novem- ber, 1963. Shell (PI. Ill, figs. 9 and 10) : large, globose, low-spired, apex often eroded ; whorls with well-marked shoulder angle, distinctly flattened below the suture, moderate carination around the umbilicus. Specimens from Lake Barombi Mbo have the shell smooth with strong spiral bands of colour which appear brownish-red externally and purple internally, those from Lobe have the colour bands less well- marked but have strong spiral ridges of periostracum which persist on adult shells ; these ridges are present on some juveniles from the lake population. The largest specimen collected (from the lake) measured 27 mm. in height, 28 mm. in width and had an aperture length of 18 mm. Radula (Text-fig. 20) : central tooth massive, median cusp flanked usually by two, sometimes three, smaller cusps all of which may be fused into a single transversely elongate blade ; lateral four-cusped and the two marginals bi-cuspid. FRESHWATER GASTROPOD MOLLUSCS OF W. CAMEROON 91 The type-locality for L. libycus is in Gaboon and the species has a wide distribution in West Africa. It has been reported from many places in the Cameroons by previous workers. The treatment of L. bernardianus as a variety of L. libycus was retained by d'Ailly (1896) because the large number of samples which he examined from Cameroon showed no intermediates between the two forms. However, the characteristic spiral ridges of bernardianus are confined to the periostracum and are easily rubbed off. This distinguishing character is therefore dependent on the nature of the substratum in the habitat. Binder (1957) has put forward a strong claim for the retention of the name libycus despite the recently demonstrated prior claim of L. intortus (Lamarck). In view of the confusion likely to be created by the use of Lamarck's name I am in entire agreement with Binder's argument. Family BITHYNIIDAE GABBIA sp. MATERIAL : Lake Wum, Bamenda Division, i specimen, yth November, 1963. Shell (PL II, fig. 12) : pale yellowish-brown with three rounded whorls, a very faint spiral microsculpture intersects the fine growth lines but the general appearance of the shell is smooth and shiny. The umbilicus is narrowly open and the spiral nucleus of the operculum occupies about half its total width. The single specimen measures 2-32 mm. in height and width and has an aperture height 1-52 mm. and aperture width of 1-36 mm. The genus Gabbia is mainly East and Central African in distribution, the only West African records being Frauenfeld's B. africana and Binder's (1955) Bithynia (Gabbia) tournieri from the Ivory Coast. Several species have been recorded from the Lake Chad region but the affinities of the present specimen appear to be nearest to G. parva Mandahl-Barth, 1954. Family THIARIIDAE Potadoma freethi (Gray) M elania freethi Gray, 1831 : n. Melania nigritina Morelet, 1848 : 345; von Martens, 1877 : 270; von Martens, 1891 : 33; d'Ailly, 1896 : 121. Melania (Nigrilella) nigritina ; Boettger, 1905 : 181. Thiara (Melanoides) dykei Spence, 1925 : 249. Potadoma graptoconus Pilsbry & Bequaert, 1927 : 276, fig. 46a and pi. XX, figs. 10 and n. MATERIAL : Stream on Bai Rubber Estate, Kumba Division. 10 specimens. 25th October, 1963. Stream on Pamol Estate, Lobe, Kumba Division. 10 specimens. 25th October, 1963. Stream near Barombi Kotto Stranger Town, Kumba Division. 18 specimens. I5th November 1963. 92 C. A. WRIGHT Stream through Barombi Mbo village, Kumba Division. 20 specimens. i6th November 1964. Shell (PI. Ill, figs. 1-8) : large, solid, turreted, adults usually black and decollate, whorls flattened, suture shallow, aperture pear-shaped, pointed above, columella gently curved outward resulting in a slight expansion of the aperture at the junction of the basal and columellar margins. There is a well-marked spiral sculpture and on the later whorls fine growth lines are more accentuated and intersect the spiral lines to give a granular pattern. Young specimens have a marked basal angle on the body whorl (PI. HI, figs. 7 and 8) but this almost disappears in the adults. The usual black colour of adult shells results from a fine covering of silt which can be removed by ultrasonic cleaning to show the basic greenish-brown colour of the young shell with at least one dark purple-brown band. Specimens from habitats with sandy bottoms are often found in this clean condition. Radula (Text-fig. 21) : central tooth complex with five small cusps on its anterior edge. The single lateral on each side is assymetrically four-cusped and the two marginals are tricuspid. The genus Potadoma is typically West African and more or less confined to the forested areas. The type locality for P. freethi is the island of Fernando Poo off the Cameroon coast, Melania nigritina was described from Gaboon, P. dykei from near Victoria in Cameroon and P. graptoconus from the Congo. All authors are in agreement concerning the identity of P. freethi and Morelet's nigritina. P. dykei was described without reference to freethi ; a specimen of dykei identified by Spence (PI. Ill, figs. 2 and 5) is in the collection of the British Museum (Natural History) and apart from being slightly more obese and having more well-marked cords around the columella it is not possible to separate it from large specimens of freethi. P. grapto- conus was distinguished by Pilsbry & Bequaert because of its well-marked spiral sculpture without the granulation said to be characteristic of freethi. However, examination of numbers of specimens from West Cameroon indicates that this granulation is not always present and is somewhat irregular on the holotype of freethi (PI. Ill, fig. 4). P. freethi was always found in gently moving water and although it is present in the stream flowing into Lake Barombi Mbo live specimens were not found in the lake itself. This species is common as a sub-fossil in the area of Lake Barombi Kotto and near Bekondo on the route to Lake Soden. DISCUSSION The collections made during this expedition fall readily into two groups, those from the low-lying forested area of Kumba Division and those from open savannah country at 4-5,000 feet in Bamenda Division. No species has been found which is common to both areas and although the search in Mamfe Division which intervenes between Kumba and Bamenda was superficial no freshwater snails of any kind were found. This suggests that an effective barrier exists between the two areas. The broad differences between the two parts of the snail fauna are those typical of altitude FRESHWATER GASTROPOD MOLLUSCS OF W. CAMEROON 93 and correspond to the differences between the coastal plain and southern plateau regions of Angola (Wright, 1963!)). The low-lying areas are characterized by an abundance of prosobranchs, the presence of Anisus and members of the Bulinus forskali group and B. rohlfsi while the highlands have Biomphalaria, Gyraulus costulatus and Bulinus tropicus. The presence of Segmentorbis kanisaensis in Lake Barombi Mbo is in keeping with its occurrence in the coastal plain of Angola and at low altitudes in the Ivory Coast and Gambia while S. angustus in Lake Wum corres- ponds with its distribution in higher crater lakes in Uganda. If the affinities of the species of Gabbia in the same lake have been correctly assessed it also points to a similarity with the Uganda fauna. Geze (1943) mentions the occurrence of a number of East African elements in the insect fauna of the mountains in Cameroon and attributes this to a more extensive temperate connection with the east along the highland chain during the pluvial periods. Sram (1955) has even gone so far as to suggest that there is evidence of glaciation on Mount Cameroon and Mannengouba during this period. The type locality of Biomphalaria camerunensis on Mount Cameroon suggests that this is a relic population now isolated from the more general highland areas by the intervening tropical rain forest. In Kumba Division the difference between the snail faunas of lakes Barombi Mbo and Barombi Kotto is interesting. The only species which they have in common is Bulinus rohlfsi, the two populations of which have distinctive characteristics. Trewavas (1962) has recently shown that the two lakes have no species of fish in common and that all but one of eight species in Barombi Mbo are probably endemic while Barombi Kotto has an endemic genus, two endemic species and one endemic sub-species out of a total of seven species. Bulinus camerunensis appears to be endemic to Barombi Kotto but the other snail species in both lakes have wide distribu- tions in West Africa. The effluent from Barombi Kotto passes into the Meme River near the mouth of which lies Lobe. Anisus coretus is common to both the lake and the streams in the Lobe area while B. camerunensis in the lake, although closely related to B. forskali, is clearly distinct from the local form of that species at Lobe. Differentiation of the two populations of B. rohlfsi may be the result of the " founder principle " and does not necessarily imply any prolonged separation. SCHISTOSOMIASIS IN WEST CAMEROON Zahra's (1953) original report has remained virtually the only source of information on schistosomiasis in West Cameroon, de Azevedo (1958) included these data in a general report covering what was at that time the whole British Cameroons but did not add anything further. One of the objectives of the present expedition was to confirm the identity of the snails responsible for transmission of the parasite in the known foci and to investigate their wider distribution in the territory. Wild specimens of Bulinus rohlfsi from Barombi Kotto were found to be shedding cercariae of Schistosoma haematobium (confirmed by exposure of hamsters with subsequent recovery of adult worms). Laboratory bred snails of this species from both of the Barombi lakes have been successfully infected by miracidia hatched from urine samples from Barombi Mbo. Bulinus camerunensis from Barombi Kotto 94 C. A. WRIGHT was not found naturally infected nor has it been successfully exposed to infection in the laboratory. Laboratory bred specimens of Bulinus tropicus from the N'dop plain were not susceptible to miracidia from Barombi Mbo. This is to be expected because so far no member of the B. tropicus complex has been demonstrated to act as an intermediate host for 5. haematobium. B. rohlfsi has not been found in any of the other places investigated in West Cameroon so there is little chance of the disease spreading unless the snails become established in other areas. The habitat for B. rohlfsi in both lakes is limited to shallow water with reasonable sun-exposure and the snails are found on dead leaves rather than on aquatic vegeta- tion. It is remarkable that the focus in Lake Barombi Mbo has remained confined to the small area on the north-west side of the lake near to the beaches where fishing canoes from the village are drawn up since there is ample opportunity for distribution of the snails both on the canoes and on fish traps which are set daily in all suitable places around the lake margin. It is probable that shading of the water's edge by fringing vegetation is the most important factor in preventing the colonization of new areas and any major clearance of the marginal bush should be viewed with concern. It is probable that transmission in both lakes is partly seasonal ; to judge from conditions in October and November it seems likely that maximum snail population densities occur in January-February and, allowing for development of the larval schistosomes there is probably a peak of transmission in February-March. An interesting aspect of these lake foci is that schistosomiasis is almost confined to members of the Barombi tribe who fish both lakes. The Barombis appear to be the only local people who do not have a profound superstitious fear of the lakes and contact with the water by people of other tribes is relatively infrequent. Lakes such as Soden and Ejaghem are known to contain fish but this is not a sufficient attraction to draw people to them regularly. On Lake Soden there are not even any canoes and only poor tracks to the water's edge. The few cases of urinary schistosomiasis encountered outside the Kumba area are usually from Nigeria or the East but some are found in people who have lived in Barombi Kotto Stranger Town and visited the lake for washing. Dissemination of Bulinus rohlfsi from the Barombi lakes might occur if develop- ment of fisheries in other lakes is undertaken using Barombi fish for stocking. There is also a slight possibility that B. rohlfsi from Barombi Kotto might gain access via the Meme river to the Lobe area where bush clearance has provided potentially good snail habitats in the streams and ditches. It seems that the West Cameroon foci of urinary schistosomiasis are effectively closed with little opportunity for spread and not much chance of importation of new strains. The strain in the Barombi foci is probably derived from the B. rohlfsi-borne form in Nigeria (Cowper, 1963). Snail control by molluscicides or habitat modification would not be practical in the lakes since the danger of damaging the fish population would be too great. The confined nature of the foci will probably lend themselves to control by drug treatment when suitable compounds are available. There is little information available concerning Schistosoma mansoni in West Cameroon. It is perhaps significant that Gaud's (1955) distribution map of this FRESHWATER GASTROPOD MOLLUSCS OF W. CAMEROON 95 parasite shows the edge of a lightly endemic zone in the east coinciding exactly with the border between East and West Cameroon in the area of the N'dop plain. Enquiries in the area near Bambalang revealed that dysentery accompanied by passing of blood is common in the area but time did not permit a survey to be carried out. Biomphalaria camerunensis is a possible host for S. mansoni but limited infection experiments with laboratory-bred specimens have so far failed. Both an Egyptian and a Liberian strain of parasite have been tested. There is at present no evidence for the existence of 5. mansoni in the Wum area despite the presence of the potential intermediate host Biomphalaria pfeifferi in Lake Wum. In conclusion brief mention must be made of two other trematode diseases, para- gonimiasis (lung-fluke) in man and liver-fluke in cattle. Zahra (1952) reported a widespread incidence of infection with Pamgonimus in Victoria and Kumba Divisions. This distribution coincides with that of the snail intermediate host, Potadoma freethi, which is common in small streams throughout the area. The incidence of Paragonimus is said to be particularly high among the Bakossi people in the region of Mount Kupe up to a height of 5,000 feet. No published reports of the occurrence of Fasciola gigantica in cattle in West Cameroon have been seen. Inquiries through the sanitary inspectors responsible for meat inspection suggest that the few cases seen are in cattle brought from outside the territory. However, around Banso and Ndu (6-7,000 feet) the parasite is scarcely ever seen while in the N'dop plain where the intermediate host Lymnaea natalensis is found it is said to be of slightly more frequent occurrence. Most of the grazing area in Bamenda Division appears to be above the altitude limit for L. natalensis and examination of apparently ideal habitats near Oku failed to yield any snails. ALPHABETICAL LIST OF COLLECTING PLACES This list includes the major places visited but not a large number of minor streams and pools in which nothing was found. 1. Babungo, N'dop Plain, Bamenda Division. 6O2' N., ioi2'E. Altitude about 4,000 feet. Slow-moving, weed-choked stream about one mile east of the village. 6th November, 1963. Bulinus tropicus, Lymnaea natalensis and unidenti- fied ancylid. 2. Bafeng (Lake), near falls on Mencham River, about 12 miles south of Wum on Bafut road, Bamenda Division. 62o' N., io02' E. Altitude unknown, probably between 5,000 and 6,000 feet. Small, very steep-sided crater lake with dense bush within the crater but surrounded by open grassland. 7th November, 1963. No snails found. 3. Bambalang, N'dop Plain, Bamenda Division. 552' N., io25' E. Altitude about 4,000 feet. Stagnant, shallow, marshy pools, heavily shaded by palm scrub. Thick, black mud bottom and mixed emergent vegetation. 6th November, 1963. Biomphalaria camerunensis. 4. Bambuluwe (Lake), south of Bamenda town, Bamenda Division. 55o' N., ioio' E. Altitude probably over 6,000 feet. Moderate-sized crater lake, slightly 96 C. A. WRIGHT irregular in shape, dense surrounding bush except near outlet of effluent stream where grassland runs down to lake shore. Few leeches and insect larvae seen. 4th Novem- ber, 1963. No snails found. 5. Barombi Kotto (Lake), Kumba Division. 42g' N., 92o' E. Altitude about 350 feet. Large crater lake with less steep margins than most and a volcanic island in the centre on which stands the village. No snails found on the lake margin at canoe beach but numerous specimens collected from dead leaves and debris on gently sloping beaches of the island. 3ist October and I5th November, 1963. Bulinus rohlfsi, B. camerunensis and Anisus coretus. 6. Barombi Mbo (Lake), Kumba Division. 442'N., g28' E. Altitude about 1,000 feet. Large crater lake, about two miles in diameter, very deep except on north-west side where entering stream has deposited a shelf of silt and gravel. Snails confined to this part of the lake. Visited six times in October and November, 1963. Bulinus rohlfsi, Segmentorbis kanisaensis, Lanistes libycus and Potadoma freethi (in entering stream). 7. Ejaghem (Lake), Mamfe Division. 545' N., 856' E. Altitude about 600 feet. Moderately large lake, shallow sloping surround with thick bush which has been cleared in patches. Natural(?) fish population but no snails found. 9th November, 1963. 8. Lobe, Pamol Estate, Kumba Division. 437' N., 9oi' E. Altitude, near sea level. Oil palm plantation on which most of the bush has been cleared and some water-management of small streams undertaken. 25th October and I5th November, 1963. Bulinus forskali, Anisus coretus, Ferrissia sp., Lanistes libycus and Potadoma freethi. 9. Soden (Lake), Kumba Division. 445' N., 9i6' E. Altitude, 1,500 feet. Moderately large, isolated crater lake with very steep, densely forested internal slope, difficult of access, very little human contact, no canoes, natural fish population. 28th October, 1963. No snails found. 10. Tiko, Victoria Division. 45' N., 92o' E. Altitude, near sea level. Heavily polluted stream flowing through the edge of the town, visited at several points after hearing reports of urinary schistosomiasis in the area. Reports later proved to be without foundation. i2th November, 1963. No snails found. 11. Wum (Lake), Bamenda Division. 628' N., io02' E. Altitude about 5,000 feet. Moderate sized lake, gently sloping bank with patches of shallow mar- ginal water and some aquatic vegetation. Lake open, surrounded by savannah, margins unshaded, yth November, 1963. Biomphalaria pfeifferi, Gymulus costu- latus, Segmentorbis angustus, Lymnaea natalensis, Gabbia sp. and unidentified ancylid. REFERENCES ADANSON, M. 1757. Histoire des coquillages. Histoire naturelle du Senegal. Paris. 1-275, 19 plates. D'AILLY, A. 1896. Contributions a la connaissance des mollusques terrestres et d'eau douce de Kameroun. Bihang K. Svenska Vet. Akad. Hand. 22 (4:2): 1-139. DE AZEVEDO, J. F. 1958. Human bilharziasis in the British Cameroons. Bull. Wld. Health Org. 18 (5 and 6) : 1052-1057. FRESHWATER GASTROPOD MOLLUSCS OF W. CAMEROON 97 DE AZEVEDO, J. F., DE MEDEIROS, L. C. M., COSTA FARO, M. M., XAVIER, M. DE L., GANDARA, A. F. & MORAIS, T. 1961. Freshwater mollusks of the Portuguese overseas provinces III. Mollusks of Mozambique. Estudos, Ensaios e Documentos, 88 : 285-394. Junta de In- vestigacSs do Ultramar, Lisbon. BINDER, E. 1955. Mollusques nouveaux de C6te d'lvoire. Prosbranches d'eau douce. Rev. Suisse Zool. 62 (suppl.) : 73-82. 1957. Mollusques aquatiques de Cote J'lvoire. Bull. Inst. franc. Afr. noire 19A(i) : 97- 125- 1958. Anatomie de Gyraulus gibbonsi Nelson de Cote d'lvoire (Moll. Gasterop). Rev. Suisse Zool. 65(io) : 301304. DE BLAINVILLE. 1826. Planorbe in Dictionnaire des Sciences Naturelles. 41 : 230. BOETTGER, C. R. 1941. Diagnose einer neuen Posthornschnecke aus Kamerun. Arch. Mollusk. 73 (2/3) : 121-123. BOETTGER, O. 1905. Beitrag zur Kenntnis der Land-, Susswasser- und Brackwasser-Mollusken von Kamerun. Nachrichtbl. D. Malakol. Gess. 37 (4) : 153-184. BROWN, D. S. 1965. Freshwater gastropod molluscs of Ethiopia. Bull. Brit. Mus. (Nat. Hist.) Zool. 12 (2) : 37-94- CLESSIN, S. 1886. Die Familie de Limnaeiden, die genera Planorbis, Limnaeus, Physa und Amphipeplea. In Martini and Chemnitz, Syst. Conch. Cab. ed. 2 (Kiister), Nurenberg, 1 (17) : 349, pi. 49. fig. 7. COWPER, S. G. 1963. Schistosomiasis in Nigeria. Ann. trop. Med. Parasit. 57 (3) : 307-322. DAUTZENBERG, PH. 1890. Recoltes malacologiques de M. le capitaine Em. Dorr, dans le Haut-Senegal et le Soudan Fran9ais de 1886-1889. Mem. Soc. zool. Fr. 3 : 123-135. EHRENBERG, C. G. 1831. Symbolae Physicae sen I cones et Descriptiones Animalium Evertebra- torum Berlin. FISCHER-PIETTE, E. 1942. Les mollusques d'Adanson. /. Conchyliol. Paris. 85 (2, 3 and 4) : 103-374- FORCART, L. 1951. Mollusques. Resultats de la Mission Zoologique Suisse au Cameroun. Mem. Inst. Franf. Afr. Noire (Centre du Cameroun) Ser. Sci. nat. 1 : 239-244. GAUD, J. 1955. Les bilharzioses en Afrique Occidentale et en Afrique Centrale. Bull. Wld Health Org. 13 (2) : 209-258. GERMAIN, L. 1907. Contributions a la faune malacologique de 1'Afrique equatoriale. X. Mollusques nouveaux du lac Tchad (Mission R. Chudeau). Bull. Mus. Hist. nat. Paris 13 (4) : 269-274. 1909. Contributions a la faune malacologique de 1'Afrique Equatoriale. XXI. Mol- lusques nouveaux du Soudan Fran9ais recueillis par M. G. Garde. Bull. Mus. Hist. nat. Paris 15 (7) : 473-478. - 1911. Contributions a la faune malacologique de 1'Afrique equatoriale. XXIV. Mollusques nouveaux de la region du Tchad et de 1'Est africain. Bull. Mus. Hist. nat. Paris 17 (3) : 133-136. - 1917. Contributions a la faune malacologique de 1'Afrique equatoriale. XLVIII. Deux gasteropodes fluviatiles du Bassin du Senegal. Bull. Mus. Hist. nat. Paris 23 (7) : 95-103. GEZE, B. 1943. Geographic physique et geologic du Cameroun occidental. Mem. Mus. National Hist. nat. Paris 17 : 1-271. GRAY, J. E. 1831. Description of some new species of freshwater shells discovered by Capt. Owen R.N. on the coast of Africa. Zool. Miscellany 10-11. - 1850. In M. E. Gray, Figures of Molluscous Animals 4 : 119, pi. 309, fig. 4. HAAS, F. 1936. Binnen Mollusken aus Inner-Afrika. Abh. senckenb. naturf. Ges. 431 : 1-156. JICKELI, C. F. 1874. Fauna der Land- und Susswasser Mollusken Nord-Ost Afrikas. Nova Ada Leap. Carol. 37 : 1-352. KRAUSS, F. 1848. Die Sudafrikanischen Mollusken. 1-140, 6 plates Stuttgart. MANDAHL-BARTH, G. 1954. Tne freshwater mollusks of Uganda and adjacent territories. Ann. Mus. Roy. Congo Belg. Tervuren ser. in 8, sciences zoologiques, 32 : 1-206. g8 C. A. WRIGHT MANDAHL-BARTH, G. I95ya. Intermediate hosts of Schistosoma. African Biomphalaria and Bulinus I. Bull. W Id Health Org. 16 : 1103-1163. I957b. Intermediate hosts of Schistosoma. African Biomphalaria and Bulinus II. Bull. Wld Health Org. 17 : 1-65. 1960. Intermediate hosts of Schistosoma in Africa. Some recent information. Bull. Wld Health Org. 22 : 565-573. VON MARTENS, E. 1877. Die von Prof. Dr. R. Buchholz in Westafrika gesammelten Land- und Susswasser-Mollusken. Monatsber. Kgl. Akad. Wiss. Berlin. 1876 : 253-274. 1891. Die von Herrn Preuss bei der Barombi-Station in Kamerun gesammelten Land- und Susswasser-Mollusken. Sitz. Naturforsch. Freunde Berlin. 1891 : 29-34. MORELET, A. 1848. Testacea quaedam Africae occidentalis terrestria et fluviatilia. Rev. Zool. Paris 11 : 351-355. 1860. Description de nouvelles especes de 1'Afrique occidentale, rapport6es par M. le capitaine Vignon. /. Conchyliol. 8 : 189-191. NELSON, W. 1878. Description of a new species of Planorbis. Quart. J. Conchology 1 (16) : 379, pl. 4. fig- 3- N0STVIK, E. 1956. Results from the Danish expedition to the French Cameroons 1949-50. VI Gastropoda terrestria. Bull. Inst. franc. Afr. Noire 18a (2) : 607-611. PILSBRY, H. A. & BEQUAERT, J. 1927. The aquatic mollusks of the Belgian Congo, with a geographical and ecological account of Congo malacology. Bull. Amer. Mus. nat. Hist. 53 (2) : 69-602. PRESTON, H. B. 1914. In Longstaff, J. On a collection of non-marine mollusca from the Southern Sudan. Appendix I. Description of new species. /. Linn. Soc. Lond., Zool. 32 : 265-266. REYMENT, R. A. 1954. The stratigraphy of the Southern Cameroons. Geol. Fdren. Forhandl. Stockholm 76 (4) : 661-683. RICHARDS, C. S. 1963. Apertural lamellae, epiphragms and aestivation of planorbid mollusks. Amer. J. trop. Med. Hyg. 12 (2) : 254-263. DE ROCHEBRUNE, A. T. 1898. Sur une collection de mollusques terrestres et d'eau douce du Kameroun. Bull. Mus. Hist. nat. Paris 4 (2) : 157-8. SMITHERS, S. R. 1956. On the ecology of schistosome vectors in the Gambia, with evidence of their role in transmission. Trans. Roy. Soc. trop. Med. Hyg. 50 (4) : 354-365. SPENCE, G. C. 1925. On a collection of Mollusca from the Cameroons. /. Conch. 17 (8) : 248-250. - 1928. Non-marine mollusca from West Africa. /. Conch. 18 (7) : 211-216. SRAM, J. 1955. Remarques sur la glaciation au Cameroun. Etudes Camerounaises Inst. Fr. Afr. Noire 49 & 50 : 53-55. TREWAVAS, E. 1962. Fishes of the crater lakes of the North-western Cameroons. Bonn. zool. Beitr. 1962 (1-3) : 146-192. WRIGHT, C. A. 1957. Studies on the structure and taxonomy of Bulinus jousseaumei (Dautzenberg). Bull. Brit. Mus. (Nat. Hist.) Zool. 5 (i) : 1-28. ig63a. The freshwater gastropod molluscs of Western Aden Protectorate. Bull. Brit. Mus. (Nat. Hist.). Zool. 10 (4) : 257-274. i963b. The freshwater gastropod Mollusca of Angola. Bull. Brit. Mus. (Nat. Hist.}. Zool. 10 (8) : 447-528. WRIGHT, C. A. & BROWN, D. S. 1962. On a collection of freshwater gastropod molluscs from the Ethiopian Highlands. Bull. Brit. Mus. (Nat. Hist.). Zool. 8 (6) : 285-312. WRIGHT, C. A. & Ross, G. C. Electrophoretic studies on some planorbid egg-proteins. Bull. Wld. Health Org. In press. ZAHRA, A. 1952. Paragonimiasis in the Southern Cameroons ; a preliminary report. West Afr. med. J. 1 (2) : 75-82. 1953. Some notes on the incidence of schistosomiasis in the southern Cameroons. West Afr. med. J. 2 (i) : 26-29. PLATE i FIGS. 1-3. Bulinus camerunensis, Lake Barombi Kotto, x 8. FIG. 4. Bulinus camerunensis, laboratory bred, x 8. FIGS. 5, 6. Bulinus forskali, Lobe, x 8. FIGS. 7, 8. Bulinus rohlfsi, Lake Barombi Mbo, x 2-5. FIGS. 9, 10. Bulinus rohlfsi, Lake Barombi Kotto, x 2-5. FIGS, n, 12. Bulinus tropicus, Babungo, N'dop plain, x 2-5. FIG. 13. Lymnaea natalensis, Lake Wum, X 2-5. FIGS. 14-16. Biomphalaria pfeifferi, Lake Wum, x 2-5. FIGS. 17-23. Biomphalaria camerunensis, Bambalang, N'dop Plain, x 2-5. Bull. B.M. (N.H.) Zoo/. 13, 3 PLATE i PLATE 2 FIGS. 1-3. Scgmentorbis angustus, Lake \Yurn, x 14. FIGS. 4-6. Anisus coretus, Lobe, x 14. FIGS. 7-9. Gyraiilus costulatus, Lake Wum, x 14. FIGS. 10, ii. Ferrissia sp., Lobe, x 14. FIG. 12. Gabbia sp. Lake Wum, x 14. Bull. B.M. (N.H.) Zoo/. 13, 3 PLATE 2 PLATE 3 Figs. 1-3 & 7-10 x 1-6, Figs. 4-6 x 6-5. FIG. i. Potadoma freethi, Fernando Poo, (Holotype). FIG. 2. Potadoma freethi, (P. dykei Spence). FIG. 3. Potadoma freethi, Barombi Mbo. FIG. 4. Microsculpture of specimen in Fig. i. FIG. 5. Microsculpture of specimen in Fig. 2. FIG. 6. Microsculpture of specimen in Fig. 3. FIGS. 7, 8. Potadoma freethi, juveniles, Lobe. FIG. 9. Lanistes libycns, Lake Barombi Mbo. FIG. 10. Lanistes libvcus, Lobe. Bull. B.M. (N.H.) Zoo/. 13, 3 PLATE 3 10 PRINTED IN GREAT BRITAIN BY ADLARD & SON LIMITED BARTHOLOMEW PRESS, DORKING T /& ' '>>, ^ ) I i?-." BARBUS (PISCES, CYPRINIDAEJ^y^ OF THE VOLTA REGION A. J. & J. HOPSON BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 13 No. 4 LONDON: 1965 BARBUS (PISCES, CYPRINIDAE) OF THE VOLTA REGION BY A. J. & J. HOPSON Fisheries Division, Accra, Ghanaf f Present address: Federal Fisheries Service, Lake Chad Research Station, P.O. Box 227, Maiduguri, Northern Nigeria. Pp. 99-149; 1 8 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 13 No. 4 LONDON: 1965 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 13, No. 4 of the Zoological series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. Trustees of the British Museum (Natural History) 1965 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued October, 1965 Price i 2s. BARBUS (PISCES, CYPRINIDAE) OF THE VOLTA REGION By A. J. & J. HOPSON CONTENTS Page SYNOPSIS . . . . . . . . . . . 101 INTRODUCTION ........... 101 CEPHALIC PIT-LINES . . . . . . . . . .103 Barbus ablabes (Bleeker) . . . . . . . . .105 Barbus spurrelli Blgr. . . . . . . . . .108 Barbus macrops Blgr. . . . . . . . . .110 Barbus par ablabes Blanc & Daget . . . . . . .114 Barbus hypsolepis Daget . . . .. . . . .117 Barbus nigeriensis Blgr. . . . . . . . . .119 Barbus lawrae Hopson . . . . . . . . .121 Barbus subinensis Hopson . . . . . . . . .124 Barbus trispilus (Bleeker) . . . . . . . . .126 Barbus sublineatus Daget . . . . . . . . .128 Barbus macinensis Daget . . . . . . . . .130 Barbus voltae Hopson . . . . . . . . .132 Barbus atakorensis Blanc & Daget . . . . . . .135 Barbus punctitaeniatiis Daget . . . . . . . .137 Barbus bawkuensis Hopson . . . . . . . . .139 Barbus pobeguini Pellegrin . . . . . . . . .141 Barbus stigmatopygus Blgr. ........ 143 Barbus leonensis Blgr. . . . . . . . . .145 KEY TO THE VOLTAIC SPECIES OF Barbus . . . . . .147 ACKNOWLEDGEMENTS .......... 148 REFERENCES ........... 149 SYNOPSIS Eighteen species of Barbus, collected recently in Ghana, are described and figured. Fifteen of the species were obtained from the Volta basin. The synonymy of West African Beirabarbus is revised and four species of this subgenus are shown to occur in Ghana. Barbus nigeriensis is redescribed using the holotype augmented with new material. A key is given for voltaic species of Barbus. INTRODUCTION Barbus is one of the most specifically numerous genera of freshwater fish in Africa and well over 200 species have been described. The genus is well represented in West Africa and fifteen species were identified from a collection of fish which we made in the Volta basin, chiefly in northern Ghana, during 1961. Most of the species are widespread and common and Barbus forms an important and characteristic element in the fish fauna of all habitats. The present investigations deal chiefly with voltaic ZOOL. 13, 4 4 102 A. J. & J. HOPSON species but observations are also made on three species of Barbus collected from the Prah basin in the forest region of southern Ghana : B. ablabes (Bleeker) B. trispilus (Bleeker) B. subinensis Hopson Environmental conditions in the rivers of northern Ghana are comparable with those described by Blanc & Daget (1957) for Haute Volta and by Holden (1963) for the Sokoto River in Northern Nigeria which are indeed typical of the whole savannah region of West Africa. Flooding occurs during a short rainy season lasting from mid-May to mid-September followed by a rapid fall in water level with the onset of the dry season. By the end of December the output of the Black Volta in northern Ghana is greatly diminished and the White Volta is reduced to a mere trickle linking a chain of sandy pools. Smaller tributaries dry out completely before the onset of the rains in May. A more stable environment is found in permanent, well-vegetated oxbow lakes which are particularly characteristic of the valley of the White Volta. Relatively stable conditions are also found in dams constructed across seasonal streams for agricultural purposes ; these are now common in Ghana. Previous observations on Barbits in Ghana have dealt almost exclusively with species occuning outside the Volta basin (Boulenger, 1911, 1916 ; Trewavas in Irvine, 1947). A detailed account of voltaic species was made only during the last decade by Blanc & Daget (op. cit.} who worked on material from Haute Volta. They recorded nine species of Barbus from the Volta basin : B. macrops Blgr. (= B. ablabes, sensu Blanc & Daget, 1957) B. parablabes Blanc & Daget B. voltae Hopson (=B. nigeriensis, sensu Blanc & Daget, op. cit.) B. macinensis Daget B. punctitaeniatus Daget B. atakorensis Blanc & Daget B. leonensis Blgr. B. stigmatopygus Blgr. Most of these are widely distributed elsewhere in the savannah region of West Africa and all, with the exception of B. parablabes and B. atakorensis, were shown by Blanc & Daget to be common to the Niger basin. Six additional species were recorded from the Volta basin during the present survey : B. spurrelli Blgr. B. hypsolepis Daget B. nigeriensis Blgr. B. lawrae Hopson B. bawkuensis Hopson B. pobeguini Pellegrin Of these, B. hypsolepis, B. nigeriensis and B. pobeguini also occur in the Niger basin. In the course of the present investigations extensive use has been made of material BARBUS OF THE VOLT A REGION 103 in the British Museum (N.H.) and in the Musee National d'Histoire naturelle, Paris. Comparisons were made with type material whenever possible. Most of our obser- vations agree with Daget (1954) and Blanc & Daget (1957). We differ mainly in the synonymy of the subgenus Beirabarbus and in the diagnosis of B. nigeriensis. Four species of Beirabarbus are shown to occur in Ghana : B. ablabes,: B. spunelli, B. macrops and B. parablabes. Barbus deserti of Daget (1954) and B. ablabes of Blanc & Daget (1957) are now regarded as misidentified examples of B. macrops, which proves to be widely distributed through the savannah regions of West Africa and the Chad basin. Barbus ablabes and B. deserti, with which B. macrops has been confused, are apparently restricted in distribution, B. ablabes to the forest rivers between south west Ghana and Liberia and B. deserti to the north central Sahara. A redescription has been prepared for B. nigeriensis, based on the holotype, on material from Western Nigeria and on specimens recently collected in northern Ghana. The species described by Blanc & Daget (1957) as B. nigeriensis was misidentified. It has been shown recently to be a new species, B. voltae (Hopson, 1965)- Methods : the snout length was measured between the verticals to the tip and the anterior margin of the eye ; the head length was taken to the posterior extremity of the bony operculum. Ledges of skin forming the dorsal margin of each orbit are included in the interorbital width. Perforated scales overlapping the base of the caudal fin are included in the lateral line count. Other data were obtained in accordance with standard practice. All specimens were fixed initially in 5% formalin for three to four months. The material was then carefully washed and transferred to an 80% solution of alcohol for permanent preservation. It was found that this procedure satisfactorily exposes the underlying melanophores, the pattern of which is often of taxonomic importance. The relationship of lateral line to horizontal myoseptum often provided characters of diagnostic value. Marked differences were noted among the Barbus under consideration in the degree to which the lateral line dipped below the myoseptum. In species of the subgenus Beirabarbus and in B. punctitaeniatus for example the dip is slight whereas in B. nigeriensis and B. lawrae it is more pronounced. The maximum distance between horizontal myoseptum and lateral line has been expressed in terms of the depth of adjacent scale rows. The point at which the myoseptum and lateral line converge also varies interspecifically. In Beirabarbus the point is usually above the base of the anal fin whereas in B. perince and in B. werneri they converge near the tip of the caudal peduncle. CEPHALIC PIT-LINES Herre (1932) first drew attention to cephalic pit-lines in Barbus palustris (== B. radiatus, vide Greenwood, 1963) and considered them to be of sufficient importance to be the criterion for establishing a new genus, Beirabarbus. Apparently unaware of Metre's work, Schultz (1942) described a new genus Mannicthys (M. lucileae Schultz = B. ma-crops Boulenger), on an identical basis. Johnels (1954) later showed that cephalic pit-lines were present in at least six West African Barbus and doubted io 4 A - J- & J- HOPSON the advisability of using them as generic or subgeneric characters. Recently, Greenwood (1962) has given a detailed discussion on the occurrence of pit-lines in African Barbus. He upholds Herre's taxon Beirabarbus at subgeneric level and recognizes two distinct patterns of pit-line distribution. (a) Pit-lines of the Beirabarbus type where the individual pits are small and very numerous, grouped tightly into lines like beads and raised above the surface as ridges. These lines sometimes branch and are present in a characteristic pattern on the side of the snout, the cheek, the operculum and on the dorsal surface of the head. On the basis of these characters, Greenwood referred the following species to the subgenus Beirabarbus : B. ablabes (Bleeker) B. deserti Pellegrin B. radiatus Peters B. aurantiacus Blgr. B. jae Blgr. B. macrops Blgr. B. aspilus Blgr. B. spurrelli Blgr. B. callipterus Blgr. (b) The pits are relatively larger than in Beirabarbus and much fewer in number. They are sometimes arranged in lines but with the pits always well separated from one another. The lines are never raised above the surface in ridges and are usually visible only on the cheek and operculum. Greenwood found this type of pit in fourteen species of Barbus : B. nigeriensis Blgr. B. trispilomimus Blgr. B. kessleri (Steindachner) B. pumilus Blgr. B. trispilus (Bleeker) B. anema Blgr. B. congicus Blgr. B. svenssoni Johnels B. pseudognathodon Blgr. B. collarti Poll B. pleiwopholis Blgr. B. leonensis Blgr. B. urostigma Blgr. B. cercops Whitehead Of the above species, only B. kessleri has a serrated last simple dorsal ray and all with the exception of B. cercops are West African in distribution. Our own observations agree with Greenwood's. Cephalic pit-lines were noted in all of the eighteen species of Barbus described in the present work. Of these, four are referable to Beirabarbus : B. ablabes B. spurrelli B. macrops B. par ablabes Blanc & Daget The remaining fourteen species all had weakly developed pit-lines corresponding to the description under (b) above. They are as follows : B. hypsolepis Daget B. voltae Hopson B. nigeriensis Blgr. B. atakorensis Blanc & Daget B. lawrae Hopson B. punctitaeniatus Daget BARBUS OF THE VOLTA REGION 105 B. subinensis Hopson B. bawkuensis Hopson B. trispiliis (Bleeker) B. pobeguini Pellegrin B. siiblineatus Daget B. stigmatopygus Blgr. B. macinemis Daget B. leonensis Blgr. Cephalic pit-lines were always most clearly seen in formalin-fixed specimens. Barbus (Beirabarbus) ablabes (Bleeker, 1863) (Text-fig, i) Puntius (Barbodes) ablabes Bleeker, 1863, Nat. Verh. Wet. Haarlem, 23 : 114, pi. 23, fig. i. Barbus ablabes : Boulenger, 1911, Cat. Afr. Fish. 2 : 156, fig. 133. Barbus ablabes : Pellegrin, 1923, Les poissons des eaux donees de I'Afrique occidentale : 134, Paris. [non Barbus albabes : Blanc & Daget, 1957, Mem. Inst. France. Afr. Noire, no. 50 : 132, fig. 13, see p. no below]. SYNTYPES : two fish S.L. 64 and 66 mm. from Dabo Crom, Guinea in the Leyden Museum. DESCRIPTION based on 30 fish, 31-53 mm. S.L. from the R. Weiwei, Kumasi, south Ghana supplemented with notes on the following material in the British Museum (N.H.) : 10 fish, 44-70 mm. S.L. from the Kotchwah River, south Ghana (reg. no. 99.12.22.46-55) ; 4 fish, 36-48 mm. S.L. from the R. Atesu, south Ghana (reg. no. 1903.4.24.94-97) ; 2 fish, 58-62 mm. S.L. from near Bunsu, south Ghana (reg. no. 1938.12.15.40-41). Body moderately compressed. S.L. 3-1 to 3-7 times the maximum depth and 3-1 to 3-6 times the length of the head. Predorsal profile convex with a slight nuchal hump. Eyes inclined slightly upwards, the diameter 2-9 to 3-4 times in the length of the head and 1-05 to 1-2 times in the interorbital width. Snout bluntly pointed, 0-6 to 0-9 times the diameter of the eye. Mouth moderate, subterminal. Anterior barbel overlapping the base of the posterior barbel, 0-4 to 0-6 times the diameter of the eye. Posterior barbel extending well past the vertical to the centre of the eye, 0-6 to 0-9 times the eye diameter. Pit-lines of the Beirabarbus type present on the side of the snout, the cheek, the operculum and on the dorsal surface of the head, between the eyes. Pectoral fin 0-6 to 0-8 times the length of the head, the tip just overlapping the pelvic origin in the four smallest individuals (31-37 mm. S.L.). Last simple dorsal ray smooth, flexible, 0-8 to i-o times the length of the head. Tip of dorsal fin slightly rounded, distal margin concave. Dorsal fin rays III 7 (f.i) or III 8 (f.29). Anal fin rays III 5. The pelvic origin lies between the verticals to the third simple and the first branched dorsal rays. Caudal peduncle i-i to 1-4 times as long as deep. The lateral line dips to a maximum of the depth of half a scale row below the horizontal myoseptum in the anterior third of the body ; the two converge at a point in the verticals to the last anal fin rays. Lateral line scales 22 to 27 (modal number 24). Three and a half scales between the lateral line and the dorsal origin, 3! scales between the lateral line and the mid-ventral line immediately in front of the pelvic origin and 2\ scales between the lateral line and the pelvic origin. Eleven (f.i) or 12 (f.29) scales round the caudal peduncle. io6 A. J . & J. HO PS OX FIG. i. Barbus ablabes ; N.B. in this and all following figures the caudal fin is simplified. Summary of morphometric data based on 30 fish, 31-53 mm. S.L. ; measurements are expressed as percentages of S.L. Maximum depth . Length head Diameter eye Interorbital width Length snout Length anterior barbel . Length posterior barbel . Length pectoral fin Length dorsal fin . Length caudal peduncle Depth caudal peduncle . COLORATION : in living specimens, greenish-brown above, silvery on the sides, whitish below. A conspicuous dark mid-lateral band runs from the snout to the posterior end of the body ; the band frequently shows a brilliant green iridescence. Dorsal and caudal fins are tinged basally with yellow which becomes more intense towards the distal margin to merge with a terminal band of orange-red. Anal fin orange-red with a colourless free margin. Paired fins yellow on the anterior rays ; the pelvic fins are also tinged with orange in brightly coloured individuals. Iris, orange-red above. In formalin fixed specimens the dorsal surface is densely pig- mented with fine melanophores free from a narrow zone near the margin of each scale ; the pigment-free zone becomes wider on more lateral scales. Dorsal and lateral scales are outlined with larger and darker melanophores, more concentrated in a vertical bar on the pocket of each scale to form a regular pattern. These markings are heaviest on the lateral line scales each of which bears a narrow vertical Range Mean Allometry 27-5-33-3 27-6-32-5 8-5-10-2 29-6 29-8 9-4 Negative Negative 9'8-H -2 6-2-8-7 10-5 7-6 3-8-5-9 6-2-8-9 19-6-23-7 4'5 7-4 21-7 25-4-30-2 27-7 17-4-21-4 19-7 14-2-15-9 15-3 BARB US OF THE VOLTA REGION 107 stripe in the shape of a cupid's bow. A black mid-lateral band, one third of a scale row in depth, runs from the side of the snout, across the operculum and along the body to the tip of the caudal peduncle. There are scattered melanophores on the pockets of the first and sometimes the second scale row below the lateral line but the ventral surface is generally pigment-free. The anterior margin of the dorsal fin appears greyish-black owing to a moderately heavy peppering of melanophores on the membrane between the second and third simple rays, along the anterior margin of the distal half of the third simple ray and to a lesser extent on the second simple ray and on the membrane between the third simple and first branched rays. Pig- mentation on the remainder of the fin is restricted to scattered and inconspicuous melanophores chiefly on the distal parts. DIAGNOSIS AND AFFINITIES : Dr. Greenwood recently compared the syntypes of B. ablabes in the Leyden Museum with material from the Kotchwah River, south Ghana (Brit. Mus. (N.H.) reg. no. 99 . 12 . 22 . 46-55) which has been used to supplement the above description. He found (pers. comm.) that the two collections were identical in barbel length and in details of pigmentation not lost in preservation. There was no sign in the type specimens of particularly heavy pigment in the region of the second simple dorsal ray (a heavily pigmented second simple ray is characteris- tic of the closely related B. macrops). Barbus ablabes resembles B. macrops in morphometric details and in pigmentation. The most obvious difference is in the distribution of melanophores on the dorsal fin. In B. ablabes moderately heavy pigment is distributed along the entire length of the second and third simple rays and the anterior margin thus appears uniformly grey ; in B. macrops the second simple ray and the tips of the first to third branched rays are densely pigmented so that the dorsal fin has a broad black tip and a black streak down the lower half of the anterior margin. The part of the third simple ray lying between these two areas of dense pigment is almost clear of melanophores. The two species also differ in the form of the pocket pigment on the lateral line scales ; in B. ablabes the markings are relatively slender in the shape of a cupid's bow, whereas in B. macrops the markings are broader and triangular. Barbus ablabes is readily separated from West African populations of B. macrops by the relatively longer posterior barbels (6-2-8-9 C - T - 2 '5 to 5-1% S.L. in Ghanaian material). The relative length of the posterior barbel in B. macrops, however, varies geographically and a population from the Tibesti shows a slight overlap with B. ablabes (5-1-6-6 in B. macrops c.f. 6-2-8-9% S.L.). The relative diameter of the eye which Boulenger (1911) originally used to separate the two species is no longer of diagnostic value. This character also varies geo- graphically in B. macrops, and in the Tibesti population the eyes are relatively smaller than in B. ablabes. The geographical distributions of B. ablabes and B. macrops do not appear to overlap. Barbus ablabes is recorded only from the forest rivers of south-west Ghana and Liberia whereas B. macrops is widespread throughout the savannah rivers of West Africa and the Chad basin. Barbus ablabes is easily distinguished from B. spurrelli by the dark mid-lateral band io8 A. J. & J. HOPSON and the denseness of the pocket pigment on the lateral line scales. In B. spurrelli the mid-lateral band is very faint and the pigmentation of lateral line scales is only slightly heavier than that of more dorsal scales. Barbus ablabes is compared with B. parablabes in the description of that species. DISTRIBUTION : in Ghana, common and widespread in forest streams and rivers in the southwest. Probably absent from the Volta Basin. Elsewhere, Dabo Crom, Guinea (type locality) ; Liberia (no locality given, Boulenger, 1911). The record of B. ablabes from the St. Paul basin, Liberia (Schultz, 1942) is doubtful ; Dr. Weitzman of the Smithsonian Institution reports that the specimens appear to have no Beira- barbus type cephalic pit-lines (pers. comm.}. Barbus ablabes probably belongs to a faunal group restricted to the forest rivers of western Ghana, Cote d'lvoire and Liberia. The group may also include Alestes longipinnis (Giinther), Petersius occidentalis (Giinther), Nannocharax seyboldi Schultz, Barbus trispilus (Bleeker) and Eutropius mentalis Blgr. among other species. Barbus (Beirabarbus) spurrelli Boulenger, 1913 (Text-fig. 2) Barbus spurrelli Boulenger, 1913, Proc. zool. Soc. Lond. : 51, pi. 3, fig. i ; Idem, 1916, Cat. Afr. Fish. 4 : 264, fig. 162. Barbus ablabes var. spurrelli Pellegrin, 1923, Les poissons des eaux douces de I'Afrique occidentale : 135, Paris. Barbus spurrelli : Daget, 1952, Mem. Inst. France. Afr. Noire, no. 19 : 316, fig. 7. LECTOTYPE : 67-8 mm. S.L.,B.M. (N.H.) reg. no. 1911.11.27.1, from near Dunkwa (Prah Basin) south Ghana. We consider this fish to be the specimen illustrated in Boulenger, 1916, fig. 162. DESCRIPTION based on 12 fish, 25-46 mm. S.L. from the River Azubone (Volta Basin), near Mpraeso, south Ghana. Body compressed. S.L. 3-25 to 3-5 times the maximum depth and 2-9 to 3-5 times the length of the head. Predorsal profile convex with a slight nuchal hump. Snout bluntly pointed, 0-85 to i-i times the diameter of the eye. Mouth moderate, subterminal. Anterior barbel 0-3 to 0-7 times the diameter of the eye, barely overlapping the base of the posterior barbel. Posterior barbel 0-6 to 0-95 times the diameter of the eye, the tip reaching the vertical to the posterior margin of the pupil. Eyes inclined slightly upwards, the diameter 3-25 to 3-85 times in the length of the head and 1-2 to 1-5 times in the interorbital width. Pit-lines of the Beirabarbus type present on the side of the snout, the cheek, the operculum and on the dorsal surface of the head, between the eyes. Pectoral fin 0-6 to 0-75 times as long as the head, sometimes (f.3) overlapping the pelvic origin. Last simple dorsal ray smooth, flexible, 0-75 to 0-85 times the length of the head. Apex of dorsal fin slightly rounded, distal margin concave. Dorsal fin rays III 7 (f.i) or III 8 (f.n). Anal fin rays III 5. The pelvic origin lies between the verticals of the first and second branched dorsal rays. Caudal peduncle 1-2 to 1-55 times as long as deep. The lateral line dips to a maximum of the depth of slightly over half a scale row below the horizontal myoseptum in the anterior third of the body ; the BARBUS OF THE VOLTA REGION 109 two converge above the posterior anal fin rays. Lateral line scales 24 to 26 (modal number 25). Three and a half to 4! scales between the lateral line and the dorsal origin, 3^ scales between the lateral line and the mid-ventral line immediately in front of the pelvic origin and 2 to 2| scales between the lateral line and the pelvic origin. Twelve scales round the caudal peduncle. 10mm. S.L. ; measure- FIG. 2. Barbus spurrelli. Summary of morphometric data based on 12 fish, 25 to 46 mm. ments are expressed as percentages of S.L. Range Mean Allometry Maximum depth . . . 28-3-30-5 29-3 Length head . . . 28-5-33-0 31-5 Negative Diameter eye . . 7 ' 4-9 '6 8-9 Negative Interorbital width . . 10-8-13-0 11-7 Length snout . . .7-8-8-8 8-2 Length anterior barbel . . 3 2-5 -8 4-7 Length posterior barbel . . 5 7-7 -9 7-2 Length pectoral fin . . 18-6-24-6 22-2 Length dorsal fin . . . 22-6-28-1 25-8 Length caudal peduncle . . 18-5-23-3 20-8 Depth caudal peduncle . . 14-3-15-1 14-6 COLORATION : in formalin fixed specimens the dorsal surface is relatively heavily pigmented with fine melanophores absent only from a narrow zone near the margin of each scale. Dorsal and lateral scales are outlined with larger and darker melano- phores ; this pigment is more concentrated in a vertical bar on the pocket of each scale to form a regular pattern. Lateral line scales are marked with similar but slightly heavier pigment ; each vertical bar is bisected horizontally by the lateral line pore. Diffuse melanophores form an indistinct band, approximately one third of a scale row in depth, running just below the horizontal myoseptum in the anterior half of the body ; the two converge over the anal fin and run together onto the caudal no A. J . & J. HOI'S ON peduncle where the band ends in front of a small and inconspicuous terminal spot, overlapping the base of the caudal fin. A narrow black band runs obliquely down- wards from the lateral line origin to the base of the pectoral fin. The ventral surface is generally pigment-free, with the exception of a few scattered melanophores on the pockets of the first and sometimes the second row of scales below the lateral line. The snout is marked with a dark lateral spot. The anterior margin of the dorsal fin is moderately peppered with melanophores which are slightly heavier on the free edge of the last simple ray. Melanophores are thinly distributed over the distal half of the branched dorsal rays. DIAGNOSIS AND AFFINITIES : these data are in close agreement (allowing for allometry) with the type description. Barbus spurrelli is readily distinguishable from other Beirabarbus occurring in West Africa either by the absence of a dark mid- lateral band or by the lack of a dark spot on or near the tip of the dorsal fin. Detailed comparisons with B. ablabes and B. pamblabes are given in the descriptions of those species. DISTRIBUTION : Ghana. The type locality of B. spurrelli is Dunkwa (Prah basin) in the forest region of south Ghana. Our material was collected from the R. Azubone near Mpraeso, a forest stream on the Volta side of the Prah-Volta watershed. Else- where, Sierra Leone (Pellegrin, 1923) ; Mount Nimba, French Guinea (Daget, 1952). The record of B. spurrelli from the St. Paul basin, Liberia (Schultz, 1942) is doubtful. Dr. Weitzman of the Smithsonian Institution has kindly re-examined the material and reports that no Beirabarbus cephalic pit-lines are visible (pers. comm.). Barbus (Beirabarbus} macrops Boulenger, 1911 (Text-fig. 3) Barbus macrops Boulenger, 1911, Ann. Mag. nat. Hist. (8) 7 : 374 ; Idem, 1916, Cat. Afr. Fish. 4 : 265, fig. 163. Barbus deserti : Pellegrin, 1921 (Ennedi and Tibesti specimens only), Mem. Soc. Sci. nat. Maroc, (1)2:143. Barbus macrops : Pellegrin, 1923, Les poissons des eaux douces de VAfrique occidentale : 136, Paris. Barbus gambiensis Svensson, 1933, Kungl. Sven. vet. Handl. (12) 3 : 67, pi. 4, fig. i. Mannichthys lucileae Schultz, 1942, Proc. U.S. Nat. Mus. 92 : 321, pi. 35, fig. 3. Barbus deserti : Daget, 1954, Mem. Inst. France. Afr. Noire, no. 36 : 192, fig. 62. Barbus ablabes : Blanc & Daget, 1957, Mem. Inst. France. Afr. Noire, no. 50 : 132, fig. 13. LECTOTYPE : a fish of 45-0 mm. S.L. (Brit. Mus. (N.H.) reg. no. 1912.4.1.110), from above the rapids, Crobal River. Portuguese Guinea. We consider this to be the specimen illustrated in Boulenger, 1916, fig. 163. DESCRIPTION based on 24 fish, 31 to 56 mm. S.L. from a tributary of the White Volta, 5 m. N. of Bawku, northern Ghana. Body moderately compressed. S.L. 3-3 to 3-9 times the maximum depth and 3-3 to 37 times the length of the head. Predorsal profile convex with a slight nuchal hump. Snout bluntly pointed, 0-7 to 0-85 times the diameter of the eye. Mouth moderate, sub-terminal. Barbels moderate, the anterior o-i to 0-35 times the diameter of the eye, the tip barely over- B A R B US OF THE VOLT A REGION lapping the base of the posterior barbel. Posterior barbel 0-25 to 0-6 times the eye diameter, the tip reaching the vertical to the centre of the eye in a few individuals. Eyes inclined slightly upwards, 2-85 to 3-1 times in the length of the head and 1-05 to 1-15 times in the interorbital width. Pit-lines of the Beirabarbus type fully developed on the side of the snout, the cheek, the operculum and on the dorsal surface of the head between the eyes. Pectoral fin 0-65 to 0-8 times the length of the head, the tip always falling short of the pelvic origin. Last simple dorsal ray, smooth, flexible, 0-9 to 1-05 times the length of the head. Apex of dorsal fin slightly rounded, distal margin concave. Dorsal fin rays III 7 (f.i) or III 8 (f.23). Anal fin rays III 5. The origin of the dorsal fin lies between the verticals to the last simple and first branched dorsal rays. Caudal peduncle 1-3 to 1-6 times as long as deep. The lateral line dips to a maximum depth of slightly over half a scale row below the horizontal myoseptum in the anterior third of the body ; the two converge in the vertical to the last anal fin rays. Lateral line scales 24 to 28 (modal number 25). Three and a half scales between the lateral line and the dorsal origin, 3-J- scales between the mid- ventral line immediately in front of the pelvic origin and 2-S- scales between the lateral line and the pelvic origin. Twelve scales round the caudal peduncle. FIG. 3. Barbiis macrops. Summary of morphometric data based on 24 fish, 31 to 56 mm. ments are expressed as percentages of S.L. Maximum depth . Length head Diameter eye Interorbital width Length snout Length anterior barbel Length posterior barbel Length pectoral fin Length dorsal fin . Length caudal peduncle Depth caudal peduncle S.L. ; measure- Range Mean Allometry 25-5-30-2 26-4-29-7 8-3-10-2 27-1 27-6 9-1 Negative Negative 9-2-12 -2 6-3-7-3 10-2 6-8 1-2-3-1 2-3 2-5-5-1 17-8-21-8 25-2-28-8 l8 I-2I -I 3'5 19-7 26-6 19-9 13-3-14-8 14-1 ii2 A. J. & J. HOPSON COLORATION : in living specimens, greenish-brown above, silvery on the sides, whitish below. A dark mid-lateral band extends from the snout to the tip of the caudal peduncle. Dorsal and caudal fins bright orange-red, the dorsal with a con- spicuous black tip. The other fins are frequently tinted orange-yellow. Iris, bright orange-red above. In formalin fixed specimens the dorsal surface is densely pig- mented with fine melanophores absent only from a narrow zone near the margin of each scale. Dorsal and lateral scales are clearly outlined with larger and darker melanophores which are more concentrated in a vertical line on the pocket of each scale to form a regular pattern. These markings are heaviest on the lateral line where each scale bears a relatively broad and triangular group of melanophores, bisected horizontally by the lateral line pore. A conspicuous black band, one-third of a scale row in depth, runs mid-laterally from the side of the snout, across the operculum and along the body to the end of the caudal peduncle. In some speci- mens an oblique dark stripe is noticeable between the origin of the lateral line and the base of the pectoral fin. The pattern of scale pocket pigment, described above for dorsolateral scales, is repeated more faintly on the first and sometimes the second scale row below the lateral line. The ventral surface is generally pigment-free. The apex of the dorsal fin bears a dense black spot extending over the distal third of the first four branched rays but heaviest on the tips of the first two. Dense pigment is also present in the vicinity of the second simple ray, particularly on the membrane between it and the third ray. The third simple ray is less heavily pigmented and closely packed melanophores are present only in the basal third along the anterior margin and near the tip ; the central section of the third simple ray is almost pigment- free. GEOGRAPHICAL VARIATION : during the present investigations, B. macrops were examined from most of the savannah region of West Africa, the Chad Basin, the Hoggar in the western Sahara and the Tibesti in the eastern Sahara. Details of pigmentation, particularly on the dorsal fin, proved to be constant throughout this wide area. Marked geographical variation was however noted in certain morpho- metric characters, especially eye diameter but also the relative length of the barbels as demonstrated below : B. macrops : geographical variations in eye diameter (N.B. negative allometry). No. Range S.L. Eye diameter Locality examined (mm.) expressed as % S.L. Tibesti 24 43-55 6-2-7-6 Niger (N. Nigeria) . 3 48-60 7-8-8-6 Sierra Leone . 2 61 8-0-8-4 Volta (N. Ghana) . . 24 31-56 8-3-10-2 Volta (S. Ghana) . . 17 21-52 8-2-10-6 Hoggar ... 4 26-39 8-6-10-0 Chad Basin (Nigeria) . 12 27-48 9-0-10-2 Port. Guinea . . 16 25-60 8-8-11-6 B. macrops : geographical variations in posterior barbel length. BARBUS OF THE VOLTA REGION 113 No. Range S.L. Length post, barbel Locality examined (mm.) expressed as % S.L. Chad Basin (Nigeria) . 12 27-48 2-4-3-3 Volta (N. Ghana) . . 24 3!~56 2-5-5-1 Volta (S. Ghana) . . 17 21-52 2-7-6-0 Hoggar . 4 26-39 3'O-5'i Port. Guinea . . 16 25-60 3'-5'3 Sierra Leone . . 2 61 4 . 5-5 i Niger (N. Nigeria) . 3 48-60 4-8-5-4 Tibesti 24 43-55 5-1-6-6 Diagnosis and Affinities : there has been considerable confusion between B. macrops and B. deserti Pellegrin. This undoubtedly originated from the inclusion of B. macrops (localities Tibesti and Ennedi) in material used for a redescription of B. deserti (Pellegrin, 1921). We have examined type material of B. deserti from the Tassili des Azdjers, central Sahara, both the syntypes in the Paris Museum (reg. no. 09.457-458) and a syntype in the British Museum (N.H.) (reg. no. 1909.12.9.7) on which Boulenger (1911) based his diagnosis. They all differ from B. macrops in the following details : (a) In B. deserti there is no well-defined mid-lateral band running from the snout across the operculum to the tip of the caudal peduncle. Barbus deserti is marked only in the posterior half of the body with relatively faint dark pigment on the horizontal myoseptum. (b) The lateral line scales are relatively narrower and deeper in B. deserti (cf. B. parablabes) , (c) The lateral line scale pockets of B. deserti are marked with narrow vertical stripes (cf. B. parablabes), never with broader triangular spots as in B. macrops. (d) The second simple dorsal fin ray (always dense black in B. macrops} is only lightly pigmented. (e) Further details of pigmentation of the dorsal fin in B. deserti are obscured by damage to the tip in all the type material. Pellegrin (1909) noted that the dorsal fin was tipped with black whereas Boulenger (op. cit.) records that a spot was present near the tip. It seems likely that Boulenger's description is the more accurate of the two. Three Barbus from the oasis of El Barkat, Fianferrari, Fezzan (within 500 miles of the type locality in the same region of the N. Central Sahara) in the British Museum (N.H.) (reg. no. 1937.7.8.4-6), similar to the types in every respect, bear a spot on the distal half of the last simple and first three branched dorsal rays, well clear of the distal margin. This should prove a trenchant character in distinguishing B. deserti from B. macrops where the black pigment extends to the tip of the anterior dorsal rays. Comparisons of B. macrops with B. ablabes and B. parablabes are included in the descriptions of those species. DISTRIBUTION AND HABITAT : in Ghana, restricted to the Volta basin where it is the most common and widespread species of Barbus. Barbus macrops occurs in rivers, well vegetated lakes and dams, over all substrates coarse and fine. Elsewhere, H4 A.J.&J.HOPSON distributed over most of the savannah area of West Africa and the Chad basin, extending into the Sahara in the Hoggar to the west and the Tibesti to the east : Portuguese Guinea (Boulenger, 1911, type locality) ; Gambia River (Svensson, 1933) ; L. Kwarko, Sierra Leone (Brit. Mus. (N.H.) reg. no. 1958.9.18.54-55) ; St. Paul and Bolor Rivers, Liberia (Schultz, 1942) ; Middle Niger (Daget, 1954) ; Arak Gorge, Hoggar (Brit. Mus. (N.H.) reg. no. 1932.5.6.5-11); R. Katagum (Chad Basin) N. Nigeria (Brit. Mus. (N.H.) reg. no. 1928.7.3.56-62) ; Sherda, Tibesti (Brit. Mus. (N.H.) reg. no. 1960.6.7.111-160). LIFE HISTORY : in the White Volta, B. macrops spawns during the rains. Gonads start to mature during April and the majority of fish are ripe by mid- June when the first post-larvae appear in the river. Spawning is over by the end of July and the size-composition of samples taken in August suggests that most fish die at the age of one year. Juvenile fish have grown to a modal length of c. 35 mm. by November. Females, which ripen at a minimum S.L. of c. 50 mm, grow slightly larger than males. Maximum size, a female of 98 mm. S.L. Barbus (Beirabarbus) parablabes Blanc & Daget, 1957 (Text-fig. 4) Barbus parablabes Blanc & Daget, 1957, Mem. Inst. France. Afr. Noire, no. 50 : 134, fig. 13. HOLOTYPE : a fish 62 mm. S.L. from between Tanguieta and Natitingou, Volta basin, N. Dahomey, Paris Mus. reg. no. 56-108. DESCRIPTION based on 24 fish, 39-52 mm. S.L. from the waterworks reservoir at Tamale (Volta basin), northern Ghana. Body moderately compressed. S.L. 3-3 to 3-9 times the maximum depth and 3-2 to 3-6 times the length of the head. Predorsal profile convex with a slight nuchal hump in the larger individuals (c. 45 mm. S.L. and over). Snout bluntly pointed, 0-6 to 0-8 times the diameter of the eye. Mouth moderate, subterminal. Barbels moderate, the anterior 0-25 to 0-4 times the dia- meter of the eye, its tip reaching the base of the posterior barbel. Posterior barbel 0-45 to 0-6 times the diameter of the eye, the tip extending to between the verticals to the anterior margin and the centre of the pupil. Eyes inclined slightly upwards, the diameter 2-95 to 3-35 times in the length of the head and i-o to 1-15 times in the interorbital width. Well developed Beirabarbus type pit-lines on the side of the snout, the cheek, the operculum and on the dorsal surface of the head between the eyes. Pectoral fin rather long, 0-7 to 0-85 times the length of the head, the tip frequently overlapping the pelvic origin, particularly in smaller males. Last simple dorsal ray smooth, flexible, 0-9 to i-o times the length of the head. Apex of dorsal fin slightly rounded, distal margin concave. Dorsal fin rays III 7 (f.i) or III 8 (f.23). Anal fin rays III 5. The pelvic origin lies between the verticals to the last simple and first branched dorsal fin rays. Caudal peduncle slender, 1-4 to 1-8 times as long as deep. The lateral line dips to a maximum of slightly more than the depth of half a scale row below the horizontal myoseptum in the anterior third of the body ; the two converge in the vertical to the anal fin, usually over the anterior rays. Lateral BARB US OF THE VOLTA REGION 115 line scales 23 to 27 (modal number 25). Three and a half scales between the lateral line and the dorsal origin, 3^ scales between the lateral line and the mid-ventral line immediately in front of the pelvic origin and 2-2 1 scales between the lateral line and the pelvic origin. Nine (f.i), 10 (f.i8) or n (f.5) scales round the caudal peduncle ; the odd row of scales, when present always lies along the mid-ventral line (N.B. Daget records only 12 scales round the caudal peduncle in the type description). 10mm. FIG. 4. Barbus parablabes. Summary of morphometric data based on 24 fish, 39 to 52 mm. S.L. ; measure m nts are expressed as percentages of S.L. Range Mean Allometry Maximum depth . . . 25-8-29-6 27-5 Length head . . . 27-3-30-7 28-9 Diameter eye . . . 8 9-9 -7 9-3 Interorbital width . . 8-9-10-4 9-8 Length snout . . . 6 4-7 -6 6-9 Length anterior barbel . . 2 5-3 -6 2-9 Length posterior barbel . . 4-1-5-7 4-9 Length pectoral fin . . 21-2-24-7 23-0 Length dorsal fin . . . 26-8-31-2 28-0 Length caudal peduncle . . 20-0-24-0 22-0 Depth caudal peduncle . . 13-1-14-4 13-7 COLORATION : in living specimens, dorsal surface greyish green, flanks silvery, whitish below. A greyish mid-lateral band is punctuated with short dark vertical streaks on the lateral line. Caudal and dorsal fins clear rose-pink, the dorsal less brightly coloured than the caudal ; all other fins colourless. Iris reddish above. In formalin-fixed specimens, dorsal surface with a moderately heavy scattering of fine melanophores, free from a narrow zone near the margin of each scale ; the clear zone becomes increasingly wide on the flanks. Dorsal and lateral scales are outlined ZOOL. 13, 4 5 n6 A. J . & J. HOPSON with larger and darker melanophores which are more concentrated in a vertical band on the pocket of each scale to form, a regular pattern. Pocket pigment is particularly heavy on the lateral line where each scale is marked with a conspicuous crescent- shaped vertical stripe traversed by the lateral line pore. A well-marked band of diffuse melanophores, one third of a scale row in depth, runs mid-laterally from the side of the snout, across the operculum, to the tip of the caudal peduncle. The ventral surface is generally pigment free with the exception of scattered melano- phores on the pockets of the first and sometimes the second row of scales below the lateral line. The dorsal fin is relatively lightly peppered with melanophores chiefly on the second simple ray and on the distal halves of the third simple and first three branched rays. This pigment is sometimes more concentrated towards the apex of the fin but never enough to form a definite spot. Fine lines of black pigment are frequently present on the anterior and posterior edges of the fin rays particularly in the basal half of the fin ; this may produce a streaky appearance. DIAGNOSIS AND AFFINITIES : Barbus pamblabes is readily distinguished from B. macrops (with which it usually occurs in North Ghana) by the absence of a heavy black spot at the tip of the dorsal fin, the relative lightness of the lateral band and the slender crescentic markings on the pockets of the lateral line scales (broader and triangular in B. macrops) . The lateral line scales themselves are narrower and deeper in B. parablabes. The caudal peduncle is more slender in B. parablabes, and in northern Ghana the lower number of scales round the caudal peduncle (9-11 cf. 12 in B. macrops} is diagnostic. However, as noted above, 12 scales round the caudal peduncle were recorded in the type material of B. parablabes from north Dahomey (Blanc & Daget, 1957). The pectoral fin is relatively longer in B. parablabes, fre- quently overlapping the pelvic origin in the smaller males (never in B. macrops.) In living specimens the pale pink of the caudal fin is strikingly different to the bright orange red of B. macrops. Most of the above characters also serve to distinguish B. parablabes from B. ablabes. Both species are, however, without a black tip to the dorsal fin and the stripes on the lateral line scale pockets are only slightly broader in B. ablabes than in B. parablabes. The shorter posterior barbel of B. parablabes (4-9-57 cf. 6-2-8-9% in B. ablabes) is diagnostic. Compared with B. spurrelli, B. parablabes has a shorter snout (6-4-7-6 cf. 7-8-8-8% S.L.), shorter posterior barbels (4-1-5-7 cf. 5-7-7-9% S.L.) and a narrower caudal peduncle (13-1-14-4 cf . 14-3-15-1% S.L.) . The dark mid-lateral band is much heavier in B. parablabes and the vertical lines of pigment on the lateral line scales are narrower and more emphasized ; in B. spurrelli there is relatively little contrast between the pigmentation of the lateral line and more dorsal scales. DISTRIBUTION AND HABITAT : in Ghana, widespread and common throughout the northern region in rivers, streams and dams. Occurs over all substrates though rarely in thickly vegetated habitats. No records from the south or from outside the Volta basin. Elsewhere, recorded from a tributary of the River Oti (Volta basin) in north Dahomey (type locality). B A R BUS OF THE VOLTA REGION 117 NATURAL HISTORY : Barbus parablabes spawns during the rains. Gonads start to mature during April and by mid- June the majority of fish are ripe. Juveniles appear in the rivers during July and by mid- August the larger mature fish have disappeared, suggesting that death occurs after breeding, at the age of approximately one year. Females grow larger than males. Maximum size : a female of 53 mm. S.L. Barbus hypsolepsis Daget, 1959 (Text-fig. 5) Barbus hypsolepis Daget, 1959, Bull. Inst. France. Afr. Noire, 21 (A), 2 : 670, fig. i. DESCRIPTION based on 21 specimens 14-27 mm. S.L. from the Black Volta, 3 miles west of Lawra, north Ghana. Body moderately compressed. S.L. 3-65 to 4-15 times the maximum depth and 3-05 to 3-90 times the length of the head. Pre-dorsal profile smooth, convex. Snout somewhat pointed, 0-4 to 0-7 times as long as the diameter of the eye. Mouth moderate, subterminal and protrusible obliquely downwards. Weakly -developed pit-lines clearly visible on the cheeks. Anterior barbel 0-03 to 0-25 times the diameter of the eye, the tip overlapping the base of the posterior barbel only in the larger specimens. Posterior barbel 0-05 to 0*45 times the diameter of the eye, the tip extending in larger specimens beyond the vertical to the anterior margin of the pupil. Eyes large, inclined slightly upwards, 2-35 to 2-95 times in the length of the head and 0-85 to 1-05 times in the interorbital width. Pectoral fin 0-55 to 0-75 times the length of the head, the tip falling short of the pelvic origin. Last simple dorsal ray smooth, flexible, 0-9 to 1-05 times as long as the head. Apex of dorsal fin pointed, distal margin concave. Dorsal fin-rays III 8. Anal fin-rays III 5. Origin of pelvic fin below the vertical to the last simple dorsal ray. Caudal peduncle slender, 1-55 to 1-9 times as long as deep. The lateral line scales are unusually deep, particularly below the origin of the dorsal fin. The lateral line dips to a maximum of the depth of a third of a scale row below the horizontal 10mm. FIG. 5. Barbus hypsolepis. "8 A. J. & J. HOPSON myoseptum in the anterior third of the body ; the two converge in the vertical to the base of the anal fin. Lateral line scales 22 to 24 (modal number 23). Two and a half scales between the lateral line and the dorsal origin, z\ scales between the lateral line and the mid- ventral line immediately in front of the pelvic origin and \\ scales between the lateral line and the pelvic origin. Eight scales round the caudal peduncle. Summary of morphometric data based on 21 fish, 14 to 27 mm. S.L. ; measurements are expressed as percentages of S.L. Range Mean Allometry Maximum depth . . . 23 -5-27 -8 25-5 Length head . . . 25-6-32-4 28-8 Negative Diameter eye . . . 9-7-12-8 10-7 Interorbital width . . 9-7-11-0 10-4 Length snout . . . 4 7-7 i 5-8 Length anterior barbel . . 0-1-2-6 1-2 Positive Length posterior barbel . . o 6-4 -6 2-4 Positive Length pectoral fin . . 16-8-22-5 19-0 Length dorsal fin . . . 25-8-30-4 27-3 Length caudal peduncle . . 19-6-23-9 21-4 Depth caudal peduncle . . 1 1-8-14-0 12-5 COLORATION : in living specimens, body nacreous silver with a pinkish flush on the dorsal surface. Scales clearly outlined with dark pigment. Dorsal and caudal fins pale pinkish-orange, anal and paired fins tinged with the same colour. Iris with an orange-red arc above. Daget (1959) noted that in B. hypsolepis from the Middle Niger, females are less brightly coloured than males and have a greyish dorsal fin. In formalin-fixed specimens scales of the lateral line and above are clearly outlined with dark pigment which is emphasized in a slender, vertical arc of melanophores on each scale pocket. A few scattered melanophores form an indistinct band on the horizontal myoseptum, more noticeable in the posterior half of the body. There is frequently a dark bar of pigment immediately behind the operculum, running obliquely downwards from the lateral line to the base of the pectoral fin. Daget (op. cit.} has observed that in B. hypsolepis this bar is present only in males. DIAGNOSIS : these data agree closely (allowing for allometry) with the original description of material from the Middle Niger (Daget, op. cit.}. Daget includes B. hypsolepis in the subgenus Clypeobarbus which differs from other Barbus in having exceptionally deep lateral line scales, particularly in the region below the origin of the dorsal fin. He observes that when more material becomes available it may be possible to consider B. hypsolepis as a subspecies of B. pleuropholis , a species occurring in the Congo basin. DISTRIBUTION : in Ghana local and uncommon, known only from the Black Volta near Lawra and from the White Volta near Bolgatanga. Elsewhere, the Middle Niger (Daget, op. cit.). BARBUS OF THE VOLTA REGION ng Barbus nigeriensis Boulenger, 1902 (Text-fig. 6) Barbus nigeriensis Boulenger, 1902, Proc. zool. Soc. London, 2 : 327, pi. 28, fig. 3. ; Idem, 1910 (type only), Cat. Afr. Fish, 2 : 154, fig. 130. [non Barbus nigeriensis : Trewavas, 1947, in Irvine, Fisheries and Fish of the Gold Coast, London ; non Barbus nigeriensis : Blanc & Daget, 1957, Mem. Inst. France. Afr. Noire, no. 50 : 135, fig. 14.] DESCRIPTION based on 20 fish, 23 to 50 mm. S.L., from the White Volta and its tributaries near Bawku, northern Ghana. Body moderately compressed. S.L. 3-2 to 4-1 times the maximum depth and 3-2 to 4-1 times the length of the head. Pre- dorsal profile convex with a slight nuchal hump. Snout rather pointed, 0-65 to 0-9 times in the diameter of the eye. Mouth relatively large, subterminal. Barbels long, the anterior 0-65 to 1-2 times the diameter of the eye, and the posterior 0-8 to i -4 times the diameter of the eye. In larger specimens the tip of the anterior barbel reaches the vertical to the centre of the eye and the tip of the posterior barbel extends beyond the posterior margin. Eyes inclined slightly upwards, relatively large, 2-5 to 3-2 times in the length of the head and 0-95 to 1-2 times in the interorbital width. Weakly-developed pit-lines are visible on the cheek and operculum. Pectoral fin 0-55 to 0-75 times the length of the head, the tip not reaching the pelvic origin. Last simple dorsal ray smooth, flexible, 0-8 to i-o times as long as the head. Apex of dorsal fin somewhat rounded, distal margin slightly concave. Dorsal fin rays III 8 (f.ig) or III 9 (f.i). Anal fin rays III 5. The origin of the pelvic fin lies between the verticals to the first and third branched dorsal fin rays. Caudal peduncle 1-35 to 1-55 times as long as deep. The lateral line dips to a maximum of approximately one and a quarter scale rows below the horizontal myoseptum in the anterior third of the body ; the two converge in the vertical to the base of the anal fin. Lateral line scales 27 to 30 (modal number 29) . Four and a half scales between the lateral lines and the dorsal origin, 3! to 4^ scales between the lateral line and the mid- ventral line immediately in front of the pelvic origin and 2| scales between the lateral line and the pelvic origin ; 12 scales round the caudal peduncle. COLORATION : in living specimens greenish-brown above, silvery on the flanks and whitish below. A dark narrow mid-lateral band is present. Dorsal and caudal fins clear pale pink. The eye is marked with a red crescent over the pupil. In formalin- fixed specimens scales of the dorsal surface with dense melanophores, free from a narrow marginal band and less dense at the centre of each scale. This pattern, which is darkest on the scales at the base of the dorsal fin, fades out rather abruptly on the flanks. Dense melanophores lying over the horizontal myoseptum form a narrow dark band running the entire length of the body. The band is generally less than a scale row in depth, becoming slightly wider on the caudal peduncle and frequently more dense at the tip to form a vague spot. The band may be interrupted below the anterior rays of the dorsal fin but never enough to form distinct spots or streaks. The scale pockets of the lateral line are each marked with a compact group of melano- phores bisected horizontally by the lateral line pore. Similar markings may appear 120 A. J. & J. HOPSON faintly on the scale row below the lateral line, but apart from an indistinct group of melanophores at the base of the anal fin the ventral surface is generally pigment-free. DIAGNOSIS AND AFFINITIES : the original description of B. nigeriensis (op. cit.} was based on a poorly preserved holotype (Brit. Mus. (N.H.) reg. no. 1902.10.25.2) lacking any obvious pigment. We believe that Boulenger (1911) later mistook other superficially similar specimens for B. nigeriensis, resulting in his observations that the markings are a " broad black lateral band which may be broken up into a series of spots or streaks ". Such a broken band is not characteristic of B. nigeriensis and the description has tended to confuse subsequent workers. Barbus nigeriensis material in the British Museum (N.H.) thus included examples of 5. werneri Boulenger, B. lawrae and B. sublineatus. The holotype is now in poor condition. Pigment is restricted laterally to microscopic melanophores in a longitudinal line on the caudal peduncle, grouped closer together at the posterior end to suggest a spot. Allowing for features lost by deterioration, the holotype corresponds closely with our own material from northern Ghana. Three fish from the Ogun River, Western Nigeria (Brit. Mus. (N.H.) reg. no. 1956.9.6.57-59) must also be regarded as authentic B. nigeriensis ; the pigmentation is identical with that of the specimens from Ghana. Data fiom the three collections are tabulated below. Measurements are expressed as percentages of the standard length. Means are given in parentheses. Number of specimens . Standard length (mm.) Maximum depth Length head* Diameter eye* Length snout Interorbital width Anterior barbel . Posterior barbel . Length pectoral fin Length dorsal fin Length caudal peduncle Depth caudal peduncle Number lateral line scales Scales lateral line to dorsal origin . Scales lateral line to vent. mid-line . Holotype Agberi River, Ogun River, White Volta, Lower Niger S. Nigeria N. Ghana i 3 20 42 45-50 24-50 25-2 32-2-35-0 (33-2) 24-5-29-7 (26-8) 23-7 22-7-23-7 (23-4) 24-5-3I-3 (28-7) 8-5 7-7-8-0 (7-8) 7-6-11-8 (9-9) 5-7 5-6-6-0 (5-8) 6-4-8-3 (7-5) 8- if 9-3-9-8 (9-5) 9-3-11-8 (10-4) 6'4t 7-9-8-6 (8-2) 7-4-10-0 (9-1) 7 -8f io-6-ii -o (10-8) 8-5-12-0 (10-8) 17-2 17-5-19-4 (18-6) 17-3-21-4 (19-6) 22-6-22-8 (22-7) 23-0-27-9 (25-0) 20-1 17-9-20-0 (19-0) 17-8-20-5 (19-3) 12-8 13-0-13-1 (13-0) 12-6-14-7 (13-7) c. 27 28-29 27-30 4i 3i 4t 4i 4i 4i * Negative allometry ; f somewhat shrivelled. Barbus nigeriensis appears to have no close relatives. It resembles B. Morotaenia Boulenger in pigmentation and barbel length but in that species the lateral line dips less markedly below the longitudinal myoseptum and the mouth is smaller and set more obliquely, with the lips folding characteristically round the base of the posterior BARBUS OF THE VOLTA REGION 121 FIG. 6. Barbus nigeriensis. barbel in the form of a socket. The combination of relatively large eyes, large mouth, long barbels and the narrow and continuous lateral band widening on the caudal peduncle should prevent the confusion of B. nigeriensis with other species. HABITAT AND DISTRIBUTION : in Ghana, an uncommon species found in small numbers in the White Volta and its tributaries in north-eastern Ghana. All examples were from running water. Elsewhere, the Agberi River, a tributary of the Lower Niger (type locality) and the Ogun River, both in south Nigeria. Barbus lawrae Hopson, 1965. (Text-fig. 7) Barbus lawrae Hopson, 1965, Rev. Zool. Bot. Afr. 71 : 245. HOLOTYPE : a female of 42 mm. S.L. from Lissa Dam on the Kamba River (a tributary of the Black Volta), 15 miles north-east of Lawra, north-west Ghana, 25.x.6i, British Museum (N.H.) reg. no. 1964.9.8.1. DESCRIPTION based on the holotype and 23 of the paratypes, 32-47 mm. S.L., from Lissa Dam. Body compressed, particularly near the dorsal surface. S.L. 2-95 to 3-65 times the maximum depth and 3-4 to 3-75 times the length of the head. Pre- dorsal profile convex with a slight nuchal hump. Snout rather pointed, 0-55 to 0-75 times the diameter of the eye. Mouth moderate, subterminal. Anterior barbel o-i to 0-4 times the diameter of the eye, the tip barely overlapping the base of the posterior barbel. Posterior barbel 0-3 to 0-8 times the diameter of the eye, the tip reaching the vertical to the posterior margin of the pupil in the largest individuals. Eyes inclined slightly upwards, 2-75 to 3-1 times in the length of the head and i-o to i-i times in the interorbital width. Weakly-developed pit-lines visible on the cheeks. Pectoral fin 0-7 to 0-8 times as long as the head, the tip falling short of the pelvic origin. Last simple dorsal ray smooth, flexible, rather long, i-o to 1-2 times the length of the head. Apex of the dorsal fin slightly rounded, distalmargin concave. Dorsal fin rays III 8. Anal fin rays III 5. The origin of the pelvic fin lies between 122 A. J. & J. HOPSON the verticals to the first and second branched dorsal rays. Caudal peduncle 1-3 to 1-6 times as long as deep. The lateral line dips to a maximum of the depth of one and a third scale rows below the horizontal myoseptum in the anterior third of the body ; the two converge in the centre of the caudal peduncle. Scales radially striated. Lateral line scales 28 to 31 (modal number 30). Four and a half to 5^ scales between the lateral line and the dorsal origin, 4^ scales between the lateral line and the mid-ventral line immediately in front of the pelvic origin and 2\ scales between the lateral line and the pelvic origin. Twelve scales round the caudal peduncle. 10mm. FIG. 7. Barbus lawrae. Summary of morphometric data based on 24 fish, 32 to 47 mm. S.L. ; measurements are expressed as percentages of S.L. Maximum depth . Length head Diameter eye Interorbital width Length snout Length anterior barbel . Length posterior barbel . Length pectoral fin Length dorsal fin . Length caudal peduncle . Depth caudal peduncle . COLORATION : in living specimens brownish above, silvery on the flanks and whitish below. A dark mid-lateral band, often broken up into spots or streaks is present. No conspicuous colour was noted on the fins. In formalin-fixed specimens Range Mean Allometry 27-2-33-3 26-6-29-3 9-1-10-0 30-3 27-6 9-7 Negative Negative 9-4-10-9 5-7-7-4 IO-I 6-5 1-2-3-9 2-2 3-3-7-8 5'5 19-5-22-4 28-4-31-6 20-8 29-5 19-3-24-0 I3-9-I5-8 21 -O 14-6 . BARBUS OF THE VOLTA REGION 123 each scale of the dorsal surface is marked with a broad band of fine melanophores, free from a narrow marginal zone and from the centre of each scale. These markings fade away quickly towards the flanks and a broad zone above the mid-lateral band is more or less pigment-free. The mid-lateral band consists of heavy pigment, approximately one third of a scale row in depth, originating at the level of the sixth or seventh lateral line scale and running along the horizontal myoseptum to the tip of the caudal peduncle. The band is frequently broken up into a series of rounded spots or streaks, connected by more lightly scattered pigment. Spots, when present, are irregular in arrangement and number ; only the first, at the level of the sixth or seventh lateral line scale, and the last, on the tip of the caudal peduncle, are fixed in position. A few scattered melanophores lie on the horizontal myoseptum anterior to the mid-lateral band. Each pocket of the anterior lateral line scales may bear a small group of melanophores, divided horizontally into two by the lateral line pore. There are several indistinct spots of black pigment at the base of the anal fin. Relatively dense pigment is present on the posterior margin of the second simple dorsal ray and on the distal half of the third simple ray. Scattered melanophores are present on the distal parts of all branched dorsal rays. DIAGNOSIS AND AFFINITIES : material identified as B. nigeriensis from the R. Katagum (Chad basin), Northern Nigeria in the British Museum (N.H.) reg. no. 1952.4.28.136-144, is now referred to B. lawrae. The Nigerian specimens differ from the types in having slightly fewer lateral line scales (28 to 31, mode 29 cf. 29 to 31, mode 30) but are in other respects identical. Barbus lawrae most closely resembles B. werneri, sewsw Daget, 1954 (from the Middle Niger) but is more compressed laterally, has more lateral line scales (28 to 31, mode 29 and 30 cf. 25 to 29, mode 26 and 27) and much shorter barbels ; in B. lawrae the anterior barbel barely extends beyond the base of the posterior barbel, whereas in B. werneri, sensu Daget, the tip reaches the anterior margin of the pupil. Furthermore, a heavier and more pronounced spot marks the tip of the caudal peduncle in B. werneri, sensu Daget. During the present studies, specimens of B. werneri, sensu Daget, 1954 from the Middle Niger (Paris Museum reg. no. 51-246) were compared with a syntype of B. werneri Boulenger from Fashoda on the White Nile (Brit. Mus. (N.H.) reg. no. 1907.12.2.1333). They were found to be a distinct species. Barbus werneri BouJenger is easily distinguishable from 5. werneri, sensu Daget, and from B. lawrae by the small terminal mouth, set very obliquely with the angle in front of the vertical to the nostril. In B. werneri, sensu Daget, and in B. lawrae the mouth is subterminal and set more horizontally with the angle overlapping the vertical to the nostril. Barbus werneri Boulenger is unusual in the form of the band of scattered melano- phores on the line of the mid-lateral row of spots ; the band veers upwards on the caudal peduncle to end slightly above the terminal spot. In B. werneri, sensu Daget, and B. lawrae the band merges horizontally with the terminal spot. Barbus lawrae closely resembles B. lepidus in form and pigment pattern. The mouth of B. lepidus is however more terminal, the barbels relatively longer and the body less compressed laterally. In B. lepidus the lateral line dips further below the 124 A. J. & J. HOPSON horizontal myoseptum and the two converge near the tip of the caudal peduncle (cf . the centre of the caudal peduncle in B. lawrae}. The lateral spots of B. kpidus, although variable in number, are always rounded and discrete ; spots when present in B. lawrae are irregular and fragmented. The terminal spot in B. lawrae is never as pronounced as in B. lepidus. Morphometrically, B. lawrae is also similar to B. perince with which it occurs in rivers of the Chad basin in Northern Nigeria (pers. record). Barbus perince is easily distinguished from B. lawrae by the three rounded, black, mid-lateral spots, one in the anterior third of the body, one below the last dorsal ray and one on the end of the caudal peduncle. Barbus perince also has longer barbels than B. lawrae (the anterior extending to the margin of the eye in B. perince) and the lateral line and horizontal myoseptum converge near the tip of the caudal peduncle (cf. the centre in B. lawrae}. DISTRIBUTION : in Ghana, probably rare and local ; known only from the type locality in the Volta basin near Lawra. Elsewhere, rivers of the Chad basin in Northern Nigeria (pers. records). Barbus subinensis Hopson, 1965 (Text-fig. 8) Barbus subinensis Hopson, 1965, Rev. Zool. Bot. Afr. 71: 249. HOLOTYPE : a female of 33 mm. S.L. collected from the River Subin (Prah basin), Juaso, south Ghana, g.xi.Gi, British Museum (N.H.) reg. no. 1964.9.8.256. DESCRIPTION based on the holotype and 19 paratypes, 25-34 mm - S.L. from the River Subin, Juaso. Body somewhat compressed. S.L. 3-1 to 3-9 times the maximum depth and 3-4 to 4-1 times the length of the head. Predorsal profile smooth, convex, becoming abruptly steep on the snout. Snout 0-6 to 0-75 times the diameter of the eye. Mouth moderate, subterminal. Barbels small, the anterior o-i to 0-3 times and the posterior 0-2 to 0-4 times the diameter of the eye. The tip of the anterior barbel falls short of the base of the posterior barbel which extends posteriorly slightly beyond the vertical to the anterior margin of the pupil. Eyes lateral 2-65 to 3-05 times in the length of the head and 1-03 to 1-16 times in the interorbital width. Weakly-developed pit-lines are visible on the cheeks and oper- culum. Pectoral fin 0-7 to 0-85 times the length of the head, the tip overlapping the pelvic origin in three smaller individuals. Last simple dorsal ray smooth flexible, 0-95 to i-i times the length of the head. Apex of dorsal fin somewhat pointed, distal margin slightly concave. Dorsal fin rays III 8 ; anal fin rays III 5. The pelvic origin lies on the vertical to the last simple dorsal ray. Caudal peduncle slender, 1-45 to 1-9 times as long as deep. The lateral line dips to a maximum depth of one and a quarter scale rows below the horizontal myoseptum in the anterior third of the body ; the two converge in the anterior half of the caudal peduncle. Scales radially striated, 27 to 30 (modal number 28) in the lateral line. Five and a half scales between the lateral l ; ne and the dorsal origin, 4 to 4^ scales between the lateral line and the mid-ventral line immediately in front of the pelvic origin and 2\ between BARBUS OF THE VOLTA REGION 125 I Omm. FIG. 8. Barbus subinensis. the lateral line and the pelvic origin. Nine (f.i), 10 (f.5), n (f.i2) or 12 (f.2) scales round the caudal peduncle. Summary of morphometric data based on 20 fish, 25 to 34 mm. S.L. ; measurements are expressed as percentages of S.L. Maximum depth . Length head Diameter eye Interorbital width Length snout Length anterior barbel . Length posterior barbel . Length pectoral fin Length dorsal fin . Length caudal peduncle . Depth caudal peduncle . COLORATION : in formalin-fixed specimens a band of melanophores close to the margin of each scale forms a regular pattern over the dorsal surface of the body, fading abruptly on the flanks. A narrow continuous mid-lateral band of black pig- ment is always present. The band, which originates on the operculum runs below the myoseptum in the anterior half of the body ; the band and the myoseptum converge between the verticals to the last dorsal ray and the anal origin and run together onto the caudal peduncle. The ventral surface is pigment-free with the exception of a group of melanophores at the base of the anal fin which continues backwards as a narrow stripe along the ventral surface of the caudal peduncle. Snout with a con- spicuous dark band running from eye to eye round the tip (cf. B. punctitaeniatus). Scattered melanophores noticeable on the anterior rays of the dorsal fin. Range Mean Allometry 25-6-32-2 28-9 24 -0-29 -2 27-1 Negative 8-5-10-4 9-4 Q-6-II -2 10-3 5-9-7-4 6-4 1-2-3.0 2-1 2-0-3-9 3-o 19-4-24-6 21-3 25-0-30-0 27-0 21 -O-25 -O 22-5 12 -4-15-2 13-9 126 A. J . & J. HOPSON DISTRIBUTION : known only from the River Subin (Prah basin), near Juaso, south Ghana. REMARKS : Barbus subinensis appears to have no close relatives. The lateral band and the markings on the snout recall B. boboi Schultz but the types of that species bear a large black spot on the caudal peduncle, have considerably longer barbels (anterior barbels 11-2-12-8 cf. 1-2-3-0% S.L. ; posterior barbels 11-8-13-0 cf. 2-0-3-9% S.L.) and only 4! scales above the lateral line compared with 5^ in B. subinensis. Barbus trispilus (Bleeker, 1863) (Text-fig. 9) Puntius (Barbodes) trispilos Bleeker, 1863, Nat. Verh. Wet. Haarlam, 23 : 113, pi. 23, fig. 3. Barbus trispilus : Boulenger, 1910, Cat. Afr. Fish. 2 : 163, fig. 141. Barbus trispilus : Daget, 1952, Mem. Inst. France. Afr. Noire, no. 19 : 314, fig. 4. DESCRIPTION based on 24 fish, 27-65 mm. S.L. from the River Weiwei (Prah Basin), Kumasi, south Ghana. Body slightly compressed. S.L. 3-0 to 3-75 times the maximum depth and 3-3 to 3-85 times the length of the head. Predorsal profile convex with a slight nuchal hump in larger individuals (over c. 40 mm. S.L.). Snout rounded, 0-65 to 0-9 times the diameter of the eye. Mouth moderate, subterminal. Barbels relatively long, the anterior 1-05 to 1-55 times the diameter of the eye, the tip usually extending beyond the vertical to the posterior margin of the eye . Posterior barbel 1-5 to 1-8 times the diameter of the eye, the tip extending beyond the posterior margin of the preoperculum. Eyes inclined slightly upwards, the diameter 2-9 to 3-6 times in the length of the head and 1-15 to 1-65 times in the interorbital width. Weakly-developed pit-lines present on the cheek and operculum. Pectoral fin 0-6 to 0-85 times the length of the head, the tip falling short of the pelvic origin Last simple dorsal ray smooth, flexible, 0-75 to 0-95 times the length of the head. Apex of dorsal fin slightly rounded, distal margin concave. Dorsal fin rays III 8. Anal I 10mm. i FIG. 9. Barbus trispilus. BARBUS OF THE VOLTA REGION 127 fin rays III 5. The pelvic origin lies between the verticals to the last simple and first branched dorsal rays. Caudal peduncle 1-2 to 1-5 times as long as deep. The lateral line dips to a maximum of the depth of slightly over a scale row below the horizontal myoseptum in the anterior third of the body ; the two converge in the first third of the caudal peduncle. Lateral line scales 26 to 29 (modal number 28). Four and a half scales between the lateral line and the dorsal origin, 4^ to 5| scales between the lateral line and the mid-ventral line, immediately in front of the pelvic origin and 2| scales between the lateral line and the pelvic origin. Twelve scales round the caudal peduncle. Summary of morphometric data based on 24 fish, 27 to 65 mm. S.L. ; measurements are expressed as percentages of S.L. Maximum depth . Length head Diameter eye Interorbital width Length snout Length anterior barbel . Length posterior barbel Length pectoral fin Length dorsal fin . Length caudal peduncle . Depth caudal peduncle . COLORATION : in formalin-fixed specimens, dorsal surface with a dense peppering of fine melanophores, more scattered towards the centre and free from a narrow zone near the margin of each scale. The lateral line scale pockets are each marked with a small, dark, triangular patch bisected horizontally by the lateral line pore. Similar but lighter markings are present on most scale rows above the lateral line and on the anterior and posterior scales of the row below. Three round, black, mid-lateral spots are always present, the first at the level of the sixth or seventh lateral line scale, the second immediately behind the vertical to the last dorsal ray and the third on the tip of the caudal peduncle. The second spot is slightly larger than the first and last spots which are approximately the depth of a scale row in diameter. Slight elonga- tion of the second and third spots was noted in a few specimens. Scattered melano- phores form a faint band over the horizontal myoseptum between the spots. The base of the anal fin is marked with a few indistinct melanophores. DIAGNOSIS : these data are in close agreement with previous descriptions of the species. Barbus trispilus is superficially similar to B. perince (from the Nile and Chad basins) which is also marked with three round mid-lateral spots. Barbus trispilus has, however, fewer lateral line scales (26-29 cf. 28-32 in B. perince), a lateral line which dips less markedly below the horizontal myoseptum (i cf. i| scale rows) and relatively longer barbels (the anterior reaching the posterior margin of the eye cf. the anterior margin in B. perince). The lateral spots are relatively larger in B. trispilus. Range Mean Allometry 26-2-32 -9 29-3 25-8-30-0 27-7 Negative 7-4-10-0 8-8 Negative IO-2-I2 -O 10-8 5-6-7-4 6-6 9-8-12-7 ii i 11-7-15-4 13-8 18-5-23-4 20-5 20-4-26-4 24-1 17-4-22-6 20 -o 13-6-15-6 14-8 128 A. J . & J . HUPSON DISTRIBUTION : in Ghana, widespread and common in forest streams and rivers (notably the Prah basin) in the south-west. Probably absent from the Volta. Elsewhere Dabo Crom, Guinea (type locality) ; Sierra Leone (Norman, 1932) ; Bolor River, Liberia (Schultz , 1942) ; Mt. Nimba, Rep. Guinea (Daget, 1952). Barbus sublineatus Daget, 1954 (Text-fig. 10) Barbus sublineatus Daget, 1954, Mem. Inst. France. Afr. Noire, no. 36 : 203, fig. 69. Barbus sublineatus : Blanc & Daget, 1957, Mem. Inst. France. Afr. Noire, no. 50 : 137, figs. 15 and 1 6. DESCRIPTION based on 24 fish 28-40 mm. S.L. from a tributary of the White Volta, 5 miles north of Bawku, north Ghana. Body slightly compressed. S.L. 3-8 to 4-4 times the maximum depth and 3-4 to 3-85 times the length of the head. Predorsal profile convex with a slight nuchal hump. Snout rather pointed 0-6 to 0-8 times as long as the diameter of the eye. Mouth moderate, sub-terminal. Barbels relatively long, the anterior 0-5 to 0-95 and the posterior 0-75 to 1-12 times the diameter of the eye. The tip of the anterior barbel normally extends beyond the centre and the posterior barbel beyond the posterior margin of the eye. Eyes inclined slightly upwards, relatively large, 2-65 to 3-1 times in the length of the head and 1-03 to 1-17 times in the interorbital width. Weakly-developed pit-lines are present on the cheek and operculum. Pectoral fin 0-65 to 0-75 times the length of the head, the tip falling short of the pelvic origin. Last simple dorsal ray smooth, flexible, 0-75 to 0-95 times as long as the head. Apex of dorsal fin slightly rounded, distal margin concave. Dorsal fin rays III 8 ; anal fin rays III 5. The origin of the pelvic fin lies between the verticals to the first and second branched dorsal rays. Caudal peduncle 1-35 to 1-7 times as long as deep. The lateral line dips to a maximum lOmm. FIG. 10. Barbus sublineatus. BARB US OF THE VOLTA REGION 129 of approximately the depth of one scale row below the horizontal myoseptum in the anterior third of the body ; the two converge in the vertical to the posterior anal rays or in the anterior third of the caudal peduncle. Lateral line scales 27 to 31 (modal number 29) . Four and a half scales between the lateral line and the dorsal origin, 4! scales between the lateral line and the mid-ventral line immediately in front of the pelvic origin and 2^ scales between the lateral line and the pelvic origin. Twelve scales round the caudal peduncle. Summary of morphometric data based on 24 fish, 28 to 40 mm. S.L. ; measurements are expressed as percentages of S.L. Range Mean Allometry Maximum depth . . . 22-5-26-3 24-5 Length head . . . 25 -8-29 -2 27-4 Diameter eye . . . 8-7-10-3 9-5 Negative Interorbital width . . 9 -7-11 -5 10-5 Length snout . . . 5 8-7 o 6-5 Length anterior barbel . 5 3-9 i 7-3 Length posterior barbel . 7 8-1 1-2 9-2 Length pectoral fin . . 18-7-20-3 19-5 Length dorsal fin . . . 22-5-25-3 23-9 Length caudal peduncle . 19-0-22-4 20-7 Depth caudal peduncle . . 12-0-14-3 13-8 COLORATION : in living specimens, yellowish fawn above, flanks silvery, belly whitish, the body marked laterally with a row of four or more dark spots. Rays of caudal fin with pale pink chromatophores, yellowish nearer the base ; dorsal and pectoral rays with pale yellowish chromatophores ; pelvic and anal fins clear. The iris is marked dorsally with an orange-red crescent. In formalin-fixed specimens dorsal surface finely peppered with melanophores free from the margin and thinning out towards the centre of each scale; the pattern fades out on more lateral rows. A basic series of four black rounded spots lies over the horizontal myoseptum, the first at the level of the seventh lateral line scale, the second below the last dorsal ray, the third starting above the last anal ray and the fourth on the tip of the caudal peduncle. The last three are usually slightly elongated. Additional spots up to four in number may occur, usually between the second and the fourth spot. Occasionally two or more of the posterior spots may coalesce to form a band. Each lateral line scale is marked on the pocket with a small compact group of melanophores. Similar pigment though less heavy is present on scales above the lateral line. The ventral surface is generally pigment-tree with the exception of a faint dark spot at the base of the anal fin continuing backwards as a slender line along the ventral surface of the caudal peduncle. DIAGNOSIS AND AFFINITIES : Blanc & Daget (1959) note variation in the pigmenta- tion of B. sublineatus. Type specimens from the Middle Niger are marked with a dark continuous mid-lateral band. The band is represented in more recent material from the Volta and from the Haute Comoe (Blanc & Daget, 1957) by a series of spots comparable with the markings of the present material from Ghana. In other 130 A. J. & J . HOPSON respects our data agree closely (with allowances for allometry) with the type descrip- tion. Barbus sublineatus is closely related to B. lineomaculatus Blgr. which occurs in East and South Africa. Barbus lineomaculatus is similarly marked with a series of lateral spots but never as few as four. In this species most of the first spot lies above the horizontal myoseptum whereas in B. sublineatus it is situated chiefly below. Barbus sublineatus also has relatively larger eyes, a narrower interorbital space and the dorsal fin is inserted more posteriorly. HABITAT AND DISTRIBUTION : in Ghana, widespread in the north, B. sublineatus occurs sparingly in most of the rivers and larger streams over a substrate of gravel or sand. It is apparently absent from lagoons, pools and dams. Elsewhere, recorded from the Haute Comoe (Blanc & Daget, 1957) and the Middle Niger (Daget, 1954). LIFE HISTORY : the occurrence of ripe females in June indicates that, as in other Barbus, spawning takes place during the rainy season. Females grow larger than males. Maximum size a female of 74 mm. S.L. Barbus macinensis Daget, 1954 (Text-fig, n) Barbus macinensis Daget, 1954, Mem. Inst, France. Afr. Noire, no 36 : 200, fig. 67. Barbus macinensis : Blanc & Daget, 1957, Mem. Inst. France. Afr. Noire, no. 50 : 137. DESCRIPTION based on 25 fish 26-32 mm. S.L. from the River Nahau, 5 miles north of Bawku, northern Ghana. Body slightly compressed. S.L. 3-75 to 4-6 times the maximum depth and 3-55 to 4-05 times the length of the head. Predorsal profile well humped; snout rounded, 0-65 to 0-8 times the diameter of the eye. Mouth moderate, subterminal. Anterior barbel 0-65 to 0-8 times the diameter of the eye, its tip reaching the vertical to the anterior half of the pupil. Posterior barbel 0-8 to i-2 times the diameter of the eye, the tip usually extending beyond the vertical to the posterior margin. Eyes inclined slightly upwards, diameter 2-85 to 3-55 times in the length of the head and 1-2 to 1-35 times in the interorbital width. Head with weakly-developed pit-lines on the cheek and operculum. Pectoral fin 0-65 to 0-8 times the length of the head, its tip falling short of the pelvic origin. Last simple dorsal ray smooth, flexible, 0-8 to i-o times as long as the head. Apex of dorsal fin rounded, distal margin barely concave. Dorsal fin rays III 8 ; anal fin rays III 5. The origin of the pelvic fin lies between the vertical to the ist and 2nd simple rays. Caudal peduncle rather stout, 1-15 to 1-5 times as long as deep. The lateral line dips to a maximum of slightly less than a scale row's depth below the horizontal myoseptum in the anterior third of the body. The two converge in the anterior half of the caudal peduncle. Lateral line scales 24 to 27 (modal numbers 25 and 26). Three and a half scales between the lateral line and the dorsal origin, 4^ scales bet- ween the lateral line and the mid-ventral line immediately in front of the pelvic origin and 2\ scales between the lateral line and the pelvic origin. Ten (f.i), n (f.i5) or 12 (f.g) scales round the caudal peduncle. BARBUS OF THE VOLTA REGION 10mm. Range Mean Allometry 24-7-27-8 26-2 21 -4-26-6 24-6 7-8-9-2 8-4 Negative 10-3-11-8 10-9 Negative 5-7-6-7 6-2 3-5-5-9 4-9 6-8-10-3 8-4 17-9-22-9 19-9 22-4-26-3 24-3 18-2-21-8 2O -2 14-4-15-9 I5-I FIG. ii. Barbus macinensis. Summary of morphometric data based on 25 fish, 26 to 32 mm. S.L. ; measurements are expressed as percentages of S.L. Maximum depth . Length head Diameter eye Interorbital width Length snout Length anterior barbel . Length posterior barbel . Length pectoral fin Length dorsal fin . Length caudal peduncle . Depth caudal peduncle . COLORATION : in living specimens, dark brownish-yellow above, silvery on the sides, pearly white below. Body marked dorsolaterally with three dark longitudinal stripes. Three lateral spots are present and the lateral line bears a series of small black dots. Basal two-thirds of the caudal fin, the dorsal, pectoral and pelvic fins pale straw-yellow. Distal third of the caudal and the anal fin clear. The iris is marked dorsal] y with an orange-red cresent. In formalin-fixed specimens body marked dorsolaterally on each side with three dark bands running along the over- lapping portion of adjacent scale rows and formed by a concentration of melanophores in the upper and lower third of each scale. The two upper bands extend along the entire length of their respective scale rows but the lowest, lying between the first and second rows of scales above the lateral line, ends in front of the vertical to the dorsal fin. The lowest band is frequently ill-defined and sometimes absent. One or two round dark spots are often present on the dorsal mid-line between the head and dorsal fin. Each scale of the lateral line, including the two overlapping the base of the caudal fin, is marked with a small but heavy concentration of melanophores, usually divided horizontally into two by the later? Uine pore. Three conspicuous black spots ZOOL. 13. 4 6 132 A. J. & J. HOPSON are constantly present on the horizontal myoseptum, the first at the level of the 7th lateral line scale, the second below the last dorsal fin ray and the third at the tip of the caudal peduncle. The second spot is somewhat elongated. Faint melanophores occasionally form a noticeable band between the second and the last spots. Scales of the row below the lateral line are often outlined with fine pigment and there is always a faint subepidermal group of melanophores at the base of the anal fin. The remainder of the ventral surface is pigment-free. DIAGNOSIS AND AFFINITIES : stumpy and thick-set in appearance, B. macinensis should be easily distinguishable from other voltaic species by the characteristic pattern of pigmentation. Our data are in close accordance with the original descrip- tion (Daget, 1954) based on specimens from the Middle Niger. Barbus macinensis has close affinities with the morphometricaUy similar B. niokoloensis Daget, 1959 from the Gambia River, which is also marked longitudinally with dark dorsolateral bands. That species, however, lacks the three lateral spots of B. macinensis and the lateral line scales bear a dark stripe, extending forward onto the operculum. Barbus macinensis is superficially similar to B. neefi Greenwood (1962) from the Upper Zambesi which is also marked with dorsolateral stripes and lateral spots. Barbus neefi, however, has more than three spots and in addition, horizontal stripes below the lateral line. Compared with B. neefi, B. macinensis has fewer scales between the lateral line and the dorsal origin (3^ cf. 4^-5^), shorter barbels (anterior 3'5-5'9 cf - 6-0-10-3% S.L. ; posterior 6-8-10-3 cf. 10-0-14-8% S.L. in B. neefi} and a stouter caudal peduncle. A comparison is made between B. macinensis and B. voltae in the description of that species. HABITAT AND DISTRIBUTION : in Ghana, B. macinensis is present in suitable localities throughout the northern regions. The distribution suggests a preference for coarse substrates and the species is particularly abundant in the sandy pools of dry season river beds. Dams are sometimes colonized if there is a sandy or gravelly bottom but the species is generally absent over muddy substrates and from areas of thick vegetation. Elsewhere, recorded from the Haute Comoe (Blanc & Daget, 1957) and from the Middle Niger (type locality). Blanc & Daget (op. cit.} note this species in the Volta basin, Haute Volta. LIFE HISTORY : Barbus macinensis breeds during the rainy season. Ripe and ripening females were observed between May and August and the first juvenile stages appeared in the rivers during July. Growth seems to be more rapid in females than in males. All fish ripen at the end of their first year and few if any survive to spawn a second time. Maximum size, a female of 5-9 cm. S.L. Barbus voltae Hopson, 1965 (Text-fig. 12) Barbus voltae Hopson, 1965, Rev. Zool. Bot. Afr. 71: 251 Barbus nigeriensis : Blanc & Daget, 1957, Mem. Inst. France. Afr. Noire, no. 50 : 135, fig. 14. HOLOTYPE : a female of 31 mm. S.L. from the River Nahau (tributary of the BARB US OF THE VOLTA REGION 133 White Volta), 5 miles north of Bawku, northern Ghana, 27.xi.6i, British Museum (N.H.) reg. no. 1964.9.8.73. DESCRIPTION based on the holotype and 23 paratypes, 27 to 33 mm. S.L. from the River Nahau, near Bawku. Unfortunately we have been unable to examine the material identified by Blanc and Daget (op. cit.} as B. nigeriensis. Body somewhat compressed. S.L. 3-3 to 3-7 times the maximum depth and 3-25 to 3-7 times the length of the head. Predorsal profile smooth, convex. Snout bluntly pointed, 0-55 to 0-7 times the diameter of the eye. Mouth small, subterminal, protrusible and opening obliquely downwards. Anterior barbel o-i to 0-25 times the diameter of the eye, the tip falling short of the base of the posterior barbel. Posterior barbel 0-15 to 0-4 times the diameter of the eye, the tip extending to between the verticals to the anterior margin of the eye and the anterior margin of the pupil. Eyes lateral, 3-35 to 3-7 times in the length of the head and i-o to 1-25 times in the interorbital width. Weakly-developed pit-lines present on the cheek and operculum. Pectoral fin 0-55 to 0-8 times the length of the head, the tip falling short of the pelvic origin. Last simple dorsal ray, smooth, flexible, 0-9 to 1-05 times as long as the head. Apex of dorsal fin slightly rounded, distal margin concave. Dorsal fin rays III 7 (f.i) or III 8 (f.23). Anal fin rays III 5. The origin of the pelvic fin lies in the vertical to the first branched dorsal ray. Caudal peduncle 1-35 to 1-55 times as long as deep. The lateral line dips to a maximum of slightly less than the depth of a scale row below the horizontal myoseptum in the anterior third of the body ; the two converge in the anterior third of the caudal peduncle. Scales radially striated, 25 to 29 (modal number 27) in the lateral line. Three and a half scales between the lateral line and the dorsal origin, 3! scales between the lateral line and the mid-ventral line immedi- ately in front of the pelvic origin, and 2 to 2\ scales between the lateral line and the pelvic origin. Eleven (f.i) or 12 (f.23) scales round the caudal peduncle. lOmm. FIG. 12. Barbus voltae. ZOOL. 13, 4 134 A. J.&J.HOPSON Summary of morphometric data based on 24 fish, 27 to 33 mm. S.L. ; measurements are expressed as percentages of S.L. Range Mean Allomctry Maximum depth . . . 26 -8-29 -6 27-8 Length head . . . 26-7-30-6 28-2 Negative Diameter eye . . . 9-4-11-1 10-0 Negative Interorbital width . 9 6-1 2-9 11-2 Length snout . . . 5-8-7-3 6-4 Length anterior barbel . . i 0-2 -6 i 8 Length posterior barbel . . i 6-3 -7 2-6 Length pectoral fin . . 16-5-22-5 19-8 Length dorsal fin . . . 25-3-31-1 27-7 Length caudal peduncle. . 20-0-25-0 21-5 Depth caudal peduncle . . 13-6-15-5 14-5 COLORATION : in living specimens brownish above, brilliant silver on the flanks, whitish below. The body is marked mid-laterally with three small black spots ; a fourth spot is present at the base of the anal fin. Caudal fin sandy-fawn, dorsal fin pinkish-fawn, base of pectoral fin rays pale yellow. The remaining fins are colour- less. Iris, red above. In formalin-fixed specimens the dorsal surface is peppered with fine melanophores, free from the centre and margin of each scale. The pigment is heaviest on the scales at the base of the dorsal fin and fades away on the second row of scales above the lateral line. A moderately heavy band of melanophores, approximately half a scale row in depth, runs along the horizontal myoseptum from the operculum to end in a small, round, black spot on the tip of the caudal peduncle. Two other mid-lateral spots are always present. They are small, usually slightly elongated and lie within the lateral band, the first at the level of the sixth or seventh lateral line scale and the second below the last dorsal fin ray. Although the band varies in intensity and is sometimes broken into a series of irregular streaks, the three spots are always discernible. Each lateral line scale is marked on the pocket with a small but conspicuous group of melanophores, bisected horizontally by the lateral line pore. Similar but lighter markings often appear on the first and second row of scales above the lateral line. A conspicuous black spot is always present on the body at the base of the anal fin. DIAGNOSIS AND AFFINITIES : Barbus voltae is the species described by Blanc & Daget (1957) as B. nigeriensis. Barbus voltae differs from B. nigeriensis Boulenger in a number of important details : fewer scales between the lateral line and the dorsal origin (3^- cf. 4! in B. nigeriensis), considerably shorter barbels (anterior 1-0-2-6 cf. 6-4-10-0% S.L. ; posterior 1-6-3-7 cf- 7'8-i2-o% S.L. in B. nigeriensis), smaller eyes and a much smaller mouth. Furthermore, B. nigeriensis Blgr. has no trace of lateral spots which are always present in B. voltae. Barbus voltae is superficially similar to B. macinensis with which it frequently occurs in the Volta basin. It is, however, easily distinguished by the smooth pre- dorsal profile (humped in B. macinensis), shorter barbels (anterior i- 0-2-6 cf. 3-5- 5-9% S.L. ; posterior 1-6-3-7 c ^- 6-8-10-3% S.L.), a smaller mouth, larger eyes BARB US OF THE VOLT A REGION 135 (9-4-11-1 cf. 7-8-9-2% S.L. in B. macinensis} and a more obvious spot at the base of the anal fin. Barbus voltae also lacks the dorsolateral stripes of B. macinensis. HABITAT AND DISTRIBUTION : in Ghana, common in rivers, streams and dams throughout the northern regions. Infrequent in well-vegetated situations. Barbus voltae is more plentiful over muddy than over sandy bottoms, indicating a preference for finer substrates. Elsewhere, recorded only from the Black Volta and its tribu- taries in Haute Volta (Blanc & Daget, 1957). Barbus atakorensis Blanc & Daget, 1957 (Text-fig. 13) Barbus atakorensis Blanc & Daget, 1957, Mem. Inst. France. Afr. Noire, no. 50 : 140, fig. 17. DESCRIPTION based on 21 fish 27-36 mm. S.L., from Dahwenia Dam, 6 miles east of Tema, south Ghana. Body moderately compressed. S.L. 3-3 to 3-9 times the maximum depth and 3-5 to 4-0 times the length of the head. Predorsal profile convex with a slight nuchal hump. Snout somewhat pointed, 0-7 to 0-95 times as long as the eye diameter. Mouth moderate, subterminal. Anterior barbel 0-25 to 0-5 times the diameter of the eye, the tip extending beyond the base of the posterior barbel. Posterior barbel 0-6 to 0-95 times the diameter of the eye, the tip extending to between the verticals to the posterior margin of the pupil and the posterior margin of the eye. Eyes lateral, the diameter 3-2 to 3-6 times in the length of the head and 1-2 to 1-5 times in the interorbital width. Weakly-developed pit-lines visible on the cheeks and operculum. Pectoral fin 0-6 to 0-8 times as long as the head, the tip falling short of the pelvic origin. Last simple dorsal ray smooth, flexible, 0-8 to 0-95 times the length of the head. Apex of dorsal fin rounded, distal margin slightly concave. Dorsal fin rays III 7. Anal fin rays III 5. The origin of the pelvic fin 10mm. FIG. 13. Barbus atakorensis. 136 A. J. & J. HOPSON lies between the vertical to the first and second branched dorsal rays. Caudal peduncle relatively slender, 1-4 to 1-8 times as long as deep. The lateral line dips to a maximum of the depth of one and a half rows below the horizontal myoseptum in the anterior third of the body ; the two converge in the anterior third of the caudal peduncle. Lateral line scales 27 to 32 (modal number 28). Four and a half to 5 1 scales between the lateral line and the dorsal origin, 4! to 5| scales between the lateral line and the mid-ventral line immediately in front of the pelvic origin, and 2! to 3 scales between the lateral line and the pelvic origin ; 12 scales round the caudal peduncle. Summary of morphomdric data based on 21 fish, 27 to 36 mm. S.L. ; measurements are expressed as percentages of S.L. Range Mean Allometry Maximum depth . . . 26 -5-30 -5 28-2 Length head . . . 24-9-27-5 26-0 Negative Diameter eye . . . 6 9-8 -5 7-5 Negative Interorbital width . . 9-7-10-9 10-3 Length snout . . . 5 4-6 -7 6-2 Length anterior barbel . . 2-1-3-3 2 '7 Length posterior barbel . 5 0-6 -8 5-7 Length pectoral fin . . 17-8-20-4 19-3 Length dorsal fin . . . 20-9-24-5 23-3 Length caudal peduncle . 20-9-24-0 21-8 Depth caudal peduncle . . 13-2-15-0 14-1 COLORATION : in living specimens body greyish green, dark above, paler on the sides and whitish on the belly. Fins colourless. The colour pattern is best seen in formalin-fixed specimens where the dorsal surface is peppered with fine melanophores, less heavy in the centre of each scale and absent from the free margin. The pig- mentation becomes lighter on the flanks and fades out on the second row of scales above the lateral line. Pockets of each lateral line scale are marked with a small group of dense melanophores split horizontally into two by the lateral line pore. Similar markings are repeated on the pockets of the first and sometimes the second row of scales above the lateral line. The horizontal myoseptum, appearing as a fine black line, is punctuated in most specimens with three black spots, the first at the level of the 7th lateral line scale, the second on the vertical to the last dorsal ray and the third at the tip of the caudal peduncle, partly overlapping the base of the caudal fin. The third spot, present in all specimens, is elongated and nearly twice as long as deep. The other two spots are variable in size and shape and one or both may be missing. Occasionally a fourth spot may occur immediately behind the vertical to the last anal ray. Scattered melanophores, usually heavier in the posterior half of the body, lie in a diffuse band over the horizontal myoseptum. Underparts unpig- mented with the exception of a faint black spot at the base of the anal fin and a thin dark line along the ventral surface of the caudal peduncle. DIAGNOSIS AND AFFINITIES : these data agree closely with the original description (Blanc & Daget, 1957) of material from the Volta basin in north Dahomey. We BARBUS OF THE VOLTA REGION 137 found 5| as well as 4! scale rows below the lateral line and a higher maximum number of lateral line scales (32 cf. 29) although the modal number, 28, was the same. The two collections differ principally in coloration. Whereas our specimens normally have three lateral spots, the type material is characterized by a dark mid- Jateral band, continuous from the operculum to the caudal peduncle. We noted, however, that in the Paris Museum paratypes the band is constricted on the caudal peduncle to form a rectangular terminal spot identical with the spot of our own specimens. Although B. atakorensis is superficially similar to numerous species of Barbus each with three lateral spots, we consider that it has no near relatives. The presence of only seven branched dorsal rays, the relatively high number of scales and the elongated spot on the slender caudal peduncle should prevent confusion with other species. DISTRIBUTION : in Ghana, at present known only from dams on the Accra plains and from the River Azubone (Volta basin) in forest country near Mpraeso. The dams are on small coastal rivers draining directly into the sea. No specimens of B. atakorensis were collected in the northern regions of the country. Elsewhere, the upper reaches of the River Oti (Volta basin), in north Dahomey (type locality), the Ogun River, south-west Nigeria (personal observation), and the Kaduna River (Niger Basin), Northern Nigeria (B.M. (N.H.) reg. no. 1936.11.24.27-29). Barbus punctitaeniatus Daget, 1954 (Text-fig. 14) Barbus punctitaeniatus Daget, 1954, Mem. Inst. France. Afr. Noire, no. 36 : 202, fig. 68. Barbus punctitaeniatus : Blanc & Daget, 1957, Mem. Inst. France. Afr. Noire, no. 50 : 137. DESCRIPTION based on 24 specimens 21-37 mm. S.L., 8 from tributaries of the White Volta near Bawku and 16 from the Black Volta near Lawra, northern Ghana. Body moderately compressed. S.L. 3-55 to 4-3 times the maximum depth and 3-55 to 4-1 times the length of the head. Predorsal profile convex with slight nuchal hump, more pronounced in the largest individuals. Snout bluntly pointed, 0-55 to 0-9 times the diameter of the eye. Mouth moderate, subterminal. Barbels relatively long, the anterior 0-4 to 0*85 times and the posterior 0-8 to 1-2 times the diameter of the eye. Tip of the anterior barbel extending to about the vertical to the centre of the eye and the posterior barbel extending beyond the posterior margin. Eyes lateral, 2-75 to 3-6 times in the length of the head and i-o to 1-35 times in the interorbital width. Cheeks and operculum with weakly-developed pit-lines. Pectoral fin 0-65 to 0-85 times the length of the head, the tip falling short of the pelvic origin. Last simple dorsal ray smooth, flexible, 0-8 to i-o times as long as the head. Apex of dorsal fin slightly rounded, distal margin concave. Dorsal fin rays III 7 (f.i) or III 8 (1.23). Anal fin rays III 5. The pelvic origin lies on the vertical to the first branched dorsal ray. Caudal peduncle 1-35 to 1-85 times as long as deep. The lateral line incomplete in all but the three largest specimens (28-37 mm S.L.) with 7 to 12 anterior scales perforated ; total scales in the row 23 to 27 (modal number 25); this row dips to a maximum of slightly more than the depth of 133 A. J . cS: J . 1IOPSOX half a scale row below the horizontal myoseptum in the anterior third of the body ; the two converge at the anterior end of the caudal peduncle. Three and a half scales between the lateral line and the dorsal origin, 3^ scales between the lateral line and the mid-ventral line immediately in front of the pelvic origin and 2-2\ scales between the lateral line and the pelvic origin. Nine (f.22) or 10 (f.2) scales round the caudal peduncle. FIG. 14. Barlus punctitaeniatus. Summary of morphometric data based on 24 fish, 21-37 mm. S.L. ; measurements are expressed as percentages of S.L. Maximum depth . Length head Diameter eye Interorbital width Length snout Length anterior barbel . Length posterior barbel . Length pectoral fin Length dorsal fin . Length caudal peduncle . Depth caudal peduncle . COLORATION : in living specimens dorsal surface light fawn, silvery on the sides, whitish below. A narrow, dark mid-lateral band runs from the tip of the snout to the tip of the caudal peduncle. Yellow chromatophores are scattered over the dorsal surface, on the operculum and most noticeably in a narrow zone above the lateral band. Pectoral fins frequently tinged with yellow ; all other fins colourless. The iris is marked dorsally with a yellow crescent. In formalin-fixed specimens pigment on the dorsal surface consists of a relatively slender band of dense melanophores near the margin of each scale fading abruptly on the flanks. A narrow black stripe Range Mean Allometry 23- 1-28 I 25 5 24- 2-28 o 26 3 Negative T I-IO o 8 7 Negative 8- 3-i i o 9 8 5' 0-6- 5 6 o 3' 2-6- 4 4 6 6- O-II 2 8 8 18- 2-24 2 19 7 22' 0-25 9 24 4 18- 8-23 6 21 3 12- 1-15 o 13 4 BARB US OF THE VOLT A REGION 139 originating on the tip of the snout (i.e. continuous round the tip from eye to eye) runs obliquely upwards across the operculum and thence along the body. The stripe lies slightly below the horizontal myoseptum in the anterior half of the body but the two converge on the caudal peduncle. The stripe is intensified by small round spots of dense pigment on the pockets of all lateral line scales. Melanophores form a faint stripe at the base of the anal fin continuing backwards as a narrow line along the ventral surface of the caudal peduncle. DIAGNOSIS : our data are in close agreement with the type description (Daget, 1954) based on specimens from the Middle Niger. We examined six of the syntypes in the Paris Museum and found that as in most of the material from Ghana, the lateral line is incomplete, a point which Daget did not observe. Barbus punctitaeniatus is readily distinguished from most West African species by the black streak on the snout running without interruption round the tip. Barbus boboi and B. subinensis are similarly marked but both have higher scale counts and a pronounced dip in the lateral line ; in B. punctitaeniatus the dip is unusually shallow. The differences between this species and B. bawkuensis are discussed in the description of that species below. DISTRIBUTION : in Ghana, widespread and reasonably common in the northern regions, usually in rivers but occasionally in dams. Elsewhere, the Middle Niger (Daget, 1954), the Volta basin in Haute Volta (Blanc & Daget, 1957) and rivers of the Chad basin in Northern Nigeria (personal records) . LIFE HISTORY : the spawning season is probably restricted to the rains. Ripe females were noted in July and juveniles appeared in the rivers during September. Females grow larger than males. Maximum size 37 mm. S.L. Barbus bawkuensis Hopson, 1965 (Text-fig. 15) Barbus bawkuensis Hopson, 1965, Rev. Zool. Bot. Afr. 71: 254 HOLOTYPE : a female of 26 mm. S.L. from the White Volta, Bazua Bridge, 10 miles west of Bawku, north Ghana, 28 . vi . 61, British Museum (N.H.) reg. no. 1964 .9.8. 278. DESCRIPTION based on the holotype and 15 paratypes 20-29 mm - S.L. from the White Volta, Bazua and on 17 specimens 21-26 mm. S.L. from the River Morago, Nakpanduri, 45 miles south of Bawku. Body moderately compressed. S.L. 3-1 to 4-1 times the maximum depth and 3-55 to 4-2 times the length of the head. Pre- dorsal profile convex with a slight nuchal hump. Snout rounded 0-65 to 0-95 times the length of the head. Mouth moderate, slightly protrusible and subterminal. Anterior barbel 0-25 to 0-7 times and the posterior barbel 0-6 to i-i times the diameter of the eye. The tip of the anterior barbel usually extends to the vertical to the anterior margin, and the tip of the posterior barbel beyond the centre of the eye. Eye moderate 2-85 to 3-8 times in the length of the head and 1-05 to 1-4 times in the inter- 140 A . J . & J . H O P S O N orbital width. Weakly-developed pit-lines visible on the cheeks and operculum. Pectoral fin 0-6 to 0-8 times as long as the head, the tip falling well short of the pelvic origin. The last simple dorsal ray smooth, flexible, 0-9 to i-i times the length of the head. Apex of dorsal fin rather pointed, distal margin concave. Dorsal fin rays III 8 ; anal fin rays III 5. The pelvic origin lies between the verticals to the first and second branched dorsal rays. Caudal peduncle slender, 1-45 to 2-0 times as long as deep. Scales radially striated, 24 to 28 in the lateral line row (modal number 26). Lateral line incomplete, perforated scales usually consecutive, 4 to 9 in number at the anterior end of the row. In a few individuals, however, up to 21 perforated scales were noted ; the additional scales were not arranged consecutively but were scattered irregularly amongst blank scales. The lateral line row dips to a maximum depth of half a scale below the myoseptum in the anterior third of the body. There are 3^ scales between the lateral line and the dorsal origin, 3! scales between the lateral line and the ventral midline immediately in front of the pelvic origin and 2 to 2\ scales between the lateral line and the pelvic origin. Eight (f.i), 9 (f.24) or 10 (f.6) scales round the caudal peduncle. 10mm. FIG. 15. Barbus bawkuensis. Summary of morphometric data based on 33 fish, 20 to 29 mm. S.L. ; measurements are expressed as percentages of S.L. Maximum depth . Length head Diameter eye Interorbital width Length snout Length anterior barbel . Length posterior barbel . Length pectoral fin Length dorsal fin . Length caudal peduncle . Depth caudal peduncle . Range Mean Allometry 24-4-31-1 27-7 23-8-28-5 25'9 7-1-9-0 7'7 8-9-11-5 9.9 5-7-6-9 6-3 1-9-4-6 3-o 4 4-8 o 6-1 15-5-21-2 18-3 22-6-28-8 25-4 19-5-22-8 21-4 12-0-14-0 12-8 BARB US OF THE VOLTA REGION 141 COLORATION : in formalin-fixed specimens dorsal surface with a light scattering of melanophores near the margin of each scale. A narrow, diffuse band of melano- phores runs mid-laterally from the operculum to the end of the caudal peduncle. An indistinct black spot lies at the base of the anal fin. The body is otherwise pigment- free. The snout is marked laterally with a dark streak which does not extend forwards onto the tip. Dark pigment is noticeable along the anterior margin of the dorsal fin. DIAGNOSIS AND AFFINITIES : morphometricalry, B. bawkuensis is very similar to B. punctitaeniatus differing only in the shorter barbels (anterior 1-9-4-6 cf. 3-2-6-4% S.L. ; posterior 4-4-8-8 cf. 6-0-11-2% S.L.) and in the slightly higher modal number (26 cf. 25) of lateral line scales. Barbus bawkuensis, however, lacks the black streak on the tip of the snout and the dark spots on the lateral line scales characteristic of B. punctitaeniatus. Other minor differences in the newly described species are heavier pigmentation on the last simple dorsal ray and a darker spot at the base of the anal fin. Barbus bawkuensis and B. punctitaeniatus are remarkable for their high variability in the number of perforated lateral line scales. This unusual feature has been noted previously in B. anoplus (Groenewald, 1958) where, as in B. bawkuensis, the lateral line is frequently interrupted, and also in B. pobeguini from Ksar Torchane, Maure- tania (Daget, 1954) where some specimens had complete, otners rudimentary lateral lines. Barbus bawkuensis and B. punctitaeniatus are unique among West African species in the combination of a incomplete lateral line and two pairs of relatively long barbels. DISTRIBUTION : at present known only from the White Volta and its tributaries near Bawku, northern Ghana. Barbus pobeguini Pellegrin, 1911 (Text-fig. 1 6) Barbus pobeguini Pellegrin, 1911, Bull. Soc. zool. France, 36 : 187. Barbus pobeguini : Daget, 1954, Mem. Inst. France. Afr. Noire, no. 36 : 205, fig. 71. DESCRIPTION based on 24 fish 29-36 mm. S.L., from Dedoro Tankara Dam, 8 miles north-east of Navrongo (Volta basin), north Ghana. Body cylindrical, only slightly compressed. S.L. 3-7 to 4-3 times the maximum depth and 3-5 to 3-8 times the length of the head. Predorsal profile smooth, convex. Snout rounded, 0-85 to i-o times the diameter of the eye. Mouth moderate, subterminal. The single pair of short barbels o-i to 0-3 times the diameter of the eye. Eyes relatively small, inclined slightly upwards, the diameter 3-45 to 4-0 times in the length of the head and 1-35 to 1-7 times in the interorbital width. Weakly-developed pit-lines present on the cheeks. Pectoral fin 0-65 to 0-8 times the length of the head, the tip falling well short of the pelvic origin. Last simple dorsal ray smooth, flexible, 0-8 to 0-9 times the length of the head. Apex of dorsal fin rounded, distal margin slightly convex. Dorsal fin rays III 7 ; anal fin rays III 5. The pelvic origin lies within the verticals to the first or second dorsal rays. Caudal peduncle 1-45 to 1-75 times as long as deep. 14.! A. J . & J. HOPSON The lateral line dips to a maximum depth of one scale row below the horizontal myoseptum in the anterior third of the body ; the two converge in the first third of the caudal peduncle. Twenty-seven to 32 lateral line scales (modal number 29). Four and a half to 5| scales between the lateral line and the dorsal origin, 5^ scales between the lateral line and the mid-ventral line immediately in front of the pelvic origin and 2\ to 3 scales between the lateral line and the pelvic origin. Twelve (f.2i) or 13 (f-3) scales round the caudal peduncle. FIG. 1 6. Barbus pobeguini. Summary of morphometric data based on 24 fish, 29 to 36 mm. S.L. ; measurements are expressed as percentages of S.L. Maximum depth . Length head Diameter eye Interorbital width Length snout Length barbel Length pectoral fin Length dorsal fin . Length caudal peduncle . Depth caudal peduncle . COLORATION : in living specimens silvery with a characteristic steel-blue iridescence. A dark lateral band ends in a smah 1 but pronounced spot on the caudal peduncle. There is an oblique black stripe on the distal half of the anterior rays of the dorsal fin. In formalin-fixed specimens dorsal surface well peppered with melanophores absent from a narrow zone close to the margin and thinning out towards the centre of each scale. Dark pigment fades out on the flanks, and scales below the lateral line are clear. Pockets of anterior lateral line scales each with a small black streak Range Mean Allometry 23-5-26-5 26-2-28-0 6-8-8-7 24-2 27-1 7-2 10-0-11-5 6-3-7-2 O 7-2 2 I7-6-22-O 10-5 6-8 i-4 19-6 Positive Negative 2I-5-25-5 23-4 21 '8-24-1 22-7 12 -9-15-2 13-9 BARBUS OF THE VOLTA REGION 143 divided horizontally into two by the lateral line pore. The upper halves of the streaks persist on more posterior scales. A narrow black lateral band slightly below the myoseptum extends backwards onto the caudal peduncle, stopping just short of a round black spot on the tip. The spot overlaps the base of the caudal rays. Dorsal fin with a narrow black stripe, parallel with the distal margin of the fin, on the upper halves of the last simple and first three branched rays. DIAGNOSIS : these data are in close agreement with descriptions of B. pobeguini from the Middle Niger (Pellegrin, 1911 and Daget, 1954). Barbus pobeguini appears to have no close relatives in West Africa. The presence of a single pair of barbels, 7 branched dorsal rays and the distinctive colour pattern should prevent confusion with other species in the area. Pellegrin (op. cit.) has pointed out the similarity of this species with B. afer Peters from South Africa. HABITAT AND DISTRIBUTION : in Ghana, rivers, streams and pools throughout the northern regions. The commonest Barbus in flood water ; often found in temporary streams after storms and one of the first colonizers of newly built dams. Never abundant in the larger rivers. Elsewhere, basins of the Niger (Daget, 1954) and Comoe (Blanc & Daget, 1957) and in Mauretania (Daget, 1954). LIFE HISTORY : Barbus pobeguini apparently breeds during the rains. Ripening fish were observed in June and juvenile stages appeared in August. Growth is more rapid in females than in males. Maximum size, a female 50 mm. S.L. Barbus stigmatopygus Boulenger, 1903 (Text-fig. 17) Barbus stigmatopygus Boulenger, 1903, Ann. Mag. nat. Hist. (7) 12 : 533 ; Idem, 1911, Cat. Afr. Fish. 2 : 185, fig. 163. Barbus stigmatopygus : Daget, 1954, Mem. Inst. France. Afr. Noire, no. 36 : 210, fig. 75. DESCRIPTION based on 24 fish, 16-19 mm - S.L. from the Black Volta, 3 miles west of Lawra, northern Ghana. Body compressed. S.L. 3-2 to 3-8 times the maximum depth and 3-45 to 3-95 times the length of the head. Predorsal profile convex. Snout short, bluntly pointed, 0-55 to 0-75 times in the diameter of the eye. Mouth terminal, small. Eyes lateral, relatively large, 2-45 to 2-8 times in the length of the head and 0-95 to 1-15 times in the interorbital width. A few weakly-developed pit- lines present on the cheek and operculum. Barbels absent. Pectoral fin 0-6 to 0-75 times as long as the head, its tip falling short of the pelvic origin. Last simple ray of dorsal fin smooth, flexible, 0-8 to i-o times the length of the head. Apex of dorsal fin rounded, distal margin concave. Dorsal fin rays III 8. Anal fin rays III 5. Origin of pelvic fin between the verticals to the last simple and first branched dorsal rays. Caudal peduncle 1-25 to 1-5 times as long as deep. Lateral line incomplete, only the first 4 to 7 scales in the row perforated. There are 23 to 25 (modal number 24) scales in the lateral line row, 4! scales between the lateral line and the dorsal origin, 2|- scales between the lateral line and the mid-ventral line immediately in front of the pelvic origin and ii scales between the lateral line and the pelvic origin. Nine (f.i) or 10 (f.23) scales round the caudal peduncle. 144 A. J. & J. HO PS ON 10mm. Range 26-1-31 -2 25-2-28-9 9-5-11-2 10-3-11-9 5-8-7-4 16-8-20-0 22-1-27-6 17-6-21 -2 I3-2-I5-0 Mean 28-5 27-3 10-3 u-i 6-7 18-6 24-9 19-9 14-2 Allometry FIG. 17. Barbus stigmatopygus. Summary of morphometric data based on 24 fish, 16 to 19 mm. S.L. ; measurements are expressed as percentages of S.L. Maximum depth . Length head Diameter eye Interorbital width Length snout Length pectoral fin Length dorsal fin . Length caudal peduncle . Depth caudal peduncle . COLORATION : in living specimens dorsal surface finely patterned with yellow and black chromatophores, scales of the lateral line and below, silvery. A round black terminal spot on the caudal peduncle and a second spot at the base of the anal fin are always present. Two additional spots frequently occur on the horizontal myoseptum in the anterior half of the body. The fins are colourless. There is a yellow arc over the pupil of the eye. Specimens from clear water are frequently transparent with the vertebral column and viscera visible through the body wall. In formalin-fixed specimens the scales of the dorsal surface are outlined with a fine lacework of melano- phores which fade away on the flanks. A small round black spot at the tip of the caudal peduncle is always present, as is a compact group of melanophores at the base of the anal fin, which continues backwards as a thin black line along the ventral surface of the caudal peduncle. Two further round black spots, smaller than the one on the caudal peduncle, usually occur on the horizontal myoseptum, the first at the level of the fifth lateral line scale, the second below the last ray of the dorsal fin. One or both of these may be absent. Pockets of anterior lateral line scales are frequently marked with a black dot above and below the canal. BARBUS OF THE VOLTA REGION 145 DIAGNOSIS : these data correspond closely with previous descriptions of B. stigmatopygus. Despite careful examination of the material described here we can find no trace of the single pair of much reduced barbels which Dr. Trewavas recently observed in specimens from the White Nile and Lake No (pers. com.). Barbus stigmatopygus is easily distinguished by its markings from other small Barbus with similarly reduced lateral lines and barbels. HABITAT AND DISTRIBUTION : in Ghana, judging from our records, B. stigmatopygus is a relatively rare and local species. With the exception of two fish from the River Nasia near Tamale all records are from backwaters and tributaries of the Black Volta in the Lawra district where it is not uncommon. The habitat is always deep, relatively still water, free from vegetation. Elsewhere, widespread in the Sudanese belt. Recorded from the White Nile (type locality), the Middle Niger (Daget, 1954), Haute Volta (Blanc & Daget, 1957) and the Geba River, Portuguese Guinea (Boulen- ger, 1911). Barbus leonensis Boulenger, 1915 (Text-fig. 18) Barbus leonensis Boulenger, 1915, Ann. Mag. nat. Hist. (8) 15 : 203 ; Idem, 1916, Cat. Afr. Fish. 4. : 273, fig. 167. Barbus leonensis : Daget, 1954, Mem. Inst. France. Afr. Noire, no. 36 : 209, fig. 74. DESCRIPTION based on 24 fish 27 to 33 mm. S.L. from the water-works reservoir, Tamale, north Ghana (Volta basin). Body somewhat compressed. S.L. 3-1 to 3-7 times the maximum depth and 3-2 to 3-7 times the length of the head. Predorsal profile smooth, convex. Snout pointed, 0-7 to 0-95 times the diameter of the eye. Mouth small, slightly subterminal. Eyes lateral, 2-9 to 3-35 times in the length of I Omm. FIG. 1 8. Barbus leonensis. 146 A. J . & J . HOPSON the head and 1-15 to 1-4 times in the interorbital width. Barbels absent. A few weakly-developed pit-lines on the cheeks and on the operculum. Pectoral fin 0-55 to 07 times the length of the head, the tip falling short of the pelvic origin. The last simple dorsal ray is smooth, flexible, 0-75 to 0-95 times as long as the head. Apex of the dorsal fin slightly rounded, the distal margin straight. Dorsal fin rays III 7 (f.2) or III 8 (f.22). Anal fin rays III 5 (f.2i) or 6 (f.3). The origin of the pelvic fin lies between the verticals to the first and second dorsal rays. Caudal peduncle 1-2 to 1-5 times as long as deep. The lateral line is always incomplete ; 6 to 10 anterior scales are perforated. There are 22 to 26 scales (modal number 24) in the lateral line scale row, 4! to 5| scales between the lateral line and the. dorsal origin, 2.\ scales between the lateral line row and the mid-ventral line immediately in front of the pelvic origin, and i| scales between the lateral line and the pelvic origin. Eight (f.i), 9 (f.i8) or 10 (f.5) scales round the caudal peduncle. Summary of morphometric data based on 24 fish, 27 to 33 mm. S.L. ; measurements are expressed as percentages of S.L. Range Mean Allometry Maximum depth . . . 27-0-31-8 29-3 Length head . . . 27-5-30-3 28-6 Diameter eye . . . 8 2-9 -6 8-9 Interorbital width . . 10-4-12-8 11-5 Length snout . . . 6 6-8 -2 7-3 Length pectoral fin . .16- 6-2 1 o 1 8 6 Length dorsal fin . . . 23-0-26-0 23-8 Length caudal peduncle . . 18-2-21-0 19-6 Depth caudal peduncle . . 13-1-15-4 14-3 COLORATION : living specimens pale lemon yellow in colour with two rounded black spots, one on the caudal peduncle the other on the anterior rays of the dorsal fin. Yellow pigment brightest in an arc over the pupil of the eye and in the vicinity of the two black spots. Living specimens frequently transparent, the vertebral column and viscera clearly visible through the body wall. In formalin-fixed specimens dorsal surface with an even scattering of melanophores thinning out and disappearing on the flanks. Scales of the lateral line row and above lightly outlined with melano- phores which are thicker in a vertical line on the pocket of each scale, thus forming a regular pattern over the dorsal surface of the fish. Traces of the pattern may persist on the row below the lateral line but the ventral surface is generally pigment- free. The black spot on the caudal peduncle is equal in diameter to the depth of one scale row and slightly overlaps the base of the caudal fin. The side of the snout is marked with a narrow, horizontal, band of melanophores touching the anterior margin of the eye. The spot on the dorsal fin lies between the last simple and the third branched ray at a point midway between the base and the apex of the fin. DIAGNOSIS : these data correspond closely with the description of the types from Sierra Leone and also with data obtained by Daget (1954) from fish in the Middle Niger. We, however, record 9 as the modal number of scales round the caudal peduncle compared with 8 in the previous descriptions. Barbus leonensis is easily BARB US OF THE VOLT A REGION 147 distinguished from B. anenia and B. stigmatopygus (both of which have incomplete lateral lines and no barbels) by the characteristic pattern of pigmentation. HABITAT AND DISTRIBUTION : in Ghana, widespread in well-vegetated pools, oxbow lakes and dams throughout northern Ghana. Rare in rivers and streams. Elsewhere, recorded from much of the savannah zone of West Africa Sierra Leone (type locality), Gambia River (Johnels, 1954), Niger basin (Daget, 1954), Comoe (Blanc & Daget, 1957), Haute Volta (Blanc & Daget, op. cit.} and Chad (Daget, op. cit.}. LIFE HISTORY : spawning is confined to the rainy season, June to September. All fish mature at the age of 8 to 10 months. Few individuals, if any, survive their first spawning season and the maximum life span is probably 15 months. Growth is slightly more rapid in females than in males. Maximum size, 33 mm. standard length. KEY TO THE VOLTAIC SPECIES OF Barbus 1 Well-developed pit-lines present ; individual pits bead-like in appearance, forming parallel ridges on the side of the snout, the cheek, the operculum and on the dorsal surface of the head between the eyes . . . subgenus Beirabarbus 2 Cephalic pits relatively few in number, sometimes orientated into lines, but with the individual pits well-separated from one another and usually visible only on the cheek and on the operculum .......... 3 2 Apex and second simple ray of dorsal fin dense black ; heavy melanophores form a conspicuous mid-lateral band running from the side of the snout to the tip of the caudal peduncle ; pockets of lateral line scales with relatively broad, triangular spots ; 12 scales round the caudal peduncle ; the tip of the pectoral fin never overlaps the pelvic origin ; snout 6-3-7-3% S.L. ; caudal fin in living specimens orange-red B. macrops No dense pigment patches on the dorsal fin ; mid-lateral band relatively light ; pockets of lateral line scales with narrow and crescentic vertical stripes ; usually 10 or ii scales round the caudal peduncle ; the tip of the pectoral fin often over- laps the pelvic origin ; snout 6-4-7-6% S.L. ; caudal fin pale pink in living specimens B. parablabes No dense black pigment patches on the dorsal fin ; mid-lateral band faint ; pockets of lateral line scales with relatively narrow vertical bars of pigment ; 12 scales round the caudal peduncle ; the tip of the pectoral fin sometimes overlaps the pelvic origin ; snout 7-8-8-8% S.L. B. spurrelli 3 Lateral line scales markedly deep, over 2-7 times as deep as long ; dorsolateral scales outlined with melanophores which are concentrated in a vertical arc on the pocket of each scale ; mid-lateral band, if present, very faint ; an oblique black stripe often lies between the origin of the lateral line and the base of the pectoral fin : subgenus Clypeobarbiis ......... B. hypsolepis Lateral line scales not markedly deep, under 2-4 times as deep as long ; pigment pattern consisting of lateral stripes or spots ; no black stripe between the origin of the lateral line and the base of the pectoral fin ....... 4 4 Barbels present ........ .5 Barbels absent ............ 13 5 Usually 8 branched dorsal rays ........ 6 Usually 7 branched dorsal rays . . . . . . . . . 12 i.fS A. J . & J. HO PS ON 6 Body marked with three or more mid-lateral spots ; one in the anterior third of the body, one below the last dorsal ray and one on the caudal peduncle are nearly always present ; the spots are sometimes partly obscured by a dark mid-lateral band .....'......... 7 Body marked with a black mid-lateral band, sometimes broken into irregular spots or streaks ............. 9 7 Body marked with three mid-lateral spots . . . . ... . . 8 Body marked mid-laterally with four or more large rounded black spots . B. sublineatus 8 Dorsolateral surface with two or three dark longitudinal stripes ; barbels relatively long, the anterior clearly overlapping the base of the posterior ; mouth not protrusible obliquely downwards ; spot at the base of the anal fin relatively faint ............ B. macinensis Dorsolateral surface without longitudinal stripes ; barbels short, the anterior barely overlapping the base of the posterior ; mouth small, protrusible obliquely down- wards ; black spot at the base of the anal fin conspicuous . . . B. voltae g Lateral line complete ; 4! or 5^ scales between the lateral line and the dorsal origin 10 Lateral line usually incomplete ; 3 1 scales between the lateral line and the dorsal origin 1 1 10 Mid-lateral band more or less continuous from the tip of the operculum to the end of the caudal peduncle ; mouth large, barbels long, the posterior extending to or beyond the posterior margin of the eye ...... JB. nigeriensis Mid-lateral band, originating between the operculum and the dorsal fin, frequently broken into irregular spots or streaks ; mouth moderate in size ; barbels moderate, the posterior not extending past the vertical to the posterior margin of the pupil B. lawrae 1 1 Black pigment on the snout continuous round the tip from eye to eye ; conspicuous black spots on the scale pockets of the lateral line B. punctitaeniatus Black pigment on the snout confined to the sides ; scale pocket pigment of the lateral line inconspicuous ........ B. bawkuensis 12 Two pairs of barbels ; mid-lateral pigmentation variable but three spots often present ; no spot on the dorsal fin . . . . . . B. atakorensis One pair of barbels ; a narrow continuous mid-lateral band ends in a spot on the tip of the caudal peduncle ; dorsal fin marked with a black streak across the distal half of the anterior rays ......... B. pobeguini 1 3 Dorsal fin marked with a round black spot ; body with a single spot on the tip of the caudal peduncle .......... B. leonensis No spot on the dorsal fin ; Body marked with two to four black spots, one at the base of the anal fin, the others on the horizontal myoseptum . B. stigmatopygus ACKNOWLEDGEMENTS We wish to thank the following people in connection with this work : Dr. P. H. Greenwood for his kind encouragement and advice, for many stimulating discussions, for his invaluable criticism of the manuscript and for providing us with facilities to work on material in the British Museum (N.H.) ; Dr. E. Trewavas for valuable help and for allowing us to use her unpublished data on B. stigmatopygus ; Dr. M. Blanc who kindly arranged for us to examine material in the Paris Museum ; Dr. S. H. Weitzman of the Smithsonian Institution who re-examined the type of Mannichthys lucileae on our behalf ; Mr. G. G. T. Harrison, formerly Chief Fisheries Officer, Ghana, who suggested taxonomic work on Ghanaian fish ; Mr. D. Ofori-Adu, Fisheries Assistant, Fisheries Division, Ghana, who helped to collect much of the data ; Mr. BARBUS OF THE VOLTA REGION 149 Kposugbe Dutanyah, Fisherman, Fisheries Division, Ghana, who caught most of the specimens ; Mr. David Barry of the Kwame Nkrumah University, Kumasi for help with the collection of the forest species. REFERENCES BLANC, M. & DAGET, J. 1957. Les eaux et les poissons de Haute-Volta. Mem. Inst. France. Afr. Noire, no. 50 : 99-169. BOULENGER, G. A. ign. Catalogue of the freshwater fishes of Africa in the British Museum (Natural History), 2, London. 1916. Catalogue of the freshwater fishes of Africa in the British Museum (Natural History), 4, London. DAGET, J. 1952. La reserve naturelle integrate du Mt. Nimba. Mem. Inst. France. Afr. Noire, no. 19 : 311-334. 1954- Les poissons du Niger sup6rieur. Mem. Inst. France. Afr. Noire, no. 36 : 391 pp. 1959. Les poissons du Niger superieur (i re note complementaire) . Bull. Inst. France. Afr. Noire, 21, s6r. A, no. 2 : 664-688. GREENWOOD, P. H. 1962. A revision of certain Barbus (Pisces, Cyprinidae) from east, central and south Africa. Bull. Brit. Mus. (Nat. Hist.), Zool. 8, no. 4 : 151-208. 1962. A new species of Barbus (Pisces, Cyprinidae) from the Upper Zambezi River, Rhodesia. Rev . Zool. Bot. Afr. 65 : 211-216. 1963. Notes on Barbus radiatus Peters 1853 and the subgenus Beirabarbus in east, south and central Africa. Rev. Zool. Bot. Afr. 67 : 20-28. HERRE, A. W. 1936. A new cyprinid genus and species and a new characin from Portuguese East Africa. Proc. Biol. Soc. Washington, 49 : 99-102. HOLDEN, M. J. 1963. The populations of fish in dry season pools of the Sokoto River. Col. Off. Fish. Pub., London, no. 19 : 65 pp. HOPSON, A. J. 1965. Four new species of Barbus (Pisces Cyprinidae) from Ghana. Rev. Zool. Bot. Afr. 71 : 245-257. IRVINE, F. R. 1947. The fishes and fisheries of the Gold Coast, London : 352 pp. JOHNELS, A. G. 1954. Notes on fishes from the Gambia River. Arkiv. f. zool. Stockholm, (2) 6 : 327-411. NORMAN, J. R. 1932. A collection of fishes from Sierra Leone. Ann. Mag. nat. Hist. (10) 10 : 180-185. PELLEGRIN, J. 1909. Description d'un Barbus nouveau du Sahara. Bull. Mus. Nat. Hist. nat. 5 : 18-19. 1911. Description d'un poisson nouveau de Guin6e Fran9aise appartenant au genre Barbus. Bull. Soc. Zool. France, 36 : 187. 1921. Les poissons des eaux douces de 1'Afrique du Nord franyaise. Mem. Soc. Sci. Nat. Maroc, 1 (2) : 216 pp. 1923. Les poissons des eaux douces de 1'Afrique occidentale, Paris : 373 pp. SCHULTZ, L. P. 1942. The freshwater fishes of Liberia. Proc. U.S. Nat. Mus. 92 : 301-348. SVENSSON, G. S. O. 1933. Freshwater fishes from the Gambia River (British West Africa). Results of the Swedish expedition, 1931. Kungl. Sven. Vet. Akad. Handl. 12 (3) : 102 pp. PRINTED IN GREAT BRITAIN BY ADLARD & SON LIMITED BARTHOLOMEW PRESS, DORKING .'a- ' CT K NOTES ON THE CUPULADRIIDAE (POLYZOA, ANASCA) PATRICIA L. COOK BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 13 No. 5 LONDON: 1965 NOTES ON THE CUPULADRIIDAE (POLYZOA, ANASCA) BY PATRICIA L. COOK British Museum (Natural History) Pp. 151-187 ; Plates 1-3 ; 6 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 13 No. 5 LONDON: 1965 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 13, No. 5 of the Zoological series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. Trustees of the British Musi urn (Natural History) 1965 Issued October, 1965 Price Seventeen Shillings NOTES ON THE CUPULADRIIDAE (POLYZOA, ANASCA) By PATRICIA L. COOK Page 1 INTRODUCTION ............ 153 2 CUPULADRIIDAE LAGAAIJ .......... 154 3 STRUCTURE AND DEVELOPMENT OF THE ZOARIUM . . . . . -155 a. Early astogeny .......... 155 b. Growth. . . . . . . . . . . .156 c. Regeneration and deformation. . . . . . . . 157 4. CHARACTERS OF SYSTEMATIC VALUE . . . . . . . .159 a. The central area .......... 159 b. The cryptocyst . . . . . . . . . .160 c. The opercula ........... 160 d. The vibracula .......... 160 e. The basal surface . . . . . . . . . .161 5 The EFFECTS OF ZOARIAL FORM ON CHARACTERS . . . . . . 1 62 6 COMPARISON WITH THE STRUCTURE OF OTHER FREE-LIVING ZOARIA . . .164 7 GROUPING OF SPECIES .......... 167 A. Cupuladria canariensis group ........ 167 B. C. guineensis group . . . . . . . . .167 C. C. owenii group .......... 167 D. Discoporella umbellata group . . . . . . . .167 8 Cupuladria pyriformis (Busk) . . 3 . . . . . .168 9 C. indica n. sp. . . . . . . . . . . . . 169 10 C. guineensis (Busk) . . . . . . . . . . .170 11 The Discoporella umbellata-complcx . . . . . . . . 175 12 D. umbellata (Defrance) ... ...... 177 13 D. umbellata subsp. depressa (Conrad) . ...... 180 14 CUPULADRIIDAE, GENERAL COMMENTS . ...... 182 15 Setosellina capriensis (Waters) . . ...... 182 1 6 ACKNOWLEDGMENTS . . . . . . . . . . .184 17 SUMMARY ............. 184 18 REFERENCES ............ 184 i INTRODUCTION RECENT work on the rich west African fauna of species belonging to the Cupulad- riidae, has led to an examination of other species and to a general review of the characters of the family. The following notes are an attempt to define the limits of variability of some of these characters, and to delimit groups of species which appear to be related. Specimens in the Zoological Department of the British Museum are referred to by registered number thus, 1899. 7. i. . . . ; in the Palaeontological Department thus, D 6474. Measurements of zooecial length have been made from the distal rim of one zooecium to that of the next succeeding radial zooecium, so including the vibraculum. ZOOL. 13, 5. 8 154 P. L. COOK Measurements are given thus : Length of zooecium Lz Width of zooecium Iz Length of vibracular opesia Lvo Length of vibracular seta Ls Length of operculum Lo Width of operculum lo Length of zooecial opesia Lop Width of zooecial opesia lop DEFINITION OF TERMS USED. The majority of the terms used may be found denned by Lagaaij (1952 : 12) and Bassler (1953 : 7). Horizontal cryptocyst lamina. A porous lamina, formed from fused cryptocyst denticles, extending from the descending cryptocyst across the opesia, parallel to, and below, the frontal membrane. Vestibular arch. The distal, cresentic, raised portion of the aperture. Vicarious vibraculum. A large vibracular individual, taking the place of a zooecium, and itself having a distal vibraculum. 2 CUPULADRIIDAE Lagaaij Cupuladriidae Lagaaij, 1952 : 31. Lagaaij denned the family to include only the genus Cupuladria. A series of species is now known to exist (Cook (in press)) which links the membraniporan forms such as C. canariensis (Busk) with the microporan forms like Discoporella umbellata (Defrance). Both genera have therefore been included in the Cupuladriidae, the diagnosis of which has been accordingly slightly modified. DIAGNOSIS. Lunulitiform Anasca with vibracula alternating with zooecia in the same radial series. Cryptocyst variously developed. Ovicells absent. Ancestrula surrounded by seven zooecia and a distal vibraculum. Lagaaij (1953 : 13) defined the term " lunulitiform " to include conical and discoid zoaria such as those of Cupuladria, Discoporella, Lunulites and Selenaria. For reasons given below, the term is here restricted to the first three of these genera, which all bud radially, although it is not assumed that the Cupuladriidae and Lunulites are closely related. The term " selenariiform " used by Harmer (1926 and 1957) for both Anascan and Ascophoran conical colonies, is here restricted to zoaria with spiral budding, e.g. Selenaria, Setosellina and Heliodoma. The term " conical " is used here for all the remaining forms described by Harmer, some of which may be attached to the substrate by rootlets, e.g. Conescharellina. It must be stressed that, whereas the combination of characters found in the Cupuladriidae is exclusive, and the limits defining the species are generally well marked, the division of the family into two genera, Cupuladria and Discoporella, is somewhat arbitrary, depending upon only one character, namely, the extent of the development of the cryptocyst. Until work can be done on the life-histories and breeding of the Cupuladriidae, the importance of this character in relation to, for example, the size of the eggs, the larvae, and the other zooecial structures, cannot be assessed, and Discoporella is therefore retained here. THE CUPULADRIIDAE 3 STRUCTURE AND DEVELOPMENT OF THE ZOARIUM 155 Although many free-living colonies are similar in appearance, they differ con- siderably in structure and development. In the Cupuladriidae both the early astogeny and subsequent growth of the zoarium are distinct from that found in other lunulitiform, selenariiform or conical colonies. a. Early astogeny. The ancestrular area is here denned as comprising the ancestrula and the primary circle of zooecia budded from it. The single ancestrula buds a distal vibraculum, a pair of proximal-lateral zooecia and then one proximal zooecium (see Marcus & Marcus, 1962 : 287). A pair of lateral zooecia follows, and a pair of distal- lateral zooecia is the last to be formed (see Text-fig. iB). Lagaaij (ig63a : 182, text-fig. 10) found that secondary proximal buds were developed before this last pair, but colonies seem to be variable in development at this stage. Evidence has been gained from very young colonies of Cupuladria sp. (Gulf of Mexico, 1959. 8. 20. 3, Lagaaij Coll.) and D. umbellata (Madeira, 1963. 2. 28. 9, and S. Africa, 1949. n. 10. 668, Burrows Coll.). Waters (1926 : 426) described a double ancestrula in C. canariensis, but none have been seen in any of the many specimens examined, and I agree with Lagaaij (ig63a : 184) that it is possible that Waters regarded the proximal bud as the second member of a paired ancestrula. In one group (C. canariensis, group A, see p. 167), the entire ancestrular area may consist of vicarious vibracula (see Smitt, 1873, pi. 2, fig. 70 ; Hastings, 1930 : 714 ; Marcus & Marcus, 1962 : 289 ; Lagaaij, 19632. : 183 ; and Cook (in press)). Although vicarious vibracula in the ancestrular area may sometimes be the result of regeneration of a normal zooecium, most of those found in C. biporosa (Canu & Bassler) and C. monotrema (Busk) are budded primarily as vibracular individuals. The " special area " described by Canu & Bassler (1923 : 81, pi. 2, fig. 18) is not an ancestrular area, but the growing edge of a regenerating, broken fragment (see below) . The pattern of budding outside the ancestrular area appears to be variable. Another variable character is found in C. Pyriformis (see 168) and a few colonies of C. canariensis, where the vibracular seta of the ancestrula, when closed, is directed towards the right side (viewed frontally) (see Text-fig. iC). In all specimens of BCD FIG. i. Early astogeny. a, ancestrula ; vibracula stippled. A. Setosellina. Spiral budding. B. Cupuladriidae. Radial budding. C. Cupuladvia pyriformis (Busk), ist pair of zooecia aborted. D. Lunulites. Radial budding: . 156 P. L. COOK other species the seta is directed toward the left side, the seta of the zooecium budded from the proximal end of the ancestrula having the same orientation (see Text-fig. iB). The seta of the right-proximal-lateral bud is directed to the right, that of the left-proximal-lateral bud to the left (see also Marcus & Marcus, 1962 : 297). b. Growth. Each zooecium and vibraculum develops radially as a unit which arises as a common bud, the vibracular chamber being divided off from the cavity of the zooecium in the later stages of development. The relationship of the vibraculum with the zooecium distal to it is comparable with that of the " recumbent " ovicell of Crepidacantha (see Brown, 1954: 244, footnote). Here the ovicell is fully formed before the development of the next distal zooecium, which grows round it from below. The vibracular chamber in the Cupuladriidae is not terminal and does not reach the basal surface of the zoarium. Thus, although apparently inserted between the zooecia at the frontal surface, the vibracula are not strictly interzooecial, nor are they adventitious or dependent. The common bud is rarely present, even in specimens preserved in alcohol, as in its peripheral position it is particularly susceptible to damage when the colony is dead, and it is no longer protected by the extended peripheral vibracular setae (see Marcus & Marcus, 1962, and Cook, 1963). The bud arises between peripheral zooecia. The membranous ectocyst is closely applied frontally and basally to the neighbouring zooecia, but over the bud it is convex and free from the underlying structures. A line of demarcation occurs basally where the ectocyst is no longer in contact with the calcareous basal walls (see Text-fig. 2A). On the basal side the bud originates from the distal end of the proximal zooecium, below the vibracular chamber. On the frontal side the upper surface of the bud appears at the distal wall of the vibracular chamber. The membranous frontal part of the bud becomes the frontal membrane of the zooecium, the basal part the endocyst lining the zooecial cavity (see Marcus & Marcus, 1962, pi. 2, fig. 5). This can be seen in well-preserved specimens of C. indica (1899. 5- I - 267) .' C. multispinata (1949. n. 10. 617) ; Discoporella umbellata (MM I 39A) ; and D. u. depressa, (Terra Nova Coll., Stn. 42). The relationships of the young common bud have also been observed in living colonies of D. umbellata from Madeira. As the bud enlarges and grows up between the two adjacent lateral zooecia, the basal wall is covered by a calcified lamina, which develops from the point of origin of the bud (see Text-fig. 2 A, B). The polypide is first differentiated at the proximal end of the bud before it is completely covered basally by the lamina. As the lamina advances distally, the cryptocyst develops beneath the frontal membrane growing inward from the lateral walls. The lamina curves upward and inward, and the vibra- cular chamber develops as an inner compartment from the distal end of the bud. The walls of the chamber are thus double-layered laterally, with connecting bars of calcification. Large pores are present in the outer layer (= " chambers " of Marcus & Marcus, 1962 : 295). These calcify further, become smaller, and eventually com- municate with the next distal-lateral zooecia to be developed. The vibracular chamber is connected with its proximal zooecium through a large pore in the dividing THE CUPULADRIIDAE 157 wall. An uncalcified area, which is left in the lamina as it curves up sharply distally to the vibracular chamber, calcines, but leaves a small pore which eventually com- municates with the succeeding radial zooecium (see Text-fig. 2). Specimens clearly showing these structures are C. multispinata (S. Africa, 1949. n. 10. 94 B, Burrows Coll.) and D-. u. depressa (Rio de Janeiro, Terra Nova Coll., Stn. 42). Transverse sections of zoaria frequently show the stem-like extension of the zooe- cium, below the vibracular chamber of the previous zooecium. Each zooecium thus has a large distal pore communicating with the vibracular chamber, and 2-3 lateral pores communicating with the adjacent distal-lateral zooecia. No pores enter the vibracular chamber laterally, but it communicates with the next distal zooecium in a radial row. Specimens showing the interzooecial pores clearly are : C. indica (1899. 5. i. 267 pt.) and C. canariensis (section, 1929. 4. 26. 85 pt.) ; see also Marcus & Marcus, 1962, pi. 2, fig. I. Large colonies of D. umbellata (e.g. coast of Ghana, 506, Achimota Coll), and C. canariensis (Senegal, MM II 3) show rings of growth (see Harmer, 1926 : 267). Apparently these zoaria have grown evenly for a period and then ceased to grow. New budding has been resumed simultaneously all round the periphery, so that a series of concentric ridges may be seen on the basal side. FIG. 2. Growth of the membranous common bud (c.b.) and calcareous basal lamina, viewed from the basal side. A. Transparent, membranous common bud developing between two peripheral zooecia. Note the uncalcified area in the distal wall of the vibracular chamber. B. Calcareous curved basal lamina developing over the growing common bud. Note the increased calcification of the previously developed proximal-lateral basal laminae, and the development of the radial median groove. c. Regeneration and deformation. A large number of colonies are regenerated from broken fragments, and the budding by which a symmetrical colony is re-formed follows a characteristic pattern. Stach (i936a : 63-65) and Dartevelle (1933 : 70-72) have discussed the fragmentation of colonies, and the ecological significance of this form of regeneration. Dartevelle (1935 : 559-561) also described the form of budding in C. canariensis. He mentioned regeneration as occurring from " calcified zooecia " (see below), and examination of many colonies shows that these closed zooecia are, in fact, the primary buds regenerated from a broken edge of a fragment. The original fragment can usually be seen on the basal side of such colonies, as secondary 158 P. L. COOK calcification at the centre does not seem to occur. The broken zooecia which were distal in position in a fragment regenerate zooecia distally, those which were proximal, regenerate zooecia proximally, in which the zooecial orientation is reversed. Zooecia growing from the lateral walls of old zooecia, are at right angles to the previous direction of growth of the fragment. Each broken zooecium produces one bud laterally from the mid-line of its basal wall, and another bud arises from beneath the vibracular chamber. These zooecia are wide and often slightly distorted (see p. 175). Subsequent budding of radial series, including intercalary rows, is exactly the same as in zoaria developing from an ancestrula. The usual radial alternation in the asym- metry of the vibracular opesia is established in the primary and secondary series of buds (see Text-fig. 3 A). The colony of C. elongata Sakakura mentioned as distorted by Marcus & Marcus (1962 : 288) is, in fact, such a regenerating fragment. Marcus & Marcus (1962 : 301) have described a second form of regenerative or " zoarial budding ", in which small, fan-shaped zoaria are produced from a single zooecium at the periphery of the parent colony. This form of budding has been seen so far only in the D. umbellata-complex (e.g. Madeira, 1912. 12. 21. 1004, Norman Coll.) and in C. multispinata (Madeira, 1963. I. 2. 3 and 4, Norman Coll.). Marcus & Marcus described the results of the fracture of the initial zooecium (" pseudo- ancestrula "), which frees the young zoarium from the parent colony. It then continues its development, eventually assuming a circular shape by the budding of intercalary series. Colonies arising in this manner have no central substrate present, e.g. a sand-grain or foraminiferan. They may also be recognized by the reversed orientation of the regenerated zooecium developed from the broken pseudoancestrula, and by their early fan-shaped budding. The pseudoancestrula does not calcify fully on the basal side, and when still attached to the zoarium has been seen to fold frontally when its vibracular seta closes (see Cook, 1963 : 411). When fracture occurs, the basal lamina is detached with the new zoarial bud, but the proximal part of the cryptocyst remains with the parent colony (see Marcus & Marcus, 1962 : 302). The distal part of the pseudoancestrula then regenerates a zooecium orientated in the opposite direction. In the Madeiran material the fracture has not always occurred, and buds have continued to develop attached to the parent colony. Deformed colonies have resulted, some of which have been described as C. lowei by Busk (1854 : 99 D. umbellata) and " C. deformis " by Waters (1921 : 413 = C. multispinata). Other deformed zoaria, especially of these two species, originate from regenerating fragments, and here, deformation is the result of failure of development of one or more radial rows of zooecia. The cause is unknown, but in some cases it appears to be correlated with the growth of associated Sponges, belonging to the Plocamiidae, on the basal side of the Polyzoan colony. The Sponge appears to have inhibited growth at the origin of the chitinous zooecial bud, just below the penultimate series of vibracular chambers. The radial rows of zooecia adjacent to the affected zooecium or zooecia, produce intercalary series until the radial slit in the colony is closed when the two sides touch. Frequently, these zooecia are unable to form interzooecial connections, and a slit or furrow continues radially through the colony, which, as the THE CUPULADRIIDAE 159 associated Sponge develops, becomes lined with spicules. No example of a spiral slit has been seen in deformed colonies of the Cupuladriidae, and the budding pattern should be compared with that of one deformed specimen of Selenaria maculata (Bass's Straits, 1899. 7. i. 1497), where the slit is distinctly spiral (compare Text-figs. 36, D). The basal surface of this zoarium also shows traces of a Sponge colony, and although it is possible that the Sponge is a secondary and unrelated growth, the correlation with this type of deformation is interesting. A B FIG. 3. Regeneration and deformation. Vibracula stippled. A. Cupuladria. Regenera- tion from lateral walls of a radial series. B. Selenaria. Spiral regeneration. C. Discoporella. Deformation (*), point where distal radial bud has not developed. D. Selenaria. Deformation (*), point where lateral distal bud has not developed. 4 CHARACTERS OF SYSTEMATIC VALUE a. The central area is here defined to include the ancestrular area and the centre of the colony where the budding of intercalary radial rows is most frequent and closed zooecia are found. In all groups of species but one (group A, see p. 167), the opesiae of the zooecia of the central area become occluded by secondary calcareous growth ; these zooecia do not then contain polypides (see Harmer, 1926 : 267). Closed zooecia are often present even in fairly young colonies, and are also found in areas of rapid budding at the edges of regenerating fragments (see above). In the membraniporan group B (see below) the lamina is formed as a thin sheet originating just below the level of the slight shelf formed by the descending cryptocyst. It develops in a proximal direction from the distal end of the opesia, later also growing out from the lateral walls. Although the lamina thickens it rarely approaches close to the frontal membrane. Irregular pores are left, especially at the lateral edges. In the denticulate forms (Group C, see p. 167), the lamina is formed by extension and fusion of the existing cryptocystal denticles and first develops proxi- mally, spreading progressively to the sub-opercular region. It is closer to the frontal membrane in origin than the lamina in Group B, and when thickened by further deposition may almost obscure the mural rims of the central zooecia. In Discoporella where the opesia is restricted to a small sub-opercular area, the central cryptocystal pores calcify first, the opesia then becoming occluded by denticulate growth. The opesiules frequently remain uncalcified or only partially closed. The lamina in these 160 P. L. COOK last two groups is rarely complete, irregular pores usually being present, but in contrast to the first group, no pores are present in the subopercular area. The existence and type of central area closure may be evidence of the group to which a species belongs, and the association of closed zooecia with ancestrular and regenerative budding in groups B, C and D is comparable to the occurrence of vicarious vibracula in group A. b. The cryptocyst is present in all species and, in well-preserved material, is a useful specific character, especially when its form can be correlated with other characters, such as those of the basal surface. Unfortunately, the cryptocyst is among the first structures to be damaged or worn, although there are usually one or two zooecia present in which it can be seen, even in fossil specimens. The type of cryptocyst forms a series in the Cupuladriidae from the membrani- poran C. canariensis to the microporan D. umbellata. In the C. canariensis and C. guineensis groups (see below), it is simple, though variously extensive. In the C. owenii group it is denticulate, the form and extent of the denticles being specific. In Discoporella it is so extensive that it has been regarded as a generic character. In development, however, the cryptocyst of D. umbellata shows great similarity to that of the C. owenii group, as it is formed from the fusion of irregular denticles (see Marcus & Marcus, 1962, pi. i, fig. 4). In Discoporella the final development of the cryptocyst results in a pair of distal opesiules or opesiular indentations, through which pass a pair of muscles. Other opesiular pores are present, but the opesia is always limited proximally by a complete bar, formed by part of the cryptocyst. c. The opercula of the majority of species of Cupuladriidae are similar to one another. They are membraniporan with a strong, curved distal sclerite. Their mean dimen- sions or proportions may give additional information toward the definition of a species, but generally their range of variation is large. Furthermore, opercula are absent not only in fossil specimens, but frequently from dead, Recent material which is a little worn. Groups C and D (see below) each include species in which the operculum has a partial or complete proximal sclerite. In C. doma (Group C), the proximal sclerite is variable, and may be incomplete. Its presence may be useful in distinguishing small fragments from those of C. owenii, which may have similar zooecial characters, but the basal surface is here more certain as a determinant of the species. Discoporella (group D), has hitherto been distinguished by the possession of a proximal sclerite in the operculum, but there is evidence (Cook (in press)) that even in this group the opercula are not all of the same type and include at least one species in which they are without a proximal sclerite. d. The vibracula are also similar in character in all species of the family. The opesia is auriform, with distinct condyles, which are alternately asymmetrical to right or left in each member of a radial row of zooecia. The vibracular setae are long and usually follow a sigmoid curve, they increase in length toward the periphery of the colony. The setae have an asymmetrical basal sclerite, to which are attached powerful muscles which fill the vibracular chamber. Although there are differences between the setae of various species, especially in the size and shape of the basal THE CUPULADRIIDAE 161 sclerite, the variation within the colony is large and they appear to be of little use in specific determination. The basal sclerites of group A seem generally to be less tuberculate and robust than those of the other groups. The detailed morphology of the vibraculum was discussed by Marcus & Marcus (1962). The seta is able to move in two directions and in one plane. Owing to its asym- metrical articulation and curvature, these movements cover nearly half of the frontal surface of the zooecia proximal and distal to any vibraculum. Live colonies examined by me have not been seen to use their setae like tweezers, as described by Marcus & Marcus (1962 : 299). The distribution of the setae, and their range of movement, is such that they rarely come into contact (see Text-fig. 4). Those near the centre of the colony may touch each other when closed, owing to the convergence of radial rows, but should any detritus be held between two setae in this position, they automatically lose all contact as soon as an opening movement begins. Particles on the frontal surface are dislodged by the movement of individual setae. They roll down the sloping surface of the colony, passing between the peripheral, supporting setae. The zoarium is thus kept clear of deposits settling upon it (see Cook, 1963 : 410). FIG. 4. Distribution of vibracular setae in Discoporella wnbellata. Membrane of vibra- cular opesiae and opercula stippled. Setae pointing distally are in the " open " position ; those pointing proximally are in the " closed " position. e. The basal surface. Lagaaij (1952 : 34, 1953 : 13) stressed the importance of the character of the basal surface in specific determination of fossil specimens, where the frontal surface may have suffered wear. In those groups (A and B below) in which rectangular or hexagonal sectors are present, the thickening consists of layers of ib2 1'. L. COOK calcification, pierced by regular series of kenozooecial chambers, or by small pores. The shape and arrangement of the sectors, and the number of pores or chambers may be specific, within defined limits. For example, the basal sectors of C. canariensis s. s. consist of several layers of long irregular units, interspersed with occasional short sectors ; whereas those of C. monotrema are single layered and invariably short and square. The sectors of C. guineensis (see Hastings, 1930 : 714, and p. 170) may vary, especially in large colonies, from square to hexagonal. In the first case the characters of the basal surface are correlated with other distinguishing features, in the second, the variability of the character is found within a single colony, and is not correlated with any other character, and is therefore not specific. In the remaining groups the basal surface consists of a series of alternating radial grooves and ridges, the characters of which may also be specific, although they are variable, affected by wear, and similar forms occur in both Cupuladria and Discoporella. The ridges and grooves are the result of the form of growth of the zooecia. Each curved basal lamina is grooved in the mid-line (see Text-fig. 2). As secondary cal- cification increases the groove may be filled in, and in some cases may be marked by a protuberant " thread " of heavier calcification. The tuberculation of the basal surface may increase with calcification, but may also become changed by wear. In some species, however, the form of tuberculation may be correlated with other charac- ters, and be specific. For example, in C. owenii, the tubercles are consistently small and regular, and in C. multispinata, they are large, coarse, and coalescent (Cook (in press)). 5 THE EFFECTS OF ZOARIAL FORM ON CHARACTERS The extreme variation in zoarial and zooecial characters of specimens attributed to D. umbellata has been discussed by several authors, and Silen (1942 : 16) noted variation in the shape of zoarium of his specimens and grouped his material into localities where one type of zoarium was dominant. The type of zoarium appears to have little effect on the zooecial characters except where the form of the colony is the same as that found normally in C. doma (d'Orbigny) or that found in zoaria described as C. peyroti by Duvergier (1924 : 19). Doma-type zoaria occur in several species of Cupuladriidae (Cook (in press)) and perhaps correspond to the Trochopora-habit of Lunulites (see Lagaaij, 1953 : 17). Specimens belonging to the D. umbellata-complex. (see p. 175), frequently produce doma-type colonies, and before discussion of specific differences, the effects of form of growth must be considered. The doma-type zoarium is small and usually high (diameter range 3-7 mm.) The zooecia are small, and the basal surface solid or nearly filled in by secondary calcifica- tion. In the majority of zoaria the peripheral zooecia are closed, and there may be from 1-3 rows of enlarged peripheral vibracula. In all lunulitiform zoaria the angle of the axis of the polypide cavity approaches a right angle with the zoarial axis, and the planar intercept of the frontal of each zooecium is thus longer in flatter colonies than in the doma-type zoarium, where there is therefore a reduction of the zooecial dimensions usually found in the species, and THE CUPULADRIIDAE 163 consequent changes in the development of the cryptocyst and the proportions of various other characters. The basal surface is also frequently non-typical, owing to the large amount of secondary calcification. Canu & Bassler (1930 : 12) stated that specimens from abyssal depths of the Pacific reported in 1929 (p. 142), were stunted. Photographs of these specimens (1963. 3. 14. 14) show that they are not, however, either particularly small, or of the doma-type. Canu & Bassler also mentioned that colonies of " var. conica " had been found in the Helvetian of Touraine, but these colonies were not specifically mentioned by Canu (1909 : 448) in his description of the zoaria from this locality. Marcus & Marcus (1962 : 304) described colonies from the coast of Brazil as D. umbellata var. conica Canu & Bassler. These are certainly the same form as those listed below as doma-type zoaria (see p. 181). The number of specimens examined of doma-type colonies is not at present sufficient to determine whether they should be considered a subspecies of D. umbellata. It appears more likely that they are a variant developed in response to specific, but unknown, ecological conditions. The specimens examined fall roughly into 3 groups, each corresponding to one of the 3 major forms in the D. umbellata-complex. (see p. 176). The ecological factors involved in the occurrence of doma-type colonies are probably not simple. There is, for example, no correlation with depth in the material examined, although C. doma itself extends in distribution to greater depths than other species of the Cupuladriidae. One doma-type specimen from Cape Frio, off Angola, is the sole record from the south west coast of Africa, where the temperatures are lower than, for example, east of the Cape of Good Hope. At Mowe point, just south of Cape Frio, in 400 m., a temperature of 9 C. has been recorded (see Hart & Currie, 1960 : 173, text-fig. 29). The conditions producing doma-type growth may perhaps be related to temperature. In the D. umbellata-complex. the doma-type colonies are very small and the number of opesiules is reduced. The size of opesiae and opercula are not, however, reduced in the same proportion. The peripheral zooecia are closed and the peripheral vibracula enlarged, as in C. doma. Generally, the effect is of a suppression of the characters normally separating the populations in the complex. In nearly all cases, however, the specimens retain traces of features consistent with those found in the larger, more normally developed colonies (see pp. 180, 182). Peyroti-type growth is similar to doma-type growth, and may also be caused by ecological conditions at present unknown. Some colonies combine features of both forms of growth (see p. 180), but whereas doma-type zoaria occur in all 3 populations of the complex, Recent peyroti-type colonies are the characteristic form of one geo- graphical area, namely the south-eastern coast of Africa. Duvergier (1924 : 19) described, as Cupularia peyroti, colonies similar in character to Discoporella umbellata, from the Miocene of France, in which the basal surface (described as " face superieur ", following Canu's theoretical orientation of lunulitiform and selenariiform colonies) was covered by kenozooecia and vibracula. Recent specimens from south and east Africa (see p. 177) show stages in the development of the same form of growth, which was briefly described by O'Donoghue (1924 : 39) and discussed by Hastings (1930 : 164 P. L. COOK 719), who doubted whether her south African specimens could be included in D. umbellata. The large number of colonies now available show that although the zoarial form is markedly different, the zooecial characters have much in common with D. umbellata s.s. The modifications to a specific character produced by peyroti-type growth are principally in the appearance of the basal surface. Zoaria are usually large, unlike do ma-type colonies, and have a smooth, slightly pitted base, which may be concave or solid. There is no sign of the tuberculate, grooved basal surface usually found in D. umbellata. In most zoaria the base is covered by a layer of kenozooecia and vibracula proliferated from the peripheral zooecia and vibracula (see PI. 3, fig. 6). The vibracular opesiae are as large as those of the peripheral vibracular zooecia but their setae are not as long. The kenozooecia have no polypides and resemble the closed zooecia of the frontal central area except that there are pores in the distal part of the lamina which in normal closed zooecia is the solid, sub-opercular region (see p. 160). The peripheral frontal zooecia are also closed as in doma-type zoaria, but the peri- pheral vibracula are not greatly enlarged nor do several rows of them occur. The basal vibracula presumably act in a similar manner to those at the periphery, and assist in stabilizing the zoarium and maintaining it just above the surface of the substrate (see Cook, 1963). The conditions determining peyroti-type. growth do not seem to be related to temperature, as similar ranges occur east of the Cape of Good Hope to those off west Africa (i.e. warm temperate to tropical, see Stephenson, 1947). No Cupuladriidae have been reported from the colder waters immediately to the west of the Cape of Good Hope. 6 COMPARISON WITH THE STRUCTURE OF OTHER FREE-LIVING ZOARIA Silen (1942 : 1-22) examined many free-living zoaria and considered that their growth was spiral i.e. that the zooecia of each apparently radial row were budded from the nearest proximal-lateral, not proximal, zooecium. Examination of many lunulitiform, selenariiform and conical colonies has confirmed Silen's observations for Setosellina, Setosella (principally an encrusting form) and Heliodoma, and Stach's observations (i936a) for Selenaria, but has also shown that the method of growth in the Cupuladriidae and the genus Lunulites is entirely different. The evidence is summarized below. a. In the spirally growing forms the ancestrula buds off both a distal and a proximal (or proximal-lateral) vibraculum, each with an adjacent zooecium. A good example was figured in Vibracellina caribbea 1 by Osburn (1947 : n, pi. I, fig. 2). Two spirals of zooecia and vibracula develop from these foci, so that although the ancestrula is eventually surrounded by zooecia, unlike that of Lunulites and the Cupuladriidae, it has not directly budded all of them (contrast Text-fig. lA with Text-figs. iB, C and 1 Lagaaij (ig63b : 172) also describes and figures the ancestrula region of this species, which he refers to Setosellina goesi (Silen) . THE CUPULADRIIDAE 165 D). The spiral budding is well illustrated in 5 zoaria of Selenaria squamosa (N. Zealand, 1947. 8. 16. i), where the spiral is anti-clockwise. All colonies of other species seen are developed in a clockwise direction. In Lunulites the budding pattern is radial, not spiral, but differs from that of the Cupuladriidae in that the ancestrula buds 6 zooecia, each of which then gives rise to a zooecium and a vibra- culum. The ancestrula does not therefore directly bud a vibraculum (see Text-fig. iD). b. Silen based his conclusions as to the spiral nature of the zoaria of the Cupulad- riidae on the assumption that morphologically the vibraculum was the distal bud formed by each zooecium, and that a vibraculum could not bud off a zooecium distal to itself. The zooecium and vibraculum of the Cupuladriidae originate as a unit, and although there is evidence that the vibraculum is a modified zooecium (see Marcus & Marcus, 1962 : 298), it is produced within the confines of the common bud, and is not itself the product of distal budding. In spirally developed colonies (e.g. Setosel- lina], the vibracular chamber can be seen to reach the basal side of the zoarium, and the distal-lateral origin of the next zooecial bud can be seen. c. In spirally developed colonies, the parent zooecium of an intercalary series forms the second bud consistently on the side opposite to the direction of spiral growth. In the Cupuladriidae the intercalary zooecia are produced from either distal-lateral wall of the parent zooecium, in a random manner. The vicarious vibracula in the C. canariensis group (see below) may be intercalary or may themselves produce intercalary zooecia or further vicarious vibracula. Their appearance is also entirely random, and has no spiral pattern. In Lunulites the radial rows of zooecia alternate with those of the vibracula. There is no direct lateral relationship between the budding of the types of individual, which are developed radially, not spirally. Successive vibracula enlarge in a series and finally one buds a distal zooecium. New intercalary rows of vibracula are produced from bifurcations of series distal to a zooe- cium, not a vibraculum. In the central area the first vibracular rows are all budded from the proximal zooecium in the same lateral relationship to the zooecial buds (see Text-fig. iD). In subsequent bifurcations the vibracular bud arises to the right or left of the zooecial bud in a random manner. d. The growing edge of Selenaria bimorphocella (off Adelaide, 1928. 9. 13. 81), shows no basal curved laminae, but a series of groups of young zooecia alternating with developing vibracula, similar in appearance to the growing edge of an encrusting colony. Similarly, the growing edges of colonies of Setosellina have no basal curved laminae. The peripheral zooecia of Lunulites arise in a similar manner to those in the Cupuladriidae and basal laminae have been seen. The resulting basal grooves mark the junction of the lateral walls and do not have the alternating pattern of the Cupuladriidae (see below). In Recent Lunulites the cavities of the vibracula reach the basal surface, in the fossil specimens described by Lagaaij (1953), they do not. e. As in the Cupuladriidae, the regenerative budding in Lunulites is at right angles to the lateral walls of broken fragments. This is in contrast to colonies of Selenaria regenerated from fragments (New Zealand, Discovery Coll., Stn. 936), which show spiral budding from the lateral walls of old zooecia. Each new bud is consistently 1 66 P. L. COOK distal-lateral, and no zooecium produces new zooecia at right angles to its long axis (see Text-fig. 36). The budding of deformed colonies of Selenaria is also spiral (see P- FIG. 5. Relationships of basal grooves. Broken lines indicate the position of the frontals of the zooecia on the upper surface. A. Cupuladriidae. Grooves running radially in the mid-line of the zooecia, branching at the beginning of an intercalary series. B. Selenaria. Grooves running between the zooecia, branching around the first zooecium of an intercalary series. f. The basal radial threads and radial sector boundaries branch frequently. The origin of a branch may be traced and found to correspond exactly with the occurrence of the first zooecium of an intercalary row on the frontal surface. The budding of an intercalary row must be of the same nature as that of all other zooecia, and thus the radial nature of the threads and sector boundaries may also be presumed to reflect the radial development of the zoarium (see Text-fig. 5 A) . The formation of the basal surface in the Cupuladriidae is exclusive, but apparently similar forms are present in other families. An unnamed species of Lunulites described and figured by Lagaaij (1953 : 18, pi. 2, figs, 6a, b), has a basal surface superficially like that of C. guineensis, and specimens of Lunulites radiata show alternating tuberculate ridges and grooves similar to those of C. owenii. Selenaria bimorphocella (S. Australia, 1928. 9. 13. 80, and off Adelaide, 1928. 9. 13. 81, Siboga Coll.), shows radial undulating grooves with large pores, which are occluded at the centre of the zoarium, except beneath the vicarious vibracula, where they remain open. In section, these pores do not show the serial kenozooecial structure found in Group A of the Cupuladriidae (see below). The grooves exactly outline the position of the zooecia above them, and, at the peri- phery, show indications of distal and proximal links, which correspond in position to the distal and proximal walls of the zooecia. Thus the grooves are entirely different in nature to, and do not have the alternating relationships of, the grooves, threads, or radial sector boundaries of the Cupuladriidae, and are only secondarily radial in appearance (see Text-fig. 56). THE CUPULADRIIDAE 167 7 GROUPING OF SPECIES The groups defined below do not include all the species which have been described, but formulate the correlation of various characters found to be common to some species. C. pyriformis and C. indica (see below) combine some of the characters of more than one of these groups. The groups are arranged in order of increasing extent of cryptocystal development. A. C. canariensis group. Cryptocyst simple, without denticles, central zooecia without closures by calcareous laminae. Vestibular arch absent. Basal surface consisting of rectangular sectors, with successive series of porous kenozooecial cham- bers. Vicarious vibracula sometimes present. (Examples : C. canariensis (Busk), C. monotrema (Busk), C. biporosa Canu & Bassler). B. C. guineensis group. Cryptocyst simple, without denticles, central zooecia closed by calcareous lamina developing proximally from below level of cryptocyst. Vestibular arch absent. Basal surface of hexagonal or quadrilateral sectors, with very small pores. Vicarious vibracula not present. (Examples : C. guineensis (Busk), C. elongata Sakaura 2 .) C. C. owenii group. Cryptocyst with denticles, central zooecia closed by lamina developing distally from fusion of denticles. Vestibular arch present. Basal surface with alternating ridges and grooves, tuberculate. (Examples : C. owenii (Busk), C. haidingeri (Reuss), C. multispinata (Canu & Bassler), C. doma (d'Orbigny), ? C. misrai (Tewari et al. s .) D. D. iimbellata group. Horizontal cryptocyst lamina complete proximally, with pores. Opesia with opesiules or opesiular indentations. Opesiaeof central area closed by a complete lamina. Vestibular arch present. Basal surface with alter- nating grooves and ridges, or short grooves and lozenges, tuberculate. (Examples : D. reussiana (Manzoni), D. umbellata (Defrance), D. umbellata subsp. depressa (Con- rad)). There are two species which cannot be included in the groups defined above. Although superficially alike, they are probably not closely related. C. Pyriformis has a similar cryptocyst to that found in groups A and B, but differs from both in the character of its basal surface, and in the possession of a distinct vestibular arch. C. indica is close to C. guineensis, but differs completely in the character of its basal surface ; its geographical distribution is also distinct. Cupularia capriensis Waters differs from the Cupuladriidae in important respects and is here referred to Setosellina (see p. 182). 2 Cupuladria elongata Sakakura (1935 : 6, pi. i, figs. 1-3) was described with extremely elongated basa sectors, with small pores, from the Pleistocene of the Bozo peninsula, Japan. Recent specimens from Japanese waters are all attributable to C. guineensis. 3 Discoporella misrai Tewari ct al. (1958 : 234, pi. 46, figs. 1-4, pi. 47 fig. 2), a Miocene fossil, was described to include specimens from Kutch, India, and those previously reported from the Miocene of Travancore as Cupularia spp. (Jacob & Sastri 1953 : 348, pi. 14, figs. 5a, b, pi. 15, figs, ya, b, pi. 16, figs. 10, n). The specimens appear to have a denticulate cryptocyst and tuberculate basal surface. 168 P. L. COOK 8 Cupuladria pyriformis (Busk) (PI. i, figs. 3, 4, Text-fig. iC) Cupularia pyriformis Busk, 1854 : 100, pi. 124, St. Vincent (Windward Is.). Cupuladria canariensis (Busk) : Sildn (part), 1942 : 14, pi. 4, figs. 15, 16, text-fig. 9, Anguilla, Leeward Is., 40-150 m. (not text-fig. 8). LECTOTYPE, chosen here, specimen on slide B.M. 1840. 10. 23. 80 (figured by Busk, 1854, pl- I2 4> basal surface), St. Vincent. LECTOPARATYPE, 1840. 10. 23. 81, on the same slide. OTHER MATERIAL. Naturhistoriska Riksmuseet, Stockholm, FBD 533, Anguilla, Leeward Is., 45-150 m., Goes Coll., i specimen. Zoarium small (3-5-4-5 mm. diameter). Cryptocyst without spines or denticles. Vestibular arch well developed. Opercula without proximal sclerite. One pair of circum-ancestrular zooecia aborted. Central zooecia closed by irregularly per- forated lamina. Vibracula raised, with salient condyles. Basal surface with deep grooves and raised ridges, with very small tubercles. DIMENSIONS. Lz 0-35-0-40 mm. ; Iz 0-20-0-30 mm. ; Lop 0-12-0-17 mm. ; lop 0-18-0-27 mm. ; Lvo 0-09-0-12 mm. C. pyriformis has been reported twice with certainty, and only 3 specimens are at present known to exist in Museum collections. The zooecia are narrow, tapering distally, with a vestibular arch distinctly raised above the zoarial surface. The vibracula are also prominent, with narrow opesiae and slightly hooked condyles. The central zooecia are closed by a thin, minutely granular lamina, originating just below the salient, shelf-like part of the cryptocyst. The lamina is perforated by 4-7 irregular pores, which in several cases are arranged parallel to the edge of the cryptocyst, but which may also be scattered. In all 3 specimens the circle of zooecia budded by the ancestrula is unlike that of other Cupuladriidae in that pair No. i is nearly aborted. Each minute zooecium has a small distal vibraculum (see Text-fig. iC). Silen (1942, text-fig. 9) showed only one abortive zooecium and did not clearly indicate its relationship to the ancestrula, but the figure shows the salient character of the vibracula. The basal surface is deeply grooved, with sharp alternating ridges. The surface is finely tuberculate but not porous. Busk described C. pyriformis with basal pores, but the faint markings are in fact very small tubercles ; his figure shows the other basal characters well, and they are the same in the specimen from Anguilla ; un- fortunately Silen's photograph does not show the ridges which are in fact present (see PI. i, fig. 4). The position of the locality given by Busk for his material, " St. Vincent " (" St. Vincents " on the slide label), is open to several interpretations. Although much of the material from "St. Vincent " described in the 1854 Catalogue was from St. Vincent, Cape Verde Islands, the specimens registered as 1840. 10. 22 ... and 23 ... consist almost entirely of slides labelled " St. Vincents, W.I.". There is a St. Vincent Island off the west coast of Florida but the most likely locality is St. Vincent Island in the Windward Islands. This supposition is strengthened by the THE CUPULADRIIDAE 169 occurrence of the one other specimen from the neighbouring Leeward Islands. Silen only doubtfully included this colony in C. canariensis , and examination of the speci- men shows certainly that it belongs to C. pyriformis. Studer (1889) listed specimens identified by Kirchenpauer from Madeira as Cupularia pyriformis Busk, on p. 7 and from near Cape Verde as Cupularia pyriformis Gray (sic), on p. 13. It is possible that Studer's specimens belonged either to C. canariensis or to Cupuladria sp. (Cook (in press)), which, superficially, have similar zooecial characters. Prof. E. Voigt (Geologische Staatsinstitut, Hamburg), has informed me that Kirchenpauer's specimens were almost certainly destroyed during the 1939-1945 war. Lagaaij (i963a : 182, pi. 25, figs. 2a, b) mentioned a species of Cupuladria, probably related to C. Pyriformis, from the Miocene of East Venezuela. The character of the basal surface and the presence of a distinct vestibular arch indicate an affinity between C. pyriformis and the C. owenii group of species ; but the absence of denticles on the cryptocyst and the form of the central closures show a similarity with the C. guineensis group. 9 Cupuladria indica sp. n. (PL 2, figs. I, 2) Cupularia umbellata Defrance : Hincks, 1887 : 125, Mergui Archipelago. Cupuladria guineensis (Busk) : Hastings (part), 1930 : 714, 719 (specimens from Mergui, Burma). Silen (part), 1942 : 8, pi. 4, figs. 13, 14 (specimens from locality i, Gulf of Aden). HOLOTYPE. Mergui, Burma, B.M. 1899. 5. i. 23A, Hincks Coll. PARATYPES, as above, 1899. 5. i. 23B-H, 1899. 5. i. 267, Hincks Coll. OTHER MATERIAL. Mergui, 1963. 3. 18. 3, Anderson Coll. " John Murray " Coll. (all specimens from a sandy or muddy substrate) : Stn. 80, S.E. of Muscat, 22 13' 30" N., 59 49' 42" E., 3O.xi.33, 16-22 m., ZigiA. Stn. 77, as above, 22 13' 30" N., 59 52' E., 421 m., ZI9&A. Stn. 178, N. of Somali Horn, 12 oo' 36" N., 50 40' 06" E., 2.V.34, 91 m., Z20iA (with Anoteropora latiroitris Silen). Stn. MB I d, Hanish Is., Red Sea, 13 39' 30" N., 42 43' E., I7.ix.33, 29 m., Zi46A and Z4iE. Stn. Mb I b, as above, Zi3gA. Stn. MB II c, Khorya Morya Is., 17 33' 30" N., 56 01' 30" E., 28.x. 33, 26 m., Zi6iA. Naturhistoriska Riksmuseet, Stockholm, FED 822, Gulf of Aden, 30 m. (with Setosellina sp.), photograph registered B.M. 1963. 3. 14. i. Zoarium frequently flattened, discoid, average diameter 7 mm. Cryptocyst simple, well developed laterally, descending gently, edge serrate but not denticulate. Oper- culum without proximal sclerite. Central zooecia closed by irregular denticles form- ing a lamina at the level of the cryptocyst. Basal surface with radial ridges and large, regular tubercles, no rectangular or hexagonal sectors or pores present. DIMENSIONS. Lz 0-40-0-50 mm., av. 0-45 mm. ; Iz 0-30-0-40 mm., av. 0-35 mm. ; Lop 0-25-0-30 mm., av. 0-28 mm. ; lop 0-16-0-21 mm., av. 0-19 mm. ; Lvo 0-12- 0-15 mm. ; Ls 0-50-0-70 mm. ZOOL. 13, 5 . 9 170 P. L. COOK Hincks described his specimens as C. umbellata and his material in the British Museum was originally labelled as this species. Hincks noted that the majority of the zooecia showed " no trace of a lamina ", but that the central zooecia were closed by one. The peripheral pores in the cryptocyst of the closed central zooecia in C. indica do somewhat resemble the opesiules of D. umbellata, as does the basal surface with its tuberculate ridges and grooves, of which Hincks gave a very full description. The specimens show little variation ; just over half of the zoaria are extremely flat, the remainder slightly more conical. The central closures differ from those of C. guineensis in that they are formed from denticles or laminae which appear to originate from the salient shelf-like edge of the descending cryptocyst, not just below it (see p. 159). This is particularly noticeable at the proximal edge of the opesia, which is the last part to be occluded. However, the later thickening results in a very similar appearance in both species. Young colonies, and those formed from regenera- ted fragments, have no closed zooecia. Generally, the cryptocyst is narrower than that of C. guineensis. The basal surface shows little variation in the size of the tubercles, and the radial grooves are marked by a slight thread in all specimens, which distinguish C. indica from those specimens of C. guineensis which have heavy basal tuberculation. The basal structure of C. indica is similar to that of the C. owenii group. There are no sectors, and no pores in the basal lamina, which is not thickened as in C. guineensis. In other respects, C. indica closely resembles C. guineensis, and its zooecial dimensions fall within the range of varation of C. guineensis ; they are, however, consistently at the smaller end of that range. The two species also appear to be allopatric, their geographical distributions being mutually exclusive. The occurrence of the conical Anoteropora latirostris and the selenariiform Setosellina sp. from the same stations as the lunulitiform C. indica illustrates the correlation between various types of free-living zoaria with a sandy or muddy substrate, as discussed by Harmer (1957 : 649). 10 Cupuladria guineensis (Busk) (PI. i, figs, i, 2, PI. 2, figs. 3, 4, Text-fig. 6) Cupularia guineensis Busk, 1854 : 98, pi. 114, figs. 1-5, New Guinea. Busk, 1884 : 206, pi. 14, fig. 6, N.E. Australia, 8 fath. Marcus, 1921 : 8, 2 30' S., 107 10' E., 15 Fd. Cupularia stellata Busk, 1854 : 99, pi. 118, figs. 1-4, Philippines. Marcus 1921 : 8. Cupularia canariensis Busk, Waters, 1885 : 308, Aldinga, Tertiary, S. Australia. 1887 : 201, Princess Charlotte Bay, N.E. Australia, 12 fath. 1921 (part) : 410. 1926 (part) : 426, pi. 18, fig. 10, Princess Charlotte Bay. Cupuladria guineensis (Busk) Harmer, 1926 : 266, text-fig. 9, pi. 16, figs. 5-7, S. of Mindanau, 0-36 m. Canu & Bassler, 1929 : 74. Hastings, 1930 : 714 (not specimens from Mergui, = C. indica). 1932 : 412, Great Barrier Reef. Silen, 1938 : 366, pi. 17, fig. 119. 1941 : 61, East China Sea, 90 m. (cf. locality 7 in 1942). 1942 (part) : 8, text-figs, 5-6, pi. 3, figs. 10-12, localities 2-7 (not locality i, pi. 4, figs. 13-14 =C. indica). Androsova, 1963 : 23, 39, pi. i, fig. 3, S. China Sea. Cupuladria intermedia Canu & Bassler, 1929 : 74. Hastings, 1932 : 412. Cupuladria brevipora Canu & Bassler, 1929 : 74. Hastings, 1932 : 412. THE CUPULADRIIDAE 171 Cupuladria transversata Canu & Bassler, 1929 : 74, 75, pi. 3, figs. 3, 4. Stn. D 5579, Borneo 175 fath., sand. Cupuladria dentifera Canu & Bassler, 1929 : 74, 75, pi. 3, figs. 5-8, Stn. D 5230, Limasaua Is. 118 fath., sand. Cupuladria tuberosa Canu & Bassler, 1929 : 76, pi. 4, figs. 1-4, Stns. D 5134, Sulu Archipelago, 25 fath., sand, and D 5144, Jolo, 19 fath., sand. Cupuladria grandis Canu & Bassler, 1929 : 77, pi. 4, figs. 10-13, Stn. D 5161, Sulu Archipelago, 1 6 fath., sand. Cupuladria granulosa Canu & Bassler, 1929 : 74, 78, pi. 4, figs. 5-9, Stn. D 5358, Jolo Sea, 39 fath., mud. Kataoka, 1961 : 224, pi. 30, figs. 2, 5, Pleistocene, Japan. Cupuladria hexagonalis Canu & Bassler, 1929 : 74 (as hexagona) 78, pi. 5, figs. 1-4, Stns. D 5141, Jolo, 29 fath., sand, and D 5147, Sulu Archipelago, 21 fath., sand. Cupuladria microdenticulata Kataoka, 1961 : 225, pi. 29, figs, i, 2, Pleistocene, Japan. HOLOTYPE, B.M. 1842. ii. 30. 14, New Guinea. OTHER MATERIAL. Where the predominant type of basal sector present is hexa- gonal (see Text-fig. 6B), the specimens listed below are marked "A", where it is nearly square (see Text-fig. 61)), they are marked " B " ; where pores have been seen, and where tubercles are present, the letters " P " and " T " are added. British Museum. New Guinea 1842. u. 30. 14 (APT). 1859. 2. 25. 11 (BP). S. of Mindanau, 15 fath. 1928. 3. 6. 74 (APT). Torres Straits 1862. 6. 5. 14 (BP), 10 fath., sand. 1882. 2. 23. 37-40 (BP). ?N. Australia 1855. 12. 27. 38 (APT), 1872. 6. 21. 21 (B). Holothuria Bank, N.W. Australia, 34 fath., 1892. i. 28. 125 (BPT). 1963. 2. 2. i (BP), 1891. 12. 16. 1-4 (BP). Cossack Is., N.W. Australia, 8 fath., 1891. 12. 16. 31-37 (BP), Baudin Is., 15 fath., 1891. 12. 16. 21 (BP), 22 (BP), 23 +24 (B + AP), 25-30 (BP + BPT), 1892. i. 28. 128. 8-15 fath., (B + APT). Cape York N.E. Australia, 10 30' S., 142 18' E., 8 fath., sand, Chall. Stn. 186, 1887. 12. 9. 819 (BP). Low Is. off Port Douglas, Gt. Barrier Reef, 1932 . 4. 20. 112 (BPT). Bass's Straits, S. Australia, 1899. 7. i. 1269. Busk Coll. (BPT) Australia 1929. 5. 10. i (AP), 1899. 7. i. 1252 Busk Coll. (labelled C. stellata). Philippines, 1854. I]C - I 5- I 47 (tyP e f C. stellata), 1899. 7. i. 1231, Busk Coll. (AT). Philippines, Stn. D. 5161, 1931. 12. 30. 9 (AP) and Stn. D 5156, off Tinakta Is., 18 fath., 1931. 12. 30. 10 (AP) (topotypes of C. grandis), Canu & Bassler Coll. Sulu Archipelago, " Albatross " Stn., 1)5134, 1963. 3. 14. 2 (AP) ; Suluade Is., Stn. D 5147, 1963. 3. 14. 3 (APT) ; and Philippines, 26 fath., Stn. D 5181, 1963. 3. 14. 4 (AT), Bassler Coll. Photographs of type specimens from the Bassler Collection, United States National Museum, all X 20. C. transversata, Holotype, 1963. 3. 14. 5 and paratypes, 1963. 3. 14. 6 (AP). C. dentifera, Cotypes, 1963. 3. 14. 7 (APT). C. grandis, Syntypes, 1963. 3. 14. 8 (AP). C. granulosa, Syntypes, 1963. 3. 14. 9 (APT). C. hexagonalis, Syntypes, 1963. 3. 14. 10 (APT). C. tuberosa, Syntypes, 1963. 3. 14. ii (AP). Manchester, Museum, Waters Coll. Princess Charlotte Bay, 12 and 13 fath., Australia, (BP). Naturhistoriska Riksmuseet, Stockholm. Forty-five miles W.S.W. off Cap Jaubert, N.W. Australia, 100 m., FBS 157 (BPT), FBS 109 (BPT). S. China Sea, between Nagasaki and Hongkong 75 m., mud, FBS 904 (BPT). Malayan Archi- pelago, Java Sea, 2 30' S., 107 10' E., 27 m., FBS 639 (AT). ZOOL. 13, 5. 9 172 P. L. COOK United States National Museum. The Museum was unable to lend material named by Canu & Bassler, but made available photographs of the type specimens (see above) and also unnamed specimens from various " Albatross " Pacific Stations, 2 of which were the same as those from which some of Canu and Bassler's species were obtained. D 5134 Sum Archipelago, 6 44' N., 121 48' E., (APT) (Locality for C. tuberosa). D 5147 Suluade Is., Sulu Archipelago, 21 fath., (APT) (Locality for C. hexagonalis). D 5145, Jolo, 23 fath. (APT). D 5159, Philippines (AP). D 5181, Philippines 26 fath. (APT). D 5314, China Sea nr. Hong Kong, 122 fath. (AT). Zoarium frequently very large (maximum observed diameter 24 mm.). Zooecia with broad tuberculate cryptocyst, opesiae often almost rectangular. Operculum without a basal sclerite. Central area of zooecia closed by growth of a lamina from beneath the shelf-like salient part of the cryptocyst. Basal surface smooth or tuberculate, divided into quadrilateral or hexagonal sectors, variable. Pores present in peripheral sectors, becoming sunk in pits, occluded or hidden by tubercles. DIMENSIONS. Lz 0-40-0-75 mm., av. 0-60 mm. ; Iz 0-30-0-50 mm., av. 0-40 mm. ; Lop 0-30-0-40 mm., av. 0-35 mm. ; lop 0-15-0-25 mm., av. 0-20 mm. ; Lvo 0-12- 0-20 mm. ; Ls 0-5-1-5 mm. C. guineensis was denned by Hastings (1930 : 714-5), who distinguished it from the C. canariensis-complex of species (Cook (in press)) by the nature of the basal surface and the extent of the cryptocyst. Of the 8 species introduced by Canu & Bassler (1929 : 74-9), two, C. intermedia and C. brevipora, were placed in the synonymy of C. guineensis by Hastings (1932 : 412), who also agreed with Marcus (1921 : 8) in regarding C. stellata Busk as the young form of C. guineensis. Before considering the remaining species introduced by Canu & Bassler, the charac- ters of the material described by Busk in 1854 an d 1884 must be re-examined. Two groups are evident in the large number of specimens available (see list above) ; these differ principally in the pattern of the basal sector boundaries, and fall very roughly, but not exclusively, into two areas of geographical distribution. The first, more northerly form (A), is that originally described by Busk. The basal sectors are hexagonal, pierced by very small pores, which are frequently obscured by tubercles. Busk described the basal surface as " divided into hexagonal areas, each corresponding to a cell ", and his fig. 3 showed irregular, tuberculate, hexagonal sectors ; he did not mention the pores, which are, however, visible at the periphery of the holotype. The sectors, as described above (p. 162) do not, in fact, each correspond to a zooecium, but do have a definite relationship with the zooecia frontal to them (see Text-fig. 6) . Busk's specimen from the Challenger Collection is of a second type (B), the basal sectors being quadrilateral, and the surface smooth. Pores are visible in the peripheral zooecia, which become immersed in pits and finally occluded toward the centre of the colony. The basal surface of his specimen was described by Busk " with shallow radial sulci, ridges quite flat, divided into moveable quadrangular areas, each of which has four or five shallow pits ". Figure 6b showed these, and (p. 207) Busk mentioned that they might " represent occluded pores ". THE CUPULADRIIDAE 173 Busk's description of " moveable " areas may have referred to a dry, loose, basal ecto- cyst, which, like the calcareous surface, shows the sector boundaries, although his specimen from the Challenger Collection has no ectocyst present. However, it may perhaps have been a reference to the variation in shape of the basal sectors of other specimens Busk had examined. Both types of basal sector may be modified by tuberculation which obscures the sector boundaries and pores. Zoaria of the B-type tend to be less tuberculated, and are also usually larger than the A-type. In several cases they show central areas with irregular and hexagonal sectors (e.g. 1891. 12. 16. 23 + 24 and 1892. i. 28. 128). The variation in zooecial size and proportion, the extent of the cryptocyst, and the shape of the zooecial and vibracular opesiae appear to include all the forms listed in the synonymy above. Canu & Bassler introduced C. intermedia for C. guineensis Busk, 1884, which they distinguished from the holotype by the possession of porous basal sectors. As noted above, pores are present in both A and B forms, and this is not a specific distinction. C. brevipora was introduced for C. guineensis Harmer, but the Siboga specimens are not distinct (see Hastings, 1932 : 412). The remaining 5 species described by Canu & Bassler all have hexagonal basal sectors and are only distinguished from each other by the form of the opesiae and the degree of basalt uberculation. Silen (1942 : 8-9) remarked on the variations within the limits of the species and on the similarity of character given in Canu & Bassler's descriptions. One species was introduced on the evidence of one fragment of a zoarium (see Canu & Bassler 1929 : 75), and examination of photographs of the type specimens, and of unsorted Philippine material, shows that all 5 species fall within a continuous range of variability. C. transversata, which was described from a single fragment, showed " a great variety of aspect of the inner face ", and had a smooth basal surface with some peripheral tubercles, and pores. The figure shows a somewhat worn, moderately tuberculate colony, and photographs of the holotype (1963. 3. 14. 5, which comprises only 32 zooecia), show that it has characters extremely like those of the photo- graphs of C. hexagonalis. C. dentifera was described as differing from C. stellata in its tuberose basal surface. Busk's fig. 3 shows C. stellata without basal tubercles, but one of the specimens numbered 1854. I][ - I 5- X 47 i s mounted on glass, and it may be seen that fine tubercles are present. Specimens from the type-locality of C. tuberosa show similar ranges of variation in size of colony, zooecia, opesiae and tuberculation of basal sur- face, as those from the type-locality of C. hexagonalis. Canu & Bassler's figure of the frontal surface of C. hexagonalis is of a worn specimen, and while some of the zoaria from Stns. D 5147 and D 5314 resemble it, others are very similar to the figures of C. transversata. Photographs of the cotype specimens of C. dentifera show that Canu & Bassler's figure (pi. 3, fig. 7) has been retouched, and that the irregular shape of the opesiae is due to the development of the cryptocyst laminae closing the zooecia. C. granulosa, in which the sector boundaries were ' scarcely visible ", is an extremely tuberculated form. The specimen FBS 639 shows this type of development, but the FIG. 6. Basal sectors in Cupuladria guineensis. Ovals in broken lines indicate the position of the frontals of the zooecia on the upper surface. A. " C. dentifera ". Sector boun- daries little developed. B. C. guineensis, " A " form. Hexagonal sector boundaries divided by radial boundaries. C. " C. transversata ". Wide hexagonal sectors where the radial sector boundaries are intermittently obscured. D. C. guineensis, " B " form. Quadrilateral sectors. sector boundaries may be seen at the periphery of the colony, and by lighting the surface obliquely. Heavily tuberculated variants of the granulosa-type are super- ficially similar to C. indica, but may be distinguished by the absence of any radial threads on the basal side (see above). The features distinguishing C. grandis were its large zoarium and smooth basal surface. The largest zoaria of C. guineensis examined are all shallow, with square basal sectors ; the topo types of C. grandis are large, strongly domed, and have hexa- gonal sectors. Domed, B-type zoaria, of similar size (e.g. FBS 157, 18 mm.) and flatter, smaller, A-type colonies (e.g. 1928. 3. 6. 74) indicate that zoarial size and shape are not correlated with type of basal sector, although there is a tendency to reduction of tuberculation in large zoaria of both types. Canu & Bassler figured, but did not comment further on, two distinct forms of vibracular seta in C. grandis ; "recurved" (text-fig. I3A) and "rectilinear" (text-fig. 136). Two kinds of seta have been found in C. canariensis s.s., the differences in which are reflected in the calcareous parts (Cook (in press)). No sign of any such dimorphism has been seen in any of the many colonies with setae, or in the opesiae of specimens of the topotypes of C. grandis and the other specimens attributed to C. guineensis. The seta in Canu & Bassler's text-fig. I3C somewhat resembles that of a regenerating seta figured by Marcus & Marcus (1962, pi. 2, fig. 7). The variation in type and degree of calcification of the curved laminae may result in differing appearances of the basal surface. Where the hollows between the laminae are not filled in by secondary calcification, and where the pores are obscured, the surface is like that described in C. dentifera (see Text-fig. 6A). Where the radial sector boundaries are intermittently obscured, large, wide, hexagonal sectors are produced, as in C. transversata (see Text-fig. 6C). Differential development of tuberculation and pitted pores, together with the effects of wear, produce many other intermediate types of basal surface, but the characters of the zooecia are similar in all zoaria. THE CUPULADRIIDAE 175 The specimens described as C. granulosa and C. microdenticulata by Kataoka (1961 : 224, 5) appear to be referable to C. guineensis. The basal surface of both species had sectors, which were porous in C. microdenticulata, and the zooecia had the well-developed cryptocyst of C. guineensis. The zoarium of C. guineensis attains a larger size than that of any other species in the Cupuladriidae. The largest colonies (Torres Straits 1862. 6. 5. 14 and S. China Sea FBS 904) are 24 mm. in diameter, and many of the other zoaria, especially those with square basal sectors, exceed 15 mm. in diameter. The size of zooecia varies considerably, but the Lz does not seem to exceed 0-75 mm. and the Iz 0-50 mm. The range of Lz given by Canu & Bassler for C. transversata includes the dimensions given for all the other species. Measurements taken from their figures indicate that the Lz for C. dentifera, C. grandis and C. gmnulosa may be as low as 0-50 mm., which is the figure given in the text for the Lz of C. transversata, C. tuberosa and C. hexagonalis. The opesiae of these last 3 species are rectangular in contrast to the opesiae of the first 3 species, which are often oval, but A-type zoaria having zooecia with both forms of opesiae are present from Stn. D 5181, and B-type zoaria with the same degree of opesial variation from the Holothuria Bank 1963. 2. 2.1, and Low Island, 1932. 4. 20. 112. Generally the more peripheral zooecia of the larger zoaria of both types tend to have more oval opesiae and more steeply descending cryptocysts. Some zooecia have distinctly serrated cryptocysts. As noted above, the vibracular setae appear to be of one kind, and are similar to those found in all other species of the family. Proportionately they are shorter than those of Discoporella umbellata, the larger zoaria of which approach those of C. guineensis in size. Two forms of C. guineensis therefore exist, but the characters of each are not exclusively correlated and cannot be regarded as specific. ii THE DISCOPORELLA U M B ELL AT A -COMPLEX The characters defining species in the Cupuladriidae, where the common zoarial form determines similarity in the development of all structures, must of necessity be judged using fine criteria. To define a species adequately there must be a recog- nizable degree of correlation between each character considered, morphologically, geographically, and if possible, palaeontologically. In each of these respects it is possible to divide records of D. umbellata into two distinct groups. The problem is whether these groups are of specific or subspecific rank, and the principal criterion applied below is that of character-consistency throughout the time range of the group considered. Canu & Bassler (1923, text-fig. 13) noted the variability of their material assigned to D. umbellata, and figured wide zooecia in Miocene material from Santo Domingo (pi. 2, fig. 18) . The part of the zoarium figured showed the growing edge of a regener- ating fragment, where the zooecia of all species are frequently wide and slightly distorted (see pp. 155, 158). Photographs of this material (1963. 3. 14. 12) show that subsequent zooecia are narrower. Canu & Bassler's figures (including fig. 17) emphasized the different appearance of Recent specimens from Oran (p. 76, text- 176 P. L. COOK fig. I3A) and the increase in size of the zooecia compared with those of Miocene material from France (text-figs. 136-6). C. robertsoniae was introduced by Canu & Bassler for specimens named C. canarien- sis by Robertson (1908 : 314). Her figures showed narrow zooecia with 5-7 opesiules and were certainly referable to the D. umbellata-complex (see Hastings, 1930 : 719). In renaming Robertson's specimens Canu & Bassler were apparently aware of basic differences between " typical " D. umbellata and their Pleistocene material from the Pacific coast, which was included in C. robertsoniae, as they gave the narrow zooecia and large, proximally curved opesia as diagnostic characters. Comparison of Canu & Bassler's figures of their Pacific Pleistocene material (1923, pi. 34, figs. 5-7) with those they gave of West Indian Miocene specimens (1923, pi. 2, figs. 15-19, as C. umbellata) shows striking similarities and indicates a strong affinity between the two groups of specimens. Osburn (1950 : 113) concluded that the wide range of variation found in American specimens justified their inclusion in one species with European D. umbellata, but apparently he did not compare his specimens with European material. Generally, fossil material from the Miocene of Europe and America shows con- sistently correlated differences in character, which have become even more divergent in Recent specimens from the two areas (see PI. 3, figs. 1-4). Applying the criteria used in defining the other species described above, the American fossils cannot be included in typical D. umbellata, and, as the Recent specimens differ still further, all American records have been separated here as D. umbellata subsp. depressa (Conrad). The constant characters of the European material are the wide zooecia, the presence of a pair of opesial denticles, the proportionately large size of the vibracular opesiae, and the numerous small opesiules. There has been an increase, in size in Recent specimens, where the proximal edge of the vibracular opesia is denticulate, and the smaller pores in the centre of the horizontal cryptocyst persist. The constant characters of the American material are the narrow zooecia, the non- denticulate opesiae, with curved proximal edge, the relatively small size of the vibra- cular opesia, which is never denticulate, and the smaller number of large opesiules. Some fossil specimens have central cryptocystal pores, but these are rarely persistent in Recent material. In contrast to D. umbellata, the zooecia of fossil specimens of D. u. depressa are, on average, larger than those of Recent colonies. It might be argued that the narrower zooecia of D. u. depressa themselves directly result in the smaller number of opesiules and the lack of central cryptocystal pores, but, as the two species exhibit similar ranges in size and shape of zoarium, there must be some inherent difference producing the narrower zooecia of D. u. depressa. The nature of the zooecial and vibracular opesiae are, however, not affected by conditions of growth or width of zooecia, and these are the most important features which show correlated character-consistency in the two forms. When the peyroti-type zoaria listed below are included, records of D. umbellata show three separate mean types of population. These forms may overlap at the ends of their ranges of variation, but the Recent forms of each group are geo- graphically distinct. By the application of a discriminant function technique to THE CUPULADRIIDAE 177 further morphometric studies, it is intended to investigate the nature and degree of relationship of the members of this complex. 12 Discoporella umbellata (Defrance) (PL i, fig. 7 ; PL 3, figs, i, 3, 5, 6 ; Text-fig. 4) Lunulites umbellata Defrance, 1823 : 361, pi. 47, figs. I, la, ib. " Lunulite en-parasol " de Blainville, 1834 : pi. 72, fig. la, ib. Discoporella umbellata (Defrance) d'Orbigny, 1853 : 473, pi. 717, figs. 1-5. Lagaaij, 1953 : 16, pi. i, fig. 3, Miocene, Netherlands. Buge (part), 1957 : J 77> pi- IO - n g- 4- Miocene, S.W. France. Discoporella berardana d'Orbigny, 1853 : 474, Recent, Algeria, 14 m. Cupularia lowei Busk, 1854 : 99, pi. 116, figs. 1-6, Madeira. C. lowei Gray (sic) Norman, 1909 : 290, pi. 37, figs. 7-12, Madeira 50-70 fath. Waters, 1921: 412, pi. 30, figs. 1-6, 26-29, Madeira. Cupularia umbellata (Defrance) : Manzoni, 1869 : 26, pi. 2, fig. 16, Pliocene, Italy. 1875 : 39, pi. 5, fig. 67, Miocene Italy. Calvet 1907 : 393. Canaries, 80 m., and Cape Verde Islands, 1900 m. Canu, 1909 : 448, pi. 16, figs. 16, 17, Burdigalien, Helvetien, Miocene, S.W. France. Waters, 1921 : 414, Pliocene, Italy. Canu & Bassler, 1923 : text-figs. I3A, Oran, Recent ; B. C. D. E France Miocene, ?I Oran, Recent (as C. multispinata) . Peyroti-type zoaria : Cupularia peyroti Duvergier, 1921 : 124. 1924 : 19, pi. i, figs. 6 10, Helvetien, Miocene, France. Cupularia umbellata Defrance : O'Donoghue, 1924 : 39, Cape St. Blaize, S. Africa. Discoporella umbellata (Defrance) part Hastings, 1930 : 719, specimens from S. Africa. MATERIAL. British Museum. Bay of Funchal, 1892. 8. 6. i (type material of C. lowei}. Madeira, 1911. 10. i. 648, 1004 and 1963. i. 16. 10 and n, Norman Coll. Madeira, 1932. 8. 23. i. Canaries, 1899. 7. i. 1250, Busk Coll. Near Cape Frio, Angola, 12 E., 18 30' S., 1899. 7. i. 1241, Busk Coll. (doma-type zoarium). Palaeontology Dept. Coll. D. 25005, 25003, Burdigalian, L. Miocene, Bordeaux. D. 39047-96, Miocene, Reinbeck, Germany. D. 38947-9, Miocene, Beeringen. D. 39884-91 (pt), Albena, Bordighera, Pliocene, Italy. Voigt Coll. Reinbeck, L. Miocene. Boring, Bockstedt, Middle Miocene (i doma- type zoarium present). Peyroti-type zoaria. MATERIAL. British Museum. Cape St. Blaize, S. Africa, 1923. 7. 26. 9, O'Donog- hue Coll. S. Africa, 1963. i. 16. 9, O'Donoghue Coll. S. Africa, 34 29' S., 21 02' E., 26. ii. 59, 73 m., sand, SCD ii4C, 1963. 1. 16. 6 ; 34 29' S., 22 18' E., 28. 8. 60, io6m., sand, SCD 1470, 1963. i. 16. 7; False Bay, 31. i. 59, 50 m., FAL 34iD (B), 1963. i. 16. 8, Univ. Cape Town Ecological Survey Coll. Cape of Good Hope, 1842. 12. 2. 20, Belcher Coll. S. Africa, Cape Infanta N.E. by N|N, 13-5 miles, 43 fath., 1949. ii. 10. 618 ; Stalwart Point N.N.W. 9 miles, 53 fath., 1949. n. 10. 652A ; Nanquas Peak N. 46 E., 19 miles, 63 fath., 1949. ii. 10. 674A, and N.E. |N., 15 miles, 59 fath., 1949. ii. 10. 675 (including 2 doma-type zoaria), N.W. N. 15 miles 49 fath., 1949. ii. 10. 668 ; Cape Infanta N.E. by Nig miles, 46 fath., 1949. ii. 10. 622A, 623A, N.E. by N.|N. 14-5 miles, 45 fath., 1949. ii. 10. 620 ; Cape St. Blaize N. 7-5 miles, 37 fath., 1949. ii. 10, 624A, 62$A, 626A, 62^A and 94A, N.E. JN. n 178 P. L. COOK miles, 40 fath., 1949. n. 10. 629 ; Tugela River mouth, N.N.W. 11-75 miles, 37 fath., 1949. ii. 10. 644 (including 2 doma-type zoaria) ; Umtwalumi River mouth N. by W. 7 miles, 50 fath., 1949. 11. 10. 651 ; False Island N. 8-5 miles, 53 fath., 1949. 11. 10. 66gA. ; Glendower Beacon N.|W. 6 miles, 39 fath., 1949. n. 10. 660, Burrows Coll. John Murray Coll. (all from sandy substrates). Stn. 103, 5 39' 30" S., 39 n' 30" E.,N. of Zanzibar, ii. 1.34. 101 m., Z. 198 C. Stn. 104., 5 37' 54" S., 39 n' 36" E., ii. i. 34, 207 m. Z. I99A. Zoarium usually large (maximum observed diameter 22 mm.). Zooecia often as wide as long. Opesia small, with straight proximal border with a pair of minute denticles. Operculum with a straight proximal sclerite. Descending cryptocyst wide, tuberculate, horizontal cryptocyst lamina with some small irregular central pores and 8-14 small opesiules, frequently slit-like. Vibracular opesia very large (X 1-75 of Lop), finely denticulate proximally in Recent specimens. Basal surface with short grooves, pits and ridges, tuberculate, or smooth, with a layer of keno- zooecia and vibracula encroaching from the frontal surface. DIMENSIONS. Lz 0-50-0-70 mm. ; Iz 0-40-0-63 mm. ; Lop 0-10-0-15 mm. ; lop 0-12-0-20 mm. ; Lo o-io-o-n mm. ; lo 0-12-0-15 mm. ; Lvo 0-18-0-25 mm. ; Ls 0-6-1-45 mm. The records of European Miocene and Pliocene specimens agree in describing and figuring D. umbellata with wide zooecia and a large number of opesiules. Although generally the zooecial measurements of fossils are smaller than those of Recent specimens, the proportions and general characters of the species seem constant, and are certainly distinct from those of D. u. depressa. The majority of zooecia are as wide or occasionally wider than long, and the number of opesiules is most frequently between 8 and 10. Defrance figured 8-9, d'Orbigny 10, de Blainville 8-9, and Manzoni (1869), 10-11, although the number given in 1875 was 7. In Recent specimens the number is often as many as 14, and the opesiules are slit-like. D'Orbigny described, as D. berardana, Recent specimens from Algeria with denticulate cryptocysts, transverse, semi-lunar orifices, and very large vibracular opesiae. D. umbellata was figured by Canu & Bassler (1923, text- fig. I3A) from Oran, and it is possible that d'Orbigny's record referred to worn speci- mens of this species. Canu & Bassler's photograph of C. multispinata (1923, text- fig. 13!.) showed opesiules and a basal sclerite in the operculum, and is perhaps also of D. umbellata . The specimens described as C. lowei were distinguished from Z). umbellata by the deformation of the zoarium (see p. 158, and Marcus & Marcus, 1962 : 294). The paired denticles on the proximal edge of the opesia are present in all well- preserved fossil and Recent material. Waters (1921 : 415) noted their occurrence, and Lagaaij (1953 : 16) included them in his diagnosis of the species. The proximal edge of the vibracular opesia of Recent specimens is finely denticulate (see pi. 3, fig. i), a character which has not been seen in fossil material. The vibra- cular opesiae are extremely large, as noted by d'Orbigny (1853 : 474, in D. berardana}, their length being significantly greater than that of the zooecial opesiae, THE CUPULADRIIDAE 179 The operculum is strongly flanged laterally, with a straight, robust, proximal sclerite. The specimen from Cape Frio has a doma-type zoarium. The basal surface is solid and there are peripheral kenozooecia, but the zooecia are worn, and the character of the opesiae cannot be clearly seen. However, the width of the zooecia (Iz 0-3-0-48 mm.) and the number of the opesiules (5-7), are nearer the higher end of the range of variation than those of many of the corresponding doma-type specimens from the west Atlantic (see below), and thus reflect the differences found in zooecia of normal zoaria of the two forms. One fossil zoarium (Voigt Coll.) is also extremely worn, but shows two rows of peripheral vibracula and a basal surface almost filled by secondary calcification. The development of the cryptocyst in the young peripheral zooecia shows the close relationship of D. umbellata both to D. u. depressa and to the denticulate species of Cupuladria (group C). Fine, irregular denticles grow first from the proximal and lateral edges of the cryptocyst and then develop progressively in a distal direction. At this stage the zooecia greatly resemble specimens of C. owenii and C. doma. Further calcification produces a distal bar cutting off the opesia, and, as the more proximal denticles fuse, leaving a series of denticulate, irregular opesiules, and a number of pores in the centre of the horizontal cryptocyst lamina thus formed (see also Marcus & Marcus, 1962, pi. i, fig. 4). The appearance of the zooecia at this stage is then very similar to that of Miocene specimens of D. u. depressa from Oak Grove, Florida, where this type of development was the final stage. Other specimens of D. u. depressa from the Miocene of Cercado de Mao, Santo Domingo, show more complete calcifica- tion, with rounded opesiules and no central pores. The calcification of eastern Atlantic fossil and Recent specimens of D. umbellata is even greater, and the opesiules are further reduced. The distribution of D. umbellata s.s. is now confined to the northern and north western coasts of Africa, and the neighbouring islands. D. umbellata and C. haidingeri are frequently associated in samples from the European Miocene. When worn they are extremely difficult to distinguish from each other. The basal surface of D. umbellata is not as regularly grooved as that of C. haidingeri, and generally the zooecia of D. umbellata are wider. Even when the central part of the cryptocyst lamina is broken, the vestibular arch surrounding the distal part of the opesia is smaller and more incurved in D. umbellata than in C. haidingeri. At the edge of the central area of closed zooecia the most recent closured frequently show the straight bar, proximal to the opesia, which is absent in C. haidingeri. Recent specimens from south and east Africa correspond almost exactly with C. peyroti as described by Duvergier. The dimensions he gave fall within the range of those now obtained from a large number of specimens, and the photograph shows kenozooecia exactly like those of the Recent material. In one respect, Duvergier's account may be confusing, as he followed Canu is describing the frontal, convex surface of the colony as basal, and the concave or planar kenozooecial surface as frontal, Q'Donoghue mentioned the invasion of the basal surface by kenozooecia 180 P. L. COOK and vibracula, and Hastings, who examined some of O'Donoghue's material, expressed doubt as to its inclusion in D. umbellata s.s. The Miocene specimen of D. umbellata figured by Buge shows a hollow basal surface covered by kenozooecia and vibracula. The specimens from south Africa show the complete range of development, those from east Africa, the young stages only. Young zoaria are flat, with a solid, smooth base, and older colonies are more domed, with the basal surface either concave (e.g. 1963. i. 16. 6-9, 1949. n. 10. 625A) or completely filled with secondary calcification (e.g. 1949. ii. 10. 66gA and 652A). In neither state is the surface ridged or tuber- culate. The budding of the basal kenozooecia appears to be random and irregular, and the calcified cryptocyst is developed like that of the zooecia, from the fusion of irregular denticles. The zooecial opesiae are often slightly curved proximally, and the denticles are often reduced or absent, but the opesiae are never sinuate as in D. u. depressa. The horizon- tal cryptocyst lamina differs from that of typical D. umbellata in the large number of frontal pores, which are often indistinguishable from the marginal opesiules. The operculum is like that of D. umbellata, and is flanged laterally, with a strong, straight proximal sclerite. A few colonies combine some characters of both doma- and peyroti-type growth. They are small (diameter 2 mm.) and high, with solid, smooth bases. The peripheral zooecia are closed, but only one row of peripheral vibracula is present, and there is no invasion of the basal surface by kenozooecia. The zooecia are small, approximately half the length of zooecia in a normal colony, but the opesiae are not reduced in proportion, being about f of the length of normal opesiae. The zooecia have an average of 8 opesiules and a number of irregular frontal pores in the cryptocyst. The proximal edge of some vibracular opesiae are finely denticulate. The doma-type colonies thus have many of the characters of the normal zoaria from the south east African coast. Their occurrence does not seem in this case to be correlated with low temperature, as all the records are confined to warm waters, extending from the warm-temperate zone, eastward to the subtropical zone, as described and discussed by Stephenson (1947 : 215). As noted above (p. 163), no specimens of typical D. umbellata have been found from the south and east coasts of Africa, nor have peyroti-type zoaria been recorded from the west African coast. Specimens of Cupuladria owenii (Gray) and C. multi- spinata (Canu & Bassler), both of which also occur from these two areas, show no examples of peyroti-type growth. 13 Discoporella umbellata subsp. depressa (Conrad) (PI. 3, figs. 2, 4) Lunulites depressa Conrad, 1841 : 348, Miocene, N. Carolina. Discoporella denticulata (Conrad) Gabb & Horn, 1862 : 142, pi. 20, fig. 25, Miocene, New Jersey to S. Carolina. THE CUPULADRIIDAE 181 Cupularia umbellata (Defrance) Smitt, 1873 : 14, pi. 3, figs. 75-80, Cape Fear River, N. Carolina, 7 fath. C. umbellata (Manz.) (sic) : Verrill, 1878 : 305. Canu & Bassler, 1918 : 118, pi. 53, figs. 2-4, 1919 : 85, pi. i, figs. 5-7, pi. 2, figs. 17-21. 1923 : 68, pi. 2, figs. 15-19, text-fig. joG, Lower Miocene, west Indies. 1928 : 64, pi. 7, figs. 1-3, Gulf of Mexico. 1929 : 142, pi. 15, figs. 5-11, text-figs. 35A-K (as C. lowei), between California and Hawaii, 2723 fath. 1930 : n, Galapagos and Hawaii. Sil6n, 1942 (part) : 15, text-figs. 10-12, localities 1-17. Cupularia punctata Canu, 1904 : 10, pi. 2, fig. 23, Miocene, Patagonia. Cupularia canariensis Busk : Robertson, 1908 : 314, pi. 24, figs. 90, 91, California. Cupularia lowei Busk : Osburn, 1914 : 194, Florida, 12 and 22 fath. Cupularia robertsoniae Canu & Bassler, 1923 : 82, pi. 34, figs. 5-7, Pleistocene, California. Discoporella umbellata (Defrance) Hastings, 1930 : 718, pi. n, fig. 54, Gorgona, Galapagos, Balboa, 15-30 fath. Mcguirt, 1941 : 65, pi. i, figs. 4, 7, 9-11, Miocene, Louisiana. Osburn, 1950 : 113, pi. n, figs. 7-10, California to Ecuador, 5-100 fath. Soule, 1959 : 34, Gulf of California, 5-46 fath. Soule & Duff, 1957 : 99> Pleistocene, California. Marcus & Marcus, 1962 : 290, pi. i, fig. 4, pis. 2-5, Brazilian coast, 4-150 m. MATERIAL. British Museum. Gorgona, 1929. 4. 26. 102, 103, 106 ; Balboa, 1929. 4. 26. 104 ; Galapagos, 1929. 4. 26. 105. " St. George " Coll. Galapagos, 1933. 12. 10. 15, Canu & Bassler Coll. Abrolhos Bank, off Brazil, 1899. 7. i. 1260, 1261 and 1265 (doma-type zoarium). Berbice, British Guiana, 25 fath. (21 zoaria, 4 of doma-type). ' Terra Nova " Coll. Stn. 42, off Rio de Janeiro. Palaeontological Dept., D 33911, D 33912, D 33916, D 33910, Miocene, San Domingo. D 41118, Bowden marl, Miocene, Jamaica. D 34392-409, Miocene, Duplin, S. Carolina. Cheetham Coll. Miocene, Oak Grove, Okaloosa Co., Florida. Philadelphia Academy Coll., Wilmington, N. Carolina. Zoarium large (maximum diameter observed 18 mm.) . Zooecia generally narrower than in D. umbellata, lateral cryptocyst descending more steeply. Horizontal crypto- cyst lamina not usually porous at centre (in Recent specimens) . Opesia large, curved proximally or slightly sinuate, denticles absent. Opesiules large, 3-9 in number. Operculum with delicate basal sclerite curved distally. Vibracular opesiae little larger than zooecial opesiae, not denticulate proximally. Basal surface tuberculate with grooves, or short furrows and pits. DIMENSIONS. Lz 0-45-0-7 mm. ; Iz 0-22-0-35 mm. ; Lop 0-11-0-13 mm - ) 1P 0-12-0-14 mm. ; Lo 0-07-0-08 mm. ; lo o-io-o-n mm., Lvo 0-15-0-18 mm. ; Ls 0-60-1-10 mm. Conrad's description of L. depressa mentioned that " many of the larger pores " (i.e. zooecia) were " filled with a minutely porous plate or diaphragm, solid in the centre ". Specimens examined from the Miocene of Wilmington, N. Carolina, have narrow zooecia, with 7 large opesiules and an opesia with a curved proximal edge. A few central pores may be seen from the underside of the cryptocyst, but most of these are occluded by secondary calcification. Natural Well, the type locality, and Wil- mington, are near the Cape Fear River, from which Smitt described the first Recent specimens from Pourtales's Collection. Canu's figure of C. punctata showed the frontal surface of a worn specimen ; it had the characters of D. u. depressa, with 3-7 opesiules, and a rounded opesia. The magnification of some of the figures of D, u. depressa (as D, umbellata} given in i8a P. L. COOK two of the works of Canu & Bassler is confusing. The same series of photographs were used throughout ; but those of 1919 (pi. 2, fig. 17 (nat. size), figs. 18 (X25), 19, 20 and 21 (xig)) are smaller than those of 1923 (pi. 2, figs. 15, 16, 17, 18, 19) which are described as having the same magnification. Osburn (1950) confirmed that Verrill's record (1878) of D. umbellata from Fort Macon, N. Carolina, was correct, as he had examined material from the same locality. The zooecia of D. u. depressa are generally narrower than those of D. umbellata. The figures of American material all agree in showing narrow zooecia with large, rounded opesiae and opesiules, and Soule & Duff described the zooecia as " elongate ". The opesia is always curved proximally, and may be distinctly sinuate ; denticles are absent. The bars between the opesiules are narrow. The proximal sclerite of the operculum is delicate and slightly curved in a distal direction. The vibracular opesiae are never denticulate proximally, as in D. umbellata , and proportionately, they are smaller ; the setae do not appear to attain the length of those of D. umbellata. Most of the 7 doma-type colonies examined have large, rounded opesiae and narrow zooecia (Iz 0-20-0-35 mm.), with 2-7 opesiules, and are thus similar to more normal specimens of D. depressa. They differ from the doma-type colony of D. umbellata principally in their narrower zooecia and larger opesiae. 14 CUPULADRIIDAE, GENERAL COMMENTS The number of species of Cupuladriidae from the Pacific Ocean is thus two, C. guineensis and D. u. depressa, which last is confined to the eastern part. Two species occur in the Indian Ocean ; C. indica in the north and west, and D. umbellata (peyroti form) in the south-west. Ten forms may be distinguished in the Atlantic Ocean (Cook (in press)), only the subspecies D. u. depressa occurring in both the Atlantic and Pacific. The family is sharply defined, its principal characters being the distinctive asto- geny, radial development and type of vibracula. Species in which any character shows a major discrepancy should be assignable to other families. This is demon- strated by examination of the characters of Cupularia capriensis Waters. HINCKSINIDAE Canu & Bassler 1927 Hincksinidae Canu & Bassler, 1927 : 3, Bassler, 1953 : G 159, Lagaaij, ig63b : 171. SETOSELLINA Calvet, 1906 : 157 15 Setosellina capriensis (Waters) (PL i, figs. 5, 6) not Setosellina roulei Calvet, 1907 : 395, pi. 26, figs. 5, 6, Cape Verde Is., 1900 m., Capo Blanco, 2330 m. " Cupularia minima (Busk MS) " Waters, 1921 : 419, as Heliodoma implicata Calvet, specimens from the Aegean Sea. (not H. implicata, see below). Setosellina roulei Calvet : Waters, 1925 : 350, pi. 21, fig. 4, Capri, 50 fath. ; Gautier, 1962 : 69, Tunisia, 150 m., Rhone Delta, 120 m. THE CUPULADRIIDAE 183 Cupularia capriensis Waters, 1926 : 432, pi. 18, figs. 8, g, Capri. Cupuladria capriensis (Waters) Marcus & Marcus, 1962 : 288. LECTOPARATYPES, chosen here, Manchester Museum, Faraglione, Capri, 150 fath., Waters Coll. OTHER MATERIAL. British Museum. Aegean Sea, 130 fath., 1899. 7. i. 1253 and 5122 ; 100 fath., 1899. 7. I. 1251, collected Spratt. Ras el-Amouch, 1899. 7. i. 4701 ; Mediterranean, 30 fath., 1899. 7. i. 1491, collected " Porcupine ". Busk Coll. Zoarium small, selenariiform, free, growing on a sand-grain. Zooecia with oval opesia. Each zooecium with a distal interzooecial vibraculum, reaching the basal wall of the zoarium. Opesia of ancestrular and central area closed by a calcined lamina with a single central pore. Ancestrula with one distal and one proximal- lateral vibraculum, and surrounded by 6 zooecia, budded spirally. DIMENSIONS. Lz 0-45-0-60 mm., av. 0-54 mm. ; Iz 0-30-0-34 mm., av. 0-32 mm. ; Lop 0-32-0-35 mm., av. 0-33 mm. ; lop 0-17-0-20 mm., av. 0-19 mm. Waters (1921) described, as Heliodoma implicate Calvet (1906 : 157, 1907 : 396, pi. 26, figs. 7, 10), some specimens from the Aegean Sea in the Busk Collection (1899. 7. i. 1253). He particularly noted that the form of the zoarial spiral differed from that figured by Calvet. The zooecia of the ancestrular area of H. implicata are not occluded by a lamina, which is however, present in the specimens on slide 1899. 7. i. 1253. These zoaria are identical with those from Capri described by Waters in 1926 as Cupularia capriensis, which has almost complete opesial closures in the ancestrular area, a small central pore only remaining open. Waters does not at any time appear to have considered the Aegean specimens either as distinct from H. implicata or as identical with his new species, C. capriensis. His quotation of Busk's MS name for them in 1921 cannot therefore be regarded as a specific designation. C. capriensis Waters is certainly referable to Setosellina Calvet. Calvet (1907 : 395) described Setosellina roulei with a salient calcareous lamina and figured (pi. 26, fig. 6) zooecia which were partially closed, with a small central pore. The early astogeny of specimens of 5. capriensis was described (as 5. roulei} by Waters (1925 : 350, pi. 21, fig. 4, Capri, 50 fath.). The budding is similar to that of Otionella and Selenaria figured by Stach (i936b : 64, text-figs. 1-7). The same type of spiral budding is found in Setosellina constricta (Borneo Bank, 1928. 3. 6. 75 and N. New Guinea, 1928. 3. 6. 76) described by Harmer (1926 : 264), and 5. goesi (Silen) des- cribed by Lagaaij (i963b : 172). The basal surface of 5. capriensis shows that the vibracular chambers are inserted between, and extend to the same depth as the zooecia. They are thus truly interzooecial, and completely unlike those found in the Cupuladriidae. Gautier (1962 : 69) commented on the probable identity of C. capriensis Waters with Setosellina roulei Calvet. The dimensions of the abyssal specimens figured by Calvet are significantly smaller thanthoseof 5. capriensis given above. They average : Lz 0-36 mm. ; Iz 0-25 mm. ; Lop 0-23 mm. ; lop 0-13 mm. Measurements taken from 184 P. L. COOK Waters's figure of 5. roulei from the Mediterranean (1925, pi. 21, fig. 4), combined with those given by Gautier, give a range of Lz 0-45-0-50 mm. ; Iz 0-32-0-35 mm. ; Lop 0-25-0-28 mm. ; lop 0-16-0-18 mm. These dimensions are far closer to those of S. capriensis given above than to those of 5. roulei as originally described. The geographical and bathymetrical distributions of the two forms are also distinct ; S. roulei occurring off the Cape Verde Islands at 1900 m., and off Capo Blanco at 2330 m. The larger S. capriensis has been recorded from the Mediterranean and Aegean only, at a greatest depth of 150 m. Until further material from both areas can be examined, all the Mediterranean records are here considered to be distinct and referable to 5. capriensis. S. capriensis differs from the Cupuladriidae in its early astogeny, and related to this, in the spiral development of the zoarium. The relationship of the vibracular chamber to the zooecium, and the type of opesial closure of the central zooecia, are also fundamentally different from those found in the Cupuladriidae. 16 ACKNOWLEDGMENTS My thanks are due to Dr. A. Anderson (Naturhistoriska Riksmuseet, Stockholm), Dr. R. S. Boardman (Smithsonian Institution, Washington) and Dr. D. E. Owen (Manchester University Museum), for their generous assistance in lending many of the specimens examined. The photographs were taken by Messrs J. V. Brown and P. Green at the British Museum (Natural History), and at the Smithsonian Institution (see PI. i for details). I am also indebted to Dr. A. Cheethan (Louisiana State University), Dr. R. Lagaaij (Shell Exploratie en Produktie Laboratorium, The Netherlands), and Herr Prof. Dr. E. Voigt (Geologisches Staatsinstitut, Hamburg), for much helpful discussion and correspondence. Lastly, I wish to thank Dr. A. B. Hastings and Dr. J. P. Harding (British Museum, Natural History) for their valuable help and encouragement throughout the course of this work. 17 SUMMARY The principal characters and mode of development of the Cupuladriidae are described and discussed, the family is defined, and groups of species with similar characters are listed. The effects of zoarial form of zooecial characters, and the occurrence and possible causes of abnormal zoaria similar to those found in C. doma, and those described as C. peyroti, are discussed. Descriptions are given of C. pyriformis, C. indica and C. guineensis. The complex of forms previously assigned to D. umbellata is examined in the light of character-consistency of fossil and Recent records. Cupularia capriensis Waters is described and referred to Setosellina. 18 REFERENCES ANDROSOVA, H. 1963. Bryozoa of the South China Sea. Studia Marina Sinica, 8, 4 : 21-45. BASSLER, R. S. 1953. Treatise on Invertebrate Paleontology part G : Gi-G253, University of Kansas. DE BLAINVILLE, H. M. 1834. Manuel d'Actinologie ou de Zoophytologie, 2 vols., Paris. BROWN, D. A. 1954. O n tne Polyzoan Genus Crepidacantha Levinsen. Bull. Brit. Mus. (N.H.) Zool, 2, 7 : 243-263. THE CUPULADRIIDAE 185 BUGE, E. 1957. Les Bryozoaires du Neogene de 1'ouest de la France. Mem. Mus. Nat. hist. Nat., M.S., C, 6 : 1-436. BUSK, G. 1854. Catalogue of the Marine Polyzoa in the . . . British Museum part 2 : i- vii, 55-120, London. 1884. Report on the Polyzoa, The Cheilostomata. Rep. Zool. Chall. Exp., 10, part 30. CALVET, L. 1906. Note preliminaire sur les Bryozoaires recueillis par les expeditions du Travailleur (1881-1882) et du Talisman (1883). Bull. Mus. Hist. nat. Paris, 12, 3 : 154-166. 1907. Bryozoaires, Exped. Sci. Travailleur et Talisman 1880-1883, 8 : 355-495. CANU, F. 1904. Les Bryozoaires du Patagonien. Mem. Soc. geol. Fr., 12, 3 : 1-30. 1909. Les Bryozoaires fossiles des Terrains du Sud-Ouest de la France. Bull. Soc. geol. Fr., ser. 4, 9 : 442-458. CANU, F. & BASSLER, R. S. 1918. Bryozoa of the Canal Zone and related areas. Bull. U.S. nat. Mus., 103 : 117-122. - 1919. Fossil Bryozoa from the West Indies. Publ. Carneg. Instn, 291 : 75-102. - 1923. North American Later Tertiary and Quaternary Bryozoa. Bull. U.S. nat. Mus., 125 : 1-302. 1927. Classification of the Cheilostomatous Bryozoa. Proc. U.S. nat. Mus., 69, 14 : 1-42. 1928. Fossil and Recent Bryozoa of the Gulf of Mexico. Ibid., 72, 14 : 1-199. 1929. Bryozoa of the Philippine Region. Bull. U.S. nat. Mus., 100, 9 : 1-685. 1930. The Bryozoan Fauna of the Galapagos Islands. Proc. U.S. nat. Mus., 76, 13 : 1-78. CONRAD, T. A. 1841. Appendix to Mr. Hodge's paper describing the New Shells, etc. Amer. J. Sci. (Silliman's Journal], 41 : 344-348. COOK, P. L. 1963. Observations on live lunulitiform zoaria of Polyzoa. Cah. Biol. mar., 4 : 407-4I3- DARTEVELLE, E. 1932. Contribution a l'6tude des Bryozoaires fossiles de 1'Eocene de la Belgique. Ann. Soc. roy. Zool. Belg., 63 : 55-116. 1935. Zoarial Regeneration of free Polyzoa. Ann. Mag. nat. Hist., (10) 15 : 559-561. DEFRANCE, J. L. M. 1823. Dictionnaire des Sciences Naturelles, 27, Paris. DUVERGIER, J. 1921. Note sur 1'Afneurement de Falun de Lalande a Mios. P.V. Soc. Linn. Bordeaux, 73 : 123-130. - 1924. Deuxieme note sur les Bryozoaires du Neogene de 1'Aquitaine. Act. Soc. Linn. Bordeaux, 75 : 5-50. GABB, W. M. & HORN, G. H. 1862. Monograph of the Fossil Polyzoa of the Secondary and Tertiary Formations of North America. /. Acad. nat. Sci. Philad., 5, 2 : 111-178. GAUTIER, Y. V. 1962. Recherches ecologiques sur les Bryozoaires Chilostomes en Mediterranee Occidentale. Theses presentees a la Faculte des Sciences de I'Universite d'Aix-Marseille, 91 : 9-434- HARMER, S. F. 1926. The Polyzoa of the Siboga Expedition, Part 2, Cheilostomata, Anasca. Rep. Siboga Exped., 28b : 181-501. 1957. Op. cit., Part 4, Cheilostomata, Ascophora. Ibid., 28d : 641-1147. HART, T. J. & CURRIE, R. I. 1960. The Benguela Current. Discovery Rep., 31 : 123-298. HASTINGS, A. B. 1930. Cheilostomatous Polyzoa from the vicinity of the Panama Canal . . . Proc. zool. Soc. Lond., 1929, 4 : 697-740. - 1932. The Polyzoa, with a note on an associated Hydroid. Rep. Barrier Reef Exp., 4, 12 : 399-458. HINCKS, T. 1887. On the Polyzoa and the Hydroida of the Mergui Archipelago. /. Linn. Soc. (Zool.), 21 : 121-135. JACOB, K. & SASTRI, V. V. 1953. Miocene Foraminifera from Chavara, near Quilon, Travan- core. Rec. Geol. Surv. India, 82, 2 : 342-353. KATAOKA, J. 1961. Bryozoan Fauna from the " Ryukyu Limestone " of Kikaijima, Kago- shima Prefecture, Japan. Sci. Rep. Tohoku Univ., ser 2, Geol., 32, 2 : 213-272. LAGAAIJ, R. 1952. The Pliocene Bryozoa of the Low Countries. Meded. Geol. Sticht., G5, 5 : 1-233- i86 P. L. COOK LAGAAIJ, R. 1953. The vertical distribution of the Lunulitiform. Bryozoa in the Tertiary of the Netherlands. Ibid., M.S. 7 : 13-19. i963a. Cupuladria canariensis (Busk) portrait of a Bryozoan. Palaeontology, 6, i : 172- 217. - 19630. New Additions to the Bryozoan Fauna of the Gulf of Mexico. Publ. Inst. mar. Sci. Univ. Tex., 9 : 162-236. McGuiRT, J. H. 1941. Louisiana Tertiary Bryozoa. Geol. Bull. New Orleans, 21 : 1-174. MANZONI, A. 1869. Bryozoi Pliocenici Italiani. S.B. Akad. Wiss. Wien, 59 : 17-28. 1875. Briozoi del pliocene antico di Castrocaro, Bologna. MARCUS, E. 1921. Indo-Pacifische Bryozen aus dem Risksmuseum in Stockholm. Ark. Zool., 14, 7 : 1-23. MARCUS, E. & MARCUS, E. 1962. On some Lunulitiform Bryozoa. Bol. Fac. Filos. Cienc. S. Paulo. Zool., 24 : 281-324. NORMAN, A. M. 1909. The Polyzoa of Madeira and neighbouring Islands. /. Linn. Soc. (Zool), 30 : 275-314. O'DoNOGHUE, C. H. 1924. The Bryozoa (Polyzoa) collected by S.S. ' Pickle '. Rep. Fish. Mar. biol. Surv. Pretoria, 3 (1922), spec. rep. 10 : 1-63. D'ORBIGNY, A. Pateontologie Fran9aise, Terrains Cretaces, 5, Bryozoaires, Paris. (Pages 473- 984 were published in 1853, see Harmer, 1957 : m8). OSBURN, R. C. 1914. The Bryozoa of the Tortugas Islands, Florida. Pap. Tortugas Lab., 5 : 181-222. - 1947. Bryozoa of the Allan Hancock Atlantic Expedition, 1939. Rep. Allan Hancock Atlantic Exped., 5 : 1-66. 1950. Bryozoa of the Pacific Coast of America, part i Cheilostomata Anasca. Rep. Allan Hancock Pacific Exped., 14, i : 1-269. ROBERTSON, A. 1908. The incrusting Chilostomatous Bryozoa of the West coast of North America. Univ. Calif. Publ. Zool., 4 : 253-344. SAKAKURA, K. 1935- Pliocene and Pleistocene Bryozoa from the Boso Peninsular, i, Bryozoa from the Dizodo Beds. /. Fac. Sci. Tokyo Univ. Geol., 4 : 1-48. SILEN, L. 1938. Zur kenntniss der Polymorphismus der Bryozoen. Zool. Bidrag. Uppsala, 17 : 149-366. 1941. Cheilostomata Anasca (Bryozoa) collected by Prof. Sixten Bock's Expedition to Japan and the Bonin Islands, 1914. Ark. Zool. 33A, 12 : 1-130. 1942. On the spiral growth of the zoaria of Bryozoa. Ibid., 34A, 2 : 1-22. SMITT, F. A. 1873. Floridan Bryozoa, part 2. K. svenska VetenskAkad. Handl., 11, 4 : 1-83. SOULE, J. D. 1959. Anascan Cheilostomata (Bryozoa) of the Gulf of California. Res. Puritan American Mus. Exped. W. Mexico, 6. Amer. Mus. Nov., No. 1969 : 1-54. SOULE, J. D. & DUFF, M. M. 1957. Fossil Bryozoa from the Pleistocene of Southern California. Proc. Calif. Acad. Sci., 4, 29 : 87-146. STACH, L. W. i936a. Correlation of Zoarial Form with Habitat. /. Geol. Chicago, 44, i : 60-65. - I936b. An Upper Oligocene Bryozoan Faunule. Proc. roy. Soc. Viet., 49, i : 62-71. STEPHENSON, T. A. 1947. The constitution of the intertidal fauna and flora of South Africa. Ann. Natal Mils., 11, 2 : 207-324. STUDER, T. 1889. Die Forschunsgreise S.M.S. Gazelle, 1874-1876. 3, Zool. & Geol. : 1-322, Berlin. TEWARI, B. S., BHARGAVA, O. N. & TANDON, K. K. 1958. Two Cheilostome Bryozoa from Lower Miocene Beds of Kutch. Journ. Palaeont. Soc. India, 3 : 233-235. VERRILL, A. E. 1878. In Cowes, E. & Yarrow, H. C., Notes on the Natural History of Fort Macon, North Carolina, and vicinity. Proc. Acad. Nat. Sci. Phil., part 2 : 297-315. WATERS, A. W. 1885 Chilostomatous Bryozoa from Aldinga and the River Murray Cliffs. Quart. J. geol. Soc. Lond., 37 : 309-347. THE CUPULADRIIDAE 187 WATERS, A. W., 1887. Bryozoa from New South Wales, North Australia, etc., part 2. Ann. Mag. not. Hist. (5) 20 : 181-203. 1921. Observations upon the relationships of the (Bryozoa) Selenariadae, Conescharel- linidae, etc., Fossil and Recent. /. Linn. Soc. (Zool.), 34 : 399-427. 1925. Ancestrulae of the Cheilostomatous Bryozoa, part 2. Ann. Mag. nat. Hist. (9) 15 : 341-352. 1926. Op. Cit., part 5, Cupularia etc. Ibid., 18 : 424-433. PLATE i Cupuladria, Setosellina and Discoporella FIG. i. Cupuladria guineensis (Busk). Philippines, United States National Museum. Syntype of " C. granulosa ", 1963. 3. 14. 9. Photograph by courtesy of the United States National Museum. x 14.5. FIG. 2. C. guineensis. Philippines, United States National Museum. Cotype of " C. dentifera " , 1963. 3. 14. 7. Photograph by courtesy of the United States National Museum. xi6. FIG. 3. C. pyriformis (Busk). Anguilla, Naturhistoriska Riksmuseet, FBD 533, frontal surface. x 10. FIG. 4. C. pyriformis. The same specimen, basal surface. x 10. FIG. 5. Setosellina capriensis (Waters). Aegean, 1899. 7. i. I253pt., Busk Coll., frontal surface. x 1 7 . 4. FIG. 6. S. capriensis. The same specimen, basal surface. XI7-4. FIG. 7. Discoporella umbellata (Defrance). Peyroti-type zoarium, S. Africa, 1949. ii. 10. 94, Burrows Coll., basal surface, showing kenozooecia and vibracula growing over the basal surface. X4-3. Bull. 13. M. (N.H.) Zool. 13, 5 PLATE 1 PLATE a Cupuladria indica and C. guineensis FIG. i. Cupuladria indican. sp. Burma, 1899. 5. i. 267 pt., Hincks Coll., paratypc, frontal surface. x 1 3 . 4. FIG. 2. C. indica. The same specimen, basal surface. xio.5. FIG. 3. C. guineensis (Busk). Sulu Archipelago, 1963. 3. 14. 2, U.S.N.M. Coll. (from type-locality of " C. tuberosa "}, frontal surface. xg. FIG. 4. C. guineensis. The same specimen, basal surface. xy. Bull. B.M. (N.H.) Zoo/. 13, 5 PLATE 2 PLATE 3 The Discoporella umbellata-complex FIG. i. Discoporella umbellata (Defrance). Ghana, Achimota Coll., II, jA. Zooecia and vibracula. x 40. FIG. 2. D. umbellata subsp. depressa (Conrad). Rio de Janeiro, Discovery Coll., SygA. Zooecia and vibracula. X55- FIG. 3. D. umbellata. S.W. France, L. Miocene, 025003. Zooecia and vibracula. X3i. FIG. 4. D. umbellata subsp. depressa. Florida, Miocene. Zooecia and vibracula. Xi8. 5 . FIG. 5. D. umbellata. Peyroti-typezoa,num.S. Africa, 1949. n. 10. 94 pt., Burrows Coll. Zooecia and vibracula. X43- FIG. 6. D. uwibellata. Peyroti-type zoarium. The same specimen, basal surface, showing kenozooecia and vibracula. X3y. Bull. B.M. (N.H.) Zool. 13, 5 PLATE 3 PRINTED IN GREAT BRITAIN BY ADLARD & SON LIMITED BARTHOLOMEW PRESS, DORKING POLYZOA FROM WEST AFRICA THE CUPULADRIIDAE (CHEILOSTOMATA, ANASCA) PATRICIA L. COOK BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 13 No. 6 LONDON: 1965 POLYZOA FROM WEST AFRICA THE CUPULADRIIDAE (CHEILOSTOMATA, ANASCA) BY PATRICIA L. COOK British Museum (Natural History) Pp. 189-227 ; Plates 1-3 ; 4 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 13 No. 6 LONDON: 1965 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. In 1965 a separate supplementary series of longer papers was instituted, numbered serially for each Department. This paper is Vol. 13, No. 6 of the Zoological series. The abbreviated titles of periodicals cited follow those of the World List of Scientific Periodicals. Trustees of the British Museum (Natural History) 1965 TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) Issued October, 1965 Price Eighteen Shillings POLYZOA FROM WEST AFRICA THE CUPULADRIIDAE (CHEILOSTOMATA, ANASCA) By PATRICIA L. COOK CONTENTS Page I INTRODUCTION AND METHODS ..... . lyi 2 CUPULADRIIDAE LAGAAIJ ..... . 192 3 GENERAL NOTES ON LUNULITIFORM COLONIES 193 4 KEY TO SPECIES ....... 196 5 Cupuladria Canu & Bassler ..... . 197 6 C. canariensis (Busk) ...... . 197 7 C. biporosa Canu & Bassler ..... . 203 8 C. monotrema (Busk) ...... 209 9 C. multispinata (Canu & Bassler) .... 210 10 C. owenii (Gray) ....... . 213 ii C. owenii subsp. disciformis n. subsp. . 215 12 C. doma (d'Orbigny) ...... . 216 13 Discoporella d'Orbigny ...... 219 14 D. reussiana (Manzoni) ...... 219 15 D. ocellata n. sp. ....... 220 16 D. umbellata (Def ranee) ...... 221 17 ACKNOWLEDGMENTS ...... 223 18 SUMMARY ........ . 223 19 REFERENCES. ....... 224 i INTRODUCTION THE Collections from which the specimens originate have been described by Cook (1964 : 44) ; they comprise the " Calypso " Collections (Collection I from Senegal to the Bay of Biafra, Collection II from the Cape Verde Islands) ; the Marche-Marchad Collections (from Senegal) ; and the Achimota Collections (from Ghana). The Collections are rich in lunulitiform colonies belonging to the family Cupuladriidae ; 8 species occur round the west African coast, and it is estimated that approximately 2,000 colonies have been available for examination. The lunulitiform and selenariiform types of zoarium are found among several unrelated genera of Polyzoa (see Harmer, 1926 and 1957, and Cook, 1965) ; they are both associated with sandy or muddy substrates, upon which other Polyzoa are usually unable to grow directly. Harmer (1957 : 649, 726, 885, 891 and 1009) discussed the correlation between substrate and selenariiform habit. A large number of the specimens in these Collections are known to be from a sandy or muddy bottom ; for example, those from the Achimota Collection are nearly all from stations close to, or included in, the " silty sand community " of Buchanan (1958 : 16, 26) and Bassindale (1961 : 492). Stations where the bottom is known to be of this type are marked with an asterisk in the lists of material examined. The lunulitiform zoarium ZOOL. 13, 6 10 iga P. L. COOK has been discussed by Waters (1921), Harmer (1926 : 261 and 1957) and Lagaaij (1952 : 31, 43 ; 1953 ; and 1963). Briefly it may be described as free, discoidal or subconical, with zooecia arranged in radial rows. In the Cupuladriidae each zooecium is associated with a distal vibraculum, which has a long seta. METHODS. Colonies were treated with eau de javelle and fragments mounted dry. Others were decalcified in dilute acid, sometimes only partially, and stained to show the relationship of muscles to the calcareous parts, etc. Dry specimens were treated with trisodium phosphate solution to restore shape to shrunken chitinous parts before preparations were made. The length of the zooecia were measured from the distal edge of the aperture to the distal edge of the next succeeding radial zooecial aperture, thus including the vibraculum. The dimensions quoted give the range of variation of 50 measurements (where possible). Generally, measurements of the central area or of very young colonies have not been included. It is hoped to make a survey of measurements useful for specific determination, with a morphometric analysis of at least one population complex, in the near future. The measurements made are : Length of zooecium Lz Length of vibracula seta Ls Width of zooecium Iz Length of vicarious seta Lvs Length of ancestrula Length of operculum Lo (including vibraculum) La Width of operculum lo Length of vibracular opesia Lvo Length of opesia Lop Length of vicarious vibraculum Lv Width of opesia lop Definition of terms and symbols. The registration numbers of specimens in the British Museum (Zoology Department) are given thus : 1899. 7. i . . . ; (Palaeon- tological Department) thus : D 6764 . . . Horizontal cryptocyst lamina. A porous lamina, formed from fused cryptocystal denticles, extending from the descending cryptocyst across the opesia, parallel to, and below, the frontal membrane. Vestibular arch. The distal, raised portion of the aperture. Usually accompanied in the Cupuladriidae by a pair of distal cryptocystal denticles. Vicarious vibraculum. A large vibracular individual, taking the place of a zooe- cium, and itself having a distal vibraculum. In the lists of material examined, the presence of vicarious vibracula is indicated thus : (V) ; of Acrothoracid Cirripedes thus : (C), see p. 194 ; and of a sandy or muddy substrate thus : *. 2 CUPULADRIIDAE Lagaaij Cupuladriidae Lagaaij, 1952 : 31. Cook, 1965: 154. The series of species described below links the membraniporan forms assigned to Cupuladria with the microporan forms represented by Discoporella, and both genera have therefore been placed in the family Cupuladriidae. POLYZOA FROM WEST AFRICA 193 Diagnosis. Lunulitiform Anasca with vibracula alternating with zooecia in the same radial series. Ancestrula surrounded by seven zooecia and a distal vibraculum. Cryptocyst variously developed. Ovicells absent. The systematic position of the two genera here included in the Cupuladriidae has been discussed by Canu & Bassler (1923 : 75), Harmer (1926 : 266), Hastings (1930 : 714, 717), Lagaaij (1952 : 32, and 1953), Marcus & Marcus (1962), and Cook (1965 : 154). Bassler (1953) placed Cupuladria in the Membraniporidae (p. G 156) and Discoporella in the Calpensiidae (p. G 171), but the small distinction between the membraniporan and microporan forms had already been indicated by Canu & Bassler (1923 : 75) who remarked " The union of spines is not a generic character. In fact it may be accidental (Cupuladria denticulata] , partial (Cupuladria reussiana) , almost complete (Cupuladria, umbellata] ". D. ocellata n. sp. links the irregular opesiular indentations and fused cryptocystal denticles typical of D. reussiana with the microporan D. umbellata (see p. 221). 3 GENERAL NOTES ON LUNULITIFORM COLONIES a. Budding. The method of budding and development of lunulitiform colonies has been discussed by Silen (1942 : 7-13), Lagaaij (1963) and Cook (1965 : 155), and the regeneration of broken colonies by Dartevelle (1935 : 559-561). A large number of the colonies in these collections are regenerated from broken fragments ; approxi- mately 80% of D. umbellata in the " Calypso " Collection are of this type, as are many specimens of the other species. It is interesting that the drawings of C. owenii by Gray (1828, pi. 3, figs. i5a) and of D. umbellata by d'Orbigny (1853, pi. 717, figs. 3, 4) are of regenerated colonies. b. Sexual reproduction. Ovicells are absent ; large eggs (" about 0-5 mm. long "), were seen in C. canariensis and C. doma (as C.johnsoni) by Waters (1921 : 404 and 414), and smaller eggs, which may not have been fully developed, by Hastings (1930 : 726) in D. umbellata depressa (as D. umbellata}. These eggs (Gorgona, 1929. 4. 26. 103 pt.), have an average diameter of 0-05 mm. (9% of Lz) ; those found in zooecia of C. multispinata (Stn. 299, Cape Verde Is., Discovery Coll.), have an average diameter of 0-30 mm. (45% of Lz). Eggs in C. owenii disciformis (" Calypso " Coll., C47A) average 0-20 mm. (60% of Lz), and one egg in C. biporosa (" Calypso " Coll., C49G) has a diameter of 0-20 mm. (50% of Lz). The date of breeding shows great variation and is probably dependent upon several unknown factors. Marcus & Marcus (1962 : 297) found no evidence of germ cells in material from the Brazilian coast collected in December, January, March, April, June and August. The " Calypso " specimen of C. biporosa was breeding on 26th May and that of C. owenii disciformis on 26th July, 1956. The Madeiran C. doma. described by Waters was received from Norman, who collected in March and May (see Norman, 1909 : 275) . The fertile C. multispinata from the Cape Verde Islands was collected on 4th September, 1927, and Hastings's D. umbellata depressa from Gorgona, was breeding in July, 1924. Specimens of D, umbellata collected in March, 1963, from Funchal, Madeira, were not breeding, HJ4 1'. L. COOK but one very young, recently settled, colony was found. This suggests that perhaps the breeding season here was in early spring. Generally, the present evidence is that the eggs of the membraniporan and denti- culate species are larger than those of the microporan species. This may be con- nected with the development of the cryptocyst and the size of the opesia, but until observations are made upon living, breeding colonies, the importance of these differen- ces cannot be assessed. The relatively large size of some of the eggs may indicate that the larvae have a yolk and therefore a short free-living existence ; but Lagaaij (1963 : 178) considered that the larval life may persist for some time. Settlement is upon a sand grain, small stone, or Foraminiferan shell, and the ancestrula buds one proximal, a pair of proximal-lateral zooecia, and a distal vibraculum. A pair of lateral zooecia follows, and the ancestrular area is later completed by a pair of distal- lateral zooecia (see Lagaaij, 1963, text-fig. 10 ; and Cook, 1965, text-fig. iB). c. Mode of life. The mode of life of lunulitiform colonies was virtually unknown, but living specimens have now been observed (Marcus & Marcus, 1962 ; and Cook, 1963). Their orientation is normally with the zooecial face upward. All the evidence at present available indicates that colonies belonging to the Cupuladriidae, at least, are not capable of free-swimming movement through the water, but are maintained by their peripheral vibracular setae in a position just above the surface of the substrate. The zooecia of the central area do not appear to have either a " hydro- static " or " radicular " function (cf. Canu & Bassler, 1920 : 75, 1923 : 238, and 1929 : 144). The single instance of 9 colonies being taken at the surface was recorded by Silen (1942 : 13, in the Atlantic, 1881, 27 16' N., 23 21' W.), who commented on the lack of any similar occurrence. These specimens have been re-examined (see p. 207). All have vibracular setae and intact frontal membranes ; the zooecia have polypides. They were therefore presumably alive when collected, but I agree with Silen, who later remarked (1947 : 10), " the free, swimming life of C. canariensis is not at present to be accepted as proved fact ". d. Associated Cirripede. West African Cupuladriidae are often hosts to individuals of an Acrothoracid Cirripede, the presence of which is indicated on the basal side of the colony by slightly thicker calcification, which in C. canariensis obscures the basal kenozooecial pores. On the frontal side a slit, surrounded by a calcified border, marks the opening through which the cirri of the Acrothoracid protrude for feeding (see PI. 3, fig. 3). Members of this group of Cirripedes have been reported from localities including Cadiz and South Africa (see Utinomi, 1950 : 5), inhabiting bar- nacle plates and corals, and producing similar slit openings in these hosts. Most of the slits are found near the central area of the colony. The majority of Acrothoracid specimens have been found in large colonies of C. canariensis and D. iimbellata, but they also occur in C. multispinata, C. doma and D. reussiana. Other conical zoaria with slits evidently made by similar Acrothoracids are those of Selenaria maculata Busk (Bass's Straits, 1854. u. 15. 52, incinerated specimen) and Stylopoma duboisii (Audouin) (Holothuria Bank, 1892. i. 28. 43, see Cook, I965a, in press). The majority of other Polyzoa of the silty sand community encrust the large Foraminiferan, Jullienella foetida, which has an argillaceous test. Neither these nor POLYZOA FROM WEST AFRICA 195 the associated large, erect branching Polyzoan colonies of Metrarabdotos unguiculatum Canu & Bassler and Cleidochasma oranense (Waters), show any evidence of being inhabited by the Acrothoracid Cirripede. Without observation of living specimens it is difficult to establish whether the association described above is one of symbiosis, commensalism or parasitism, but it should be noted that many of the zooecia surrounding the slit and thus directly overlying the cavity containing the Acrothoracid appear to be unaffected, and have well-developed polypides. The successful functioning of the Cirripede suggests that the Polyzoan zoarium has a degree of stability surprising in an unattached organism. Inhabited specimens are marked in the lists of material examined thus, (C) . e. Epifauna. The colony may provide a substrate for the settlement of larvae of species of Polyzoa which would otherwise not be able to establish themselves success- fully in a sandy or muddy habitat. Osburn (1914 : 190) described Beania cu-pularien- sis and (1950 : 176) Membraniporella pulchra, growing on the basal side of colonies of Cupuladria, and Soule (1959 : 22) reported Chaperiella condylata on C. canariensis and D. umbellata. Smittipora levinseni (Canu & Bassler) has been found on C. canariensis from the " Calypso " Coll. (C 726), and a specimen of Onychocdla angulosa (Reuss) budded from a central ancestrula completely covers the frontal surface of a colony of C. biporosa from the Canaries (1962. 10. 8. 7), the peripheral zooecia of which can be seen beneath the growing edge of the Onychocella. The colonies are frequently the substrate for other groups, notably small barnacles, tube-worms, hydroids and sponges. The barnacles and worms are found on the basal side of zoaria with a well domed cavity, where there is presumably sufficient room for their cirri and branchiae to be protruded. Hydroid stolons are found on the frontal surface, running between the zooecia, and sponges on the basal side. Apparently these last may cause deformation of the zoaria, by the suppression of growth of one or more radial rows of zooecia. C. multispinata (see p. 210) was referred to by Waters (1921 : 413) under Busk's MS. name, Cupularia deformis, and specimens labelled "C. deformis ' ' by Busk nearly all show fragments of sponges covering the area where the zooecia have failed to develop. Deformed colonies of D. umbellata were described as C. lowei by Waters, and his type-material and other specimens in the British Museum Coll. have sponge colonies on the basal side. Zoaria of D. umbellata (" Calypso " Coll. C56G and Marche-Marchad Coll. I 390) are also covered basally with a sponge belonging to the Plocamiidae. Here the sponge appears to prevent the connection of the intercalary rows of zooecia produced on either side of the abortive row ; the sponge eventually lines the radial slit thus produced. This may, of course, be a secondary effect, but the correlation of the occurrence of deformity and sponges is significant (see Cook, 1965 : 158). The geographical, bathymetrical and palaeontological distribution of the Cupula- driidae is very wide. Recent specimens of C. canariensis have been found from the S.W. Mediterranean to the Gulf of Mexico ; " D. umbellata " has been reported from a depth of 2,723 fathoms (Canu & Bassler, 1929 : 144) ; and the majority of the species have also been recorded from the Tertiary, D. reussiana and D. umbellata having originally been described as fossils. 196 P. L. COOK 4 KEY TO THE SPECIES OF CUPUL AD RI I D AE DESCRIBED BELOW Note on the identification of worn specimens. Lunulitiform colonies are frequently found to be worn and identification under these conditions is extremely difficult. Lagaaij (1952 : 34 and 1953 : 13) stressed the importance of the character of the basal surface, which is less susceptible to wear, and this is emphasized in the following key. Characters not found in fossil or worn specimens are placed in parentheses. All other features have been found to be present in a few zooecia at least of moderately worn Recent and fossil specimens examined. 1 Basal surface divided by radial and tangential boundaries into sectors ... 2 Basal surface not divided into sectors ........ 4 2 Basal sectors nearly always (95%) small, square, with 1-6 pores, vicarious vibracula frequently present (Operculum longer than wide) . . . . . . ' . 3 - Basal sectors irregular, majority (75%) long with 6-20 pores. Vicarious vibracula infrequent, zooecial vibracula of two kinds. (Operculum wider than long) C. canariensis (p. 197) 3 Basal sectors with 1-3 pores per sector, one layer only of basal kenozooecia, vicarious vibracula present throughout colony. (Operculum within a thickened area of the frontal membrane) ....... C. monotrema (p. 209) - Basal sectors with 1-6 pores per sector, basal surface frequently filled in by many layers of kenozooecia, vicarious vibracula near central area. (Operculum longer than wide) .......... C. biporosa (p. 203) 4 Colonies small (3-7 mm. diameter), steep-sided, solid basally, or with a small con- cavity lined with spinous tubercles. Peripheral zooecia closed, several rows of en- larged peripheral vibracula present. (Operculum with incomplete proximal sclerite) .......... C. doma 1 (p. 216) - Colonies flatter, concave basally. Cryptocyst with denticles and spinules, or with horizontal cryptocyst lamina formed of fused denticles ..... 5 5 Cryptocyst denticles fused to form horizontal cryptocyst lamina, with pores (Discoporelta) 6 Cryptocystal denticles not fused to form horizontal cryptocyst lamina ... 8 6 Opesia small with a pair of closed opesiules ....... 7 - Opesia with a pair of opesiular indentations, trifoliate. Zoarium high. Basal surface with large tubercles ....... D. reussiana (p. 219) 7 Opesia sinuate proximally, cryptocyst with a few large pores. Basal surface tuberculate with radial threads. (Operculum without a proximal sclerite) D. ocellata (p. 220) 8 Opesia straight proximally with a pair of small denticles, basal surface tuberculate with pits and grooves. (Operculum with a proximal sclerite) D. umbellata (p. 221) 9 Zooecia small Lz 0-37-0-47 mm., sides of vestibular arch convergent. Cryptocyst with large distal denticles. Basal surface with very small, regular tubercules, or smooth and glassy. (Operculum longer than wide) C. owenii (p. 213) - Sides of vestibular arch straighter, distal denticles not well developed. (Operculum not longer than wide) ........... 9 10 Zooecia large, Lz 0-55-0-76 mm., denticles spinulose. Basal surface with large, irregular tubercles and salient threads. (Operculum as long as wide) C. multispinata (p. 210) 1 Small, doma-type zoaria may occur in C. biporosa, C. owenii and D. umbellata (see Cook, 1965 : 162). C. biporosa may be distinguished here by the basal surface (fork i), C. owenii by the absence of closed peripheral zooecia and of a proximal sclerite in the operculum (fork 4), and D. umbellata by the presence of a complete horizontal cryptocyst lamina (fork 4). POLYZOA FROM WEST AFRICA 197 Zoarium flat, zooecia smaller, Lz 0-37-0-53 mm., denticles small, simple, few. Basal surface with small regular tubercles. (Operculum wider than long) C. owenii subsp. disciformis (p. 215) 5 CUPULADRIA Canu & Bassler Cuptiladria Canu & Bassler, 1919 : 77, 1920 : 103. Hastings, 1930 : 718. Lagaaij, 1952 : 32. Cheetham & Sandberg, 1964 : 1020. TYPE SPECIES. Cupularia canariensis Busk, Madeira, Recent. Zoarium lunulitiform. Zooecia with cryptocyst variously developed, frequently with denticles, which do not normally fuse. Vibraculum distal to each zooecium. Vicarious vibracula and basal sectors present in some species. Vicarious vibracula have been found to occur only in the C. canariensis group of species (group A, Cook, 1965 : 167). They are marked in the lists of material examined thus : (V). 6 Cupuladria canariensis (Busk) (Pl. i, figs. lA, B, PL 3, fig. 4, Text-figs. la-f) Cupularia canariensis Busk (part), i85ga : 66 (not pi. 23, figs. 7, 8 C. biporosa see p. 203), Madeira, Canaries. Cupularia canariensis Busk : Manzoni, 1869 : 26, pi. 2, fig. 17, Pliocene, Italy. 1877 : 24, pi. 17, figs. 5a, b, c, Miocene, Austria and Hungary. IMembranipora canariensis (Busk) Smitt (part), 1888 : 79, text-fig. 326, 120 fath. and over, Florida. ICupularia canariensis Busk : Angelis (part), 1899 : xxxiii, pi. B, figs. 7, 8, Pliocene, Spain (not fig. 6 = ?C. biporosa). Cupularia canariensis Busk : Neviani 1891 : 130, Post Pliocene, Italy. 1895 : 101, Miocene, Italy. ICupularia canariensis Busk : Calvet (part), 1907 : 393, Canaries and Cap Blanc, 80-259 m. Cupularia guineensis Busk : Norman (part), 1909 : 289, pi. 37, figs. 3 and 6, Madeira (not figs. 4, 5 = C. biporosa). ?CupTtlaria guineensis Busk : Osburn (part), 1914 : 195, Tortugas Islands, Florida, 10 fath. ^.Cupularia canariensis Busk : Faura & Canu (part), 1916 : 133, Miocene, Spain (not pi. 3, fig. 8). Cupularia canariensis Busk : Canu, 1917 : 137, pi. 3, figs. 4-6, Burdigalien, L. Miocene, France. Cupularia canariensis Busk : Waters (part), 1921 : 410 pi. 30, figs, n, 12 Liberia, (not pi. 29, fig. 15 = C. biporosa). Cupuladria canariensis (Busk) : Canu & Bassler (part), 1928 : 16, pi. i, figs. 7-9, Recent, Gulf of Mexico (not text-fig. 2 = C. biporosa). Cupuladria canariensis (Busk) : Dartevelle, 1935 : 560, pi. 19, figs, i, 2, Ras el-Amouch, Algeria. Cupuladria canariensis (Busk) : Sil6n (part), 1942 : 13, Pliocene, Italy ; Recent, Azores, West Indies (not text-fig. 8 = C. biporosa ; not text-fig. 9, pi. 4, figs, 15, 16 = C. pyriformis). Ciipuladria canariensis (Busk) : Lagaaij, 1952 : 33, pi. 2, figs. la, b, Pliocene, Netherlands. T 953 : T 5. pl- I . fig- T Miocene, Pliocene, Netherlands. Cupuladria canariensis (Busk) : Buge, 1957 : I 39> pl- 9> fig- 5- Miocene, S.W. France (not pl. 10, fig- 3)- Cupuladria canariensis (Busk) : Gautier, 1962 : 53, Algeria, 100-300 m. Cupuladria canariensis (Busk) : Lagaaij, (part), 1963, pl. 25, figs, i, 3-5, pl. 26, figs. 2, ?3. Cupuladria canariensis (Busk) Annoscia, 1963 : 225, pl. 9, fig. i, pl. 10, fig. i, pl. n, figs, ra, ib, pl. 12, figs. la, ib, Quaternary, Italy. Cupuladria canariensis (Busk) Cheetham & Sandberg, 1964 : 1021, text-figs, n, 13, Quaternary, Louisiana, ig8 P. L. COOK MATERIAL. LECTOTYPE (chosen here), the top left-hand specimen on the slide B.M. 1899. 7. i. 4697 (A), Busk Coll., M' Andrew, Canaries. Lectoparatypes, the top right-hand specimen on the same slide and 1962. i. 24. 3, Busk Coll., M'Andrew, Canaries. "Calypso " Coll. I. Stn. 7*, 9 40' N., 13 53' 5" W., 17^.56, 18 m., C4F. Stn. 17*, 5 N., 5 28' 30" W., 2i.v.56, 27 m., C$6] (C). Stn. 29*, 4 3' N., 6 12' E., 26.V.56, 32 m., C49B (V). Stn. 45*, o 25' N., 9 o' E,, 8.vi.56, 73 m., C48B (V). Coll. II. Stn. 24, 15 16' 34" N., 23 47' 44" W., i8.xi.59, 55-6 m -> C86A. Stn. 26, He Sao Tiago, 15 16' 30" N., 23 47' 31" W., i8.xi.59, 50-65 m., C65G. Stn. 73, C 7 2B (V). Marche-Marchard Coll. I. Konakrey, Guinee Ise., iF, 2E. Flor de la bouteille, Guinee Ise., 21.1.53, 8 m., 36 (V). Baie de Goree, 1954, 46-48 m., 6B. Sud de Goree, 27.x. 53, 38-42 m., 76 (CV) ; and 24.11.53, 40-41 m., nH (CV). S.W. Madeleines, i5.ix.53, 48 m., 2oB ; 45-46 m., 26G ; and 21.1.54, 46-48 m., 46E (V). Sud de presque File du Cap Vert, 18.11.54, 95 m -> 2 9^ (CV), 33C (V). Baie de Seminoles, Goree, 8.xii.53, 38 m., 3gC (V). Coll. II. Baie de Goree, g.ix.55, 190-220 m., 3A (V) ; and 210-220 m., 33A (V). S.W. large du Cap Vert, 9.ix.55, 100 m., uA (V), I4E (V). Banque de Fagaque (Joal), J-5-V.53, 5 m., I5A. S.W. Madeleines, 9.1.54, 47-5 m., 3iD ; and i5.ix.53, 48 m., 4oA (V). Devant le Cap Manuel, Oct., 1952, 35 m., 37A (CV). Coll. III. Sud de presque File du Cap Vert, i8.ii.54, 4^-5 m -> I C (V). Sud de Goree, I3.xi.53, 33-35 m., 96 ; and 34-37 m., 176. Either as above, 34-37 m., or S.W. Madeleines, I5.ix.53, 48 m., i6E. Dragage i " Gerard Freca " i8.ii.54, 23F, dragage 4, 97 m., 28G ; dragage, 5, 156. 27.xi.53, 33-34 m., 29C. No infor- mation 24G. Achimota Coll. I, Stn. 69, Dredge haul No. 5, 22.1.51, 22 m., 90. III. C. British Museum Coll. Madeira, 1912. 12. 21. 1002 (V), 1911. 10. i. 647 (V) and 1962. i. 28. 2 (V), Norman Coll. Bay of Funchal, Madeira, 1962. i. 28. i (V), Norman Coll ; 50 m., 1963. 2. 28. 10 Cook Coll., mud*. Tangier Bay, 35 fath., 1899. 7. i. 1225-1229, Busk Coll. Ras el-Amouch, Algeria, 45 fath. and upward, " Porcupine ", 1899. 7. i. 84, 41 fath. 1226 ; 45 fath., 1222, 1264 and 1228 ; 4693 and 4694, Busk Coll. Mediterranean, 45 fath., 1899. 7. i. 1223 and 1230, Busk Coll. Cape Rosa, Algeria, 95 fath., 1899. 7. i. 1227, Busk Coll. Cape Sagraas, Portugal, 1899. 7. i. 1221, Busk Coll. West Indies, " Blake ", 1879,1911. 10. I. 1722, Norman Coll. Barbados, 73 fath., 9.iii.79, and 23 13' N., 89 10' W., 84 fath., " Blake " 1911. 10. i. 644, Norman Coll. St. James Coast, Barbados, 1962. i. 26. 2, Saunders Coll. Gulf of Mexico, 28 58' N., 89 9' W., mud lump SP 5, 1961. n. 2. 4gA, Cheetham Coll. " Discovery " Coll., Stn. 279, off Cape Lopez, io.vii.27, 58-67 m., mud and fine sand*, (CV). Waters Coll. Manchester Museum. Four slides from Oran, i slide from " Post Pliocene ", Pisa. Naturhistoriska Riksmuseet, Stockholm Coll. (numbers in Silen, 1942 : 13-14 in parentheses). St. Agata, Piemonte, Italy, Pliocene, FED 737 (No. 2). Azores, off POLYZOA FROM WEST AFRICA 199 Punta Delgada, FBS, 964 (No. 6). West Indies, 52 miles off Florida, 40 m., FBD 548, i specimen only (No. 16). Specimens from Dr. R. Lagaaij. Beeringen, Netherlands, Miocene. Karsy, Poland, Miocene. Antwerp, Belgium, Pliocene. Gulf of Mexico " Atlantis " Stn. 161, 22 fath., Recent. Specimens from Prof. Voigt. Rostej Banat, M. Miocene. Hamburg, Miocene. Pinneberg, nr. Hamburg. U. Miocene. Zoarium frequently large (maximum diameter measured, 23 mm.). Cryptocyst narrow, descending steeply, finely tuberculate, not thickened in central zooecia. Operculum slightly wider than long with no proximal sclerite. Zooecial vibracula of two kinds, one with a short hooked seta. Vicarious vibracula sometimes present surrounding ancestrula, and in the central area, with a short, stout hooked seta. Basal surface divided into irregular rectangular sectors, the majority long, with from 4-20 pores (usually 6-12). Tentacles 17-19 (see p. 203). DIMENSIONS. Lz 0-50-0-80 mm., Iz 0-30-0-47 mm., La 0-45-0-60 mm., Lvo 0-12-0-19 mm., Lv 0-32-0-45 mm., Ls 0-70-1-50 mm., Lvs 0-17-0-30 mm., Lo 0-10-0-13 mm., lo 0-12-0-15 mm., Lop 0-35-0-47 mm., lop 0-21-0-32 mm. Lagaaij (1952 : 33) first chose Busk's slide (B.M. 1899. 7. i. 4697) as lectotype of C. canariensis, but did not indicate upon which of the 6 specimens present his description was based. The two specimens chosen above as lectotype and lecto- paratype are definitely C. canariensis as described by Lagaaij, and as generally understood by European authors ; although only part of Busk's description and neither of his magnified figures (iSsga, pi. 23, figs. 7, 8) refer to this species. I have here followed the Recommendation 74A of the International Code of Zoological Nomenclature (1961 : 79), with regard to the " agreement with previous restriction " of the species. Smitt 2 (1888 : 79, text-fig. 326) recorded Membranipora canariensis from 120 fathoms and over, from Florida. The figured colony is 18 mm. in diameter, and as C. biporosa does not appear to attain this size, is perhaps C. canariensis. Specimens sent to Norman by Smitt, from the " Blake " Collection from the West Indies, consist of a mixture of C. canariensis and C. biporosa, and presumably Smitt's record refers to both species, as perhaps does Osburn's from the Tortugas Islands, where some of the specimens were very large (0-75 in.). Calvet (1907 : 393) included " C. guineensis Kirkpatrick " 3 (sic) in his synonymy ; it may be presumed that part of his material, at least, was true C. canariensis. Waters (1921 : 399) received material from Norman, whose figures (1909, pi. 37, figs. 3 and 6) are of C. canariensis. Norman's specimens from Madeira in the British Museum consist of a mixture of C. canariensis and C. biporosa. The material described by Dartevelle is C. canariensis (Ras el-Amouch, 1899. 7. i. 84, Busk Coll.), and shows the method of budding of zooecia from a broken frag- ment particularly well. 2 For Smitt's authorship see Agassiz, 1888 : xxi. 3 Kirkpatrick (1890 : 612) only listed C. guineensis from the Torres Straits, his record is certainly not C, canariensis (see Hastings, 1930 : 714). 200 P. L. COOK Much of the material listed by Silen (1942 : 13-14) has been re-examined, namely specimens from localities 2, 6, 7, 8, 9, 10, 15, 16 and 18. Of these, specimens 2 and 6, and one of the zoaria from 16, are C. canariensis. The remaining specimens belong to three other species, C. biporosa and C. owenii (see pp. 203 and 213 below), and C. Pyriformis (Busk). Silen expressed doubt as to the identity of specimen 18, from Anguilla, West Indies ; it is certainly referable to C. pyriformis (see Cook 1965 : 168). The figures given by Canu & Bassler (1928) of specimens from the Gulf of Mexico are all of C. canariensis with the exception of text-fig. 2, of an operculum, which shows the elongated appearance typical of C. biporosa. Presumably their Recent material consisted of both species. The figures of C. canariensis given by Manzoni (1869, Pliocene, Italy), Manzoni (1877, Miocene, Austria and Hungary) and Canu (1917, Lower Miocene, France) all show steeply descending cryptocysts and long basal sectors. Neviani (1895) recorded large colonies (2 cm. diameter) from the Italian Miocene, and (1891) had mentioned the elongated basal sectors with numerous pores of his post-Pliocene specimens. Both the specimens from the European Pliocene and Miocene lent by Dr. Lagaaij, and his description and figures (1952 and 1953) have the same consistent characters and thus contrast significantly with the figures of American Miocene specimens given by Canu & Bassler (see p. 206), all of which are attributable to C. biporosa. Buge (1957 : 177) stated that the figure of the basal side (as " face superieure non zoeciale ") of D. umbellata given by D'Orbigny (1853, pi. 717, fig. 4) was in reality of C. canariensis, and showed " secteurs radiaux avec pores ". D'Orbigny's figure shows radial grooves typical of D. umbellata, and irregular tubercles, with no sign of tangential sector boundaries or biserial pores. It is, moreover, in outline, a mirror image of figure 3, which shows the frontal surface of D. umbellata, and the two figures were obviously drawn from the same colony with great attention to detail, as the budding pattern of the regenerated fragment from which the zoarium originates (see p. 193), may be seen to correspond on the frontal side, with the shape of the fragment visible on the basal side. D'Orbigny's explana- tion of pi. 717 (p. 473), however, gives the Plate number as " 747 " (which contains drawings of Cyclostomes only), and the descriptions of figures 2 and 3 have been interchanged. Buge found that d'Orbigny's material of D. umbellata contained one zoarium of C. canariensis. The frequent association of this pair of species has been noted, (see p. 209), and may have occurred in Buge's material from the Redonian, as his figure (pi. 10, fig. 3) of the frontal surface of his " C. canariensis " shows a well- developed vestibular arch and wide zooecia, and is perhaps a worn fragment of D. umbellata or C. haidingeri ; it has not the character of C. canariensis, but the basal surface of a specimen figured on pi. 9, fig. 5, has irregular sectors with 4-8 pores, and is similar to other Miocene specimens of C. canariensis. The figure given by Faura & Canu (pi. 3, fig. 8) shows evidence of a worn denticulate cryptocyst in some zooecia and is not of a specimen of C. canariensis. Their text- figures 4a, b, are reproductions of those of Pliocene specimens given by Angelis (1899), an d do not represent their own material. Canu & Lecointre (1925, pi. 3) figured C, canariensis from the Burdigalian of the Gironde, but the photographs have POLYZUA FROM WEST AFRICA 201 been retouched (in fig. n sector boundaries have been drawn in through basal pores), and little of the character of the specimens can be deduced. The measurements given by Gautier (1962 : 53) show that the zooecia of his speci- mens had wide opesiae. The zoaria were large (12-5 mm. in diameter), and were thus very probably of C. canariensis. Lagaaij's extremely full and detailed study of C. canariensis (1963), includes a bibliography which combines references to both C. canariensis and to C. biporosa, as limited here. Zoaria from west Africa are generally larger than those from the Gulf of Mexico, and of a flattened cone shape. In all the specimens examined the sides of the opesia are curved and the lateral cryptocyst descends steeply, and is so little salient that in cleaned specimens the pores in the lateral walls can be seen from a frontal view. The ancestrula is frequently slightly larger than the zooecia of the primary circle budded from it (see Canu & Bassler, 1928, pi. i, fig. 7), and the zooecia of the ances- trular area never show any thickening of the cryptocyst as in C. biporosa and C, monotrema. The zooecial vibracula of Recent specimens of C. canariensis are of two distinct kinds. The majority are similar to those found in all other species of the family ; they have auriform opesiae, rounded proximally, and long vibracula setae. A proportion of between I : 10-1 : 30 individuals are small, have short, hooked setae, and more symmetrical opesiae, with prominent narrow rostra pointing proximally (see pi. 3, fig. 4). This kind of zooecial vibraculum has been found to occur in the lectotype, the majority of west African colonies, and in a few zoaria from the Gulf of Mexico. The occurrence thus differs from that of the vicarious vibracula with hooked setae (see below) . No other species of Cupuladriidae has been seen in which two such types of vibracula occur. Canu & Bassler (1929 : text-fig. 136) figured a second type of seta in C. grandis, but did not describe it. Their figure is somewhat similar to that of a regenerating seta in D. umbellata (subsp. depressa), given by Marcus & Marcus (1962, pi. 2, fig. 7). The opesiae of regenerated vibracula are not of the narrow, symmetrical type described above. Vicarious vibracula are commonest in the central area of zoaria in which they occur, but may also be present among zooecia budded from a regenerating edge. They completely replace a zooecium and each has an ordinary small, distal vibraculum, like the zooecia. The opesia of the vicarious vibraculum of C. canariensis is quite unlike that of C. biporosa, which is auriform (see below). In C. canariensis a long, acute, raised, beaked rostrum protrudes proximally beyond the short gymnocyst. It supports a wide, short, strongly hooked seta, which is slung between the asym- metrical condyles. The basal sclerite differs slightly from that of the normal vibra- cular seta, but the individual is still basically a vibraculum (see Hastings, 1963 : 180). The basal sectors of C. canariensis are unlike those of C. biporosa, being irregular in length and rarely square in outline. The majority are elongated rectangles, with 6-12 pores, and occasionally more (see Text-fig. 21). Interspersed with these sectors are short wide ones with 3-6 pores, these occur at the origin of an intercalary row of zooecia. In basal view colonies of C. canariensis do not show the regular concentric series of sectors which is so striking in C. biporosa. At the growing edge, the minute 202 P. L. COOK pores in the basal wall can be seen in the developing zooecia to be without any sur- rounding kenozooecial chamber. Further toward the centre of the colony, a cresent of calcification with the concavity pointing distally grows round the pore ; when the arms of the cresent meet, this forms the first kenozooecial chamber. In young colonies the tangential divisions between the radial sector boundaries are not developed until at least 4 radial series of zooecia have been budded. Thus even young colonies FIG. i. Cupuladria canariensis and C. biporosa. a-f, C. canariensis. a. Vicarious vibraculum and normal distal vibraculum, treated with eau de javelle. Marche-Marchad Coll., I, 39 C. b. Hooked seta of vicarious vibraculum. c. Seta of normal vibraculum, central part omitted, d. Operculum. b-e, " Calypso " Coll., I, C 56 J. f. Basal surface, treated with eau de javelle, showing elongated sector with porous kenozooecial chambers. Marche-Marchad Coll., I, 39 C. g-h, C. biporosa. g. Vicarious vibraculum and normal distal vibraculum, treated with eau de javelle. 1962. i. 26. i. h. Scimitar-shaped seta of vicarious vibraculum, in lateral view. i. Operculum. j. Basal surface, treated with eau de javelle, showing small, square sectors with porous kenozooecial chambers. Marche- Marchad Coll., Ill, 24 G. POLYZUA FROM WEST AFRICA 203 may be distinguished from those of C. biporosa where the tangential sector boun- daries are seen as soon as the zooecia grow out beyond the original substrate. Waters (1921 : 411) gave the number of tentacles as 14, but this is the number found in sections of C. biporosa (1929. 4. 26. 85 pt.). Waters's material included both C. canariensis and C. biporosa, and the number of tentacles in C. canariensis sections (MM II, 3A) is 17-19. C. doma (see Waters, 1921 : 411, and Cook, 1963 : 409) and the D. timbellata-complex (see Marcus & Marcus, 1962 : 295, and Cook, 1963 : 409) have 13-16 tentacles, and the same number has been seen in sections of C. owenii disciformis (" Calypso " Coll. C47A). Thus C. canariensis differs from C. biporosa chiefly in the character of the basal surface, the cryptocyst, the number of tentacles, in the possession of two kinds of zooecial vibracula, and the form of the vicarious vibracula. The distribution of the two species also appears to have been completely distinct in the Miocene (see below) . C. canariensis does not occur south of Barbados in Recent collections, all Brazilian records being of C. biporosa or C. monotrema. 7 Cupuladria biporosa Canu & Bassler (PL i, figs. 2A, B, 3A, B, 4A, B, 5, 6A, B, Text-figs, ig-j) Cupularia canariensis Busk (part) i85ga : 66, pi. 23, figs. 7, 8, Madeira and Canaries. Cupularia canariensis Busk : Busk, 1859 : 87, pi. 13, figs. 2 a-e, Coralline Crag, Pliocene, Britain. Membranipora canariensis (Busk) Smitt, 1873 : 10, pi. 2, figs. 69-71, 10-14 fath. Florida. (Ppart), 1888 : 79, 1 20 fath. and over, Florida. ICiipularia canariensis Busk : Angelis (part), 1899 : 33, pi. B, fig. 6 only, Pliocene, Spain. Cupularia canariensis Busk : Canu, 1908 : 275, pi. 5, figs. 8, 9, 10, Pampean, Pliocene, Argentina. Cupularia giiineensis Busk : Norman (part), 1909 : 289, pi. 37, figs. 4, 5, Madeira (not figs. 3, 6 = C. canariensis). ICupularia guineensis Busk : (part) Osburn, 1914 : 195, Tortugas Islands, Florida, 10 fath. Cupularia canariensis Busk : Canu & Bassler, 1918 : 119, pi. 53, figs. 5-7, Miocene, Costa Rica and Jamaica. Cupuladria canariensis (Busk) Canu & Bassler, 1919 : 78, pi. i, figs. 8-10, Lower Miocene, Costa Rica, 1920 : 103, text-fig. 240. 1923 : 28, pi. i, figs. 7-9, Lower Miocene, Florida, Jamaica, Santo Domingo, Costa Rica ; Miocene and Pliocene, Florida. Cupularia canariensis Busk : Waters (part), 1921 : 410, pi. 29, fig. 5, Madeira. Cupuladria biporosa Canu & Bassler, 1923 : 29, pi. 47, figs. 1-2, Miocene, Santo Domingo. Cupuladria canariensis (Busk) : Canu & Bassler, 1928 : 16, text-fig. 2, Pliocene Panama ; and Recent, Gulf of Mexico. Cupuladria canariensis (Busk) : Hastings, 1930 : 714, pi. 8, figs. 38, 40, Gorgona, Colombia, 15-30 fath. Cupuladria canariensis Busk (sic) McGuirt, 1941 : 46, pi. i, figs. 1-3, 5-6, 8, Miocene & Pliocene, Louisiana. Cupuladria canariensis (Busk) : Silen (part), 1942 : 13, text-fig. 8, West Indies. Cupuladria canariensis (Busk) : Osburn, 1950 : 33, pi. 3, figs. 2, 3, Lower California to the Galapagos Islands, shallow water to 40 fath. Cupuladria canariensis (Busk) : Soule, 1959 : 8, California, 7-40 fath. Cupuladna canariensis (Busk) : Galopim de Carvalho, 1961 : 97, pi. i, figs. 1-3, Pliocene, Portugal. Cupuladria canariensis Marcus & Marcus, 1962 : 285, pi. i, figs. 1-3, off Sao Paulo, 150 m., near Cabo Frio, 3 m., off mouths of the R. Amazon, 70 1^-71-5 m., dead. Cupuladria canariensis (Busk) : Lagaaij, part, 1963, pi. 26, figs. 4, 5. Cupuladria sp. Cheetham & Sandberg, 1964 : 1021, Quaternary, Louisiana. 204 !' 1- COOK MATERIAL. Holotype U.S.N.M. 68425. (Photographs, B.M.N.H. 1963. i. 2. i, 2) Miocene, Bowden Marl, Santo Domingo. " Calypso " Coll. I. Stn. 7*, 9 40' N., 13 53' 5" W., 17^.56, 18 m., C4l(V). Stn. 29*, 4 3' N., 6 12' E., 26^.56, 32 m., C49G (V). Coll. II. Stn. 26, 15 16' 30" N., 23 47' 31" W., i8.ii.59, 50-65 m., C65R. Stn. 75, 16 04' 20" N., 22 58' 10" W., 45 m., CiogA. Marche-Marchad Coll. I. Flor de la bouteille, Guinee Ise, 21.1.53, 8 m., 3E (V). Sud de Goree, 27.x. 53, 38-42 m., jD. Sud de presque Tile de Cap Vert, 95 m., 33Q. S.W. Madeleines, 21.1.54, 46-48 m., 46G (V). Coll. II. S.W. Madeleines, 9.1.54, 77-5 m., 3iR. Coll. III. Dragage 5, "Gerard Freca ", iS.ii. 54, 15! ; Dragage 4, 97-98 m., 28K (V). 33-34 m., 27.11.53, 29! (V). No information, 24R (V). British Museum Coll. Canaries, M'Andrew, 1899. 7. i. 46976, Busk Coll. Canaries, M'Andrew, 1899. 7. i. 4702, 1962. i. 24. 4 and 5, Busk Coll. Madeira, 1879. 5- 28. 6. Bay of Funchal, Madeira, 1962. i. 24. 6, 1963. i. 16. 73, Madeira, 1962. i. 24. 7, and 9 (V), Norman Coll ; Funchal, 50 m., 1963. 2. 28. 6, Cook Coll., mud.* Ras el-Amouch, Algiers " Porcupine ", 4 fath. and upwards, 1962. i. 24. i. Tangier Bay " Porcupine ", 1911. 10. i. 646, Norman Coll. Barbados, 73 fath., 9.iii.79, and 23 13' N., 89 10' W., 84 fath., " Blake ", 1962. i. 24. 8, Norman Coll. (V) West Indies, 1962. i. 24. 2, Norman Coll (V). Gulf of Mexico, 91 W., 28 6' 25" N., 37 fath., 1959. 8. 20. 3, Lagaaij Coll. (V). 89 9' W., 28 58' N., 1961.11.2. 496, Cheetham Coll. Gorgona, Colombia, 1929. 4. 26. 85. 86, 263 (V), 87 (V). St. James coast, Barbados, 1962. i. 26. i,SaundersColl.(V). Miocene, Santo Domingo, 1899. 7. i. 1259, Busk Coll. (V). B.M. Palaeontological Dept. Coll. Coralline Crag, Pliocene, Britain, D 39968, 39969, D 6767, 6762, 6764, B 1624. Astigiana, Pliocene, Italy, 1848. 6. 12. 3-7. U.S. National Museum Coll. Holotype, see above (V). 80747, Miocene, Bowden Marl, Jamaica (V). Dr. A. Cheetham Coll. Forest Hill Clay, Alabama, L. Miocene. Oak Grove, Yellow River, Okaloosa County, Florida, L. Miocene. Cercado de Mao, Santo Domgino, L. Miocene (V). Tamboo Trace, Shell Bed, nr. Talparo, Trinidad, L. Miocene (V). Manzanilla, Manzanilla coast, Trinidad, Miocene (V). Dr. R. Lagaaij Coll. Miocene, Limestone Creek, Missouri. Miocene, Cubagua, Venezuela. Miocene, Bowden Marl (V), Jamaica. Pleistocene, Gulf of Mexico " Neptune " i. Gulf of Mexico, " Atlantis " Stn. 163, 20 fath. (V). Naturhistoriska Riksmuseet, Stockholm Coll. (Nos. in Silen, 1942 : 13-14 in parentheses). Atlantic, 27 16' N., 23 21' W. at the surface, FBS 321 (No. 7). (One of the 9 specimens is C. owenii qv) West Indies, S.W. off Tortugas 40 m., FED 239 (No 8) (V). As above, East Key, 20 m., FED 258 (No. 9) (V). As above, nr. Rebecca shoal, 15 m., FED 281 (No. 10) (V). Yucatan, 35 miles, N.N.E. off C. Catocha, 40 m., FED 696 (No. 15) (V). 52 miles off Florida, 40 m., FED 548 (No. 16) (V) (One colony of C. canariensis present). C. biporosa is not present in the Achimota Collection. Zoarium generally smaller than that of C. canariensis, not exceeding 16 mm. in POLYZOA FROM WEST AFRICA 205 diameter, zooecia also smaller. Operculum frequently longer than wide with no basal sclerite. Cryptocyst descending gently, sides of opesia straight. Vicarious vibracula frequent, usually present in ancestrular area, occasionally throughout zoarium ; seta straight, elongated, broad, slightly hooked, scimitar shaped. Basal surface divided into concentric series of small, nearly square sectors with 1-6 pores, usually 4. Basal concavity flattened and filled by many layers of kenozooecial chambers. Tentacles 14-15 (see above). DIMENSIONS. Lz 0-40-0-65 mm., Iz 0-25-0-35 mm., La 0-40-0-47 mm., Lvo 0-07- 0-13 mm., Lv 0-45-0-55 mm., Ls 0-40-1-40 mm., Lvs 0-25-0-30 mm., Lo o-io-o-n mm., lo 0-09-0-10 mm., Lop 0-25-0-32 mm., lop 0-15-0-19 mm. Dr. R. S. Boardman (in litt. 13.11.1962) has found that the only specimen referred to C. biporosa in the U.S. National Museum Collection, is the holotype-fragment of which part was figured by Canu & Bassler (1923, pi. 47, figs, i, 2). Enlarged photo- graphs of the entire fragment show that one vicarious vibraculum is present, and that the opesiae of the zooecia in the published figure (fig. i) have been enlarged by re- touching of the photograph. This was presumably done to erase the image of detritus lodged in the opesiae, but has unfortunately given the impression that the cryptocyst descends more steeply than it in fact does. Specimens of complete young zoaria labelled " C. canariensis ", from the same horizon in Jamaica as the holotype (U.S.N.M. 80747) are strikingly similar in all characters to young Recent colonies from the West Indies (see pi. i, figs. 4, 6). Busk's description (i85ga : 66) of the occurrence of 2-4 pores in the sectors on the basal side of his specimens, and his figures (pi. 23, figs. 7, 8) showing the small square basal sectors (fig. 7), and well-developed cryptocyst (fig. 8), are certainly of C. biporosa. The major part of the material from the Canaries collected by M' Andrew in the Busk Collection (see Busk, i85ga : 67) is also of C. biporosa, but specimens of C. canariensis are mixed with it, and it is probably these which Busk described as 0-5 in. in width, as none of his specimens of C. biporosa reach this size. The speci- mens on slide 1899. 7. i. 4697 A -f B, of C. canariensis and C. biporosa, are mounted on glass, and the distinctive characters of both surfaces of the colonies of the two species can be seen clearly. Busk's figures of specimens from the British Crag (1859 : 87, pi. 13, figs. 2a-e), show typical C. biporosa characters, namely the straight sides of the opesiae and the gently descending cryptocyst. The figure of the basal side shows some irregular sectors, but specimen D 6764 has square sectors, with 4 pores, and many layers of kenozooecial chambers, as well as the frontal character of C. biporosa. Busk's specimens are particularly interesting as the. only other records of C. biporosa from the Pliocene of Europe are from Spain, Portugal and Italy. Pliocene specimens from the Netherlands are all referable to C. canariensis. Smitt (1873 : 10) described the " inner lamina " (i.e. cryptocyst) of his specimens as being similar to that of Farcimia cereus (p. 3, pi. i, figs. 55, 56), which has a well- developed cryptocyst. His figure 69 shows the slightly elongated operculum, figure 70 the cryptocyst, and the vicarious vibracula in the ancestrular area (see Hastings, 1930 : 714). Figure 71 shows the basal sectors, some of which have 4 ZOOL. 13, 6 n 206 P. L. COOK pores, although others are shown bearing more. The specimens listed by Silen (1942, Nos. 8, 9 and 10) have been examined ; they include those described and figured by Smitt. No. 8 (FED 239) is the original of figure 71 and the actual sectors have been recognized. The basal sectors are square and more regular than in Smitt's drawing and the long irregular sectors do not occur, but are composed of several smaller sectors. Figure 70 is also drawn from this specimen and confirms that the irregular opesiae shown in the figure are those of vicarious vibracula. In fact, not only the ancestrular, but all the 7 individuals surrounding it are of this type. This development is not unusual in C. biporosa and also occurs in colonies from FED 696, FED 548, Dr. Lagaaij's specimen from the Gulf of Mexico, and 1962. i. 26. i. from Barbados. Specimen FED 281 is the original of Smitt's figure 69 ; it is a large colony, 10 mm. in diameter. The figures given by Angelis (1899, pi. B, figs. 7, 8) of the basal surface of his specimens from the Pliocene of Catalonia, show long irregular sectors, and are attributable to C. canariensis. Figure 6 (in which the zooecia are shown upside- down) has the appearance of C. biporosa, with a wide cryptocyst and straight-sided opesia. The figure is very similar in appearance to that of Busk's Crag specimens. Angelis's descriptions however, are taken from that of Manzoni (1877) and refers to C. canariensis. No fossil Spanish material has been seen, but specimens figured by Galopim de Carvalho (1961, pi. i, figs. 1-3, Pliocene, Portugal) also show the straight- sided opesia and small basal sectors of C. biporosa. Pliocene material from Italy also consists of both C. canariensis and C. biporosa, (see pp. 198 and 204), and the cor- relation of the characters present is respectively consistent with the diagnoses of the two species. The specimens figured from the Pliocene of Argentina by Canu (1908, pi. 5, figs. 8, 9, 10) were extremely worn, but figure 8 perhaps shows 2 vicarious vibracula (in the bottom right-hand corner). Both Norman (1909, pi. 37, figs. 4, 5), and Waters (1921, pi. 29, fig. 5), who received his Madeiran material from Norman, show 2-4 pores on the basal sectors of their specimens. Examination of the plentiful Madeiran colonies in the Norman Collection shows that they, too, consist of a mixture of both C. biporosa and C. canariensis. Norman and Waters both stated that they considered C. canariensis andC. guineensis (Busk) to be synonymous. In view of the mixed nature of their material it should be noted that the salient lateral cryptocyst and small square basal sectors of C. biporosa are very similar to some specimens of C. guineensis, which was, however, distin- guished and defined by Hastings (1930 : 714). The material from Costa Rica figured by Canu & Bassler (1918, 1919) as C. canariensis, belongs to C. biporosa. The same set of photographs was reproduced in both these papers and was again used to illustrate American fossil " C. canariensis " in 1920 and 1923. The photographs show i) small colonies, natural size, with flat- tened bases ; 2) zooecia with well-developed lateral cryptocysts, and with one vicarious vibraculum at the lower right-hand edge ; 3) the basal sectors, each with 2-4 pores. The introduction of C. biporosa was illustrated by material from Santo POLYZOA FROM WEST AFRICA 207 Domingo (1923). The fossils from the Gulf of Mexico region (1928), were illustrated by photographs of Recent C. canariensis from the same area. The specimens described by Hastings (1930 : 714) from Gorgona, consist of colonies, the majority of which have regenerated from broken fragments. Where an ances- trula is present, the vicarious vibracula near it have smaller, more auriform opesiae than those occurring elsewhere in the colony. McGuirt (1941 : 46) mentioned the difficulty in distinguishing his Miocene material from Louisiana, attributed to C. canariensis, from C. biporosa. The specimens figured show vicarious vibracula (pi. i, figs, i, 3) and the small, square basal sectors (figs. 2, 5), and are certainly of C. biporosa. The zooecia in figure 8 have unusually narrow cryptocysts, and may have been worn. Of the colonies collected at the surface of the Atlantic off the west coast of Africa (see Silen, 1942 : 13, specimen FBS 321 above, and p. 194), 8 colonies belong to C. biporosa. The ninth is a specimen of C. owenii (see p. 213). Osburn's figure (1950, pi. 3, figs. 2, 3) shows small, square sectors with 1-3 pores (fig. 3), the zooecia in figure 2 have the well-developed lateral cryptocyst of C. biporosa, and Soule (1959 : 8) mentioned vicarious vibracula in his Californian material. Marcus & Marcus (1962) gave a short description of the central vicarious vibracula, and, on pi. i, fig. 2, illustrated the basal surface of one of their colonies with small segments arranged concentrically. Their material was collected from the northern and southern coasts of Brazil, no specimens being found in the intermediate area. The shallow depths at which some specimens occurred (3 m. near Cabo Frio) were at " sheltered localities ", where there must have been little turbulence. Lagaaij's figures (1963, pi. 26, figs. 4, 5) of material from the Gulf of Mexico are of C. biporosa, as is the unnamed species mentioned by Cheetham and Sandberg (1964). The zoarium rarely attains a diameter of more than 10 mm. ; fragments of large colonies estimated at 16 mm. diameter were found in fossil material from the U.S. Nat. Mus. Collection, 80747, and a colony of n mm. diameter, covered with 3 small colonies of Onychocella angulosa (Reuss) occurs in material from Madeira, 1962. i. 24. 7. Unlike C. canariensis, the sides of the opesiae are almost straight, and in the great majority of specimens the cryptocyst descends gently forming a salient lateral shelf and obscuring the pores in the lateral walls, which are not seen in frontal view as in C. canariensis. In some fossil specimens the cryptocyst descends more steeply, but this seems to be confined to zooecia of fragments of large zoaria, and does not occur in the smaller, whole colonies. The cryptocyst of the zooecia of the central area are sometimes slightly thickened by secondary calcification, but a lamina is never formed as is found in C. guineensis (see Cook, 1965 : 172) ; a similar thickening is found in C. monotrema. The ordinary vibracula are of one kind only and vicarious vibracula are nearly always present, especially near the centre of the colony ; their opesiae are auriform, and there is a distinct, prominent gymnocyst. The setae of Recent specimens have the same type of basal sclerite and are similar to those of the ordinary vibracula (see also Hastings, 1930 : 714) ; they have no sigmoid curve, however, and in lateral view 2o8 P. L. COOK they are broad, scimitar-shaped, and hooked at the tip (see Text-fig, ih). They thus differ from the short, hooked setae of the vicarious vibracula of C. canariensis. Both Marcus & Marcus (1962 : 288, pi. i, fig. 2) and Lagaaij (1963 : 183) described the vicarious vibracula of the ancestrular area as regenerated. Although the original zooecia may regenerate as vibracula in some cases, many specimens (notably those from Gorgona and the Gulf of Mexico) , show that the vicarious vibracula are budded primarily as part of the normal development of a colony. Vicarious vibracula are often associated with areas of regenerating growth (see Cook, 1965 : 160), and here also, they are some of the first individuals to be budded, not the results of regeneration of the older zooecia. The development of the basal porous sectors in C. biporosa differs from that of C. canariensis in that the chambers appear to develop more quickly, and large pores are rarely seen at the growing edges of young colonies. The tangential sector boundaries are always present on the basal side of the young peripheral zooecia, unlike C. canariensis (see above). The layers of kenozooecial chambers are more numerous than C. canariensis, even in young colonies. They frequently conceal the original substrate, and in many colonies are so thick that the basal surface is flattened. In section these zoaria show successive rows of chambers up to 14 deep. A slide of sectioned colonies of C. biporosa, where the species is mounted with C. guineensis for comparison (1879. 5- 2 ^- 6, Madeira + 1929. 5. 10. i, Australia), was figured by Hastings (1930, pi. 8, fig. 38, as C. canariensis}. The identity of the species in no way affects her description of the chambers as a distinction from C. guineensis, as those of C. canariensis are exactly the same in appearance, although they rarely reach a comparable number of layers, even in large colonies of this species. One of the specimens from the Miocene of Santo Domingo (U.S.N.M. Coll.) shows a form of growth similar to normal colonies of C. doma (see Cook, 1965 : 162). The zoarium is small and high, the basal side is half filled by secondary calcification, and there are 3 rows of peripheral vibracula, the last of which encroaches on the basal side (see PI. i, fig. 5). The geographical distribution of specimens of C. canariensis and C. biporosa, with and without vicarious vibracula, falls into distinct groups. The first, northerly area, consists of records from the S.W. Mediterranean, Spain, Portugal, and the Canaries ; no vicarious vibracula have been found in colonies of either species from these localities. Material of both species from Madeira, the Cape Verde Islands and the coast of Senegal has colonies both with and without vicarious vibracula ; whereas all the specimens from the southerly localities in the Gulf of Guinea and the Bay of Biafra have vicarious vibracula. Recent and fossil specimens of C. biporosa from the Gulf of Mexico are rarely without vicarious vibracula, whereas none have been found in the Recent colonies of C. canariensis from this area. C. biporosa differs from C. canariensis consistently in the character of the basal surface, of the cryptocyst, in the number of tentacles, in the absence of a second kind of ordinary vibraculum, and in the form of the vicarious vibraculum. The fossil record shows whereas there are apparently no certain descriptions of C. biporosa POLYZOA FROM WEST AFRICA 209 occurring before the Pliocene in south-western Europe, the only American records of C. canariensis, from the Gulf of Mexico and Florida, are Recent. The Miocene faunas of the eastern and western Atlantic were distinct. C. canariensis and C. biporosa have similar geographical distributions, live in similar depths, under the same ecological conditions, and are thus frequently associ- ated in Recent collections. There is a similar correlation between their occurrence and that of the D. umbellata-complex (see Canu & Bassler, 1918 : 119), and as many as 5 lunulitiform species occur together at Konakrey and Cape Verde, and Stn. 45, " Calypso " Coll. I, off Cape Lopez, in the Bay of Biafra. The confusion between C. canariensis and C. biporosa, and the occurrence of vicarious vibracula in both species, make it advisable to compare them wi