/" GENERAL > ( 29 JAN 1981 Bulletin of the vgSf British Museum (Natural History) Zoology series Vol 38 1980 British Museum (Natural History) London 1981 Dates of publication of the parts No 1 . .27 March 1980 No 2 . .27 March 1980 No 3 . 29 May 1980 No 4 . 25 September 1980 No 5 30 October 1980 ISSN 007-1498 Printed in Great Britain by Henry Ling Ltd, at the Dorset Press, Dorchester, Dorset Contents Zoology Volume 38 Page No 1 Miscellanea Shell structure of three littoral species of testate amoebae from the Black Sea (Rhizopodea : Protozoa) By V. Golemansky & C. G. Ogden . ... 1 Notes on the hydroid, Synthecium evansi By P. F. S. Cornelius 7 The life history and ecology of the aberrant bathypelagic genus Bentho- misophria Sars, 1909 (Copepoda: Misophrioida) By G. A. Boxshall & H. S. J. Roe ... 9 Abyssal benthic Amphipoda (Crustacea) from the Iceland Basin. 1. The genus Rhachotropis By M. H. Thurston ... . ... 43 Abyssal benthic Amphipoda (Crustacea) from the Iceland Basin. 2. Lepechinella and an allied new genus By M. H. Thurston ... . .69 A new species of Holplophorella (Acari, Cryptostigmata) from Java By B. W. Parry ... 89 On a new species of Rousettus Gray, 1821, from Sumatra and Borneo (Mammalia : Megachiroptera) By W. Bergmans & J. E. Hill 95 No 2 Relationships between hummingbirds and flowers in the Andes of Colombia By D. W. Snow & B. K. Snow . . 105 No 3 Miscellanea A new species of small Barbus (Pisces, Cyprinidae) from Tanzania, East Africa By R. C. Bailey .141 A new species of Barbus (Pisces, Cyprinidae) from Africa By K. E. Banister . ... .145 A new blind cyprinid fish from Iraq By K. E. Banister and M. K. Bunni . .151 A new species of cichlid fish from the Malagarasi swamps and rivers (Tanzania, E. Africa) By P. H. Greenwood .... .159 A new catfish from Sierra Leone By G. J. Howes ... .165 A new genus of cheline cyprinid fishes By G. J. Howes . . 171 No 4 Miscellanea A redescription of Trochammina nana (Brady) (Protozoa: Foramini- ferida), with observations on several other Recent Trochamminidae in the Collections of the British Museum (Natural History) By P. Bronnimann & J. E. Whittaker . . .175 Crinoidea collected by the Meteor and Discovery in the NE Atlantic By A. M. Clark 187 Page A new abyssal genus of the family Ophiuridae (Echinodermata : Ophiuroidea) By G. L. J. Paterson . . . . . . . . ':"!. 211 A revision of the spider genus Macopaeus (Araneae : Salticidae) By F. R. Wanless . . ... 219 A revision of the spider genus Orthrus (Aranae : Salticidae) By F. R. Wanless ... 225 A note on Lonchophylla (Chiroptera: Phyllostomatidae) from Ecuador and Peru, with the description of a new species By J. E. Hill 233 No 5 Mites of the subfamily Parasitinae (Mesostigmata : Parasitidae) in the the British Isles 237 GENERAL ulletin of the British Museum (Natural History) Miscellanea Zoology series Vol 38 No 1 27 March 1980 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum's resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself, available separately, and individually priced. Volumes contain about 300 pages and are not necessarily completed within one calendar year. Subscriptions may be placed for one or more series. Subscriptions vary according to the contents of the Volume and are based on a forecast list of titles. As each Volume nears completion, subscribers are informed of the cost of the next Volume and invited to renew their subscriptions. Orders and enquiries should be sent to: Publications Sales, British Museum (Natural History), Cromwell Road, London SW7 5BD, England. World List abbreviation: Bull. Br. Mus. not. Hist. (Zool.) Trustees of the British Museum (Natural History), 1980 ISSN 0007-1498 Zoology series Vol 38 No 1 pp 1-104 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 27 March 1980 Miscellanea Contents Page Shell structure of three littoral species of testate amoebae from the Black Sea (Rhizopodea: Protozoa). By V. Golemansky and C. G. Ogden ... 1 Notes on the hydroid, Synthecium evansi. By P. F. S. Cornelius .... 7 The life history and ecology of the aberrant bathypelagic genus Benthomisophria Sars, 1 909 (Copepoda: Misophrioida). By G. A. Boxshall and H. S. J. Roe . 9 Abyssal benthic Amphipoda (Crustacea) from the East Iceland Basin. 1 . The genus Rhachotropis. By M. H. Thurston 43 Abyssal benthic Amphipoda (Crustacea) from the East Iceland Basin. 2. Lepechinella and an allied new genus. By M. H. Thurston 69 A new species of Hoplophorella (Acari, Cryptostigmata) from Java. By B. W. Parry 89 On a new species of Rousettus Gray, 1821, from Sumatra and Borneo (Mammalia: Megachiroptera). By W. Bergmans and J. E. Hill 95 Shell structure of three littoral species of testate amoebae from the Black Sea (Rhizopodea: Protozoa) Vassil Golemansky Institut de Zoologie, Sofia, Bulgaria and Colin G. Ogden Department of Zoology, British Museum (Natural History), London SW7 5BD, England Introduction Ultrastructural studies have been reported on both the shell and cytoplasm of testate amoebae from freshwater, moss and soil habitats. Since the first reports of Thomas (1958) and Thomas & Hovasse (1962) on the shell of the genus Trinema, several reports have dealt with these structures in other genera, for example Euglypha (Mercier et al, 1964; Hedley & Ogden, 1973; Hedley et al., 1974); Arcella (Cambers al., 1963;Netzel, 1975; Moraczewski, 1971); Centropyxis (Netzel, 1976a; Hedley et al., 1976); Dijjlugia (Netzel, 19766; Ogden, 1979) and Cryptodifjlugia (Hedley et al., 1977). The diversity and complexity of the shell for most of the common freshwater species has been illustrated in the recent publication of an Atlas of Freshwater Testate Amoebae by Ogden & Hedley (1980). The present work is a contribution to the ultrastructure of the shell of three species of testate amoebae which are found specifically amongst sand of the supralittoral region border- ing marine waters. Materials and Methods Individual specimens were selected from samples collected from the shores of the Black Sea, Bulgaria during 1977. The animals were fixed in 3%'glutaraldehyde in 0-1 M cacodylic acid buffer and post-fixed in 1 % osmium tetroxide, after fixation for about 1 0 minutes they were washed several times in distilled water. Each shell was manipulated using a single-hair brush onto small drops of Araldite arranged on an otherwise clean cover slip. The prepared cover slips were mounted onto standard Stereoscan stubs with Araldite, prior to coating evenly with a conducting layer of gold. The stubs were examined on a Stereoscan Mk II operating at lOkV and the results recorded on Ilford HP4 film. Results Pomoriella valkanovi Golemansky, 1970 This species is found in underground water littorals on sandy beaches of the Black Sea, where the salinity varies between 15-17-2%o. The shell is colourless or transparent. In outline it is pyriform with anterior region curved, and in transverse section it is circular (Figs 1 and 3). The aperture is circular or oval and bordered by organic cement (Fig. 4). The surface is covered with shell plates that abut against each other, without overlapping, and are joined by an abundance of organic cement (Fig. 2). The shell plates in the aboral region are long, oval, about 9-lOjim in length and 6-7um in width, and in the apertural region they are smaller, about 4-5um long and l-2um in width. Unfortunately the specimens have bacteria and Bull. Br. Mus. nal. Hist. (Zool.) 38 ( 1 ): 1-6 Issued 27 March 1980 V. GOLEMANSKY AND C. G. OGDEN Figs 1-4 Pomoriella valkanovi Fig. 1. Lateral view x 2300 Fig. 2. Portion of shell surface, note the organic cement and adhering debris bordering the shell plates x 4800 Fig. 3. View to illustrate the slight curvature of the shell near the aperture x 1 900 Fig. 4. Apertural view x 4900. Fig. 5 Apertural surface of shell of Centropyxiella arenaria, to show the arrangement of flat particles x 3600. LITTORAL SPECIES OF TESTATE AMOEBAE Figs 6-10 Centropyxiella arenaria Fig. 6. Lateral view, note that most of the shell surface is smooth x 1350 Fig. 7. Aboral view of apertural disc-like extension to show the rough surface x 1400 Fig. 8. Latero-apertural view to illustrate the apertural extension x 1600 Fig. 9. Apertural view x 1950 Fig. 10. Detail of irregular particles forming the apertural extension x 4300. 4 V. GOLEMANSKY AND C. G. OGDEN debris adhering to portions of the shell surface, which obscured some detail. The two speci- mens examined measured 39; 45um in length, 24; 21 urn in diameter and the diameter of the aperture 13-5; 10-3um respectively. In the original description Golemansky (1970) described the shell plates as being rectangu- lar, square or polygonal, sometimes rounded at the corners. It would seem, from the present observations, that the shell plates are usually rounded at the corners, but in all other respects agree with the earlier description. 80 r 60 40 20 - \, X v X >C x X X 50 60 70 80 90 length ( pm) Fig. 11 Measurements of breadth (points) and diameter of aperture (crosses), for individual specimens of Centropyxiella arenaria, plotted against the shell length. Centropyxiella arenaria Valkanov, 1970 This species appears to have a cosmopolitan distribution and is found in salinities varying from 26-l-37-8%o. The shell is colourless, ovoid in outline, sometimes having almost parallel lateral margins which curve evenly in the aboral region but often curve irregularly around the aperture (Fig. 6). In lateral view it is elliptical, tapering rapidly towards the aperture (Fig. 8). The aperture is oval and invaginated (Fig. 9). Most of the shell surface is smooth, and composed of overlapping, flattish particles of quartz held together by organic cement (Fig. 5). It would appear that the apertural surface is made of small flat particles, whilst the curved surface of the remainder of the shell is made from a mixture of small and large, flattish particles. The aperture is surrounded by a thickened, often rough, irregular disc-like extension composed of an assortment of quartz particles (Figs 6, 7 and 1 0). In the original description Valkanov (1970) noted that the quartz particles bordering the aperture were larger than those of the shell, whilst Golemansky (1976) described some small specimens, 28-3 lum long, with transparent organic shells and particles principally around the aperture. Centropyxiella elegans also described by Valkanov (1970) appears to differ in size alone, varying between 70-80um in length, whereas C. arenaria reaches only 60um. LITTORAL SPECIES OF TtESTATE AMOEBAE 5 Specimens of C. arenaria collected from the shore at Ostend, North Sea, varied between 35-60um (Chardez, 1977), whilst those from the Guinea Coast, Atlantic Ocean, were 53-56|im in length (Golemansky, 1976). Measurements of specimens in this study suggest that the length varies between 54-89(im. Individual dimensions of the breadth and the diameter of the aperture are expressed graphically against shell length in Fig. 11. It is possible that both C. arenaria and C. elegans are present in our sample, but the number of specimens examined are insufficient to reach any conclusions and we have therefore referred to them collectively as C. arenaria. Fig. 12 Collapsed specimen of Cyphoderia compressa showing the arrangement of the shell plates x 820. Fig. 13 Detail of kidney-shaped shell plates of Cyphoderia compressa x 9 100. Cyphoderia compressa Golemansky, 1979 This species was described recently from underground water littorals of marine beaches, and is similar to the common freshwater species C. ampulla (Ehrenberg, 1 840). It is distinguished from C. ampulla by the strong lateral compression of the shell. In transverse section C. compressa is oval and appears to taper evenly along most of the shell length, whereas in C. ampulla the transverse section is circular and has the greatest diameter around the mid-point of the shell and tapers rapidly at both extremities. Unfortunately, as the shells of this species are fragile, they collapsed when air dried in preparation for scanning (Fig. 1 2). In species of Cyphoderia there appears to be two main arrangements of the siliceous shell plates, these are (i) individually separate plates that lie close together and (ii) those that over- lap. These arrangements have been described in detail by Penard (1902), Cash et ai, (1915) and more recently by Ogden & Hedley (1980). The shell plates of C. compressa (Fig. 13) are similar to those of C. ampulla, although the number of kidney-shaped plates -about 1 -Sum long and 1 4um wide - and the amount of visible organic matrix appears to be greater in the former species. 6 V. GOLEMANSKY AND C. G. OGDEN References Camber, R., Thomas, R. & Le Blanc, M. 1963. Recherches sur la constitution de la theques des Arcelles (Genre Arcella, Rhizopoda testace); observations au microscope electronique. C. r. Acad. Sci, 256: 1364-1366. Cash, J., Wailes, G. H. & Hopkinson, J. 1915. The British Freshwater Rhizopoda and Heliozoa. Vol. 3 Rhizopoda, part 3. The Ray Society, London. 1 56pp. Chardez, D. 1977. Thecamoebiens du Mesopsammon des Plages de la Mer du Nord. Revue verviet. Hist nat. 34 Nos 4 a 6. Golemansky, V. 1970. Rhizopodes nouveaux du psammon littoral de la mer Noire. Prostistologica 6 : 365-371. 1976. Rhizopodes psammobiontes (Protozoa, Rhizopoda) du psammal supralittoral des cotes guineennes de 1'Atlantique. Acta zool. bulg4 : 23-29. 1979. Cyphoderia compressa n.sp. (Rhizopoda: Arcellinida)-un nouveau thecamoebien psammbionte du supralittoral des mers. Acta Protozool. 18 : (in press). Hedley, R. H. & Ogden, C. G. 1973. Biology and fine structure of Euglypha rotunda (Testacea: Protozoa). Bull. Br. Mus. nat. Hist. (Zool.) 25 : 1 19-137. Hedley, R. H., Ogden, C. G. & Krafft, J. I. 1974. Observations on clonal cultures of Euglypha acanthophora and Euglypha strigosa (Testacea: Protozoa). Bull. Br. Mus. nat. Hist. (Zool.) 27 : 103-111. Hedley, R. H., Ogden, C. G. & Mordan, N. J. 1976. Manganese in the shell of Centropyxis (Rhizopodea: Protozoa). Cell. Tiss. Res. Ill : 543-549. 1977. Biology and fine structure of Cryptodifflugia oviformis (Rhizopodea: Protozoa). Bull. Br. Mus. nat. Hist. (Zool.) 30 : 3 1 1-328. Mercier, M., Le Blanc, M.. Thomas, R. & Cambar, R. 1964. Observations, en microscopic electronique, sur la constitution de la theque de quelques Euglyphidae (Rhizopodes, testaces). C. r. Acad. Sci. (Paris) 258 : 5967-5968. Moraczewski, J. 197 1 . Structure et formation de la coque d' Arcella. Acta Protozol. 8 : 423-437. Netzel, H. 1975. Struktur und Ultrastruktur von Arcella vulgaris var. multinucleata (Rhizopoda, Testacea). Arch. Protistenk. Ill : 219-245. 1976a. Die Abscheidung der Gehausewand bei Centropyxis discoides (Rhizopoda, Testacea). Arch. Protistenk. 118 : 53-9 1 . 19766. Die Ultrastruktur der Schale von Difflugia oviformis (Rhizopoda, Testacea). Arch. Protistenk. 118:321-339. Ogden, C. G. 1979. Comparative morphology of some pyriform species of Difflugia (Rhizopoda). Arch. Protistenk. 122 : 143-153. Ogden, C. G. & Hedley, R. H. 1980. An Atlas of Freshwater Testate Amoebae. British Museum (Natural History) & Oxford University Press, London & Oxford. Penard, E. 1 902. Faune Rhizopodique du Bassin du Leman. Geneva. 700pp. Thomas, R. 1958. Observations sur le revetement des Trinema. Bull. Microsc. appl. 8 : 105-108. Thomas, R. & Hovasse, R. 1962. Sur la constitution des theques des Thecamoebiens. I. Le genre Trinema et Trinema lineare Penard. Bull. Microsc. appl. 12 : 1 1 7-1 19. Valkanov, A. 1970. Beitrag zur Kenntnis der Protozoen des Schwarzen Meeres. Zool. Anz. 184 : 241-290. Manuscript accepted for publication 25 June 1979 Notes on the hydroid, Synthecium evansi P. F. S. Cornelius Zoology Department, British Museum (Natural History), London SW7 5BD, England Synopsis The Mediterranean hydroid, Synthecium evansi (Ellis & Solander, 1 786), has been recorded twice from British waters but neither record can be proved. As the species has not been reported from Britain in the past 1 50 years it is now removed from the British list. British records Synthecium evansi (Ellis & Solander, 1 786) is a large and distinctive hydroid recorded widely in the Mediterranean Sea. However, its first description was based on material said to have been 'brought from' Great Yarmouth, Norfolk, England. But the species has not been recorded from British waters for 1 50 years and the Norfolk record is in doubt. It is true that the collector, John Evans, worked for the East India Company and was based at Great Yarmouth. Further, George Johnston (1838) recorded that Evans sent specimens to James Petiver's once famous Museum in London and there seems no reason to doubt that he could have supplied the S. evansi material to Ellis. Indeed, Ellis often received specimens from traders and based several new species on them. For example Mediterranean material is mentioned in Ellis & Solander's book in five places (pp. 15, 27, 88, 101, 149) and other specimens, from still farther afield, were brought back by East India Company ships (p. 58, twice on p. 107). But as Ellis & Solander's book was posthumous to them both it is possible that the locality data did not receive Ellis' usual careful attention. Further, the illustrations of S. evansi were left out of the book and are to be published only now (Wells & Cornelius, in prep.). There was clearly some confusion about the first description and the Norfolk locality cannot be accepted without further proof. Unfortunately some authors (Lamouroux, 1816; Johnston, 1838, 1847; Landsborough, 1 852) repeated the Norfolk record and although it was overlooked by Hincks (1 868) and later compilers its accuracy has not been questioned. The only other British record is also suspect. Lamouroux (in Lamouroux, Bory de Saint- Vincent & Deslongchamps, 1824) stated that he had received material from 'the coast of England'. But this material was not mentioned in his two earlier books (Lamouroux, 1816, 1821); and no specimens were found in Lamouroux' collections by Billard ( 1 909) and Redier (1967). Once again there is no proof that any specimens were collected in British waters. Conclusion There have been no records of S. evansi from Britain or anywhere else in NW Europe for 1 50 years. There is only a slim chance that such a large species would have been missed by the many collectors of this period and it seems right to take S. evansi off the British list. References Billard, A 1909. Revision des especes types d'hydroi'des de la collection Lamouroux conservee a 1'Institut Botanique de Caen. Annls Sci. nat. Zool. (9) 9 : 307-336. Ellis, J. & Solander, D. C. 1786. The natural history of many curious and uncommon zoophytes collected from various parts of the globe. London. Hincks, T. 1 868. A history of the British hydroid zoophytes. London. Johnston, G. 1838. A history of the British zoophytes. London. Bull. Br. Mus. nat. Hist. (Zool.) 38 ( 1 ): 7-8 Issued 27 March ' 98° 8 P. F. S. CORNELIUS - 1 847. A history of the British zoophytes. 2nd edn. London. Lamouroux J. V. F. 1816. Hisloire des polypiers corattigtnes flexible*, vulgairemenl nommes zoophytes. Caen. 1 82 1 . Exposition methodiques des genres de I'ordre des polypiers. Paris. Bory de Saint- Vincent, J. B. G. M. & Deslongchamps, E. 1824. Histoire naturelle des zoophytes, ou animaux rayonnes, faisant suite a I'histoire naturelle des vers, de Bruguiere. In: Encyclopedie methodique(supp\.). Paris. Landsborough, D. 1 852. A popular history of British zoophytes, or corallines. London. Redier, L. 1967. Revision de la collection du Museum des hydraires de Lamouroux. Bull. Mus. natn. Hist. nat. Paris (2) 39 : 38 1-4 1 0. Wells, J. W. & Cornelius, P. F. S. (in prep). On The unpublished plates of Ellis & Solander, 1 786. Manuscript accepted for publication 13 March 1979 The life history and ecology of the aberrant bathypelagic genus Benthomisophria Sars, 1909 (Copepoda: Misophrioida) G. A. Boxshall Department of Zoology, British Museum (Natural History), South Kensington, London SW7 5BD, England H. S. J. Roe Institute of Oceanographic Sciences, Brook Road, Wormley, Godalming, Surrey Introduction The copepod order Misophrioida comprises the littoral genus Misophria Boeck, 1864 and the rare bathypelagic genus Benthomisophria Sars, 1909. The only published records of the latter genus are those of Sars (1909), Hulsemann & Grice (1964) and Tanaka (1966 -as Misophria}. Little is known of the life history and ecology of either of the two Bentho- misophria species, primarily because of the difficulty in obtaining large samples of these bathypelagic forms. In 1976 the Institute of Oceanographic Sciences commenced a total water column sampling programme, integrating midwater, near-bottom and benthic hauls. A new technique for taking near-bottom samples was employed and this is described in detail below. Analysis of the first complete series taken under this programme revealed large numbers of Benthomisophria, including many developmental stages. These specimens were found to belong to B. palliata Sars, 1909 and to B. cornuta Hulsemann & Grice, 1964. The 5 copepodid stages of B. palliata, 3 copepodid stages and male of B. cornuta are described for the first time. Notes on the vertical distribution of both species are presented, and phylo- genetic relationships within the Misophrioida are examined. Materials and methods Sampling methods The samples were taken with the acoustically controlled RMT 1-1-8 net system described in Baker, Clarke & Harris (1973). The small mouth area RMT 1 has a mesh size of 320 um and the larger RMT 8 has a 4-5 mm mesh. The effective mouth areas of these nets vary with tow- ing speed (Baker et al, 1973; Roe & Baker, in prep.). The nets are opened and closed acoustically using a telemetering net monitor that activates a mechanical release gear. The monitor continuously telemeters the depth, temperature, speed and distance travelled (collectively called flow) and an indication of whether the nets are open or closed. This infor- mation is displayed on the ship on a modified Mufax facsimile recorder (Roe & Harris, in press). The sampling programme required samples to be taken in discrete depth layers from just above the sea bottom to the surface. In order to fish the nets to within approximately 5m of the sea bed in depths of 4000m a system of bottom indicator switches was used. This con- sisted of 2 mercury switches, each encased in a streamlined protective housing, connected to the net monitor by a single protected conducting cable. The cable was fixed down the side wires of the nets to the weight bar, and the switches hung below this bar so that their estimated distances below the bottom bar when fishing were 5m and 20m respectively. The near-bottom hauls were made by paying out the net until it was within 20m of the bottom, as shown by the lower indicator switch hitting the sea bed and tilting so that the mercury within the switch made an electrical contact. This contact was telemetered back to the ship where it Bull. Br. Mm. not. Hist. (Zool.) 38 ( 1 ): 9-41 Issued 27 March 1 980 10 G. A. BOXSHALL AND H. S. J. ROE was displayed on the Mufax as a displacement of the net indicator trace. Paying out more warp caused the second switch to hit the sea bed, resulting in a further displacement of the net trace. The first shift in this trace therefore indicated that the net was about 20m above the sea bed and the second that it was about 5m above the bottom. The net was then opened and maintained within this 5-20m layer by adjusting the amount of warp so that the 20m indicator was continuously registering with the 5m indicator appear- ing intermittently. Two hauls, Station numbers 9541 # 18 and 19 (Table 1), were made in this way. The remaining hauls sampled progressively further from the bottom. Table 1 Station data of the hauls in which Benthomisophria spp. were found. Station No. Sounding range (m) Mean Sounding (m) Depth off bottom of sample (m) Depth from surface of sample (m) Duration of haul (H) 9541#18 #\9 #22 #24 #25 #26 #21 9131#18 #19 #23 3974^036 4008-4060 3856-3914 4000-4079 4008-4105 4005 4031 3883 4035 4058 5-20 5-20 20-100 100-500 500-1000 20-100 100-500 500-1000 2510-3000 2000-2500 6 6 6 6 6 4 4 4 5 4 23856^105 3918-3945 3839-3889 3870-3896 3957-4036 3865-3934 4002 3934 3860 3884 3998 3905 Since the net was a considerable distance behind the ship during these deep hauls the bottom depth over which it was actually fishing was derived as follows. The sounding beneath the ship was recorded on a Precision Echo Sounder (PES) continuously during the paying out and retrieval of the nets and periodically throughout the fishing periods. From these data a bathymetric chart of the area was subsequently constructed. The bottom was found to be essentially flat with the soundings in each haul varying by less than 100m (Table 1) in a total distance travelled over the ground of between 19-1-21-9 Km. The position of the net whilst fishing was calculated by triangulation - knowing at a given moment the ship's course, the depth below the ship and the amount of wire out. Because the gradient of the bottom was so slight it was possible to ignore errors in triangulating the net position caused by a changing depth and by lateral displacement of the net behind the ship. This calculated net track was then superimposed upon the bathymetric chart and the mean bottom depth over which the net had fished was derived from the soundings taken at every one minute of longitude on the chart. Station number 9541*22 (Table 1) fished between about 20 and 100m off the bottom. The 20m indicator switch was used to determine the lower limit of the haul whereas the upper limit was obtained from the monitor depth. The net was fished so that depth according to the monitor was within 100m of the bottom sounding at the position derived from triangulating back along the ship's track. All subsequent hauls up to 1000m off the bottom were fished using the monitor in this way. Table 1 gives the sounding ranges and mean soundings for these hauls. The depths of samples taken at up to 1000m off the bottom are expressed in terms of distance off the bottom. The depths of hauls shallower than 1000m off the bottom were taken directly from the monitor, i.e. measured from the surface. The absolute depths fished using the techniques described are probably inaccurate, as both PES and monitor depths are subject to error due to the changes in water density and temper- ature, but they are accurate relative to both the bottom and each other. Any effect which the LIFE HISTORY OF BENTHOM1SOPHRIA 11 bottom may have on the bathypelagic animal populations is presumably proportional to their distance away from it, and these relative distances are consequently much more impor- tant ecologically than a possible error of a few metres in absolute depth. The changeover depth to conventional surface measurements was arbitrary. It may be shown subsequently that a distance of 1000m off the bottom is not the optimum changeover point although Brewer, Spencer, Biscaye, Hanley, Sachs, Smith, Radar & Fredericks (1976) have found that this distance is approximately the top of the nepheloid layer, at least further north in the Atlantic. Most of the samples used here (Table 1) were taken at Station 9541 in April 1977, within an area 20°0-6'N to 20°254'N and 21°9-6'W to 21'564'W. These are the hauls described above. Station 9131 was fished in the same area in November 1976 as a preliminary to the detailed programme of April 1977. The near-bottom samples from this station were taken in a similar way to that described above but only a single bottom indicator switch was used. Material The samples were preserved on the ship in 5% formalin. Subsequently they were transferred to a preserving fluid based on that of Steedman (1974) and the Benthomisophria spp. were picked out of the total RMT 1 and RMT 8 samples. A small triangular net (the DN) of mesh size 61|im was added to the RMT 1+8 combination net to sample small zooplankton. The DN catch was examined for one haul (9541 # 24) and contained 3 specimens of B. palliata and a single B. cornuta. All the material came from the hauls shown in Table 1 except for a single adult female B. palliata which was found in a RMT 8 sample fished between 3000 and 3500m in a position 44°27'N 12°44'W. The numbers of B. palliata and B. cornuta caught are given in Tables 2 and 3. No specimens of either species were found in hauls shallower than 2000m. Undamaged specimens were measured from the tip of the cephalothorax to the end of the caudal rami. The numbers measured are also given in Tables 2 and 3. There was no differ- ence in the size distribution of either species between stations 9131 and 9541 and data from both stations have been combined in the frequency distributions of body length (Figs 14 & 15). Table 2 The numbers of each stage of Benthomisophria palliata found and measured. Net RMT1 RMT 8 Stage I II III IV V 9 d IV V 9 cf Station 9541*18 38 13 41 14 16 11 1 5 4 19 7 10 63 17 39 38 1 25 17 22 5 9 12 13 22 4 2 5 7 1 24 7 44 17 15 15 13 17 5 4 2 25 3 8 26 3 6 1 1 26 1 5 6 11 6 5 2 4 27 1 9131*18 8 16 26 10 8 4 2 13 2 19 5 37 33 22 12 17 4 1 4 9 1 23 n/"\t £» v Q TTI i n ^H 9 3 8509#20 IlOL CAallllllCU 1 Total of RMT1, RMT8+DN 14 144 107 201 132 203 117 No. measured 14 133 98 151 103 186 93 12 G. A. BOXSHALL AND H. S. J. ROE Table 3 The numbers of each stage of Bentho- misophria cornuta found and measured. Net RMT 1 Stage III IV V 9 d Station 9541#18 1 2 2 2 19 1 1 22 1 1 1 24 2 1 1 25 1 1 26 3 3 9131#18 4 4 1 19 1 3 2 Total inc. DN 4 9 11 10 6 No. measured 4 9 11 10 6 The copepods were stained in chlorazol black in lactophenol, dissected and examined in lactophenol and mounted as permanent preparations in polyvinyl lactophenol. They were examined by phase contrast microscopy and drawings were made with the aid of a camera lucida. Descriptions Family MISOPHRIIDAE Genus BENTHOM1SOPHRIA Benthomisophria palliata SARS, 1909 BenthomisophriapalliataG. O. Sars, 1909: 1-4, Fig. 1. MisophriajaponicaO. Tanaka, 1966: 51-55, Figs 1-2. The descriptions of the adults of both sexes and of the copepodid stages of B. palliata are based on the examination of a total of 918 specimens. The specimens were sorted, on morphological criteria, into copepodid stages I to V, adult male and adult female. COPEPODID i Prosome (Fig. 1A) large, 5-segmented and comprising 86% of total body length; rounded anteriorly, with greatest width at posterior border of cephalosome. Last prosome somite with parallel sides and small postero-lateral projections. Urosome 1 -segmented, sub-rectangular. Caudal rami apparently bearing only 1 dorsal and 2 distal armature elements. First antenna (Fig. IB) 6-segmented; with incomplete armature, elements present as follows: 1-2+1 spine, II - 1 aesthete, III - 0, IV - 0, V - 1 , VI - 6. Second antenna (Fig. 1 C) biramous with 5-segmented exopod and 3-segmented endopod, first segment of which is partially fused to basis. Labrum large, fused to rostrum. Mandible (Fig. ID) blade with 4 large teeth and several other elements; mandibular palp biramous, with 2-segmented endopod and 4-segmented exopod. First maxilla (Fig. IE) gnathobase with 2 naked and 10 spinulate setae; rudimentary outer lobe on basal segment bearing at least 3 setae; inner lobes 1 and 2 about equal in size and armed with 4 and 3 setae respectively; distal endopod seg- ment with at least 3 distal setae (some elements missing), proximal segment also with 3 distal setae; exopod 1 -segmented, armature incomplete. Second maxilla (Fig. IF) apparently 5-segmented: first segment with 2 inner lobes both armed with 3 setae; second segment with LIFE HISTORY OF BENTHOM1SOPHRIA 13 Fig. 1 B. palliata Copepodid I. A, dorsal; B, first antenna; C, second antenna; D, mandible; E, first maxilla; F, second maxilla; G, maxilliped; H, Leg 1 ; I, leg 2; J, leg 3; K, leg 4. Scales 1 00(jm unless otherwise indicated. 14 G. A. BOXSHALL AND H. S. J. ROE 2 inner lobes bearing 3 and 4 setae; third segment extended medially into a large curved pro- cess carrying 2 setae basally; two distal segments bearing a total of 10 setae. Maxilliped (Fig. 1G) 5-segmented, armature incomplete in all available specimens. Thoracic legs 1 to 4 present; legs 1 and 2 biramous, leg 3 bilobed and leg 4 rudimentary. Leg 1 (Fig. 1H) with 2-segmented sympod and 1 -segmented rami; exopod bearing 4 outer margin spines, an apical spine and 3 inner margin setae; endopod with 1 outer, 2 distal and 4 inner margin setae. Leg 2 (Fig. 1 1) segmentation as for leg 1 ; exopod missing in all specimens; endopod with 1 outer, 2 distal and 3 inner margin setae. Leg 3 (Fig. 1J) a broad flattened structure, bilobed distally with each lobe bearing 2 apical armature elements. Leg 4 (Fig. 1 K) rudimentary, comprising a spinous process on a broad base. Body length ranging from 0-69 to 0-95mm, with a mean of 0-79mm (based on 14 specimens). COPEPODID ii Prosome (Fig. 2A) large, 4-segmented; first pedigerous somite incorporated into cephalo- some, but with some indication of surture line marking the boundary. Urosome (Fig. 2B) 2-segmented; first segment with small processes at postero- lateral angles, second segment sub-rectangular. Caudal rami with 1 dorsal, 1 lateral and 2 distal elements. First antenna (Fig. 2C) 9-segmented; armature incomplete, elements present as follows: 1 - 4, II - 0, III - 1 spine, IV - 1 aesthete, V - 1 , VI - 1 , VII - 0, VIII - 2, IX - 6 + 1 aesthete. Second antenna as in copepodid stage I but with exopod apparently comprising 5 distinct segments (Fig. 2D) and possibly a sixth segment fused to the basis; each of first 4 distinct seg- ments bearing 1 seta, terminal segment with at least 1 element. Labrum, mandibles, maxillae and maxilliped as in adult. Leg 1 (Fig. 2E) biramous, with 2-segmented rami; exopod segment 1 with 1 outer spine, inner seta missing; exopod segment 2 missing; endopod segment 1 with 1 inner seta, segment 2 with 1 outer, 2 distal and 4 inner margin setae. Leg 2 (Fig. 2F) as for leg 1 except endopod segment 2 with only 3 inner margin setae. Leg 3 (Fig. 2G) with 2-segmented sympod, biramous but with both rami missing in all specimens. Leg 4 (Fig. 2H) a flattened structure, bilobed distally with each lobe carrying 2 apical armature elements. Body length ranging from 0-76 to l-33mm, with a mean of l-03mm (based on 133 specimens). COPEPODID in Prosome (Fig. 21) large, 4-segmented, with first pedigerous somite fully incorporated into cephalosome. Urosome (Fig. 2J) 3-segmented; first segment with leg 5 present, second seg- ment with rudiment of leg 6 present, third segment sub-rectangular. Caudal rami with 6 armature elements. First antenna (Fig. 2K) 1 2-segmented; armature incomplete, elements present as follows: 1-1, II-7, HI-2, IV-0, V-l, VI- 1 spine, VII-1 + 1 aesthete, VIII- 1, IX- 1, X-0, XI - 1 , XII - 6 + 1 aesthete. Second antenna, labrum, mandible, maxillae and maxilliped as in adult. Leg 1 (Fig. 3A) biramous, with 2-segmented rami; exopod segment 1 with 1 outer spine and 1 inner seta, segment 2 with 3 outer spines, 1 apical spine and 4 inner margin setae; endopod segment 1 with 1 inner seta, segment 2 with 1 outer, 2 distal and 5 inner margin setae. Leg 2 (Fig. 3B) as for leg 1 except exopod segment 2 missing in all specimens, and endopod segment 2 with only 4 inner margin setae. Leg 3 (Fig. 3C) with both rami 2-segmented but incomplete in material available. Leg 4 (Fig. 3D) with 2-segmented sympod, biramous but with both rami missing in all specimens. Leg 5 (Fig. 2L) 1 -segmented, with 1 inner proximal and 3 distal setae. Leg 6 (Fig. 2M) represented by an unarmed spinous process. Body length ranging from 0-99 to 1-7 1mm, with a mean of l-29mm (based on 98 specimens). LIFE HISTORY OF BENTHOMISOPHRIA 15 Fig. 2 B. palliata Copepodid II & III. A, Copepodid II, dorsal; B, urosome, ventral; C, first antenna; D, exopod of second antenna; E, leg 1; F, leg 2; G, leg 3; H, leg 4; I, Copepodid III, dorsal; J, urosome, ventral; K, first antenna; L, leg 5; M, leg 6. Scales lOOum unless otherwise indicated. 16 G. A. BOXSHALL AND H. S. J. ROE Fig. 3 B. palliata Copepodid III & IV. A, Copepodid III leg 1; B, leg 2; C, leg 3; D, leg 4; E, Copepodid IV, dorsal; F, first antenna; G, urosome, ventral; H, leg 1 ; I, leg 3; J, leg 2; K, leg 4; L, leg 6. Scales 200|am unless otherwise indicated. LIFE HISTORY OF BENTHOMISOPHRIA 17 COPEPODID IV Prosome (Fig. 3E) large, 4-segmented. Urosome (Fig. 3G) 4-segmented; first and second seg- ments bearing legs 5 and 6 respectively; segments 1 to 3 all with conspicuous reticulate surface markings; urosome segment 4 without surface markings. First antenna (Fig. 3F) 15-segmented; armature elements present as follows; I - 1, II - 12, III -2, IV -3, V-2+1 aesthete, VI -2, VII -3 + 1 aesthete (missing from figured specimen), VIII-2, IX-1 + 1 spine, X-l, XI- 1, XII- 1, XIII-2, XIV -2, XV-7+1 aesthete. Second antenna, labrum, mandible, maxillae and maxilliped as in adult. Leg 1 (Fig. 3H) and leg 2 (Fig. 3J) as in Copepodid III stage, except endopod segment 2 of leg 2 now with 5 inner margin setae. Leg 3 (Fig. 31) exopod segment 1 with 1 outer spine and 1 inner seta, segment 2 missing in all specimens; endopod segment 1 with 1 inner seta, segment 2 with 1 outer, 2 distal and 4 inner margin setae. Leg 4 (Fig. 3K) biramous, both rami 2-segmented, armature and rami incomplete in all available specimens. Leg 5 (Fig. 3G) 2-segmented; basal segment with indication of subdivision laterally and armed with 1 inner and 1 outer seta; second segment with 3 apical setae. Leg 6 (Fig. 3L) a small process with 1 long and 1 short apical seta. Body length ranging from 1-4-1 to 2-55mm, with a mean of l-88mm (based on 151 specimens). COPEPODID v Prosome (Fig. 4A) large, 4-segmented, with epimeral plates of last free thoracic somite not acutely pointed. Urosome 5-segmented; segments 1 to 4 with reticulate surface markings, segment 5 without surface markings. Caudal rami (Fig. 4B) with 6 armature elements. First antenna (Fig. 4C) 1 7-segmented, armature elements present as follows: I - 1 , II - 10, III - 2, IV - 8, V - 1 + 1 aesthete, VI - 2, VII - 1 + 1 aesthete, VIII - 2, IX - 2, X - 2, XI - 2, XII -2, XIII -1, XIV -0, XV -2, XVI -2+1 aesthete, XVII-7+1 aesthete. Second antenna, labrum, mandible, maxillae and maxilliped as in adult. Legs 1 to 4 (Figs 4D-G) biramous, with 3-segmented rami; armature formula as follows: coxa basis endopod exopod legl 0-1 1-0 0- 1 ; 0-2; 1,2,3 I- 1 ; I - 1 ; missing Ieg2 0-1 1-0 0-1 ;0-2;1.,2,3 I- 1 ; I- 1 ; 111,1,5 leg 3 0-1 1-0 0- 1 ; 0-2 ; missing I- 1*; I- 1; missing leg 4 0-1 1-0 0- 1 ; 0-2*; missing I- 1 ; I- 1 ; missing *elements missing but presumed present. Leg 5 (Fig. 4H) as in adult. Leg 6 (Fig. 41) a process armed with an outer plumose seta and 2 unequal apical elements. Body length ranging from 2-01 to 3-75mm, with a mean of 2-71mm (based on 103 specimens). ADULT MALE Prosome (Fig. 5A) comprising 70 to 75% of total body length; epimeral plates of last free thoracic somite produced posteriorly and provided with expanded hyaline areas laterally. Urosome 6-segmented; segments 1 to 5 with reticulate surface markings (not figured), seg- ment 6 with surface denticulations. First antenna (Fig. 4J) apparently 1 6-segmented and geniculate, with the articulation between segments XIV and XV; aesthetes present on segments I, II (2), III, V, VII, X, XI, XIII, XV and XVI (2); segment XV also with a long spiniform process originating near base of aesthete; other armature elements present but omitted from figure. All other appendages as for adult female except leg 6. Leg 6 (Fig. 5E) a process with a plumose outer seta, a long plumose apical seta and a short inner spiniform element. Body length of males ranging from 2-70 to 4-72mm, with a mean of 3-75mm (based on 93 specimens). 18 G. A. BOXSHALL AND H. S. J. ROE Fig. 4 B. palliata Copepodid V & adult male. A, Copepodid V, dorsal; B, caudal ramus, ventral; C, first antenna; D, leg 1 ; E, leg 2; F, leg 3; G, leg 4; H, leg 5; I, leg 6; J, Adult male first antenna (with setae omitted). Scales 200um unless otherwise indicated. LIFE HISTORY OF BENTHOMISOPHRIA 19 Fig. 5 B. palliata Adult male & female. A, Male, dorsal; B, Female urosome, ventral; C, Female, dorsal; D, leg 5; E, Male leg 6; F, Female first antenna. Scales 200um unless otherwise indicated. 20 G. A. BOXSHALL AND H. S. J. ROE ADULT FEMALE Prosome (Fig. 5C) large, 4-segmented with reticulate surface markings (not shown in Fig. 5C); segments 1 to 3 with strips of hyaline membrane dorsally along their posterior borders; segment 4 with posteriorly produced epimeral plates and with expanded hyaline areas laterally. Urpsome (Fig. 5B) 6-segmented; segments 1 to 5 with reticulate surface markings, segment 6 with surface denticulations. Labrum (Fig. 7B) strongly developed, fused to rostrum and projecting postero-ventrally from ventral body surface. First antenna (Fig. 5F) 18-segmented; armature elements as follows: I- 1, II -7, HI -2, IV - 2, V - 10, VI - 2 + 1 aesthete, VII - 2, VIII -2 + 1 aesthete, IX - 2, X - 2 + 1 aesthete, XI -1, XII -2, XIII-1 + 1 aesthete, XIV- 1, XV -1, XVI -2, XVII -2+1 aesthete, XVIII -6+1 aesthete. Second antenna (Fig. 6 A), basis with 1 outer distal seta; endopod 3 -segmented, exopod with 5 distinct segments and possibly a sixth fused to basis. Endopod segment 1 with 1 outer margin seta, segment 2 with 3 outer margin setae and segment 3 with 6 long plumose apical setae. Exopod with 1 unilaterally pinnate seta on each of the first 4 distinct segments, and with 3 similar setae on terminal segment. Mandible (Fig. 6B) blade with 8 main elements on distal margin (numbered 1-7 on Fig. 6B as element 8 is identical to element 7). Mandibular palp biramous, with 2-segmented endopod and 4-segmented exopod; endopod segments 1 and 2 bearing 1 and 2 setae respec- tively; exopod segments 1 to 4 bearing 1,1,1 and 3 long plumose setae respectively. First maxilla (Fig. 6C) gnathobase with 15 elements; inner lobes 1 and 2 with 6 and 4 plumose setae respectively; outer lobe rudimentary, represented by 7 plumose setae on sur- face of segment; endopod apparently 2-segmented with distal segment positioned at an angle of 90° to proximal segment; proximal segment with 4 plumose setae apically, distal segment with 4 naked setae medially, 1 naked and 4 plumose setae apically; exopod 1 -segmented, with 2 plumose setae medially and 5 long plumose setae distally. Second maxilla (Fig. 6D-E) apparently 6-segmented; first segment with 5 plumose setae on proximal inner lobe and 3 on distal inner lobe; second segment with 2 inner lobes, each bearing 3 apical setae; third segment produced medially into a curved claw armed with 2 slender naked setae near its base; segments 4 to 6 bearing a total of 1 1 elements, 7 strong curved setae armed with 1 or 2 rows of pinnules and 4 slender naked setae. Maxilliped (Fig. 6F) apparently 5-segmented; first segment with a single pinnate seta on a raised denticulated area situated proximally on the medial surface, a group of 1 small naked seta and 2 pinnate setae on another raised denticulated area in the middle, and 2 pinnate setae distally; second segment with a medial row of pinnules, 3 pinnate setae distally on medial margin and 3 areas of denticulations; third segment small and armed with a single pinnate seta; fourth segment partially subdivided into 3 portions each bearing 1 seta; fifth segment very small, bearing 4 slender setae. Legs 1 to 4 (Figs 7C-F) biramous, with 3-segmented rami; armature formula as follows: coxa basis endopod exopod legl 0-1 1-0 0-1 ;0-2;1,2,3 I- 1 ; I- 1 ; 111,1,4 Ieg2 0-1 1-0 0-1 ;0-2; 1,2,3 I- 1 ; I- 1 ; 111,1,5 Ieg3 0-1 1-0 0-1 ;0-2; 1,2,3 I- 1 ; I- 1 ; 111,1,5 leg 4 0-1 1-0 0 - 1 ; 0 - 2 ; missing I - 1 ; missing Pinnule rows present on lateral margins of all endopod segments and medial margins of all exopod segments. Lateral margins of all exopod segments armed with strips of serrated membrane. All outer margin spines bilaterally serrate, apical spines on exopods of legs 2 and 3 bilateraly serrate, that on leg 1 unilaterally serrate. Leg 5 (Fig. 5D) 2-segmented; proximal segment with indication of suture line laterally, bearing a long outer seta apically and a medium length seta on inner margin; distal segment LIFE HISTORY OF BENTHOMISOPHRIA 21 Fig. 6 B. palliata Adult female. A, second antenna; B, mandible - showing armature elements 1-7 enlarged; C, first maxilla -with rami separated from limb; D, second maxilla -showing segmentation; E, second maxilla with segments separated from each other; F, maxilliped. Scales 200um unless otherwise indicated. 22 G. A. BOXSHALL AND H. S. J. ROE Fig. 7 B. palliata Adult female. A, leg 6; B, labrum, ventral; C, leg 1; D, leg 2; E, leg 3; F, leg 4. Scales 0-5mm unless otherwise indicated. LIFE HISTORY OF BENTHOMISOPHRIA 23 bearing 3 unequal setae along oblique distal margin. Leg 6 (Fig. 7A) a broad plate produced into a small process laterally, bearing a long plumose seta on its apex, and with 2 short spini- form elements on lateral portion of posterior margin. Body length of females ranging from 2-88 to 5-70mm, with a mean of 4-Omm (based on 186 specimens). MATERIAL EXAMINED 19 , Id1 and ICo.V; S.A.S. Le Prince Albert de Monaco, Stn 2738 Cote W. du Portugal, 1908; donated by G. O. Sars, stored in the Musee Oceanographique de Monaco. 20399, H7cf c? , 132 Co.V, 201 Co.IV, 107 Co.III, 144 Co.II, 14 Co.I; 'Discovery' Stns 9541, 9131 and 8509 (see Table 2); BM(NH) registration numbers 1979.1-10 (99), 1 979.1 1-20 (cfcf), 1979.2 1-30 (Co.V), 1979.3 1-40 (Co.IV), 1979.41-50 (Co.III), 1979.51-60 (Co.II) and 1979.6 1-70 (Co.I). REMARKS A certain amount of confusion has been caused by inconsistencies in the original description of B. palliata (Sars, 1909). The type 'specimen' figured by Sars exhibits features charac- teristic of two different life cycle stages, i.e. the 6-segmented urosome of an adult female and the 1 7-segmented first antenna of a Copepodid V stage. It is probable that Sars added the first antennae to his drawing of the body of an adult female. In an attempt to resolve this situation Sars' material of B. palliata, stored in the Musee Oceanographique de Monaco, was examined. This material comprises an adult female (with a 6-segmented urosome and 18-segmented first antenna), an adult male and a Copepodid V stage (with a 5-segmented urosome and 1 7-segmented first antenna). If this is the type material it would provide confir- mation that Sars' figured 'specimen' combines features of both adult female and Copepodid V. However, it is not certain that this is the type material because the label with the material gives only one station (No. 2738) whereas Sars (1909) states that his 3 type specimens were taken at 3 separate stations (Nos. 819, 1479 and 2 108). Misophria japonica Tanaka, 1966 was described from one specimen of each sex and does not differ significantly from B. palliata. It is therefore regarded as a junior synonym. Tanaka's specimens were taken in a haul made between 1000-Om depth in the Pacific Ocean adjacent to the Izu region of Japan. This is the shallowest record of B. palliata. Benthomisophria cornuta Hulsemann & Grice, 1964 Benthomisophria cornuta K. Hulsemann & G. D. Grice, 1964: 259-260, Figs 1-15. Misophria maculata. O. Tanaka, 1966: 55-56, Fig. 3. The descriptions of the adults of both sexes and of the later copepodid stages of B. cornuta are based on the examination of a total of 40 specimens. The specimens were sorted, on morphological criteria, into copepodid stages III to V, adult male and adult female. COPEPODID in Prosome (Fig. 8 A) large, 4-segmented, with the epimeral plates of the last two free thoracic somites produced posteriorly into acute projections. Urosome (Fig. 8B) 3-segmented; seg- ments 1 and 2 bearing rudiments of legs 5 and 6 respectively. Caudal rami with 6 armature elements as in adult. First antenna (Fig. 8C) 12-segmented; armature incomplete, elements present as follows: I - 1 , II - 6, III - 2 + 1 aesthete, IV - 1 , V - 1 + 1 aesthete, VI - 1 spine, VII - 0, VIII - 1 , IX- 1, X- 1, XI- 1 + 1 aesthete, XII -6+ 1 aesthete. Second antenna, labrum, mandible, maxillae and maxilliped as in adult. 24 G. A. BOXSHALL AND H. S. J. ROE 200 p M Fig. 8 B. comma Copepodid III & IV. A, Copepodid III, dorsal; B, urosome, ventral; C, first antenna; D, leg 1; E, leg 2; F, leg 3; G, leg 4; H, leg 5; I, leg 6; J, Copepodid IV, dorsal; K, urosome, ventral; L, leg 5; M, leg 6. Scales 1 OOum unless otherwise indicated. LIFE HISTORY OF BENTHOMISOPHRIA 25 Legs 1 to 3 (Figs 8D-F) biramous, with 2-segmented rami; leg 4 (Fig. 8G) biramous with 1 -segmented rami. Armature formula as follows: legl leg 2 leg 3 leg 4 coxa 0-1 0-1 0-1 0-1 basis 1-0 1-0 1-0 1-0 endopod 0-1; 1,2,5 0-1 ; 1,2,4 0-1 ; 1,2,3 1,2,3 exopod 1-1 ; 111,1,4 1-1 ; 111,1,5 I - 1 ; missing missing All exopod segments bearing strips of serrated membrane laterally and all outer margin spines bilaterally serrate. Leg 5 (Fig. 8H) a rounded process situated postero-laterally on first urosome somite, armed with a single rudimentary element subapically, another element may have been present but was missing from all specimens available for examination. Leg 6 (Fig. 81) a small spinous process. Body length ranging from 0-72 to 0-73mm, with a mean of 0-72mm (based on 3 specimens). COPEPODID iv Prosome (Fig. 8J) large, comprising about 75% of total body length; epimeral plates of somites 3 and 4 strongly produced posteriorly. Urosome (Fig. 8K) 4-segmented. First antenna (Fig. 9 A) 15-segmented, with armature elements present as follows: I- 1, 11-11, III-l + l spine, IV -2, V-3 + 1 aesthete, VI -1, VII-1 + 1 aesthete, VIII -2, IX-1 + 1 spine, X-l (+ ?), XI- 1, XII- 1, XIII-2, XIV- 1 + 1 aesthete, XV-5+1 aesthete. Second antenna, labrum, mandible, maxillae and maxilliped as in adult. Legs 1 to 4 (Figs 9B-E) biramous, with 2-segmented rami; armature formula as follows: legl leg 2 leg 3 leg 4 coxa 0-1 0-1 0-1 0-1 basis 1-0 1-0 1-0 1-0 endopod 0-1 ; 1,2,5 0-1 ; 1,2,5 0-1 ; 1,2,4 0-1 ; 1,2,3 1-1 1-1 1-1 1-1 exopod 111,1,4 111,1,5 missing missing Leg 5 (Fig. 8L) a sub-rectangular plate, incompletely subdivided by a transverse suture line; distal margin with 3 armature elements. Leg 6 (Fig. 8M) a broad, flat process with 2 armature elements. Body length ranging from 0-77 to 0-92mm, with a mean of 0-85mm (based on 7 specimens). COPEPODID v Prosome (Fig. 9F) with epimeral plates on last two free thoracic somites strongly produced posteriorly; dorsal and lateral surfaces of prosome with reticulate surface markings as shown in Fig. 9G on the right hand lateral portion of the cephalosome. Urosome (Fig. 10A) 5-segmented; segments 1 to 4 with reticulate surface markings, segment 5 without surface markings. First antenna (Fig. 10B) 16-segmented, with armature elements as follows: I- 1, II- 10, III -9, IV -2 + 1 aesthete, V-2, VI-2+1 aesthete, VII -2, VIII -2+1 aesthete, IX -2, X-2, XI -2 + 1 aesthete, XII -1, XIII -1, XIV -2, XV -2+1 aesthete, XVI -6+1 aesthete. Second antenna, labrum, mandible, maxillae and maxilliped as in adult. G. A. BOXSHALL AND H. S. J. ROE Fig. 9 B. cornuta Copepodid IV & V. A, Copepodid IV first afitenna; B, leg 1 ; C, leg 2; D, leg 3; E, leg 4; F, Copepodid V, dorsal; G, lateral portion of cephalosome - showing reticulate surface markings. Scales lOOum unless otherwise indicated. LIFE HISTORY OF BENTHOM1SOPHRIA 27 Fig. 10 B. cornuta Copepodid V. A, urosome, ventral; B, first antenna; C, leg 1 ; D, leg 2; E, leg 3; F, leg 4; G, leg 5. Scales lOOum unless otherwise indicated. 28 G. A. BOXSHALL AND H. S. J. ROE Legs 1 to 4 (Figs 10C-F) biramous, with 3-segmented rami; armature formula as follows: coxa basis endopod exopod legl 0-1 1-0 0-1 ;0-2; 1,2,3 I- 1 ; I- 1 ; 111,1,4 Ieg2 0-1 1-0 0-1 ;0-2; 1,2,3 I- 1 ; I- 1 ; 111,1,5 Ieg3 0-1 1-0 0-1 ;0-2; 1,2,3 I- 1 ; I- 1 ; 111,1,5 Ieg4 0-1 1-0 0-1 ;0-2; 1,2,2 I- 1 ; I- 1 ; 111,1,5 Lateral margins of all endopod segments except segment 3 of leg 1 with rows of pinnules; lateral margins of endopod segment 3 of leg 1 and all exopod segments armed with strips of serrated membrane. Leg 5 (Fig. 10G) a single segment, with no indication of any subdivision, armed with a long naked apical seta and 2 short naked setae distally on the medial margin; some denticulations also scattered over surface. Leg 6 (Fig. 10A) represented by 1 short and 2 long setae. Body length ranging from 0-85 to l-10mm, with a mean of 1-Omm (based on 10 specimens). ADULT MALE Prosome (Fig. 1 1 A) 4-segmented; epimeral plates of last two free thoracic somites strongly produced posteriorly; lateral indentations present on cephalosome (as in all other stages) allowing for movement of setae on mandibular palp and second antenna. Prosome with con- spicuous reticulate surface markings both dorsally and laterally. Urosome (Fig. 11B) 6-segmented; segments 1 to 5 with reticulate surface markings, segment 6 without surface markings. First antenna (Fig. 11D) apparently 1 6-segmented; geniculate,with articulation between segments XIV and XV; aesthetes present on segments I, II (2), III, V (2), VIII, IX, XI, XIII, XV and XVI (2); segment XV also with a long spiniform process originating near base of aesthete; other armature elements omitted from basal region in Fig. 1 1 D for greater clarity. All other appendages as in adult female. Body length of males ranging from 1-06 to l-37mm, with a mean of l-23mm (based on 6 specimens). ADULT FEMALE Prosome (Fig. 1 1G) 4-segmented, epimeral plates of last two free thoracic somites strongly produced posteriorly; surface of prosome with conspicuous reticulate surface markings (not figured) extending onto rostrum and labrum (Fig. 11C). Major reticulations on prosome divided into minute networks of fine markings, as shown on rostrum (Fig. 1 1C). Urosome 6-segmented, segments 1 to 5 with reticulate surface markings, segment 6 without markings. Labrum (Fig. 1 1C) strongly developed, fused to rostrum; projecting postero-ventrally from ventral body surface, and with reticulate markings on ventral surface. First antenna (Fig. 12A) 1 6-segmented; armature elements present as follows: I - 1 , II - 9, III - 5, IV - 1 + 1 spine + 1 aesthete, V - 1 , VI - 1 + 1 aesthete, VII - 0, VIII -2 + 1 aesthete, IX-1,X-1,XI-1 + 1 aesthete, XII -1, XIII -0, XIV -2, XV -2+1 aesthete, XVI - 7 + 1 aesthete. Second antenna (Fig. 12B) basis with 1 outer distal seta; endopod 3-segmented and exopod with 5 distinct segments and possibly a sixth fused to basis. Endopod segment 1 unarmed, segment 2 with 4 outer margin setae, segment 3 with 1 outer margin seta and 6 long plumose apical setae. Exopod with 1 seta on each of first 4 distinct segments and with 3 setae on terminal segment. Mandible (Fig. 12C) blade armed with 5 main teeth, 3 hirsute setae and a naked seta; mandibular palp biramous, with a 2-segmented endopod and 4-segmented exopod; endopod segments 1 and 2 bearing 2 and 6 setae respectively; exopod segments 1 to 4 bearing 2,1,1 and 2 long plumose setae respectively. LIFE HISTORY OF BENTHOMISOPHRIA 29 Fig. 11 B. cornuta Adult male and female. A, Male, lateral; B, urosome, ventral; C, Female labrum, ventral; D, Male first antenna (with setae omitted from proximal portion); E, Female leg 5; F, leg 6; G, Female, dorsal. Scales lOOum unless otherwise indicated. 30 G. A. BOXSHALL AND H. S. J. ROE First maxilla (Fig. 12D) gnathobase with 1 5 armature elements; inner lobes 1 and 2 with 5 and 4 setae respectively; outer lobe rudimentary, represented by 6 plumose setae on surface of segment; endopod apparently 2-segmented with small distal segment positioned at an angle of 90° to proximal segment, proximal segment with 4 setae apically, distal segment with 2 naked setae medially and 1 naked and 2 plumose setae apically; exopod large, 1 -segmented, with 6 plumose setae apically and 3 naked setae subapically on medial margin. Second maxilla (Fig. 12E) apparently 6-segmented; first segment with proximal inner lobe bearing 4 plumose setae and distal inner lobe with 3 plumose setae; second segment with 2 inner lobes, each bearing 3 distal setae; third segment produced medially into a strong, slightly curved claw armed with 2 slender naked setae basally; distal 3 segments bearing a total of at least 8 elements. Maxilliped (Fig. 12F) 5-segmented; first segment with 1 small naked seta and 2 plumose setae proximally and 2 plumose setae distally on medial margin; second segment with 3 plumose setae on medial margin; third segment small, bearing 1 seta; fourth segment with 4 setae; fifth segment reduced, armed with 5 setae. Legs 1 to 4 (Figs 1 3A-D) biramous, with 3-segmented rami; armature formula as follows: coxa basis endopod exopod legl 0-1 1-0 0-1;0-2;1,2,3 I- 1;I- 1;III,I,4 Ieg2 0-1 1-0 0-1;0-2;1,2,3 I- 1;I- 1;III,I,5 Ieg3 0-1 1-0 0-1;0-2;1,2,3 I- 1;I- 1;III,I,5 leg 4 0-1 1-0 0- l;0-2;missing I - 1 ;I - 1 ;missing Pinnule rows present on lateral margins of all endopod segments and medial margins of exopod segments; lateral margins of all exopod segments and of endopod segment 3 of leg 3 armed with strips of serrated membrane; all outer margin spines bilaterally serrate, apical spine on exopods of legs 2 and 3 bilaterally serrate, that on leg 1 unilaterally serrate. Leg 5 (Fig. 1 IE) 1 -segmented, without trace of suture line; armed with 1 long naked apical seta and 2 short subapical setae on medial margin; denticulations scattered over surface of segment. Leg 6 (Fig. 1 1 F) a simple process bearing 2 plumose setae on distal margin and a short spine at the disto-medial angle. Body length of female ranging from 1 • 18 to 1 -33mm, with a mean of 1 -25mm (based on 9 specimens). MATERIAL EXAMINED Holotype 9 ; Discovery Stn 4768 (40'03'N 19°57'W), BM(NH) registration number 1964.7. 10. Hulsemann and Grice, 1964. 1099, 6rf rf , 11 Co.V, 9 Co.IV, 4 Co.III; Discovery Stns 9541 and 9131 (see Table 3); BM(NH) registration numbers 1979.71-80 (99), 1979.81-86 (d-d1), 1979.87-97 (Co.V), 1979.98-106 (Co.IV) and 1979. 107-1 10 (Co.III). REMARKS Hulsemann & Grice (1964) described the first antenna of this species as being 17-segmented, with the apical segment very small. The first antenna of the holotype on the prepared slide of the appendages (BM(NH) No. 1964.7.10) appears to have a seventeenth segment but examination of the other first antenna of the pair (figured in Fig. 12 A) shows that the seventeenth segment of Hulsemann & Grice is actually the base of the apical aesthete and that this appendage is only 1 6-segmented. This species is readily distinguishable from B. palliata on the basis of its small size, distinc- tive surface ornamentation, prolonged epimeral plates on the last 2 free thoracic somites, leg 5 and the segmentation of the first antenna. There are no significant differences between B. cornuta and Misophria maculata Tanaka, LIFE HISTORY OF BENTHOMISOPHRIA 31 Fig. 12 B. comma Adult female. A, Holotype 9 (BM(NH) Reg. No. 1964.7.10), first antenna; B, Female second antenna; C, mandible; D, first maxilla; E, second maxilla; F, maxilliped. Scales lOO^m. 1966. The slight differences between them can be accounted for largely by missing or damaged armature elements. M. maculata is therefore regarded as a junior synonym of B. cornuta. Tanaka's single female specimen was taken in a haul made between 1000-Om depth in the Pacific Ocean adjacent to the Izu region of Japan. This is the shallowest record of B. cornuta. 32 G. A. BOXSHALL AND H. S. J. ROE Genus M1SOPHRIA Misophria pallida Boeck, 1864 Misophria pallida A. Boeck, 1864: 248. Gurney (1933) collected ovigerous female M. pallida and successfully hatched the eggs from them. He described the nauplius and copepodid I stages of this species. The descriptions given below are based on Gurney ( 1 933). Fig. 13 B. cornuta Adult female. A, leg 1; B, leg2;C, leg 3; D, leg 4. Scales lOOum. NAUPLIUS Body rounded anteriorly, with maximum width at level of mandible. Eye absent. Mouth absent. Body filled with yolk. Appendages without gnathobases. First antenna uniramous and 2-segmented. Second antennae and mandibles biramous, with 4-segmented exopods. Rudiments of legs 1 and 2 visible through cuticle. COPEPODID i Body apparently as in Benthomisophria palliata copepodid I. Urosome 1 -segmented. Caudal rami with 6 setae. First antenna 6-segmented. Labrum large. Second antenna, mandibles, first and second maxillae and maxillipeds present and similar in structure to those of the adult. Legs 1 and 2 biramous, with 2-segmented sympods and 1 -segmented rami; armature formula for rami as follows: legl leg 2 endopod 1,2,4 ' exopod IV, I, 3 111,1,2 LIFE HISTORY OF BENTHOM1SOPHRIA 33 Leg 3 a broad flattened structure, bilobed distally, bearing 2 small apical setae on outer lobe; inner lobe apparently unarmed. Leg 4 rudimentary, comprising a spinous process on a broad base. REMARKS The adult male and adult female are described in detail by Sars (1903). The similarities between the Copepodid I and the adult stages of Misophria pallida and Benthomisophria palliata suggest that development in these two genera follows a similar pattern through the copepodid stages. The nauplius of Benthomisophria is unknown but it may well resemble the yolky nauplius of Misophria. Development in the Misophrioida appears to be characterized by only a single nauplius stage remaining in the life cycle. The yolk-filled non-feeding nauplius is short lived in M. pallida, moulting into the Copepodid I stage within about a day (Gurney, 1933). A similar type of lecithotrophic nauplius is found in many associated or parasitic copepods belonging to the orders Siphonostomatoida, Poecilostomatoida and Cyclopoida. The lecithotrophic nauplii possess more simply constructed appendages than those of the nauplii of free-living forms (Dudley, 1966). The lecithotrophic nauplius may represent an apomorphic character state, derived independently in several copepod orders as an adaption to an associated or parasitic mode of life. The adaptive significance of the presence of a single lecithotrophic nauplius in the life cycle of M pallida (Misophrioida) will remain unknown until more data on the feeding biology and ecology become available for this species. Variation Body length All the complete specimens of B. palliata were measured (Table 2) and the body length frequency distributions were calculated for all stages (Figs 14 & 15). The size range for each stage is extremely wide. In both adults and in the later copepodid stages the frequency distri- butions appear to be bimodal, indicating that two size groups are present within the population. In the Copepodid V the modes are at 2-28 (2-24-2-32) and 2-96 (2-92-3-00)mm, in the adult male at 3-22 and 3-97mm and in the adult female at 3-30 and 4-35mm. There also appear to be traces of similar bimodality in the preceding copepodid IV and III stages, but more data are required to confirm this pattern. If the existence of a bimodal distribution at each stage from Copepodid III onwards is accepted, modes can be recognized at 1 -22 and l-52mm (Co.III), 1-75 and 2-05mm (Co.IV), 2-28 and 2-96mm (Co.V), 3-22 and 3-97mm (male) and 3-30 and 4-35mm (female). The proportional increase in body length from one stage to the next (for example, Co.III to Co.IV, or Co.V to female) of the small and/or the large form varies from 1-30 to 1-47 for each moult. These results indicate that the small and large size morphs persist from one moult to the next and, therefore, that a small Co.V would moult into a small adult male or adult female. The small and large morphs differ only in size, no other morphological differences were observed between them. There is also a difference in the vertical distribution of the two morphs (page 37). Body shape In addition to size dimorphism there is also considerable variation in the body shape of B. palliata individuals. Some specimens have a bloated, inflated appearance whereas others are much thinner. Swollen specimens are inflated both laterally and dorsally and have a distinct hump-backed appearance in lateral view (Figs 16D-E). Slim individuals have a much flatter dorsal profile (Figs 16B-C). These changes in body proportions occur in both small and large individuals and do not correspond to changes in body length. In all other aspects of external morphology indi- viduals of both types are identical. The gut, however, is very distended in swollen specimens 34 G. A. BOXSHALL AND H. S. J. ROE 0-76 0-84 0-91 0-99 1-O6 1-14 1-22 1-29 0-99 V05 1-14 1-22 1-29 1-37 1-44 1-52 1-60 1-67 N BODY LENGTH (mm) Fig. 14 Size frequency distribution (in percent) of B. palliata Copepodid I - IV stages, using data from all measured specimens (see Table 2). and is visible through the thin carapace virtually filling the entire prosome (Figs 16D-E). In slim individuals the gut is much smaller and cannot be seen through the carapace (Figs 16B-C). The gut comprises a thin walled tube which is relatively wide in the prosome and is con- stricted just before entering the urosome. There is an anterior lobe extending into the head and two lateral lobes. In swollen individuals the gut is usually full (Figs 16F-G), in slim individuals it is empty (Fig. 16A). The gut contents are mainly small copepods, with a sub- stantial amount of unidentifiable residue, mostly in the form of convoluted strings of tissue. The latter are possibly the remains of chaetognaths or coelenterates. A single radiolarian was also found. The combination of a very elastic gut and a thin flexible integument enables B. palliata to gorge until its entire prosome is distended. The ornate pattern of surface markings on the exoskeleton of this species may facilitate this stretching. A similar gorging phenomenon has been observed in some species of abyssal amphipods (Shulenberger & Barnard, 1976, Thurston, 1979) and in the leptostracan Nebaliopsis typica (Rowett, 1943, Cannon, 1946). The ability to eat as much as possible when the opportunity arises is presumably a consider- able advantage in areas, such as deep oceans, where food is scarce. LIFE HISTORY OF BENTHOMISOPHRIA 35 ca es T3 OC C 73 C cd -p o E ex c <" E o. — o U £ a \— rj CD -2 z = LU a 00 -O 'C '•o si 36 G. A. BOXSHALL AND H. S. J. ROE Similarly it is advantageous in such situations to be able to accurately locate a food supply. There is some evidence (Thurston, 1979) that copepods were attracted to a baited fish trap laid in a water depth of 4885m. The numerous large aesthetes on the first antennae of B. palliata probably have a chemosensory function. A chemosensory system will be necessary for the location of food in all post-naupliar stages. The large number of aesthetes on the first antennae of the adult male suggests that they also have a reproductive function, possibly enabling the male to locate a female. Other male copepods have been shown to respond to sex pheromones produced by the females, thereby demonstrating the existence of a chemo- sensory system (Katona, 1973). Fig. 16 Variation in body shape in B. palliata. Slim specimen (d1 T.L. 3-22mm, Stn 954 1 # 1 8) A, gut; B, prosome, dorsal; C, prosome, lateral. Swollen specimen (9 T.L. 4-27mm, Stn 9541 # 18) D, prosome, dorsal; E, prosome, lateral - both with gut visible through body wall. Swollen speci- men. (9 T.L. 4-5mm, Stn 954 1 # 22) F, gut, dorsal; G, gut, lateral. Scales 1 mm. Vertical distribution Benthomisophria palliata The vertical distributions of the copepodid and adult stages of B. palliata caught in the RMT 1 series at Station 9541 are shown in Figure 17. The two hauls that fished between 5-20m off the bottom (#18 and 19, Table 1 ) have been combined, and as there was no sexual difference in distribution the adults have also been combined. B. palliata is too small to be adequately sampled by the RMT 8 and the Copepodid I and II stages, with mean body lengths of 0*79 and l-03mm respectively, are probably not taken effectively by the RMT 1 either. The results from the RMT 1 series at Station 9131 are similar to those for Station 9541 but this series was incomplete. The numbers of specimens in each haul have been corrected for an equal volume of water filtered. The depth range of all stages is very wide, extending from 2500m to the bottom. In addition, a single adult was caught between 2000 and 2500m. There are no apparent differ- ences in depth distribution between the various stages or between the sexes. At no depth is the species abundant, but there is a general increase in numbers in the deepest samples taken within 500m of the bottom. After correction for the volume of water filtered, the greatest number of specimens caught in a single haul (all stages combined) was 1 36, in the 5-20m off the bottom sample. Assuming an even distribution this represents 1 animal per 106m3 of water. The largest number of any particular stage caught was 45 adults, also in the 5-20m LIFE HISTORY OF BENTHOMISOPHR1A 37 sample. This represents 1 animal per 322m3. In terms of biomass these figures are infini- tesimal. The mean body length at each depth of each stage of B. palliata was calculated, using the same RMT 1 series of hauls from Station 9541 (Fig. 18). The mean length of each copepodid stage is more or less uniform throughout the depth range. Both males and females, however, are apparently larger in the samples taken within 500m of the bottom. 2000 ADULTS 2500 ~ 3000 3500 4000 10 10 10 20 10 30 10 30 10 30 CORRECTED NUMBERS OF INDIVIDUALS Fig. 17 Vertical distributions of the Copepodid and adult stages of B. palliata, based on RMT 1 series at Station 954 1 . Some stages (Co. V, d and 9) show distinct large and small size morphs (page 33). In Co. V stage both small and large morphs occur throughout the depth range. In the adults small morphs are found at all depths but large ones are not. Females longer than 3t67mm and males longer than 3'45mm in total body length were found only in the deepest hauls taken within 500m. of the bottom. It is, therefore, only this deep population that shows the total range in body length and the bimodal size frequency distribution (Figs 14 & 1 5). Benthomisophria cornuta The numbers of B. cornuta caught in each haul are shown in Table 3. These numbers are insufficient to permit detailed analysis and it is likely that the RMT 1 does not effectively sample the copepodid stages of this small species. The catch data suggest that, as with B. palliata, the vertical range occupied by this species is very wide, extending from 2500m to the bottom. 38 G. A. BOXSHALL AND H. S. J. ROE Phylogenetic relationships The order Misophrioida comprises two genera, Misophria and Benthomisophria, and three species M. pallida, B. palliata and B. cornuta. The species and genera are most clearly separated on the basis of the segmentation of the leg 5. In Misophria pallida the leg 5 consists of 3 distinct segments. In Benthomisophria palliata the leg 5 is 2-segmented. The basal seg- ment is notched laterally and bears a small suture line indicating that it probably represents two incompletely fused segments, derived from the 3 -segmented type of leg 5 found in M. pallida. In B. cornuta the process effusion of the segments has progressed further and the leg 5 is 1 -segmented in the adult. The condition found in the Copepodid IV stage of this species (Fig. 8L) suggests that the single segment is derived from the fusion of two distinct segments. 2000 2500- -B 3000 3500- 4000 01 2345 BODY LENGTH (mm) Fig. 18 Mean body length, at each depth, of the Copepodid stages and the adult male (M) and female (F) of B. palliata. (Copepodid I line is hatched). Based on data from RMT 1 series at Station 9541. There appears to be a phylogenetic transformation series from the plesiotypic 3 -segmented condition, via the intermediate 2-segmented condition, to the apotypic 1 -segmented condition. The cladogram (Fig. 19) shows the probable phylogenetic relationships between the 3 species included in the Misophrioida. Hulsemann & Grice (1964) reviewed the historical aspects of the classification of the Family Misophriidae within the Copepoda. Misophria has long been recognized as repre- senting a distinct group within the Copepoda. Giesbrecht (1892) placed Misophria in a separate family, the Misophriidae and, according to Gurney (1927), Giesbrecht's families are now largely regarded as equivalent to orders. Gurney (1933) proposed that Misophria should be regarded as the type of a separate order, the Misophrioida. Lang (1948) erected the sub- order Propodoplea, to contain the family Misophriidae. In a recent work on the classification of the Copepoda, Kabata (1979) regarded the Misophrioida as an early offshoot from the podoplean line and this system has been adopted in the present account. LIFE HISTORY OF BENTHOMISOPHRIA 39 Misophria Benthomisophria Benthomisophria pallida palliata cornuta Fig. 19 Cladogram showing the probable phylogenetic relationships between the three species comprising the Misophrioida. (1 ,2 or 3 denotes the number of segments in the adult leg 5). Discussion The samples analysed are the first to have been taken close to the bottom in the total water column programme initiated by the Institue of Oceanographic Sciences in 1976. The new near-bottom sampling technique employs bottom indicator switches that are rather inaccurate and difficult to use. A near bottom echo sounder is being developed to facilitate this type of sampling. B. palliata has rarely been caught but it has a wide geographical distribution. It was recorded by Sars (1909) from 3 stations in the northeastern Atlantic in hauls made with open nets fished from depths of 2585 (also given as 5285m), 1414 and 3465m. Tanaka (1966) found a single specimen of each sex between 1000-Om depth in the Pacific adjacent to the Izu region of Japan. The present material extends its geographical range southwards in the eastern North Atlantic. B. palliata has not previously been caught with closing nets. The new data indicate that B. palliata is bathypelagic but that its vertical distributions show little affinity with the sea bottom. It occupies a wide depth range and has no well defined optimum depth. Sars' (1909) specimens were presumably caught in the deep parts of his tows and Tanaka's (1966) record is the shallowest of this species. Only two specimens of B. cornuta have been recorded, both female. One was taken in the northeastern Atlantic at 40°03-5/N 19°57'W, between 4750 and 4000m depth (Hulsemann & Grice, 1964), and the other off Izu, Japan between 1000-Om (Tanaka, 1966 -as M. maculatd). The present record extends the known geographical of B. cornuta southwards in the North Atlantic. Too few specimens were taken to clarify its vertical distribution but it seems similar to that of B. palliata, with a wide depth range. The analysis of body length frequency distribution in B. palliata indicated that two size morphs of B. palliata were present in the population. There was a slight difference in the vertical distribution of the two morphs in the adults of both sexes but there were no other morphological differences. This kind of size dimorphism has been observed in several other planktonic copepods, such as Oncaea venusta Philippi, 1843 (Boxshall, 1977). The explanation of this dimorphism must await additional seasonal, geographical and ecological data. The ability to gorge and distend the body wall when the opportunity arises has now been found in three distinct crustacean groups, as well as in several midwater fish (Murray & Hjort, 1912). In Benthomisophria and the amphipod Paralicella Chevreux, 1908 the gut is 40 G. A. BOXSHALL AND H. S. J. ROE very elastic but has no other obvious structural modification. Nebaliopsis typica G. O. Sars, 1887 is specialized in having a very distensible mid gut sac (Rowett, 1943). When this is filled the cephalothorax is so distended that its posterior region protrudes behind the carapace (Cannon, 193 1 & 1 946). N. typica is bathypelagic and is believed to feed by sucking eggs. The mid gut sac is presumed to be a storage organ enabling the animal to take maximum advantage of an irregular food supply (Rowett, 1943; Cannon, 1946). Except for the amphipods all the animals that are known to have the ability to gorge live in midwater at depths where food items are likely to be scarce. It is perhaps to be expected that modifi- cations to enable opportunistic feeding exist in such an environment. Bottom living animals possibly have a more continuous supply of food available to them, both from the benthic community and from the 'rain' of debris. The ability to utilize infrequent large particles of food is, however, clearly advantageous and Paralicella is adapted accordingly. Acknowledgements Mr M. J. Harris and Mr R. Wild devised and made the bottom indicator switch system. Mr A. de C. Baker drew the bathymetric chart and plotted the net tracks. We would also like to thank Mr P. M. David and Dr R. J. Lincoln for stimulating discussion of the paper. Dr G. Testa of the Musee Oceanographique de Monaco kindly arranged the loan of the B. palliata material from the S.A.S. Le Prince Albert de Monaco. References Baker, A. de C., M. R. Clarke & M. J. Harris. 1973. The N.I.O. combination net (RMT 1 +8) and further developments of rectangular midwater trawls. J. mar. biol. Ass. U.K. 53 : 167-184. Boeck, A. 1864. Oversigt over de ved Norgs Kyster iagttagne Copepoder henhovende tie Calanidernes, Cyclopidernes og Harpacticidernes Familiar. Fork. VidenskSelsk. Krist. 1864 : 226-28 1 . Boxshall, G. A. 1977. The planktonic copepods of the northeastern Atlantic Ocean: some taxonomic observations on the Oncaeidae (Cyclopoida). Bull. Br. Mus. nat. Hist. (Zool.) 31 : 103-1 55. Brewer, P. G., D. W. Spencer, P. G. Biscaye, A. Hanley, P. L. Sachs, C. L. Smith, S. Radar & J. Fredericks. 1976. The distribution of paniculate matter in the Atlantic Ocean. Earth planet. Sci. Lett. 32 : 393-402. Cannon, H. G. 193 1 . Nebaliacea. 'Discovery' Rep. 3 : 199-222. 1946. Nebaliopsis typica. 'Discovery' Rep. 23 : 2 13-222. Dudley, P. L. ( 1 966). Development and Systematics of some Pacific marine symbiotic copepods. Univ. Washington Press: Seattle & London. 282 pp. Giesbrecht, W. 1 892. Systematik und Faunistik der pelagischen Copepoden des Golfes von Neapel und der angrenzenden Meeresabschnitte. Fauna Flora Golfo Napoli. 19 : 1-83 1 . Gurney, R. 1927. Zoological results of the Cambridge Expedition to the Suez Canal, 1924. XXXIII. Report on the Crustacea:- Copepoda (Littoral and Semi-parasitic). Trans, zool. Soc. Lond. 22 : 451-577. 1933. Notes on some Copepoda from Plymouth. J. mar. biol. Ass. U.K. 19 : 229-304. Hulsemann, K. & Grice, G. D. 1964. A new bathypelagic species of Benthomisophria (Copepoda: Misophriidae) from the North Atlantic. Zool. Anz. 173 : 259-264. Kabata, Z. 1979. Parasitic Copepoda of British Fishes. Ray Society: London. 459 pp. Katona, S. K. 1973. Evidence for sex pheromones in planktonic copepods. Limnol. Oceanogr. 18 : 574-583. Lang, K. 1948. Copepoda 'Notodelphyoida' from the Swedish West Coast with an outline on the syste- matics of the copepods. Arkiv. Zool. 40 A (14) : 1-36. Murray, J. & Hjort, J. 1912. The Depths of the Ocean. Macmillan &Co. Ltd.: London. 821 pp. Roe, H. S. J. & Harris, M. J. in press. A new accoustically telemetering deep sea photometer with some observations on underwater light in the North East Atlantic. Deep-Sea Res. Rowett, H. G. Q. 1943. The gut of Nebaliacea. 'Discovery' Rep. 23 : 1-17. Sars, G. O. 1903. An account of the Crustacea of Norway. V. Copepoda Harpacticoida. Bergen Museum. Pts. I-II : 1-28. 1 909. Note preliminaire sur trois formes remarquables de copepodes provenant des Campagnes de S.A.S. Le Prince Albert de Monaco. Bull. Inst. Oceanogr. Monaco. 147 : 1-8. LIFE HI STORY OF BENTHOMISOPHR1A 41 Shulenberger, E. & Barnard, J. L. 1976. Amphipods from an abyssal trap set in the north Pacific gyre. Crustaceana. 31 : 241-258. Steedman, H. F. 1974. Laboratory methods in the study of marine zooplankton. /. Cons. perm. int. Explor.Mer. 35 : 35 1-35 8. Tanaka, O. 1966. Some rare species of Harpacticoida from the Izu region. Mar. biol. Ass. India. symposium ser. 2. Proc. Symp. on Crustacea. 1 : 5 1-56. Thurston, M. H. 1979. Scavenging abyssal amphipods from the North-East Atlantic Ocean. Mar. Biol. Berlin. 51 : 55-68. Manuscript accepted for publication 27 June 1979 Abyssal benthic Amphipoda (Crustacea) from the East Iceland Basin 1 . The genus Rhachotropis Michael H. Thurston , Institute of Oceanographic Sciences, Wormley, Godalming, Surrey Introduction Cruise 39 of R.R.S. Discovery, made in April-June 1971, was concerned with biological pro- jects in the north-east Atlantic Ocean. Major programmes involved intensive investigations of the vertical distribution and diurnal migration of planktonic and nektonic organisms at 60°N 20°W and 53"N 20°W. As an adjunct to the midwater work at the northern station, five hauls were made with a large epibenthic sledge at depths of 2600-2800m on the floor of the East Iceland Basin. The present paper, describing species of the eusirid genus Rhachotropis, is the first to deal with the 120 species of gammaridean amphipods recovered from these benthic samples. A considerable body of data is available on the bathyal and abyssal gammarid amphipods of the temperate north Atlantic Ocean. Among the more important contributions are those of Bonnier (1896), Chevreux (1900, 1927, 1935), Stephensen (1915, 1923, 1925, 1931, 1944a) and Mills (1967, 1971). Evidence summarized by Briggs (1974: 390) suggests that the bathyal and abyssal fauna of the Greenland and Norwegian Seas (and the Arctic Basin as a whole) is fairly distinct from that of the North Atlantic, but data on the deep-dwelling amphipods of these areas are relevant to the present study. Major works are those of Boeck (1876) and Sars (1885). Regional studies dealing mainly with shallow water faunas, but including data on slope species include the works of Sars (1890-1895), Stephensen (1933, 1935, 1938, 1940a, b, 1942, 1 9446) and Gurjanova( 1951). The past twenty years have seen an upsurge in interest in the fauna of the deep sea. Investigations have shown that the bathyal and abyssal biomass is, in general, lower by several orders of magnitude (see, for example, Zenkevitch, 1963: 723) than that of the continental shelf and upper continental slope. Despite the low biomass, the number of species represented is frequently high (see, for example, Barnard (1962), Menzies (1962), Sanders et al (1965)), so much so that the species diversity is higher at these depths than in many other shallower and more productive marine environments (Hessler & Sanders (1967), Sanders (1968)). The nature and causes of high diversity have been discussed by Sanders ( 1 969), Sanders & Hessler ( 1 969), Dayton & Hessler ( 1 972) and Grassle & Sanders ( 1 973). The collecting effort which has been devoted to sampling the deep sea benthos in recent years has made available large quantities of biological material. Problems of sorting and identification of organisms has led to much of this material being studied in general terms only, identification being made to phylum, class or perhaps order. One of the major problems in detailed studies is the lack of taxonomic effort currently available. Barnard (1958) highlighted this problem for the Gammaridea. He showed that despite the great increases in material available for study, the systematic activity was considerably less than during the 1930s. Although Barnard himself has made massive contributions to amphipod systematics, the problem is still acute, and effort is not commensurate with the volume of specimens being obtained. Many of the major faunules are still very poorly known (witness Barnard (1972, 1974) working on non-domicolous, non-fossorial amphipods from shallow water environments in warm temperate Australia, who discussed 47 genera of which 18 were Bull. Br. Mm. nat. Hist. (Zool.) 38 (1): 43-67 Issued 27 March 1980 43 44 M. H. THURSTON new, and 124 species of which 84 were new). That systematic effort is still a prime requisite in the study of deep sea amphipods is shown by a series of four papers by Barnard (1961, 1962, 1964, 1967). In the four collections dealt with 64%, 63%, 59% and 52% respectively of the species discussed were new to science. The relatively low figure of 52% reflects previous work in the same small area by Barnard himself (1966). Brook (1974), in discussing the zoogeography of the deep sea, has collated data from many sources which indicate that most invertebrate groups, including amphipods (see Barnard (1962)), contain a high proportion of species which appear to be endemic to single ocean basins. Although future work may well reduce the apparently high endemicity in the deep sea, there is no indication that the descriptive effort within the Gammaridea, and probably most other major invertebrate groups, has reached a level of diminishing returns. Taxonomy, although a legitimate end in itself, should also serve as a tool for further studies in the ecology, physiology, zoogeography and related disciplines of the organisms concerned. It is in this light that the present paper is presented. Material and methods The five epibenthic sledge hauls made in the vicinity of 60°N 20°W produced 267 specimens ofRhachotropis. These specimens represented 7-5% of the total amphipod catch. The epibenthic sledge, designated BN2-4, which was used to obtain these samples, is a modified beam trawl with a mouth area of 2-4m2. The framework consists of two vertical, oval structures, each equipped with a broad skid to prevent sinking in soft bottoms and con- nected at the top by two transverse metal bars. The net, which is 6- 10m long and with a mouth 2-44m, by 0-9 1m, is made of 4-5mm mesh. It is lined for 2-44m in front of the canvas cod-end with 1mm netting. The net is laced to the rearmost of the two transverse bars on the framework, clipped to vertical side wires and laced to a solid bottom bar which is connected by weak links to the heel of the skid of each vertical frame. The sides and top of the main framework are lined with 6mm polypropylene mesh to prevent the escape of animals dis- turbed by the tickler chain. A depth telemetering pinger akin to that discribed by Clarke (1969), but incorporating a mercury switch, was attached to the framework of the net. This device monitored the changes in angle of attack of the sledge which marked arrival on and departure from the bottom. A subsequent modification of this sledge which incorporates photographic and opening/closing capabilities is designated BN 1-5 and has been described byAldred£*fl/(1976). Catches consisted of animals and a variable, but usually large, volume of sediment. Large and delicate organisms were picked out of the sediment prior to sieving through 2mm and 500um meshes. The separated fractions were fixed in 5% formalin and transferred to 70% alcohol 48-72 hours later. Some amphipods were picked out during initial shipboard sort- ing, but most were recovered in the laboratory. The positions listed in Table 1 are best estimates of the first and last bottom contact for each haul. Positions, depths and lengths of haul have been calculated from satellite navi- gation fixes and precision echosounder records. Thoracic limbs are referred to as gnathopods 1-2 and peraeopods 3-7. Lengths have been taken on straightened animals and were measured from the tip of the rostrum to the tip of the telson. Males have been identified by the presence of genital papillae and females by oostogites; all other specimens being regarded as juveniles. The specific epithets applied to new species are derived from personal names occurring in the rich Icelandic literature of the Viking period. Systematic* Over 3500 specimens of gammarid amphipod belonging to about 120 species were obtained from the five epibenthic sledge hauls. Although several species, particularly some lysianas- sids, were probable contaminants from meso- and bathypelagic midwater zones, a prelimi- nary examination suggests that 94% of the species and over 99% of the specimens were BENTHIC AMPHIPODA RHACHOTROPIS 45 benthic in origin. These data suggest that contamination of catches by midwater organisms is of little significance. This observation is supported by the very small catches made during other hauls which were terminated without bcttom contact. Over 50% of the benthic species appear to be undescribed. This proportion of new species corresponds closely with the levels found in similar collections studied by Barnard (see above). The bottom sampled by all five hauls consisted of dark grey-brown mud containing small quantities of larger particles together with some pteropod shells (mainly Diacria trispinosd). Table 1 Station List Station Date Position Gear Length of haul Depth (m) 7709 #62 1 May 1971 59 °58-8'N 19°58- 2'W BN2 4 2714(-0)m 1010 -59°58- 8'N 19 °58-8 'W 7709#66 3 May 1971 59 °59-4'N 19°53- 5'W BN2 4 2712(-0)m 2150 -59°58- 3'N 19 °53-5 'W 7709 #72 5 May 1971 60 °05-7'N 19°43- 3'W BN2 4 2663-2649(-0)m 2450 - 60°06- 8'N 19 °42-5 'W 7709#73 5 May 1971 60 °07-5'N 19°32-4'W BN2 4 2646-2636(-0)m 9460 -60°06- 4'N 19 °26-6 'W 7709#85 7 May 1971 59 °58-6'N 19°53-4'W BN2 •4 2708(-0)m 1690 -59°58- 1'N 19 °54-3 'W Rhachotropis proximo Chevreux, 1911 (Figs. 1-2) Rhachotropis proximo Chevreux, 1911:11-13, fig. 5; 1935: 1 10-1 1 1, pi. 12, fig. 3.— Stephensen, 1944: 15.— Gurjanova, 1955: 184 (in key).— Barnard, 1957: 14. Rachotropisproxima(err.) — Schellenberg, 1955: 194. — Belloc, 1960: 13. MATERIAL EXAMINED. Sta. 7709 #62 19 9mm, 4 damaged specimens; Sta. 7709 #66 1 damaged specimen; Sta. 7709#72 5rf 12-14mm, 1 ovig. 9 17mm, 79 14-17mm, 1 juv. 8mm, 2 damaged specimens; Sta. 7709 #73 8cT 10-1 5mm, 289 8-1 5mm, 7 juv. 7-9mm; Sta. 7709 #85 2d'9-12mm, 79 14-1 9mm, 2 damaged specimens. DESCRIPTION. Male, 15mm, from Sta. 7709^ 73. Pleon, each segment weakly carinate, produced posteriorly into low, acute tooth; segments 1 and 2 with small dorso-lateral teeth on posterior margin. Epimera, first rounded ventrally; second produced anterior-ventrally, posteriorly sub-rectangular; third similar in shape to second but with six small teeth at posterior distal angle. Urosome, first segment with acute dorsal tooth on posterior margin, lacking lateral carinae. Head, rostrum long, deep and down-curved, as long as head; eyelobes produced, rounded, eyes not seen; postantennal angle bilobed, weak; epistome just more prominent than upper lip. Antenna 1, as long as head, peraeon and first pleon segment combined; first article of penducle rather stout, equal to second and third combined, shorter than rostrum; second and third articles more slender than first, second three times length of third; flagellum more than twice as long as peduncle, basal third somewhat inflated, articles short, bearing many elongate aesthetes medially, distal articles long and slender; accessory flagellum of one article, minute, bearing three setae and a spine apically. Antenna 2, as long as head, peraeon and first two pleon segments combined; fifth article of peduncle long and slender, as long as first four peduncle articles combined; flagellum nearly twice as long as peduncle, consisting of 44 slender articles. Upper lip entire, truncate distally. Mandible, incisor process well Fig. 1 Rhachotropis proxima. Male, 15mm. (a) habitus; (b) head; (c) antenna 1; (d) antenna 1, detail; (e) accessory flagellum; (0 antenna 2; (g) upper lip; (h) right mandible; (i) left mandible, detail; (j) lower lip; (k-1) maxillae 1 & 2; (m) maxilliped. Female, 1 5mm. (n-o) antennae 1 & 2. BENTHIC AMPHIPODA RHACHOTROPIS 47 developed with a single tooth and long chitinized cutting edge, lacina mobilis of right mandible with three teeth, of left mandible broader and with six teeth; molar well developed, ridged and bearing marginal teeth; palp longer than body of mandible, third article falciform, densely setose along most of anterior margin, Lower lip, inner lobes well developed. Maxilla 1, inner plate rather narrow with two subapical setae; outer plate bearing nine slender dentate spines; palp article 2 with serrate, very oblique apex, armed with anterior and posterior rows of setae. Maxilla 2, inner plate rather broad, apex rounded, truncate, medial angle armed with stiff setae; outer plate narrower than inner, apically rounded and armed with stiff setae. Maxilliped, inner plate small; outer plate narrow with many spine teeth medially; palp strong, second article longest, second and third articles armed with setae and long stout spines, dactyl as long as third article, strongly curved. Gnathopod 1, coxa strongly produced anteriorly, nearly three times as long as high; basal expanded distally, carpus short, lobed posteriorly; propod oval, palm long, armed with setae; dactyl slender, as long as palm. Gnathopod 2, coxa pentagonal; basal expanded distally, armed with group of stout setae at anterior-distal corner; carpus short, with long, slender, heavily spined lobe posteriorly, lobe extending as far as palmar angle of propod; propod oval, rather longer than that of gnathopod 1 ; dactyl slender, as long as palm. Peraeopod 3, coxa small, articles slender; basal as long as articles 3 to 5 combined; carpus nearly twice as long as article 4, subequal in length to propod and to dactyl. Peraeopod 4, coxa longer than high, excavate posteriorly, distal articles as in peraeopod 3. Peraeopod 5, coxa weakly bilobed, posterior lobe stronger, distally rectangular; basal short, expanded distally; merus elongate, slender; carpus shorter and more slender than merus; propod very long and slender, length more than 20 times width; dactyl very slender, slightly curved, about 40% of length of propod. Peraeopod 6, coxa with posterior lobe stronger, rectangular at posterior-distal angle; distal articles similar to those of peraeopod 5. Peraeopod 7, coxa just acute posterior-distally; basal nearly twice as long as that of peraeopod 6, posterior distal angle acute; fourth article twice as long as basal, slightly produced posterior-distally; carpus more slender than merus and 20% longer; propod elongate and extremely slender, length 25% greater than that of carpus, maximum width less than 3% of length; dactyl straight, very slender, 25% length of propod. Uropod 1, inner ramus just longer than outer, shorter than peduncle. Uropod 2, extends back to apex of uropod 1; inner ramus just shorter than peduncle, with four small spines on medial margin, outer ramus unarmed, 75% as long as inner ramus. Uropod 3, rami lanceo- late, subequal, margins minutely pectinate, twice as long as peduncle, inner ramus armed on medial margin, outer on both margins. Telson as long as uropod 3, broad, sides parallel, tapering distally, cleft 12% of length, armed dorsally with minute spines. Female, 15mm, from Sta. 7709 #73. Females attain a larger size than males, reaching 18mm, but differ morphologically only in the broader peraeon and shorter antennae. Antenna 1, peduncle and flagellum subequal in length; peduncle article 1 stouter and just longer than article 2, article 3 60% length of article 2; flagellum of 12 articles. Antenna 2, peduncle article 5 longer and more slender than article 4; flagellum shorter than peduncle articles 4 and 5 combined, of 1 5 articles. REMARKS. Barnard (1957) omitted R. proxima from his key to the genus on the grounds of uncertainty concerning the dorsal armature of this species. The present material, which agrees precisely with Chevreux's illustrations, reveals an error in the original description (Chevreux, 1911:11). The reference to the dorsal armature of the mesosome in fact applies to the metasome. No species of Rhachotropis has yet been described in which peraeon and urosome are carinate and dentate, but the pleon unarmed. R. proxima belongs to the group of species characterized by a strongly produced coxa 1 , peraeonite 7 without teeth, pleonites 1-3 with three, three and one tooth respectively and urosomite 1 dentate. Two species, R. kergueleni Stebbing, 1888 and R. hunteri Nicholls, 1938, agree with these basic characters of R. proxima. R. proxima is closely related to R. kergueleni, but the 48 M. H. THURSTON Fig. 2 Rhachotropis proximo. Male, 15mm. (a-b) gnathopods 1 & 2; (c) gnathopod 2, detail of palm; (d-h) peraeopods 3-7; (i-k) epimera 1-3; (1-n) uropods 1-3; (o) uropod 3, margin of outer ramus; (p) telson. Male, 14mm. (q) peraeopod 7. BENTHICAMPHIPODA/?///lC7/077?0m 49 southern hemisphere species has a large tooth on urosomite 1 , broadly rounded and strongly serrate epimeron 3, a less deeply cleft telson with divergent apices, and basal articles of peraeopods 5-7 with serrate posterior margins. The distinction between those two species is obscured somewhat by specimens doubtfully assigned to R. kergueleni by Stephensen (1944a). In the dehiscent apices of the telson and broadly rounded, strongly serrate epimeron 3, the klngolf material from the Davis Strait agrees with R. kergueleni while the weakly serrate basal articles of peraeopods 5-7 are closer to R. proxima. The degree of variation of these characters in the present specimens is minimal so until the status of Stephensen's speci- mens and the type material of R. kergueleni can be clarified, it seems best to exclude both these entities and maintain R. proxima as distinct. Two new species described later in this paper are related to R. proxima. One possesses the same basic characters as R. proxima and the other differs in the condition of coxa 1 . The diagnostic characters are discussed on p. 53 and p. 57 respectively. A number of other species are also related to R. proxima. R. levant is Barnard, 1961 and R. anoculata Barnard, 1962 both lack teeth on urosomite 1, have telsons cleft more than 30% of their length, and bear calceoli on the antennae. In addition the shape of the basal articles of peraeopods 5-7 is markedly different in R. levantis and R. proxima. The dorsal posterior margin of pleonite 3 is prominent but not dentate in R. anoculata, thus contrasting with R. proxima in which this segment bears a distinct tooth. R. leucophthalma Sars, 1893, R. sibogae Pirlot, 1934, R. integricauda Carausu, 1949, R. cervus Barnard, 1957 and R. flemmingi Dahl, 1959 all differ from R. proxima in possessing dorso-lateral teeth on pleonite 3 in addition to the median tooth. R. integricauda is unique in the genus in having an uncleft telson. R. leucophthalma and R. cervus can also be separated from R. proxima by the telson which is cleft 50% and 30% of its length respectively as opposed to 14-17%. R. sibogae is further distinguished from R. proxima by the broadly rounded coxa 1 , the form of epimeron 3 and the divergent apices of the telson. Peraeopod 7 of R. flemmingi has an extremely long propod, which is as long as the fourth and fifth articles combined, and a dactyl half the length of the propod. These proportions are significantly different in R. proxima as is the condition of epimeron 3. R. antarctica Barnard, 1932 and R. clemens Barnard, 1967 are separable from R. proxima in that they lack teeth on pleonite 3. R. antarctica has eyes, dorso-lateral keels on pleonite 1 and on urosomite 1 , and a broadly expanded basal article of peraeopod 7, features not present in R. proxima. A short rostrum, weakly p.roduced coxa 1 , slender basal articles of peraeopods 5 and 6 and a large tooth on urosomite 1 are additional characters separating R. clemens from the present species. R. platycera Barnard, 1931 is much smaller than R. proxima, adults reaching only 4'5mm as opposed to 12-20mm,and has calceolate antennae, eyes, a dorsal tooth on peraeonite 7 and a deeply cleft telson. Rhachotropis thordisae sp. nov. (Figs. 3-4) MATERIAL EXAMINED. Sta. 7709*62 2§ S 5 £ "0 c £ a x' co E 1. 00 CO E C _j_ 1 " 00 (~H QJ C S c o. JH uo — CO CO X — t/3 CO 00 -g E JU ">> c r? C 1 •^ o 0 in p ^t Os OS — oo fN Os pi. Os rn Tt m TJ- rn rn rn fN (N SO — « * rn SO OS — rn p m 7 p ^ fN •— fN oc 7 rn rn 0 OS Os O Tf fN fN rn ^1 "* rn rn TT ^- •^r *~r — — r^i — — — _ — fN •n r- _ oo oc f*. oo (-1. rn rn m m 1 1 1 1 — ^ p T' r~^ Os fN so sO m in in vC •^j- in m m m m rn m rn rn rr. rn — fN 00 — — — m — m m fN sO 00 SO in m m rn 1 1 fN Os GS os oo m o r~- in s6 in m -rj- so T-j- r<-, m m m m m rn rn — in — — fN — ' — • fN oo Os 0 p Os — , — , fN fN pi 1 1 | | i m ^9 T "9 f? "?' SO - fN fN OS - — — in » SO OS ^ m _- _L. fN fN 1 1 1 1 rn — ^ oo O O Tf m sO •— ' fN — — r-'l "- fN fN — fN — — — — ' fN so — • — fN — — — • — m rn in Os T r~- Os "? ~ m T 'n 1 — fN O oo fN 00 00 — 00 — CD fN fN fN — — — fN C] — fN 01 — — * _)- * * * Os m ""^ c in ^2 in ri ^ Os °l' m Os O fNrn in m m m rn fN E- •a O OS o Os r^ in O r~- m o >n in r3 0 | 'I £ E -J3 • - O CO H 03 h- 2? 0 0 CL 3 r3 3 c_ New Guinea un U E '5, a IE a. pecimens in ale C/3 104 W. BERGMANS & J. E. HILL From its measurements, this specimen, which is the first of this species recorded for Sumatra, seems to link the western nominate race with the larger, eastern subspecies shortridgei Thomas & Wroughton, 1909, from Java. Rousettus species According to Boeadi (in lit., 22 March, 1979) there are in the Museum Zoologicum Bogoriense at Bogor about 27 Sumatran Rousettus specimens from Lampong, South Sumatra, and upper Siak-Kampar rivers, Central Sumatra, of which the specific identity has not yet been determined. Of course, actual sympatry of spinalatus with either ample 'xicaudatus or leschenaulti on Sumatra has yet to be proven. The specimen of Rousettus spinalatus from Sarawak, Borneo was first recorded as R. amplexicaudatus by Harrisson (1967: 229). This specimen, taken 'low off floor' in the West Mouth of Niah Great Cave, is now BM(NH) 75.589. The other specimen that he records, a male, taken about six weeks previously on the same spot and now BM(NH) 69.22 1 , actually represents R. amplexicaudatus. Thus, at Niah at least both species occur sympatrically. R. amplexicaudatus has also been recorded from Sabah and West Kalimantan (Medway, 1977 : 39). Acknowledgements We thank Dr K Bauer, Vienna, for drawing our attention to the Sumatran specimens of Rousettus spinalatus and allowing us to study these. We gratefully acknowledge the help of Ms Dr A. Zuiderwijk of the Instituut voor Taxonomische Zoologie in Amsterdam, who, with the assistance of Mr Tranier of the Museum National d'Histoire Naturelle in Paris, tried to locate the holotype specimen of Rousettus amplexicaudatus. Dr C. Smeenk of the Rijksmuseum van Natuurlijke Historic at Leiden was helpful in giving access to the impor- tant collections of Indonesian Rousettus in that museum. Dr H. Hackethal of the Zoologisches Museum in Berlin kindly allowed us to study the Rousettus specimens in his care. We are also indebted to Dr F. Dusbabek of the Parasitologicky Ustav, Ceskoslovenska Akademie V£d, Praha, who identified the mite from the paratype of Rousettus spinalatus. References Andersen, K., 1912. Catalogue of the Chiroptera in the Collection of the British Museum. I. Megachiroptera:i-ci, 1-854. Trustees British Museum, London. Bergmans, W., 1979. Taxonomy and zoogeography of the fruit bats of the People's Republic of Congo, with notes on their reproductive biology (Mammalia, Megachiroptera). Bijdr. Dierk. 48 (2): 161-186. Dobson, G. E., 1878 Catalogue of the Chiroptera in the Collection of the British Museum: i-xlii, 1-567, pis. I-XXX. Trustees British Museum (Natural History), London. Harrisson, T. H., 1967. Bats netted in and around Niah Great Cave, 1965-6. Sarawak Mus. J. 14 (28/29): 229-233. Jentink, F. A., 1877. In Schlegel, H. & Jentink, F. A., Museum d'Histoire Naturelle des Pays Bas. Revue methodique et critique des Collections deposees dans cet Etablissement. Tome 9. Catalogue Osteologique des mammijeres. E. J. Brill, Leiden. Medway, Lord, 1977. Mammals of Borneo. Field keys and an annotated checklist. Monogr. Malay. Brch. R. Asiat. Soc. 7: i-xii, 1-172, 10 figs., frontis., 24 pis. Ostwald, W., 1939. Die kleine Farbmesstafel; Ausgabe B: 1-7, pis. 1-5. Muster-Schmidt, Gottingen. Temminck, C. J., 1825. Monographies de Mammalogie, ou description de quelques genres de mammijeres, dont les especes ont ete observees dans les differens musees de {'Europe. 1, livr. 5 : 1 57-204. Van de Hoek, Leiden, and Dufour & d'Ocagne, Paris. Zimmerman, K., 1952. Vergleichende Farbtabellen. Comparing colour plates: 1-47. Schops Verlag, Frankfurt/Main. Manuscript accepted for publication 24 May 1979 An Important New Title From British Museum (Natural History) Monographs British Marine Amphipoda: Gammaridea by R. J. Lincoln 658pp 2,300 figures 4to hard bound ISBN 0 565 00818 £50.00 Amphipods are both numerous and diverse in numbers of genera and species in British coastal waters, but in the absence of any form of modem systematic synopsis or key this group of crustaceans has acquired the reputation of being notoriously difficult to identify. This monograph, which is the first comprehensive and illustrated text on British gammaridean amphipods to be published in more than a century, should go a long way towards solving the problem. The systematic section of the book contains descriptions and figures of all 271 species of marine and brackish water amphipods, in 123 genera and 36 families, recorded from British coasts and the adjacent continental shelf to a depth of 200 metres. Keys are provided at all levels, as well as relevant synonymies and diagnoses of genera and families. The text is illustrated with about 2,300 separate figures which have been drawn by the author from Museum and other material, in many cases with reference to type specimens. The work has been carefully edited to bring corresponding descriptions, keys and figures into close proximity within the text. The systematic section is supported by chapters dealing with morphology, systematics, geographical distribution, biology and ecology, the latter being presented in the form of an annotated subject index of research literature. Finally, there is an extensive bibliography of about 1,200 references that includes most of the British marine amphipod literature published to date. Titles to be published in Volume 38 Miscellanea Relationships between hummingbirds and flowers hi the Andes of Columbia. By David W. Snow & Barbara K. Snow Miscellanea Mites of the subfamily Parasitinae (Mesostigmata : Parasitidae) in the British Isles. By K. H. Hyatt Miscellanea Printed by Henry Ling Ltd, Dorchester Bulletin of the British Museum (Natural History) Relationships between hummingbirds and flowers in the Andes of Colombia David W. Snow & Barbara K, Snow Zoology series Vol 38 No 2 27 March 1980 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum's resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself, available separately, and individually priced. Volumes contain about 300 pages and are not necessarily completed within one calendar year. Subscriptions may be placed for one or more series. Subscriptions vary according to the contents of the Volume and are based on a forecast list of titles. As each Volume nears completion, subscribers are informed of the cost of the next Volume and invited to renew their subscriptions. Orders and enquiries should be sent to : Publications Sales, British Museum (Natural History), Cromwell Road, London SW7 5BD, England World List abbreviation: Bull Br. Mus. nat. Hist. (Zool.) Trustees of the British Museum (Natural History), 1980 ISSN 0007-1498 Zoology series Vol 38 No 2 pp 105-139 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 27 March 1980 GENERAL Relationships between hummingbirds and flowers 3 APR 19 in the Andes of Colombia WUBRAR* David W. Snow .and Barbara K. Snow ^ •"•• fx. Department of Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD Synopsis An account is given of the feeding ecology of hummingbirds in the Eastern Andes of Colombia, based on nine weeks of field work in 1978. Three hummingbird communities were studied, two in the temper- ate and one in the subtropical zone. The 25 species of hummingbirds observed showed very different degrees of specialization for exploiting the nectar of particular kinds of flowers, from an extreme in Ensifera, which appears to have coevolved with and to be dependent on species of Passiflora, to gener- alists which share their nectar sources with other hummingbirds and insects. In the light of these obser- vations we discuss the 'syndrome of ornithophily', the alternative foraging strategies of trap-lining and territoriality, nectar characteristics, and the evolution of ornithophily in the Andean flora. Introduction The ecology of hummingbirds has been the subject of much research in the last ten years, with the emphasis on three related aspects: energetics, the structure of local hummingbird guilds (with which may be included flower-piercers and other nectarivores), and co- evolutionary relationships between hummingbirds and their food plants. Undoubtedly a prime reason for the popularity of hummingbirds for this kind of research is the fact that under favourable conditions, in the field, quantitative assessments can be made both of their expenditure of energy, of the availability of nectar from different plants, and of the birds' efficiency in obtaining nectar. As a result, tremendous advances have been made in our un- derstanding of the biology of hummingbirds. Stiles (1980) gives a review, with key references. In comparison with the energetic aspects of hummingbird ecology and with studies ot interactions between guild members, revolutionary relationships have tended to take a subordinate place, a notable exception being the analysis by Stiles (1975) of the relationship between hermit hummingbirds and Heliconia species in a lowland forest in Costa Rica. Most field studies have been carried out at high altitudes in Central America, where a relatively small number of hummingbird species coexist. Nothing has been published so far on relationships between hummingbirds and their food plants in what may be called the core area of hummingbird evolution, the subtropical and temperate forests of the Andes from Colombia to Peru, where the greatest number and variety of hummingbird genera and species coexist. It was for this reason that we attempted a survey of the feeding ecology of hummingbirds in the Eastern Andes of Colombia in July and August 1978, the results of which are presented in this paper. It must be stressed at the outset that our findings represent the merest scratching at the sur- face of a subject of great complexity. Our observations extended over only two months, with a span of only three weeks in each of the three areas (two forest, one farmland) in which we worked; but even in that time in all three areas there were significant changes in the avail- ability of nectar, as some plants finished and new ones came into flower. In any area, field work must obviously be continued over at least a complete year in order to get an adequate idea of the food resources available to hummingbirds. In addition, the vegetation was almost entirely different in the two forest areas, about 150 km apart, and the hummingbird species were largely different. Many areas would have to be sampled in order to get an adequate idea Bull. Br. Mus. not. Hist. (Zool.)38 (2) : 105-1 39 Issued 27 March 1980 105 106 D. W. SNOW & B. K. SNOW of the overall distributional pattern of vegetational communities and their associated hummingbirds, even in a small section of the Andes. Recent research on hummingbird ecology has revealed an important distinction between two fundamentally different feeding strategies. A hummingbird may defend a feeding territory, usually a concentrated source of nectar of small extent such as a single profusely flowering tree, or it may make the rounds of many scattered sources of nectar which it does not attempt to defend, a strategy generally known as 'trap-lining'. Territorial hummingbirds are usually unspecialized feeders, not closely adapted to feeding at particular kinds of flower; trap-lining hummingbirds are typically specialists, adapted to visiting a few kinds of flowers. This division of feeding strategies will be discussed later with reference to our observations in the Andes; we mention it here as we refer to trap-lining and territorial feeding in the species accounts that precede the discussion. Study areas Our original intention was to spend the whole of our time in one area, on the eastern slopes of the Eastern Andes, quite close to Bogota. Owing to unforeseen circumstances we were forced to leave this area half way through the period, and moved to another locality about 1 50 km to the NNE, on the western slopes of the Eastern Andes, where observations were made in two areas several kilometres apart. The three study areas (referred to in what follows as Fonte, Carare and Togui) are briefly described below. The first two are in the lower part of the Temperate Zone, as defined by Meyer de Schauensee (1964), and the third is near the lower limit of the Subtropical Zone. Slopes of Cerro Fonte, Vereda de Ferralarada, Municipio de Choachi, Cundina- marca (4° 32' N, 73° 51' W), 2400-2550 m; 3-25 July. Gently sloping, western- facing mountainsides, partly wooded, partly cleared for primitive agriculture (pasture, small fields of potatoes and maize), cleared areas in places reverting to bushy secondary growth. Woods mostly with trees not over 1 5 m in height (perhaps secondary, resulting from earlier clearances). Slopes of Cerro Carare, 9 km ESE of Togui, Mun. Togui, Boyaca (5° 54' N, 73° 25' W), 2300-2500 m; 29 July-3 August, 14-22 August. Natural forest on gentle or moderate slopes, thinned in places by extraction of larger trees, with open areas (some cleared for farming, some probably naturally unforested, due to poor drainage). Hacienda Versalles, near Togui, Boyaca (5° 55' N, 73° 32' W), 1700 m; 27-30 July, 4-14 and 23-24 August. Cattle farm with sugar cane, maize and 3 small patches of coffee, natural vegetation almost entirely gone; trees confined to strips along river banks and between fields; Eugenia jambos (introduced) more abundant than any other tree. Methods A special effort was made to find all the flowers that were being fed at by hummingbirds in each area at the time of our visits. Flowers were measured (internal length of tubular part of corolla, from opening to base), and measurements were made of nectar concentrations and rates of nectar production. Nectar was extracted with disposable micropipettes (manu- factured by Camlab, Cambridge, England) of various capacities (5 to lOOul), and sugar concentrations were measured with a Bellingham & Stanley pocket refractometer calibrated from 0 to 50%. Percentage concentrations were not compensated for temperature, but at the temperatures and concentrations recorded compensations are all below 0-4% and thus negligible. To measure nectar production rates, flowers were drained of nectar and then bagged, so as to be inaccessible to birds over the period of the measurement. Observations of the feeding behaviour of hummingbirds were made with 8x and lOx binoculars, usually at fairly close quarters. Some observations were 'casual', i.e. made during RELATIONSHIPS BETWEEN HUMMINGBIRDS & FLOWERS 107 random walks in the area, during which all observed feeding was recorded; but most were made during timed watches at nectar sources which we had found to be, or suspected to be, important. Each visit by an individual hummingbird to one plant or (in the case of some herbaceous plants, small bushes and clumped trees) group of plants was counted as a unit for purposes of tabulation. When practicable, stop-watch timings were made of the rate at which flowers were visited, and the number and duration of probes into individual flowers. A limited number of hummingbirds were caught in mist-nets, weighed, measured, and marked with quick-drying paint (dope), mainly on the central tail-feathers, or with coloured inks on white or pale parts of the plumage, for future identification in the field. Specimens of food plants were collected, and have been deposited in the Herbarium of the Institute de Ciencias Naturales, Universidad Nacional de Colombia. The hummingbirds On the basis of Meyer de Schauensee's (1964) statements of ranges and altitudinal zones of hummingbirds in Colombia, a potential total of 2 1 species might occur in the Temperate Zone in the vicinity of Carare and Fonte. Five of these were not to be expected in our study areas, as they are birds of open country or of the upper part of the Temperate Zone, above the level where we worked. Of the remaining 16 species, we recorded 14 in one or both of our study areas. The two species not recorded were Coeligena bonapartei, which might have occurred at Carare, and Ramphomicron microrhynchum, which might have occurred in both areas. Again on the basis of Meyer de Schauensee's statements, there are 1 8 species that reach their upper limits in the Subtropical Zone and, from their geographical ranges, might have been expected in at least one of our study areas. Of these, five were recorded in the Carare and/or Fonte study areas, and four more at Togui. Hence our sampling of the more strictly subtropical hummingbirds was much less complete than our sampling of the temperate species, as would be expected from the fact that our two main study areas were at lower temperate levels and the subtropical Togui study area had very little natural vegetation. All the species recorded are listed, together with weights and measurements, in Table 1 . Observations at Fonte Vegetation, and available nectar sources As already mentioned, the woods in the Fonte study area consisted mainly of rather small trees, not over 1 5 m in height, and may mostly have been secondary, resulting from earlier clearances. Some tracts, consisting almost entirely of one species of Melastomataceae, were certainly secondary. In some places the undergrowth of unfenced woods had been badly damaged by cattle infiltrating from adjacent pastures. In some rocky and steep places the trees were higher and more massive, and the vegetation probably approximated to the natural condition. Very few of the canopy trees had flowers attractive to hummingbirds at the time of our visit; the only two that were found, Symplocos mucronata and an unidentified species, were both uncommon. Epiphytic bromeliads of several species were numerous on the larger trees, but were not in flower at the time of our visit. The under-storey of the woods in the lower part of the study area consisted largely of slender, straggling trees of a single species, Palicourea angustifolia, which were in flower at the time of our visit and were the most important nectar source for the unspecialized hummingbirds. This tree was much less abundant in the upper part of the area, where another, larger species of Palicourea was fairly common. This latter species, P. cf. anacardifolia, had just finished flowering at the time of our visit and most trees had unripe fruit; almost certainly it had been an important nectar source in the period immediately before our arrival. The only other under-storey trees and shrubs seen to be fed at by hummingbirds were: Symbolanthus tricolor, a sparsely distributed straggling shrub; Psychotria cf. aschersonianoides, a small locally common shrub; and Cinchona cf. 108 D. W. SNOW & B. K. SNOW Table 1 Wing-lengths, bill-lengths and weights of hummingbirds recorded at Fonte (F), Carare (C) and Togui (T) Area Wing mm Bill (exposed culmen) mm Billas% of wing Weight g Doryfera ludoviciae C 60-0 32-9 54-8 5-7 Phaethornis guy T 61-3 44-3 72-3 5-2 Colibri thalassinus C 65-6 20-8 31-3 5-3 Colibri coruscans FC 77-2 24-2 31-3 7-8 Anthracothorax nigricollis T 65-7 24-8 37-7 7-0 Chlorostilbon gibsoni T 46-1 16-2 35-1 3-1 Chlorostilbon poortmani FC 42-5 16-5 38-8 — A mazilia franciae T 53-6 24-3 45-2 5-0 Amazilia cyanifrons T 54-5 17-8 32-6 5-0 Amazilia tzacatl T 55-0 20-4 37-1 4-8 Adelomyia melanogenys CT 52-8 15-2 28-8 4-2 Lafresnaya lafresnayi F 61-7 25-2 40-8 5-9 Coeligena prunellei C 77-8 29-8 38-3 7-4 Coeligena helianthea F 71-6 29-9 41-8 — Coeligena torquata C 75-5 33-0 43-7 7-0 Ensifera ensifera F 76-5 83-1' 108-6 12-5 Boissoneauaflavescens C 76-7 17-9 23-3 8-6 Heliangelus amethysticollis FC 64-3 18-0 28-0 5-9 Eriocnemis vestitus F 59-3 18-3 30-9 4-9 Eriocnemis cupreoventris F 61-6 18-6 30-2 5-6 Ocreatus underwoodi C 43-1 14-0 32-5 3-1 Lesbia nuna F 50-7 10-0 19-7 — Metallura tyrianthina F 53-8 11-0 20-4 3-4 Aglaiocercus kingi C 53-5 15-2 28-4 4-4 Acestrura mulsant FCT 40-4 16-7 41-3 3-2 Note: All figures are means. Sexes are combined. Measurements are from museum specimens from the Eastern Andes of Colombia, supplemented by data from this study. Weights are from this study, supplemented by data from Greenewalt ( 1 962), Miller ( 1 963), Carpenter ( 1 976) and Feinsinger et al. ( 1 979). 1 Bill-length very variable, up to c. 105 mm. Effective length for probing tubular flowers is about 10 mm greater than bill-length, owing to marked narrowing of the skull towards the base of the bill. pitoyensis, a rather uncommon tree in the upper part of the study area. Among the vines, Pentadenia strigosa was common in all the woods with undisturbed undergrowth. Its large orange flowers, borne almost from ground level up to about 5 m, were conspicuous in the woody and mossy undergrowth. Along wood-edges and in more open, rocky areas an ericaceous shrub, Cavendishia cordi- folia, was locally abundant and in flower. As a nectar source for several hummingbird species it was second only to Palicourea angustifolia. Only two other ericaceous plants were found in flower, Disterigma alaternoides, which was rare, and a single example of an unidentified species. The poverty of the Ericaceae was in marked contrast to their richness and abun- dance at Carare (p. 1 1 7). In bushy second growth, especially along edges, several hummingbird flowers were fairly abundant. Castilleja fissifolia, a scrambling shrub, was common in such places. Passiflora mixta, a vine with very long tubular flowers, was more local but conspicuous where present. An abundant low scrambling shrub, Siphocampylus bogotensis, began to come into flower about half way through our visit. Manettia coccocypseloides, a climber with small white flowers, was more local. Blackberries, Rubus spp., were abundant and in all stages of flower and fruit. Cuphea dipetala, a slender shrub, was generally distributed in bushy areas but sparse. Rubus and Cuphea also grew as low plants out in the open in neglected fields. u. SO ^ O m m rs ^ £] — r«-) (N r* -^ C _o ^ ^ 0 CN) ^ *~* 3 fi C-J ^T . -O t. f"J *^f r^ 0 3 X X X 03 O..C tj u" 03 03 IO ^^ 03 Z a O • 06 r^- r*") • cd CH 2 — (N — ^N ^ •- f i—^ ^ 'c ^" — ro — (N| OO QJ f 03 .'ti .'ti >- i % s^ s> ^ ^ ^ > £> >, aj» U ^ w uj o o c E ^ ^ ^ = c c 15 c "S '_? _O 3 .ti U o o 'S-'cx^j ^ ^ | o ^ o ^ "« c"f >. c ? c c (U 'a 'a t« 00 <*- "P 0 C 09 •£§ ^^ ^f ^o r — — -> oc ^t vO ON OO u~> (N ^^ r*^ ^O C — — — " ^H — CN m — O — C^l ^^ 2- C?_Q — £ c - 5 3 o C3 u. ^ U c/f to Co .5$ "^S w ^ ristics of hummingbird flowei ^i^ ->~ ^ s: sf -^ 3 -2^ 1 fe ^^ i ^1 §i aSig !s|| ^|-||||g >J ••^ rs rZl ^"-* •>» <:^J " •** ^ PN s^8-2 S 5 ^ ^ H J S ^s|« -HPHs 8 Si S ci^i §.« 5 S 2 ^^1^- § l"i=l^ ss C^^UQ^O^C^U^CO^ Palicourea cf. anacardifbi Palicourea angustifolia Psychotria aschersoniano ntal; P = pendent; U = upright, nean. Single figures indicate singl maximum recorded) recorded; for further details, see t osed tube; see text. o3 S^-gs^ •g^u 03 .C U 03 O (1J gj J; 0 C 0 o 2 CO ' "T? TO cd r" ' ^ i— i 0^ .j QJ c3 3 o f- Q. UJ OJ J H O on O U &. ^ 109 HO D.W. SNOW & B. K. SNOW No other plants were important nectar sources at the time of our visit. Over 99% of all hummingbird feeding records were from the plants mentioned above. Flower and nectar characteristics of the majority of them are given in Table 2, and further details of the most important of them are given in the following section. The most important nectar sources Palicourea angustifolia The tubular mauve flowers are borne in racemes. Most racemes did not hold more than 6 open, nectar-producing flowers at a time. Individual flowers lasted 2-3 days. The figures for nectar concentration given in Table 2 are based on 46 out of 50 readings, four very low readings being omitted (5-9%). Concentrations tended to be higher in the early morning (14-18%) than at midday or in the afternoon (10-16%). There was also some decrease in concentration with the age of the flower: the last readings for 6 flowers, immediately before the end of their nectar production, averaged 1 1% (range 6-13%). Nectar production rates also were highest in the morning, the average for 12 morning rates being 1-23 ul/hr, while rates for the whole day averaged 0-97 ul/hr. Hourly rates for the night (which included the early evening and very early morning) averaged 0-4 1 ul/hr. The maximum figure for nectar production over 24 hours given in Table 2 (29-5 ul) was approached by two others (28-5, 27-9), while three others were considerably lower (21*5, 13'8, 9-l). Age of the flowers might largely have accounted for these differences. Because of its abundance, the considerable number of flowering racemes on even quite small plants, and its short corolla tube, Palicourea angustifolia offered an abundant nectar supply to the shortest-billed of the hummingbirds, as well as to those with longer bills. Several timings of Eriocnemis vestitus showed that it visited Palicourea flowers at a rate of one flower every 1-3 s (combined figures, 141 flowers in 178 s), and this rate seemed typical of the smaller hummingbird species. As mentioned later, four species of hummingbirds maintained feeding territories at Palicourea clumps. Cavendishia cordifolia The tubular, pale pink and white flowers of this heath are borne in terminal clusters, pro- tected by scaly sheaths of pink bracts. Individual flowers last 2-3 days. The figures for nectar concentration given in Table 2 are based on 32 of 36 readings, the four omitted being abnormally low (2-6%). The nectar concentration of individual flowers was rather consistent over 2 or more days (i.e. some had consistently high, and some low concentrations), with a tendency for concentrations to decrease very slightly in the course of a day (by 0-6%, average of 1 0), and from one day to the next (by 1-5%, average of 6). As already mentioned, Cavendishia was second in importance only to Palicourea as a nectar source for several hummingbird species. The richest feeding areas for these birds were along wood-edges, where both Palicourea and Cavendishia frequently grew side by side. Castilleja fissifolia The flowers are borne in terminal spikes, the pale green corolla being ensheathed along most of its length by the calyx whose distal half is red. The corolla is slightly curved along its whole length; it is tubular basally (for the proximal 20 mm), but distally is split along the under-side (for the distal 25 mm). In spite of many attempts, we never found it easy to extract much nectar from Castilleja flowers. Recorded nectar concentrations were 23, 17 and 16%. The maximum figure for nectar production over 24 hours was 1 1 ul; other figures were 8, 6 and 1 |il. Castilleja was one of the preferred flowers of Lafresnaya lafresnayi, whose curved bill fits the corolla exactly. Almost certainly Lafresnaya is the plant's chief pollinator, as was indi- cated by the pale pollen that conspicuously coated their foreheads when they were feeding on RELATIONSHIPS BETWEEN HUMMINGBIRDS & FLOWERS 1 1 1 it. The only other hummingbird seen to visit Castilleja flowers, Lesbia nuna, has a much shorter, straight bill. It either probed the flowers at the base of the split in the corolla, about 19 mm from the nectary, or pierced the base of the calyx, on the upper side. Neither method would bring it into contact with the anthers or stigma. Pass (flora mixta The long tubular flowers, with a ring of pink petals round the entrance to the tube, arise singly from leaf axils near the growing tip of the long climbing and sprawling stems. Along each stem only one or two flowers are open and producing nectar at any time, those that are distal being in bud and those that are proximal in fruit. Each flower usually lasts 4-5 days (extremes of 3 and 6 recorded). The figures for nectar concentration in Table 2 are based on 19 out of 22 readings, the three omitted being very low (17-18%). Two of the very low readings, and perhaps the third, were from flowers that were nearly over. (A reading of 12% from a flower that was already withered is also omitted). Nectar concentrations remained rather steady throughout the day; but measurements of individual flowers on successive days indicated that concentration rises and then falls during the life of the flower. Nectar pro- duction is copious, but also very variable, the maximum amount produced in 24 hours shown in Table 2 (506 jil) being a good deal more than the next highest (397 ul). A younger flower on the same plant as the latter produced only 8 1 jil on the same day in the same 24-hour period. Another flower produced 98 ul in one 24-hour period, and 1 50 ul in the next. Ensifera ensifera is the only hummingbird that we saw taking nectar 'legitimately' from Passiflora mixta, i.e. by entering through the opening of the corolla tube, and in fact is the only one that could possibly do so, its effective bill length (Table 1) being almost the same as the Passiflora corolla tube. Practically all the Passiflora flowers in the study area were, how- ever, pierced at the base by the flower-piercer Diglossa albilatera, and hummingbirds of several species extracted nectar from the holes made by Diglossa. Consequently, unprotected Passiflora flowers usually contained very little nectar. The effect of this nectar 'thieving' on Ensifera is discussed later (p. 1 1 5). Pentadenia strigosa This climbing plant was abundant in shady places, especially among rocky outcrops. It flowered as low as a few cm from the ground in dense second growth and up to 5 m where it had clambered up trees in older woodland. The flowers grow in pairs along the stem, both members of a pair opening at about the same time; they last 7-8 days and are protandrous. The hairy orange corolla is in the form of a wide tube (28 mm long, 18 mm deep), decurved distally and opening downwards, and is borne horizontally, suspended at the end of the long, vertically hanging pedicel. Only Lafresnaya females were seen feeding at Pentadenia flowers; no flower-piercers were seen to visit them and no pierce holes were found on the flowers examined. Several features of the flower suggest adaptations protecting the nectar against flower-piercers: the delicate suspension of the flower at the end of the long pedicel, which would make it difficult to perch on and inaccessible from other perches, and the wide, rounded corolla tube, which would be difficult for a flower-piercer to grasp and pierce. Symbolanthus tricolor The pink flowers, with darker pink streaks, are borne very sparsely, only 1-4 being open at a time on the few plants that were seen. The nectaries are protected by the tough thick calyx base. Only a single reading of nectar concentration was made, and one measurement of rate of nectar production (Table 2). Two hummingbird species, Coeligena helianthea and Lafresnaya, were seen visiting these flowers. Siphocampylus bogotensis Little information was obtained on this plant, as its flowering season was only just beginning 112 D. W. SNOW & B. K. SNOW when we left the area. The nectar concentration and production rates given in Table 2 are based on single measurements. It seemed very likely that Siphocampylus Was going to be an important nectar source for Lafresnaya, the only hummingbird species seen to visit it. The corolla length and curvature fit LafresnaycCs bill closely. The flowers are soft, with no pro- tection of the nectaries, and preliminary observations indicated that Diglossa albilatera regularly pierced them at the base. Feeding ecology of the hummingbirds Of the 1 1 species of hummingbirds recorded on Fonte, all but two were regularly present and seen whenever watches were kept at appropriate flowers. The two exceptions were Acestrura mulsant, which was seen only once, and Colibri coruscans, which was abundant at lower levels (below c.2200 m) but only once seen at the lower edge of our study area. The nine resi- dent species can be divided into three groups: four medium-sized to small woodland species with short, straight bills, which held territories centred on food plants with short corolla tubes (Heliangelus amethysticollis, Eriocnemis cupreoventris, E. vestitus, Metallura tyri- anthina); three larger, long-billed species of woodland or woodland edge, feeding mainly or entirely on specialized hummingbird flowers and non-territorial (Lafresnaya lafresnayi, Coeligena helianthea, Ensifera ensifera); and two small open-country species (Lesbia nuna, Chlorostilbon poortmani). The feeding records for these hummingbirds are summarized in Table 3. Three species of flower-piercers (Diglossa spp.) occurred in the study area, two of which exploited hummingbird flowers. One of them, D. albilatera, was abundant and had a signifi- cant effect on the feeding ecology of the hummingbirds with which it competed for nectar. Table 3 Hummingbird feeding records, Fonte C.p. L.l.d L./.9 CM. E.e. H.a. E.v. E.c. L.n. M.t. Shrubs and scramblers Cavendish ia cordifolia 1 1 16 5111343 Rubussp. 3 6 16 Cuphaea dipetala 1 1 171 Svmbolanthus tricolor 56 1 ' Castillejafissi folia 1 17 35' Siphocampylus bogotensis 1 ,6 I1 Psychotria aschersonianoides 6 4 Vines/climbers Passiflora mixta 29' 10 I1 3' 29' Pentadenia strigosa 8 Manettia coccocypseloides Trees Cinchona cf. pitoyensis 1 Palicourea cf. anacardifolia 1 1 Palicoureaangitsti folia 1 23 4 5 21612369 Other plants Insect-hawking Totals (453) 5 2 31 2 44 1 61 10 26 19 7 36 16 74 5 147 Hummingbirds abbreviated as follows: C./J., Coeligena helianthea: C.p.. Chlorostilbon poortmani: E.c.. Eriocnemis cupreoventris; E.e., Ensifera ensifera; E.v., Eriocnemis vestilus; H.a., Heliangelus amethysticollis; L.I.. Lafresnaya lafresnayi; L.n.. Lesbia nuna; M.t., Metallura tyrianthina. Note: ' nectar taken 'illegitimately', through hole in corolla base. RELATIONSHIPS BETWEEN HUMMINGBIRDS & FLOWERS 1 1 3 Heliangelus amethysticollis Eriocnemis cupreoventris Eriocnemis vestitus Metallura tyrianthina These four straight-billed, medium-sized to small hummingbirds are treated together, as all fed largely on Palicourea angustifolia and (especially Eriocnemis spp.) on Cavendishia cordifolia, and individuals of all four held feeding territories centred on these plants. They frequently came into conflict with one another, dominance relations being in accordance with their size, i.e. in the order in which they are listed above. A small wood, measuring about 100 by 30 m (0-3 ha), surrounded on three sides by pasture and on the fourth by scrub, with an under-storey consisting largely of Palicourea while Cavendishia grew abundantly along the edges, contained an extraordinary number of hummingbirds of these four species for the first two weeks of our visit. Towards the end, the flowering season of Palicourea appeared to be coming to an end and the number of hummingbirds decreased. Accurate censusing was not possible in the dense growth within the wood, but probably at least 25 birds were present around the middle of July. Trapping also gave some idea of the numbers. Nine different individuals were caught in two 13-m mist-nets set within the wood on 10-1 1 July (2 Heliangelus, 4 E. cupreoventris, 1 E. vestitus, 2 Metallura), and on 18 July nets set in the same positions caught six different birds (1 Heliangelus, 3 E. cupreoventris, 2 Metallura), all except one Metallura different from the individuals caught a week earlier. Several hours were spent in looking for marked birds 2-4 days after the netting operations, but only three marked individuals were seen among the many that were examined. It seemed that individuals were holding very small feeding terri- tories, many of them perhaps rather briefly. There was much trespassing and chases were very frequent, especially early in the period, but were almost always impossible to follow for more than a second or two. Two Heliangelus territories (one in an adjacent wood with similar vegetation) consisted of (1) several Palicourea bushes spread over a space of 10 by 9 m, and (2) two Cavendishia bushes about 10 m apart. In an even smaller copse, roughly circular and about 25 m in diameter, with sparse Palicourea in the under-storey and a few Cavendishia bushes round the edge, a single male E. vestitus held a territory from 7 to 2 1 July, effectively excluding other hummingbirds from the area. It also chased intruding White-sided Flower-piercers Diglossa albilatera, which were present in all the larger woods, so persistently that it effectively excluded them, thus maintaining its nectar resources intact. On 22 July, when the number of Palicourea flowers had declined somewhat, it disappeared and its place was taken almost immediately by a Metallura, which occupied the area and patrolled it much as E. vestitus had done, but was unable to prevent occasional visits by Lafresnaya and Lesbia nuna. It seemed that the nectar supply had become insufficient to maintain E. vestitus but could still maintain the smaller Metallura. There were some differences in habitat preference between these four hummingbirds. We did not record Heliangelus outside more or less closed woodland, whereas E. cupreoventris and Metallura showed some preference for woodland edges, at times coming well into the open. Rather few observations were made of E. vestitus, the least common of the four species (as the trapping figures indicate), but from the records it appeared to favour more open habi- tats than the others. This is consistent with its general distribution: according to Meyer de Schauensee (1964) it is a bird of the temperate and lower paramo (i.e. more open) zones, whereas E. cupreoventris is a bird of the subtropical (more densely wooded) and temperate zones. Among the species that we observed, they were the only examples of overlap between two closely related species with different altitudinal preferences. Some of the differences between the feeding records for the four species probably resulted from these differences in habitat preference; for instance the marked preponderance of feeding at Palicourea by Heliangelus, which fed within the woods, compared with the slight preponderance of feeding at Cavendishia by the two Eriocnemis species, which fed more in the open. The lack of records of Heliangelus feeding at Passiflora may have the same 114 D. W. SNOW & B. K. SNOW explanation. The very few records of feeding at Cavendishia by Metallura, however, indi- cate active avoidance of Cavendishia flowers, whose 15 mm corolla tubes may be too long for efficient probing by Metallura's 1 1 mm bill. Metallura showed much the greatest variety in its nectar feeding of all the species in the area (Table 3), especially in visiting small flowers not seen to be visited by other species (Psychotria, Manettia, Symplocos). Lafresnaya lafresnayi This was the only hummingbird with a markedly decurved bill in the Fonte study area. It is, in fact, the only species with such a bill that occurs at temperate levels anywhere in the Andes. Our feeding records for males and females were very different, and so are treated separately. It is probably significant that females, although shorter-winged, have on average longer bills than males. The mean difference in bill-length is only about 2 mm, but a long- billed female may have a bill 4 mm longer than a short-billed male, and such a difference may well affect their relative efficiency in feeding at different flowers. Females were recorded feeding mainly at flowers of four species whose long corolla tubes fitted their bills rather closely (Castilleja, Symbolanthus, Pentadenia and Siphocampylus). Three of these were not seen to be visited 'legitimately' by any other hummingbird species and probably depended on Lafresnaya for pollination (see also above, under Castilleja). In their behaviour, females seemed to be typical trap-liners, reminiscent of hermit humming- birds Phaethornis, moving between scattered nectar sources and showing no tendency to defend a feeding territory. As mentioned later, a female that regularly visited a large Castilleja clump at which a female Lesbia nuna, a much smaller bird, held a territory was regularly chased by the latter. They fed silently, but during longer flights between nectar sources regularly uttered a monosyllabic 'seep' (reminiscent of the flight call of some Phaethornis species). Watches at places that females visited on their foraging rounds gave some idea of the timing of their feeding routine. Of 17 intervals between successive visits of what was probably the same bird to a nectar source, 1 5 were from 10 to 3 1 min, with a mean of 2 1 min. This figure is in good agreement with observations made at a feeding site which a female regularly visited, where three of its main food plants (Castilleja, Symbolanthus and Pentadenia) were flowering close together along the edge of a wooded strip. At each visit it remained in the area for 2 to 1 1 min, during which it went to a favourite perch between feeding bouts. It would then leave, and was absent for 1 5 to 38 min. The rather brief observations suggested that, in contrast to the females, male Lafresnaya tend to hold feeding territories. On 22 July, a male held a small territory within a wood in which there were many small Palicourea trees in the under-storey and 16 open Pentadenia flowers near the ground. This bird was seen feeding only at the Palicourea. Feeding bouts were frequent, mostly at intervals of 2-16 min. Once it chased a female Lafresnaya which intruded, and once a Diglossa albilatera. It was probably present for almost the whole of two watches, totalling 2 h 40 min. It was silent when perched in the territory between bouts of feeding, but it uttered the monosyllabic 'seep' when flying round the territory and between probes when feeding. Another male that was watched for an hour in part of a wood that con- tained only Palicourea and a little Castilleja seemed to be attempting to infiltrate a territory previously held by a male Metallura. It fed silently, mainly on Palicourea, and was occasionally chased by another hummingbird, probably the Metallura. In a place where woodland bordered an area of semi-open bushy second growth, with plants mostly not more than 3 m tall, there was intensive activity on three successive mornings among a group of at least 5 Lafresnaya. Watches totalling 5 h were insufficient to clarify the nature of the activity, but it was not related to feeding. One male spent much of the time (at least 77 out of 150 min of one watch) on a small group of perches about 10 m up in a tree on the woodland edge overlooking the bushy area. He was regularly visited at this perch by birds in female plumage and occasionally by male-plumaged birds, and these visits usually led at once to long flight chases over the bushy area in which up to three other birds might join. On the perch the male was silent; flight chases were accompanied by bursts of RELATIONSHIPS BETWEEN HUMMINGBIRDS & FLOWERS ] 1 5 rapidly repeated calls, 'see-see-see-see .'. .', apparently uttered by the bird being chased. This activity suggested an initial stage of some kind of courtship display. Perhaps males group themselves in a 'dispersed lek'; but if so, one would not expect them to be silent on their perches. Coeligena helianthea Nearly all our feeding records were of adult males. Females were seen feeding on only two occasions (4 feeding records: Cavendishia 2, Symbolanthus 1, Tropaeolum 1). Males fed predominantly by trap-lining. A few observations suggested that some, at least, spent a good deal of their time at nectar sources that were not within the feeding territories of other hummingbirds, perhaps using them as a base from which to make trap-lining excursions. Aggressive behaviour was seen only once, when a trap-lining male Coeligena chased a female Metallura that came to a clump of Passiflora at the same time as it did. Sixteen intervals between successive visits to the same flowers by what we presumed (from the behaviour) to be the same bird ranged from 16 to 58 min, with an average of 37 min. Although we had some records of their feeding on small flowers, this fairly large, long- billed hummingbird probably feeds most efficiently from large flowers with straight corolla tubes. At the time of our visit, Symbolanthus and Passiflora were the only such flowers avail- able. The latter has a corolla tube so long that Coeligena cannot probe it legitimately, but instead uses the holes that Diglossa albilatera pierces at the base of the calyx. Timed visits of Coeligena to Symbolanthus flowers lasted from 2 to 9 s, during which up to three probes were made, and visits to Passiflora flowers lasted 3-3j s. Ensifera ensifera We saw Ensifera feeding only at the flowers of Passiflora mixta, and in fact never saw it except when we were watching at clumps of Passiflora. The extraordinarily long bill of Ensifera, which exceeds in length the head and body combined, closely matches the corolla tube of Passiflora. No other plant that was in flower in the study area approached Passiflora in the length of the corolla tube. Ensifera must be a most pronounced trap-liner. Usually a bird suddenly arrived, fed at the Passiflora flowers, and flew straight off. Occasionally it perched briefly near the flowers on arrival, or before flying off. Its feeding visits were few and far between. We watched at Passiflora clumps for a total of 23 morning hours (0540-1200 h), in which time we saw 6 visits by Ensifera. Four other visits were seen in the morning during watches that were not timed (because attention was not being concentrated solely on the Passiflora). Although six of the timed watches were longer than 2 h, and two exceeded 4 h, only once did we see two visits by Ensifera during a watch (with an interval of 31 min). Nine of the ten visits were in the periods 0600-0800 and 1000-1 100 h. As has been mentioned, practically all the Passiflora flowers that were examined had been pierced at the base by Diglossa albilatera. In addition, at least four hummingbird species (and at least one butterfly) took nectar from the holes made by Diglossa. Visits by the nectar thieves were regular, and the combined visits were frequent. Thus at the main Passiflora clump that was studied a male Diglossa albilatera made a thorough round of all the flowers about three times per hour, a female Metallura about every half-hour, and a male Coeligena about every 40 minutes. Doubtless as a consequence of these frequent visits, Passiflora flowers that we sampled at this and other clumps yielded very small quantities of nectar; many gave none that was measurable - in spite of the fact that Passiflora flowers produce nectar copiously (Table 2). In order to determine to what extent Ensifera's food supply was being depleted by the nectar thieves, we protected the bases of a number of Passiflora flowers by wrapping them with polythene strips while they were still in bud. The experiment was unfortunately termin- ated prematurely when we had to leave the Fonte area, but the first results were striking. Visits by Ensifera to unprotected flowers were brief, 4 being the maximum number of probes that we recorded. On 19 July, 6 days after several flowers had been protected at the clump 116 D. W. SNOW & B. K. SNOW where most observations were made, an Ensifera arrived at the clump, went first to three high, unprotected flowers, probed each of them 2 or 3 times, then flew straight down to one of the wrapped flowers (ignoring several flowers that were nearer), probed it ten times, flew straight to another wrapped flower, probed it 9 times, and then flew off. The two protected flowers had opened 1-2 days previously. Even more striking was a visit by Ensifera to the same clump on 22 July. On this occasion it went first to an unprotected flower, fed at it for about 2 seconds, then flew to one of the wrapped flowers (not one of those visited on the 19th, which were now over), made 27 probes, perched briefly, and flew off. It seemed that this bird had learnt which flowers were most rewarding, and that one or two flowers could supply as much nectar as it could take on one visit. Lesbia nuna The few males that were seen appeared to be holding territories in open areas with scattered trees, not far from woodland edge. They were mostly seen feeding by making aerial sallies for flying insects from tree perches, and only a single visit was seen to a flower (a low-growing Rubus). Many more records were obtained for females, which were seen visiting flowers of five species (Table 3). Particular attention was paid to one female, which defended a small feeding territory centred on a large clump of Castilleja. This bird fed mainly within an area measuring c. 25 by 15m, containing the Castilleja clump and, a little separate, a patch of bushy growth including Castilleja, Palicourea and Cavendishia. At times it moved out onto an adjacent open field and fed on low Rubus bushes, and occasionally it moved about 50 m away to feed on Palicourea along a wood-edge. Although it ranged over a 25 x 15 m area for most of its feeding, it defended only the large Castilleja clump, not attacking the hummingbirds that often came to feed in the bushy patch about 10 m away. Within the defended area it regularly uttered a short, monosyllabic 'zit, zit' while feeding and immediately after returning to the perch. Between bouts of feeding it perched for much of the time on a few favourite perches close to the Castilleja clump. During a 2-hour watch from 0730 to 0930 h it spent 39% of the time on these perches, and during a 2-hour watch from 1 345 to 1 545 h, 48% of the time. The only other hummingbird seen feeding at the Castilleja clump was a female Lafresnaya. Although a considerably larger bird, it was regularly chased by Lesbia nuna and on several occasions was either prevented from feeding or had its feeding interrupted. Several timings of feeding rates showed that Lesbia was consistently less efficient than Lafresnaya in feeding at Castilleja. On average Lesbia visited a flower every 2-7 s (combined figure from records on 3 days, averaging individually 2-8, 2-3 and 3-1), compared with l-3s for Lafresnaya. This relative inefficiency seemed to result from the fact that Lesbia could not hover and probe the Castilleja flowers by inserting its bill along the curvature of the corolla, as Lafresnaya could. Instead, it probed the flower either by inserting the bill on the under- side, at the point where the corolla tube divides, or by piercing a hole (or using an existing hole) on the upper side of the calyx near the base. Both methods frequently involved clinging to the flower with beating wings. Chlorostilbon poortmani This species was seen in only two places, in one of which a male held a feeding territory and was once seen engaged in what appeared to be courtship display with a female. Both were open areas near woodland edge, and the four kinds of plants at which the birds were seen feeding were low-growing herbs or stunted shrubs. The flower-piercers Three species of Diglossa occurred in the study area, D. albilatera, D. caerulescens and D. carbonaria. The first two of these occurred mainly in woodland and along woodland edges, while the third was a bird of more open, bushy country. Far the most important of these species, in its effect on the ecology of the hummingbirds, was D. albilatera, the smallest of the three and the one whose bill is most highly modified for piercing the base of flowers. RELATIONSHIPS BETWEEN HUMMINGBIRDS & FLOWERS 1 1 7 D. albilatera fed very largely on the nectar of Palicourea angustifolia. They are small agile birds, and the rate at which they are able to exploit Palicourea flowers, about one flower per second, is remarkable. They are regularly attacked by hummingbirds, which clearly recog- nize them as ecological competitors, and consequently are very furtive, feeding silently and keeping as far as possible to the interior of the plants on which they are feeding. As already mentioned, the flowers of Pass (flora mixta were almost invariably pierced by D. albilatera; but the proportion of the D. albilatera population that fed on Passiflora flowers cannot have been very high, as Passiflora clumps were sparsely distributed. In contrast to their behaviour at Palicourea, they remained for several seconds at Passiflora flowers, often piercing them in more than one place, apparently staying until they had extracted all the nectar that they could reach. Although their general behaviour was still furtive, we never saw one attacked while feeding at Passiflora, probably because no hummingbirds maintained feeding territories centred on Passiflora clumps. The effect of their exploitation of Passiflora flowers must have been to reduce greatly the nectar available to Ensifera, as dis- cussed above, not only by taking it themselves but also by making available to other hummingbirds a nectar source that would otherwise have been unavailable. Other flowers seen to be visited by D. albilatera were Cavendishia, Castilleja and Symbolanthus (once each), Siphocampylus (twice), a small and apparently rare Passiflora lacking a corolla tube, and the introduced foxglove (Digitalis). The flowers of Siphocampylus, which were just beginning to open when we left the area, are soft and unprotected at the base of the corolla tube, and it seemed likely that their exploitation by Diglossa would affect the amount of nectar that they would provide for Lafresnaya, their probable main pollinator. Diglossa caerulescens has a much less specialized bill than D. albilatera, and very different habits. Several times we saw single individuals or pairs of D. caerulescens accompanying mixed foraging parties of tanagers, flycatchers and other birds. We also saw them eating fruits of Cavendishia and Rubus. They were recorded visiting flowers of Palicourea and Rubus, but when they could be observed closely they were seen to enter the flower 'legitimately'. We never definitely saw one piercing a flower. Diglossa carbonaria, with a bill nearly as specialized as that of D. albilatera, was recorded piercing the flowers only of Palicourea, Cavendishia and Manettia. An individual that held a territory embracing the large Castilleja clump at which we spent watches totalling 1 1 h, showed no interest in the Castilleja flowers though it sometimes perched close to them. This species is considerably larger than D. albilatera and is not so subordinate to hummingbirds, though its behaviour in approaching and working through vegetation in search of flowers is similarly furtive. Once one was seen to drive an Eriocnemis vestitus from a Cavendishia flower, and once one was attacked by E. vestitus while feeding at Cavendishia but was not dislodged. Observations at Carare Vegetation, and available nectar sources Most observations at Carare were made along Ukm of forest path at a height of 2300 to 2450 m. The path ran through an extensive area of forest some 2{ km wide, extending from the ravine of the Rio Uvasa (2000 m) to the top of a steep cliff (Las Alturas, 2800 m) and probably beyond. There had been recent and old selective felling in the forest but apparently no clear felling. In addition observations were made in adjacent open country, some of which was naturally unforested being poorly drained and boggy but most had been cleared and probably regularly burnt. These open areas were at 2200 to 2300 m, with a small area of about 2 ha at 2450 m. In the forest 20 species of plants and in the open areas 4 species were seen to be visited for nectar by hummingbirds in 10l days of observation. Of these 24 species, 9 were climbers including the scrambling tree heaths, 7 were woody shrubs, 4 were herbaceous ground-living plants, 3 were trees, and one was an epiphyte (Table 5). Considered by families, the 118 Bg.f /•S oo'c -o 2 u a — OO VO OO ON rf O rl OOO I«N (-s) Tt — Ill (N (N — ^ cO ^ ^^ rs — X5 oo— ' — a a o o ed U x> oo 1 •E 3 JS ^ I) § ,0 Ci, to ':~ ^r O _« O (U 03 CO •a cu E" S J3 J3 03 O 2 .5 « a ca OuJ i 22 £ fc! <0 4 "co o "5 I 3 'I 'c D T3 Q . gjj S S -n .3 -— t± ^^ CO fe-Sg -e 3-0 15 « "•sl 1 ^ UT3 '- -^.9 S g|>« -«e •a«— o. c'G c S 9 RELATIONSHIPS BETWEEN HUMMINGBIRDS & FLOWERS 1 J 9 Ericaceae with 6 species (27% of total nectar records), the Rubiaceae with 5 species (16% of total nectar records) and the Bromeliaceae with 3 species (15% of total nectar records) were making the biggest contribution to hummingbird nectar resources during our visit. Only single members of other families were utilized for nectar. Herbaceous plants In the unforested areas two herbaceous ground plants were important and rich sources of nectar: Guzmania cryptanta, a bromeliad associated with boggy areas, and an orchid, Elleanthus smithii, which grew on dry rocky ground. Both had inflorescent spikes with flowers presented in an upright or horizontal position (Table 4), but the Guzmania, with a much longer corolla, produced over ten times more nectar per 24 hours. At the first visit (3 1 July to 3 August) the Guzmania growing at 2300 m was half over but it was in full flower (average of c. 7 flowers per plant) at 2450 m. Here it grew quite densely, a sample area of c 30 x 20 m containing 1 39 plants with flowering spikes. By 1 5 August only 1 7 of these 139 spikes were still flowering, with a total of only 42 flowers. Elleanthus was only just beginning to flower at the first visit but was in full flower at 2200 m between 14 and 21 August. Here it grew extremely densely: an area measuring 1 x 5 m contained 24 flowering spikes and many other equally dense stands were nearby. The two other open country plants were the ground bromeliad Pitcairnia, growing rather sparsely at the edges of bog and forest, and the heath Cavendishia cordifolia which grew as a very low shrub (30-60 cm) and was fruiting rather than flowering during our visit. All the remaining 20 species were associated with the forest except the runner bean Phaseolus coccineus which was a weed in recent forest clearing. Trees The flowering strategies of the three trees were completely different; probably only Huilaea macrocarpa has coevolved with hummingbirds. Posoqueria sp. is a small under-storey tree 6-10 m tall. Its large, sweet-scented white flowers with long, 67 mm corollas are presumably adapted for pollination by night-flying moths with long probosces. When a small tree was cut down at 1340 h in order to examine the flowers, four anthers pressed together blocked the entrance to the tubular corollas but these anthers sprang open when pushed by a pipette. Posoqueria flowers were only visited by hummingbirds early in the morning, the short-billed species obtaining their nectar through pierce holes at the base of the corolla. Clusia grew in patches of secondary forest, reaching a height of 8-12 m. Trees in full flower bear masses of abundantly staminate, cream-coloured flowers, whose corollas are open shallow cups. Individual flowers open early in the morning and are over by evening. On opening, flowers were found to contain an average of c. 20 ul of nectar. Many insects and different species of hummingbirds visited the Clusia flowers, particularly in the early morning. Huilaea was a common under-storey tree growing to a height of 14m. Seven trees examined showed all stages of flowering and fruiting on each tree, and it looked as if the species might have a continuous flowering regime throughout the year or at least for the greater part of the year. The flowers are large, open, pendent red bells; the pedicel is long and the calyx is a hard woody hemisphere. Flowers in situ were too high to be reached; but three obtained by cutting down the limb of a small tree contained 9-5-22 ul (average 16-5 ul) of nectar. One, bagged at 1 1-00 h, contained 15-5 ul of 13% nectar 2 h later. Individuals of Boissoneaua flavescens were territorial at all the Huilaea trees that were watched. No insects or flower-piercers were seen to visit the flowers. The following characteristics of Huilaea indicate that it is adapted for hummingbird pollination: the large pendent red flowers (p. 134) with copious nectar of low sugar concen- tration (p. 137), the woody calyx (protecting the nectaries from nectar-thieves- p. 134), and the long flowering season (which ensures a long-term nectar supply for a reliable, long-lived pollinator). It is noteworthy that Huilaea is a monotypic genus known only from the 1 20 D. W. SNOW & B. K. SNOW Colombian Andes, and belongs to a family (Melastomataceae) whose species are typically insect-pollinated. The Ericaceae The Ericaceae showed examples of different flowering strategies, both for attracting pollin- ators and for avoiding flower-piercers. Within the forest the abundant Disterigma sp., a shrub up to 2 m in height, was flowering plentifully at both visits. Its many small white flowers are evenly scattered all over the densely growing shrub; the flowers are presented horizontally and produce small amounts of concentrated nectar (Table 4). It is much visited by bees, which are presumably its pollinators. Although short-billed hummingbirds, particularly the very small Adelomyia, fed at it, this heath appears not to have evolved away from the hymenopterophilous syndrome. D. albilatera was once seen at it, but only 2 out of 25 flowers examined had pierce holes. Presumably such dispersed, small amounts of nectar, also available to insects, are not an attractive food source for flower-piercers. No insects were seen to visit the two scrambling tree heaths Psammisia falcata and Thibaudia rigidiflora. Both have pendent bright pink or orange flowers with long corollas which bloom in dense ranks along horizontal woody stems or twigs. The corollas are unpro- tected and 71% of 58 flowers of P. falcata and all of 10 T. rigidiflora flowers examined had been pierced at the base. D. albilatera was frequently observed piercing the corollas of both plants. These two heaths appeared to have different flowering seasons, a characteristic of related ornithophilous plants (Stiles 1975, 1978). Thus P. falcata was in full flower at the lowest forest levels (2350 m) during both visits but not flowering at higher levels, while T. rigidiflora was not flowering inside the forest at either visit but at the second visit a few flowers were out at the forest edge at 2350 m and an isolated bush in cleared land at c. 2250 was in full flower. Psammisia was an important source of nectar for C. prunellei and Doryfera, and the former was also seen to feed at Thibaudia. A characteristic of these heaths is the dense packing of stamens, particularly in Psammisia, so that in the latter it requires some pressure to push a micropipette more than 19 mm down the corolla tube. Possibly this serves to exclude all but a few hummingbird species, and so increases the likelihood of out- cross pollination. The three Cavendishia species present their flowers clustered together in groups of from 5 to 13 flowers, with the base of the corollas enclosed by several layers of leaf-like bracts. In C. pubescens the bracts are thick and densely covered with hairs, the clusters are pendent and the corollas long (31 mm). This species was an important nectar source for Doryfera and Coeligena torquata, both long-billed trap-lining hummingbirds. Only one out of 13 flowers examined had a pierce hole in the corolla and no flower-piercers were seen feeding at them. C. guatapeensis had nearly finished flowering at the time of our visit, so observations were few. Its flower clusters are pendent and the corollas short (10 mm); they were not examined for pierce holes. Both long- and short-billed hummingbirds were seen to feed at them, but no insects were noted. C. cordifolia was found only as a low-growing shrub on cleared rough land and on the forest fringes; it was mainly fruiting at the time of our visit and had few flowers. C. cordifolia flowers have shorter bracts than C. pubescens and the corolla tubes are readily pierced by Diglossa albilatera; only a small number were examined but all were pierced. The flowers are orientated between horizontal and pendent, and the corolla tube averages 15 mm in length. At Carare two short-billed hummingbirds were seen to feed at it; at Fonte all the hummingbirds except Ensifera fed at it. While all three Cavendishia species are visited and probably pollinated by hummingbirds, C. pubescens shows the most advanced adaptations for hummingbird pollination, i.e. the highest nectar production per flower, restriction of nectar to long-billed hummingbirds (indicating coevolution with specialist pollinators- see p. 134), and adequate protection of the nectar from flower-piercers. RELATIONSHIPS BETWEEN HUMMINGBIRDS & FLOWERS 121 The Rubiaceae Hummingbirds visited five species of Rubiaceae, one of which (Posoqueria) has already been described. Two species, Manettia aff. sabiceoides and Palicourea cf. vagans, were visited by bees as well as by hummingbirds. Both have features typical of bee-pollinated flowers: white corollas 9-10 mm long presented horizontally, each flower containing a small volume of nectar of relatively high concentration (23-25-5%). The Manettia, which also grew at subtropical altitudes at Togui, has no ornithophilous features. The small white flowers with green calyces are scattered evenly all over the plant. Only two very small hummingbirds were seen feeding at it, Ocreatus underwoodi at Carare and Chlorostilbon gibsoni at Togui. Both were only seen to feed early in the morning and there was good evidence that C. gibsoni stopped feeding when bees became active and began to feed at it (see p. 133). Palicourea cf. vagans shows some ornithophilous characters: the flowers are visually more conspicuous as the calyx is yellow, and they grow in a loose panicle. Spatial concentration of the flowers probably makes them less energetically costly for a hummingbird to exploit. This Palicourea was an extremely common small shrub and seven species of hummingbird were seen to feed at it at various times of day; 82% of nectar records were from species with beak lengths of 18 mm or less. The two remaining species of Rubiaceae, both Palicourea, show many more ornitho- philous characters. Palicourea sp. 43 was a large shrub or small tree growing to a height of 5 m. The flowers are borne in loose panicles, orientated between horizontal and pendent. The tubular corolla is yellow, averaging 154 mm in length, and the calyx is red. The nectar could only be sampled by cutting down a small limb. If the nectar secretion which accumu- lated 2 hours after the limb was cut is typical (Table 4), the quantity per flower is high. Two shrubs formed the feeding territory of a male Heliangelus. During watches totalling 3 h on two mornings it fed almost exclusively at these shrubs and drove off a Diglossa albilatera which attempted to feed at them. Palicourea demissa shows the most advanced ornithophilous characteristics. The purple flowers, borne in groups, have long pendent tubular corollas and abundant nectar of low concentration. The corolla is thick and fleshy, probably for the protection of the nectar from flower-piercers, which were not seen to visit the flowers in 4 h of watching. P. demissa is a woody scrambler and was found flowering as low as 1 m in disturbed forest and at 10 m in undisturbed forest. Many of these characteristics suggest convergence with tree-heaths such as Psammisia and Thibaudia, and in fact they share the same Coeligena pollinators (Table 5). There was good evidence that short-billed hummingbird species such as Adelomyia, Boissoneaua and Heliangelus did not visit P. demissa, nor was it visited by insects. Other plants Of the remaining plants, three climbers and two shrubs, only one appeared to be adapted for hummingbird pollination, the climber Bomarea cf. carderi. This was abundant and in flower at 2350 m, and was also flowering but less abundant at higher levels. The inflorescence is a cluster of pendent flowers with long pedicels. Each flower is tripetalous, with nectaries at the base of each petal. The nectar is abundant and of low concentration. The basal portion of the petal is rolled up and joined to form a long thin tube, which is so narrow that it is impossible to insert a fine pipette (c. 0-6 mm external diameter) into it. The nectar is thus available only to hummingbirds with long thin bills that can hover and exert some upwards thrust, an ability similar to that needed for feeding at the heath Psammisia. Only Coeligena torquata was seen to feed at it. C. prunellei, a thicker-billed bird, is probably unable to reach the nectar as it was frequently seen to fly past and ignore it. Aphelandra sp., a climber with conspicuous orange flowers, was visited by large bees as well as by hummingbirds, and may be primarily bee-pollinated. It has the following characteristics associated with bee rather than hummingbird pollination: flowers pointing more or less upwards; tubular corolla enlarging at the mouth, with a lower landing lip; and small quantities (5-5 ul) of high concentration (25%) nectar. In spite of its long corolla its nectar is available to such short-billed hummingbirds as Adelomyia, Heliangelus and 122 D. W. SNOW & B. K. SNOW Aglaiocercus, probably because the wide end of the corolla enables these species to insert part of their heads into the mouth of the flower. Feeding ecology of the hummingbirds Of the 12 species of hummingbirds recorded at Carare, all but two were resident and seen every day that watches were kept on appropriate flowers. The two exceptions were Acestrura mulsant, which was seen on four occasions between 1 and 3 August but not at all between 14 and 2 1 August, and Aglaiocercus kingi, which was seen on only three of the possible 10^ days of observation. Nine of the 10 regular species can be subdivided into four groups: two small forest species with short bills and wings, Adelomyia melanogenys and Ocreatus underwoodi, most of whose nectar flowers had short corollas and were also visited by insects, mainly Hymenoptera; two behaviourally similar forest species with long wings and short bills, Boissoneaua flavescens and Heliangelus amethysticollis, which combined territorially over concentrated nectar resources with a high percentage of insect-feeding by hawking; three long-billed woodland species, Coeligena prunellei, C. torquata and Doryfera ludoviciae, all of which were trap-liners, largely feeding on specialized hummingbird flowers; and two non- woodland species, Colibri coruscans and Chlorostilbon poortmani, only seen feeding in open areas. The tenth species, Colibri thalassinus, was seldom seen feeding, although a total of seven individuals with tree-top song posts along forest-edges were regularly present and sang throughout the day. The few feeding records were from forest openings or edges. The feeding records for hummingbirds at Carare are summarized in Table 5. Adelomyia melanogenys Ocreatus underwoodi Both species were confined to the forest. Adelomyia was probably the most abundant forest hummingbird; Ocreatus appeared to be considerably less abundant, as indicated by the number of feeding records (Table 5), but the difference may be due in part to the height at which if fed, 4-7 m (mean 3-2) compared with 0-6-7 m (mean 1-4) in Adelomyia, as small hummingbirds feeding high up, especially in the crowns of trees, are difficult to see. Ocreatus was also seen hawking for insects at a height of 1 0 m. Eighty-three per cent of the nectar records for Adelomyia were from flowers at which insects, mostly Hymenoptera, also fed (the probably night-flowering Posoqueria, which it exploited through pierce-holes, is excluded). The seven flowers with tubular corollas that it visited had effective corolla lengths of 8-30 mm, (mean 14-2), and none of them were pen- dent. Adelomyia was not basically territorial but appeared to be trap-lining over consider- able distances, mainly to flowers with small amounts of nectar (e.g. Disterigma) which would be uneconomic for the larger trap-lining hummingbirds to visit, or to plants where the amount of nectar per flower was higher but only a few flowers were blooming per plant (e.g. Palicourea sp. 43). Timed intervals between return visits to the same flower were 1 3 min for Aphelandra and 26 min for Palicourea cf. vagans. The times spent at each flower were gener- ally extremely short. Territorial behaviour (calling while feeding) was noted only from a bird feeding in a small forest clearing at Phaseolus flowers that were also being visited by Chlorostilbon poortmani, and from a bird feeding in the early morning at the pierce holes of Posoqueria flowers in company with conspecifics and Ocreatus. An Adelomyia was repeatedly driven away from profusely flowering Palicourea sp. 43 by a territorial Heliangelus. Three out of the four flower species visited by Ocreatus (omitting Posoqueria, which it exploited through pierce holes) were also visited by insects. Ocreatus visited flowers with short corolla tubes that were not pendent (effective corolla lengths 9-15 mm, mean 11-7) and open cup flowers such as Clusia sp. It was only seen visiting Clusia, where there was much territorial activity between many large hummingbirds, after the early morning peak of feeding activity. The only encounters seen were with conspecifics at Clusia, and when it was chased by Adelomyia at Posoqueria. RELATIONSHIPS BETWEEN HUMMINGBIRDS & FLOWERS 123 Boissoneauaflavescens Heliangelus amethysticollis These two species have short bills in relation to their wing-lengths, especially Boissoneaua whose bill is only 23% of its wing-length (Table 1). As discussed later (p. 136), this is characteristic of territorial species, andiit may also be associated with efficiency in hawking for insects, a method of feeding that was important for both species. Boissoneaua and Heliangelus were to a large extent separated by the height at which they fed. The average height of the 56 nectar-feeding records for Boissoneaua was 8-1 m (range 3-12 m), and that of the 34 records for Heliangelus 2-8 m (range 0-6-6 m). These height differences were mainly due to the fact that the feeding territories of Boissoneaua were based on Huilaea trees, whose flowers were mainly at 6-12 m, whereas one of the two feeding terri- tories of Heliangelus was based on a patch of ground bromeliads (Guzmania cryptanta) and the other on shrubs of Palicourea sp. 43, whose flowers were at 1-5-5 m. The heights at which they hawked for insects differed in the same way, Boissoneaua usually making long sallies of up to several metres from high perches (mean perch height for 96 sallies 8-6 m, range 2-14 m) while Heliangelus made short sallies of 1 m or less from low perches (mean perch height for 20 sallies 2-8 m, range 2-5-3 m). Boissoneaua and Heliangelus have in common two behavioural characteristics associated with territorial defence. One is to hold both wings open and vertical above the back for about half a second after landing. Thus a Boissoneaua, after an encounter with a conspecific, fed at five Huilaea flowers on one tree, perching briefly with wing display after each feed. When no encounter had taken place for the previous 40 min, a Boissoneaua perched four times with wing display during a feeding bout lasting 1 min 9 s, during which 2 1 Huilaea flowers were visited. In a typical instance a male Heliangelus, 8 min after chasing an Adelomyia from its feeding territory, perched briefly with wing display eight times during a feeding bout at a single bush of Palicourea sp. 43 lasting 1 min 39 s. Secondly, both species periodically circle part of their feeding territory in level flight uttering a trilling call. These circular flights were made at heights of 4-6 m by Boissoneaua and about 3 m by Heliangelus. Boissoneaua fed mainly at mid to canopy levels in the forest, including the canopy- flowering Clusias that grew in places along the forest edge. As well as conspecifics, it was seen to drive Heliangelus, Colibri coruscans and Chlorostilbon poortmani from its feeding territories. It failed to drive off a Coeligena prunellei (from Huilaea} but made vigorous attempts to do so. The four flowers with tubular corollas which it visited had effective corolla lengths of 8-1 5 mm (mean 10-7). Three out of the seven flower species which it visited were also visited by insects, but these three accounted for only 29% of nectar records. During our visit the under-storey tree Huilaea macrocarpa was undoubtedly its most important source of nectar (66% of all nectar records). Eight different Huilaea trees found flowering in four different areas were all centres of Boissoneaua feeding territories and except for one record for Coeligena prunellei no other hummingbirds were seen feeding at them in a total of 5 h of watching. Some individual Boissoneaua, apparently with compact territories, continually uttered a repeated sip sip, at an average rate of 30 sips per minute, when in their territories except while they were on the wing nectar-feeding or hawking for insects. One such individual, watched for 1 h on a sunny but rather windy afternoon, called for 91% of the time. For the remainder of the time, except for 1 min 34 s when it was not in view, it made 9 hawking sallies and 10 visits for nectar to a total of 13 Huilaea, 9 Disterigma and 8 Palicourea cf. vagans flowers. Another 1 h afternoon watch on the same individual, when the weather was overcast and still in the first half of the watch and slightly breezy in the second half, gave the following results: first 2 h, 41 insect-hawking sallies and 4 nectar visits to a total of 5 Huilaea flowers, average interval between visits 8-7 min; second ih, 33 insect-hawking sallies and 5 nectar visits to a total of 10 Huilaea flowers, average interval between nectar visits 5-8 min. It was evident that the still overcast weather was the more suitable for insect hawking and an indication that even a slight breeze may decrease the profitability of such an activity. The 124 3 O U U ON OO — o Ci. u -g u a u p o £ WJ X) 00 c 1 E 3 DC — ,03 ound •S. * cryptan p. smithii Guzmani Pitcairnia s Elleanthu •^j- ON VO !D CU a a — c — C C (U w « 2 ,, « J-c « ^ 5- 3 III I l|ll UJ > 9-0^ < U c/o 129 o •c a -§ a U = up P = pe 2, U "s ° u I sac 2 x art of day. . Herbarium pomoea and s^s — U <2 •C— ' O ^•^s iCi^.S i S?o >> ' CS ^ ^, 1 30 D- W. SNOW & B. K. SNOW partially so (all except Castilleja). Castilleja communis differs from the others in having hidden closed green flowers beneath a rosette of red leaves. Of these five native species, which are probably hummingbird-pollinated, only Psammisia cf. pendulijlora was seen to have its corolla slit by nectar-thieving Bananaquits Coereba jlaveola, which at Togui took the place of the flower-piercers (Diglossa spp.). There was evi- dence that such plundering was reduced by hummingbird territoriality. Thus 43% of the 28 flowers examined were slit in two small separate groups of 16 and 22 flowers, which were not defended by a hummingbird; but in a large plant with c. 300 flowers within the territory of an Amazilia tzacatl, none of 22 flowers examined were slit in this way. Table 7-Hummingbird feeding records, Togui P.g. A.n. A.f. A.c. A.t. C.g. Herbaceous plants, ground Castilleja communis 1 Musa sp. (banana) 6 17 11 1 ' Heliconia sp. 1 Canna indica 2 9 1 ' Herbaceous plants, epiphytic Guzmania sp. 5 Climbers Psammisia pendulijlora 2 5 Mandevilla afif. mollissima 2 Ipomoea sp. 2 3 Manettia afif. sabiceoides 10 Shrubs Cavendishia pubescens 6 7 Hamelia patens 3 1 4 14 Trees Calliandra purdiei 5 Eugenia jambos 9 27 38 67 Insect-hawking 19 23 20 34 Insect-gleaning 1 2 1 9 Totals (368) 10 47 81 65 135 30 Hummingbirds abbreviated as follows: A.c., Amazilia cyanifrons; A.f, Amazilia franciae; A.n., Anthracothorax nigricollis; A.t., Amazilia tzacatl; C.g. Chlorostilbon gibsoni; P.g., Phaethornis guy. Note: ' nectar taken 'illegitimately', through pierce holes. Feeding ecology of the hummingbirds Six hummingbird species were resident on the hacienda at the time of our visit (Table 7), of which four (A. tzacatl, A. cyanifrons, C. gibsoni, P. guy) were seen nest-building. Two other species (A. melanogenys, A. mulsant) were each seen once. In spite of the degradation of the indigenous flora and the predominance of introduced plants, the six resident hummingbirds largely occupied different feeding niches. The three Amazilias consisted of two territorial species (A. tzacatl, A. cyanifrons) and a non-territorial trap-liner (A. franciae). The main difference in their feeding ecologies is in RELATIONSHIPS BETWEEN HUMMINGBIRDS & FLOWERS 131 accordance with their relative bill-lengths, as measured by the bill/wing index (for further discussion see p. 136). In weight and wing-length the three species are similar (Table 1), but franciae has a considerably longer bill (bill/wing 45%, cf. 33% in cyanifrons and 37% in tzacatl). A. franciae fed at all levels, from herbaceous plants at O5 m to the tree canopy. At Eugenia trees it was dominated by the two other Amazilia spp., and consequently fed frequently at small, undefended trees with few flowers. The greatest amount of overlap was found in the feeding niches of A. tzacatl and A. cyanifrons, both of which held feeding territories at Eugenia trees. A. cyanifrons, however, tended to keep to higher levels and exploited a more limited range of plant species. Anthracothorax nigricollis was territorial over high insect-hawking perches; a considerably larger bird than the Amazilias, it was able to feed unmolested at the ubiquitous Eugenia but did not defend its nectar. Phaethornis guy was a trap-line feeder on herbaceous plants with long corollas. Chlorostilbon gibsoni was also a trap-liner, feeding at low-growing flowers of herbs, shrubs and vines with short corollas, most of its nectar being shared with insects. Amazilia tzacatl A. tzacatl was probably the most abundant hummingbird on the hacienda. Single birds were observed feeding in 16 well separated areas, all except one of which were centred on Eugenia trees. The exception, which was not occupied until near the end of our visit, was centred on a climber, Psammisia cf. pendulijlora, which had just come into flower. Conspecifics, A. franciae, and Bananaquits were chased from these territories. Although most feeding records were from Eugenia, tzacatl fed on a greater variety of flowers than any other hummingbird, from as low as 0-5 m to tree tops at 1 5 m. It was only once seen attempting to feed on banana flowers, when it went to a pierce hole probably made by a Bananaquit. When feeding at Eugenia, tzacatl frequently (62% of records) uttered a monosyllabic chack between probes, but at small Eugenias with few flowers and after 1630 h such calling was rare. At two heaths over which tzacatl held feeding territories, Psammisia on the hacienda and Cavendishia about 6 km away, it also regularly called between feeding probes. Frequently it gleaned insects from low, thick vegetation, at heights of 0-3 to 3 m (mean 1-9 m), e.g. from moss and the backs of leaves inside thick hedges and bushes, and along the underside of the midribs of banana leaves. It hawked also close to or inside thick vegetation, mainly (27 observations) from perches 4 to 1 1 m high, less often (6 observations) low over herbaceous vegetation from perches 1-2 m high. Hawking sallies were usually within 1-5 m of the perch and were directed at groups of flying insects, the bird often darting about and catching more than one insect per sally. Amazilia cyanifrons Single birds were observed feeding in 12 well separated areas on the hacienda, in nine of which the occupier regularly sang. All were centred on Eugenia trees. When taking nectar in these territories cyanifrons frequently uttered a monosyllabic sip, or sick, between probes. Unlike tzacatl, it visited Eugenia trees standing alone in pastures or along fences, 54% of Eugenia feeding records being at such trees. Except for one record at Hamelia, cyanifrons was not seen to descend from tree canopy level for nectar foraging. During 2 h watching at several Cavendishia pubescens bushes in a tree-shaded ravine 6 km from the hacienda, a tzacatl and two franciae frequently fed at the Cavendishia, but a cyanifrons remained in the trees above. Watches at the one flowering Guzmania plant, when it had 17 to 19 flowers in bloom, showed a cyanifrons silently visiting all flowers at each visit and returning to feed between 1 5 and 27 min later. From its Eugenia territories cyanifrons displaced franciae and attempted, often unsuccess- fully, to displace Bananaquits. At a territory that overlapped the feeding area of a male Anthracothorax nigricollis, the visits of the latter were not disputed. The encounters which presumably established the division of the Eugenia trees between tzacatl and cyanifrons were not witnessed; where the two species had adjoining territories, the occasional silent 132 D. W. SNOW & B. K. SNOW incursions of one bird into the territory of the other always coincided with the temporary absence of the owner. The choice of Eugenia territories by these two Amazilia was probably related to their insect-foraging methods, those of tzacatl requiring thicker vegetation at canopy, shrub and herbaceous levels while cyanifrons foraged for insects in more open places. Between one and two hours after dawn, boundary encounters occurred between two, three and occasionally four cyanifrons, which sang from bare twigs 3-5 m up in an open area and also hawked for insects from the same perches. The insect-hawking recorded during the rest of the day was largely from higher perches (average of 12 records, 8-0 m), nine of them from trees with sparse foliage standing alone in pastures. Hawking sallies were either out from the crown or between the twigs and branches of the crown. There was only one record of cyani- frons gleaning for insects. A mazilia franciae This was the least abundant Amazilia on the hacienda. Only one singing male was located, whose territory, in a coffee and banana patch, a second male attempted to take over. Other males and females were recorded nectar-feeding along the tree-lined river banks and in a second coffee and banana patch. Except for Eugenia, the flowers that franciae visited had tubular corollas from 15 to 35 mm in length, on average longer than those visited by tzacatl (15-19 mm) and cyanifrons (15-18 mm). No defence of a feeding territory was observed; all nectar-feeding was silent, in contrast to the territorial tzacatl and cyanifrons, which frequently called between probes. Most feeding records were from a male which was trapped and marked with a red dye, and was watched (both before and after being trapped) for a total of 18 h. This bird did much insect-hawking (28% of feeding records). Most sallies were short, about 30-60 cm from the perch, a single insect being captured. Perches used for hawking ranged from 1 to 5 m above ground level (average of 23, 2-5 m), and insects were usually caught close to vegetation, sometimes between the twigs of shrubs or the branches of low trees. Beside insect-catching, this male took nectar from Canna, Hamelia, Eugenia and banana flowers within the coffee patch, and also crossed 55m of rough pasture to visit some Eugenia trees which were divided between adjacent territories of a tzacatl and a cyanifrons. Both of the territory-holders were seen to drive off the franciae. Before 5 August the marked male franciae was heard to sing only once in 6 h of obser- vation. Over a 4-day period, 5-8 August, another male franciae contested the possession of the territory with the owner. The encounters occurred almost entirely within an area measuring 26 x 14m and at heights of 2-5 m. During this period, and for some days after the intruder had been ousted, the song output of the marked male was very high, averaging 44 song phrases an hour during the first 3 hours of the day. The intruder also occasionally sang while he was contesting possession. It seemed that the marked male's territorial behaviour was not defence of an insect-foraging area, because a tzacatl occasionally hawked for insects unmolested within 2-3 m of where he was perched. Presumably he was defending a potential mating site. Anthracothorax nigricollis This species was scarce at the time of our visit. In the first half of the period only a single territorial male was located on the fmca; in the second half two more males were present and one or two females. The first located male was watched for a total of 18 h. It occupied a territory measuring c. 55 x 15 m, containing shade trees 10-13 m tall, old citrus trees, banana plants and a rich herbaceous undergrowth. The territory overlapped the territories of a male Amazilia franciae and two A. cyanifrons. This bird fed largely by making frequent hawking sallies from exposed perches, often from the highest perch in the territory (13 m) or in the early morning and late evening from lower perches (mean height 10-2 m). In a typical sally it would fly 2-4 m out from the tree and make a number of darts from side to side before returning to its RELATIONSHIPS BETWEEN HUMMINGBIRDS & FLOWERS 133 perch. Its territorial activities were almost entirely directed to tne protection of the airspace where it hawked. Except for one instance, when it displaced a Phaethornis guy from banana flowers, it was not seen contesting nectar sources. All three species ofAmazilia came to feed at an Eugenia about 9 m from its main hawking perch but were not chased off; once it fed at Eugenia flowers within 1-5 m of the male franc iae without any interaction. The male A. franciae whose territory overlapped that of the male nigricollis also frequently hawked for insects, from perches l-5m high, without interference; but on three occasions when it perched about 6 m up on the top of a half-dead citrus tree it was at once chased from the perch by the male nigricollis. It was ignored on one of its main perches 4-3 m up in the same tree. On eight occasions the male nigricollis chased off conspecific males which perched in his territory or flew over it. A. nigricollis was seen to visit only two plants for nectar, both introduced (Eugenia and banana). Seven intervals between successive visits to the same banana flowers ranged from 19 to 55 min (mean, 37 min). Females were not recorded nectar-feeding, but on two occasions were seen gleaning for insects in thick hedges. A. nigricollis was one of the nine species of hummingbirds whose feeding habits were studied at tropical levels in Trinidad (Snow & Snow, 1972). Insect-foraging, almost entirely by hawking from high perches, accounted for 38% of the Trinidad records, compared with 45% of all records at Togui. In both places this species fed more on insects than any other resident species of hummingbird. Chlorostilbon gibsoni C. gibsoni was the smallest hummingbird species regularly present on the hacienda. All but one of the feeding records were of female-plumaged birds, as were the two birds trapped. It was recorded nectar-feeding low down (0-3^4 m, mean, 1-8 m) in open unshaded pastures and along hedges. It was silent when feeding and did not appear to be territorial. Of the five plant species visited, Ipomoea and Manettia were also visited by bees and their concentrated nectar (Table 6) is typical of bee-pollinated flowers. Three morning watches (0605^0805, 0700-0930 and 0935-1035 h) on a hedge where Manettia (15-20 flowers blooming) and Hamelia (45 flowers blooming) grew side by side, provided evidence that C. gibsoni stopped visiting Manettia when bees started feeding at it, between 0720 and 0755 h. In all gibsoni made 14 visits, returning at intervals ranging from 14 to 34 min (mean, 23 min). On the 8 visits up until 0800 h it visited both plants on 7 occasions and the Manettia only on one occasion; on the 6 visits after 0800 h only the Hamelia was visited, except at 0816 when the Hamelia and one bloom only of the Manettia were visited. Bees were not seen to feed at Hamelia but very occasionally butterflies did so. Timed feedings showed that C. gibsoni was able to exploit Manettia flowers more rapidly than Hamelia flowers; the time spent probing a flower (about 0-4 s) was almost the same for both plants, but as less time was taken in manoeuvring between Manettia flowers the rate of visiting them (one flower per 0-7 s) was higher than the rate for Hamelia flowers (one per 1-0 s). During these observations Amazilia tzacatl was often in view foraging for insects, and fed at Hamelia and at Ipomoea growing near by in the same hedge, but made no visit to Manettia. Phaethornis guy There were extremely few of this species on the hacienda, apparently only one or two females, and the two plants at which they were recorded feeding (Musa and Canna) were both introduced. A clump ofHeliconia sp. in a ravine 5 km from Togui was also fed at by P. guy. These three herbaceous plants all have long (25-35 mm) curved tubular corollas, matching the bird's long curved bill. The two nests found were on banana leaves and no other plants with suitable leaves were seen on the hacienda. P. guy is essentially a forest hummingbird, specialized both in its feeding and in its nest-site (Snow, 1974). Its survival in this intensively farmed area is heavily dependent on the banana, which is planted to provide shade in young coffee plantations, and may be retained or cleared away after the main shade trees have grown. 1 34 D. W. SNOW & B. K. SNOW Discussion Faegri & Fiji (1966) list a number of morphological features of flowers that characterize the 'syndrome of ornithophily'. Stiles (1979) has extended them to include the phenology, nectar production, spacing on the plant, and other features of floral biology that may impinge on a plant's avian pollinators and thus be subject to coevolutionary selection. On the bird's side, a complex cf characters, structural, physiological and behavioural, are related to nectar feeding and are presumed to have evolved in the course of development of the mutual relationship between bird and flower. An extensive and specialized literature has resulted from recent research into nectarivory in birds, with hummingbirds the main sub- jects. Emphasis has been especially on energetic aspects of the interaction, and the different feeding strategies that have resulted therefrom. It would be inappropriate to attempt here a thorough review of a fast-growing subject, and we confine the discussion to those aspects of the subject on which our observations provide new data or suggest modification of previous opinions. The syndrome of ornithophily Faegri & Fiji considered that the toughness that is characteristic of bird flowers is a protective adaptation against damage by the beaks of their avian pollinators, and this opinion seems to have been generally accepted. So far as hummingbird flowers are concerned, however, we think that the protective layers at the base of the flower are generally an adaptation pro- tecting the nectar from nectar thieves and making it available only to legitimate visitors. The hard woody calyx of the pendent flowers of Huilaea (p. 1 19), very different from the rather delicate, insect-pollinated flowers of other melastomes, is a good example. Within the Ericaceae, the flowers of Cavendishia pubescens seemed to be effectively protected against nectar thieves by the bracts round the base (p. 1 20). There seems to be no good evidence that the delicate and exact probing of the bill of a hovering hummingbird damages even soft- tissued flowers. Protection of a flower's nectar against thieves need not be absolute in order to be effective. At Fonte we had only a single record of the flower-piercer Diglossa albilatera piercing the flowers of Cavendishia cordifolia, which was very common in places where the flower- piercer was also abundant. The flower of C. cordifolia is moderately protected at its base, but is not immune to exploitation by D. albilatera (p. 120). It seemed likely that this flower- piercer was generally avoiding Cavendishia flowers at Fonte because they could exploit much more efficiently the unprotected flowers of another abundant plant, Palicourea angustifolia. The orientation of the corolla tube is often a good indicator of the degree to which the flower is adapted for hummingbird pollination. This is well illustrated by the three species of Palicourea at Carare: in P. cf. vagans, the least specialized (corolla 9 mm, white; calyx yellow), the corolla tube is horizontal; in P. sp. 43 (corolla 15mm, yellow; calyx red) between horizontal and pendent; and in P. demissa, the most specialized (corolla 31 mm, purple), pendent. Similarly Huilaea differs from other melastomes in having pendent flowers, and the specialized tree-heaths, e.g. Psammisia falcata and Thibaudia rigidiflora, from the unspecialized Disterigma sp. (Table 4). The sugar concentration of the nectar, another important aspect of the syndrome of ornithophily, is discussed later in a separate section. Specialized coevolutionary relationships Just as a flower may be specialized to a greater or lesser degree for pollination by particular kinds of hummingbirds, so different hummingbird species are specialized to varying degrees for feeding at particular kinds of flower. As the most advanced mutual specialization one might imagine a flower adapted for pollination by only one species of hummingbird, which in turn might be dependent on the nectar of that one kind of flower. In fact it seems that, while very specialized flowers may indeed depend on a single hummingbird pollinator, even the most specialized hummingbirds usually feed from a number of different kinds of flower. RELATIONSHIPS BETWEEN HUMMINGBIRDS & FLOWERS 1 35 The most striking of the revolutionary adaptations involving the hummingbirds that we observed is that between Ensifera and Passiflora mixta. Other species of Passiflora, e.g. P. vitifolia (Skutch, 1964) and P. longiracemosa (Snow, 1973), have flowers with long tubular corollas that are fed at, and probably pollinated, by long-billed hummingbirds, but the present case represents the extreme. It can hardly be doubted that the flowers of P. mixta evolved to their present length in step with the bill of Ensifera. The range of P. mixta, which has its centre of abundance at temperate levels in the Andes of Colombia and Ecuador, cor- responds broadly with the range of Ensifera (Andes of north-western Venezuela south through Colombia, Ecuador and Peru to northern Bolivia); but P. mixta now occurs also in a few scattered localities in the coastal range of northern Venezuela beyond the range of Ensifera. This is, however, almost certainly due to introduction by Amerindians, who have long made use of P. mixta's highly edible fruit (S. Tillett, pers. comm.). We recorded Ensifera feeding at no other kind of flower; indeed there were no plants in flower at the time of our visit approaching Passiflora mixta in length of corolla. Ensifera is, however, generally said to be adapted to feeding at Datura flowers. It certainly feeds at Datura flowers, but it may be doubted whether this is a case of coevolutionary adaptation. Datura flowers have few ornithophilous features; they are widely open at the mouth, so that shorter-billed hummingbirds than Ensifera can feed at them by getting their heads well inside the corolla tube (pers. obs.), the anthers and stigma are not exserted so as to come into contact with Ensifera's forehead (as they are in P. mixta), and the corolla is thin-walled and easily pierced. We saw no Datura growing wild in the area where we worked, but they were common ornamental plants round houses. Datura is a small genus of only 10 species, occurring in tropical America but widespread elsewhere in tropical and warm temperature areas. Its pollination has not been thoroughly studied. D. meteloides is visited by a hawk moth which enters the corolla tube (Baker, 1961), and the open vase shape of the flowers suggests that they are adapted to facilitate the bodily entry of pollinators rather than to restrict their nectar to pollinators with very long bills or probosces. The white colour of most Datura flowers suggests adaptation to a crepuscular pollinator. Two species have red flowers, a colour typical of hummingbird flowers but also found in some moth-pollinated flowers (Baker, loc. cit.). We provisionally conclude that Ensifera is adapted primarily to Passi/lora mixta and other species of Passiflora with very long corolla tubes, and that its habit of feeding at Datura flowers is a result of man's alteration of its habitat. But it would be desirable to study Ensifera's feeding habits in an area, if there be any, where a red-flowered Datura grows wild. Also needing further study is the effect of exploitation of the nectar of Passiflora mixta by flower-piercers and the hummingbirds that make use of the holes made by flower-piercers. In the Fonte study area nectar-thieving reduced the amount of nectar available to Ensifera to such an extent that it must surely have affected Ensifera'?, abundance. Yet there is every reason to suppose that these or similar nectar thieves have been present for as long as Ensifera and P. mixta have existed. Our experimental protection of Passiflora flowers showed that Ensifera could obtain all the nectar that it could ingest from one or two flowers of the dozen or so available on a plant. Had the experiments continued, it would have been of great interest to have seen whether individual Ensiferas would have switched from trap- lining, over what must have been considerable distances, to territoriality at single Passiflora clumps. The effect of territoriality, from the plant's point of view, would be greatly to increase the incidence of self-pollination. Trap-lining by Ensifera, on the other hand, must promote out-crossing (Linhart, 1973). Hence it might be advantageous for the plant not to evolve increased protection of the nectar from nectar thieves, but to maintain its pollinator at low population densities; but such a course would be hazardous, since it would carry the risk of occasional local disappearance of the pollinator. Lafresnaya, the only temperate-level Andean hummingbird with a long curved bill, was recorded feeding mainly at three kinds of flowers with long, curved corolla tubes that fitted its bill closely (Pentadenia strigosa, Castilleja fissifblia, Siphocampylus bogotensis). None of these flowers was seen to be fed at 'legitimately' by any other hummingbird species. It is 136 D. W. SNOW & B. K. SNOW reasonable to conclude that Lafresnaya is the main pollinator of these three plants, and that their flower structure coevolved with Lafresnaya^ bill. The situation is not simple, how- ever: Lafresnaya males and females have distinctly different bill-lengths, that of the male being shorter although the male's wing is longer fhan the female's. This sex difference suggests that the male is likely to be more territorial than the female, and also to be less specialized in its feeding habits (see next section). Our limited observations did in fact suggest that Lafresnaya females were trap-liners and that some, at least, of the males held feeding territories. The only other specialized coadaptations between particular hummingbird species and flowers suggested by our observations were those between the three hummingbirds with long straight bills at Carare, Coeligena prunellei, C. torquata and Doryfera ludoviciae, and some of the flowers at which they fed. C. prunellei and Doryfera were the only two hummingbirds seen feeding legitimately on the heath Psammisia falcata, whose long corolla tube is filled with densely packed stamens, necessitating a considerable upward thrust by a hovering hummingbird in order to reach the nectar; and C. torquata (a smaller species, subordinate to C. prunellei, but with an even longer and markedly finer bill) appeared to be the only species capable of exploiting the specialized and very narrow-tubed flowers ofBomarea cf. carderi. Trap-lining vs. territoriality The broad distinction between trap-lining and territoriality as alternative foraging strategies for hummingbirds has been mentioned earlier. Feinsinger & Colwell (1978) have recently made a more refined classification, defining four community roles available to humming- birds. They distinguish two kinds of trap-liners. 'High-reward' trap-liners are species with either particularly long or curved bills, which have coevolved with certain plant species that offer high nectar rewards per flower, the nectar being inaccessible to other hummingbirds. These are the species usually referred to as trap-liners. 'Low-reward' trap-liners are small, shorter-billed species, which are excluded from clumped flowers by territorialists and instead visit dispersed, unspecialized flowers. Hummingbirds in this second category are not found in all communities. Territorialists form a third category, and the fourth consists of 'territory-parasites', which may either be large birds with moderate-length beaks that can forage with impunity among flowers defended by smaller territorialists, or very small, short- billed species which can infiltrate into territories, often feeding in parts of the territory not much used by the owners. In addition to beak characteristics and body size, wing-disc loading has been shown to be important in fitting a hummingbird for a particular foraging strategy (Feinsinger & Chaplin, 1975). Species with high wing-disc loading (high weight in relation to wing-length) are more manoeuvrable, have greater acceleration, but are less efficient at hovering than species with low wing-disc loading. The former tend to be terri- torialists (being better in aggressive encounters and having less need for sustained flight), and the latter trap-liners. The hummingbirds in our study areas fitted into Feinsinger & Colwell's four categories well; but on the basis of available weights and measurements there is no clear relationship between foraging strategies and wing-disc loading. Thus Ensifera, the most pronounced trap- liner, has the highest wing-disc loading (-0489) and Metallura, a territorialist, the lowest (-0257). Coeligena torquata, a typical trap-liner, and C. prunellei, a more territorial species, both have wing-disc loadings of -028 1 . It is likely that the lack of relationship is due in part to the small samples of weights available to us, some of which are from other areas. We draw attention instead to the ratio between bill-length and wing-length, given as a percentage in Table 1. This proves to be a good indicator of feeding strategy, and has the advantage that it can be consistently derived from museum specimens. There were 10 species with bill/wing percentages of 38 or over (Table 1), all but two of which were high-reward trap-liners. The two with the highest percentages (Ensifera and Phaethornis guy) show the highest degree of feeding specialization. The exceptions were Acestrura mulsant, which although apparently a trap-liner was not seen to visit any flowers to which it had exclusive or nearly exclusive access. Observations were, however, few. The RELATIONSHIPS BETWEEN HUMMINGBIRDS & FLOWERS 1 37 other exception, Chlorostilbon poortmani, was a low-reward trap-liner which exploited the less specialized hummingbird flowers and some bee-pollinated flowers. Eight of the 12 hummingbirds with bill/wing percentages between 20 and 32 were terri- torial over high-density, morphologically less specialized ornithophilous flowers (e.g. Guzmania cryptanta, Palicourea sp. 43) or over high-density flowers also available to insects (e.g. Clusia, Eugenia). They also frequently hawked for insects. Three forest species (Adelomyia, Ocreatus, Aglaiocercus), however, were not seen to be territorial. Excluding pierce-hole feeding, 80% of nectar-feeding records for these three species were at flowers also visited by insects, mostly large bees. Some large bees are trap-line feeders at dispersed specialized flowers, and possibly these hummingbirds are adapted in some way to share this nectar source. There were insufficient observations to determine the feeding strategy of Colibri thalassinus, but in Costa Rica Wolf et al. (1976) found that it defended a specialized hummingbird flower, mostly from conspecifics. The three Amazilias at Togui, the only three congeneric species resident in the same locality, have bill/wing percentages closely related to their observed feeding strategies: A. franciae, a trap-liner -45-2; A. cyanifrons, almost entirely territorial over one concentrated nectar source (Eugenia) - 32-6; and A. tzacatl, territorial over the same nectar source, also over hummingbird flowers with moderate-length corollas, and in addition exploiting a variety of more dispersed plants- 37-1. The two Coeligena species at Carare showed slight niche differences which can be related to bill/wing percentages, prunellei (38-3) showing some territorial tendencies and torquata (40-8) being a trap-liner with possibly exclusive access to one plant species. Nectar characteristics The sugar concentration of the nectar of many hummingbird-pollinated flowers is relatively low, averaging about 20% (weight/total weight of solution) (Baker, 1975). Bee-pollinated flowers have more concentrated nectar. The greater dilution of hummingbird nectar is at first sight puzzling, since it has been shown experimentally that, given a choice, hummingbirds prefer the most concentrated solution (Hainsworth & Wolf, 1976). Bolten & Feinsinger (1978) have argued that the dilute nectar typical of hummingbird flowers may have evolved to deter bees, which need more concentrated nectar, and that once a flower has evolved adaptations to exclude bees, selective pressure from feeding hummingbirds may lead to a return to more concentrated nectar. They present evidence from Trinidad in support of this hypothesis. The flowers whose nectar concentrations are given in Tables 2, 4 and 6 can be divided into three groups: flowers visited by bees as well as hummingbirds (Aphelandra, Disterigma, Ipomoea, Manettia, Palicourea cf. vagans, Phaseolus); the less specialized hummingbird flowers, with corolla lengths of 1 1-20 mm (10 species); and the more specialized humming- bird flowers, with corolla lengths of 21-1 14 mm (9 species). The ranges and means of nectar concentrations for these groups were, respectively, 23-38, mean 29-2%; 14-21, mean 16-8%; and 14-27, mean 19-5%. These figures are in good agreement with Bolten & Feinsinger's hypothesis. It may be noted that two of the morphologically most specialized species, Passiflora mixta (25-5%) and Pentadenia strigosa (27-0%), had the highest concentrations among the hummingbird flowers. Ornithophily in the Andean flora Our data are far too incomplete for any survey of the incidence of ornithophily among Andean plants, and the literature on the subject is very limited; but a few points deserve mention. It seems clear that three plant families, the Rubiaceae, Ericaceae and Bromeliaceae, are of prime importance, the first mainly for short-billed, unspecialized hummingbirds, the second and third both for unspecialized and for some of the more specialized hummingbirds with long straight bills. In the richest of our three study areas these three families accounted for nearly two-thirds (14 out of 22) of the plant species providing nectar at the time of our visit; 1 38 D. W. SNOW & B. K. SNOW in the two other areas, with poorer vegetation modified by agriculture, the proportions were lower (6 out of 15, and 4 out of 13). Members of the Rubiaceae and Ericaceae provide examples of different degrees of specialization for ornithophily, from such unspecialized flowers as Palicourea cf. vagans (Rubiaceae) and Disterigma spp. (Ericaceae), which have white flowers with short corolla tubes that were visited by insects as well as short-billed hummingbirds, to specialized flowers such as Palicourea demissa (Rubiaceae) and Psammisia falcata (Ericaceae), which were visited only by two long-billed species. The bromeliads as a family seem to be adapted for pollination by hummingbirds (McWilliams, 1974). To what extent pollination by hummingbirds has affected speciation and interspecific differences in phenology within these families is a matter for future research. Palicourea is a large and taxonomically difficult genus, with about 200 known species (Willis, 1973). Specialization of the flower for pollination by different hummingbird species or groups of species may have led to divergence between taxa (in geographical isolation ?), resulting in reproductive isolation when they later became sympatric. Competition for pollinators between morphologically similar sympatric species may have led to the staggering of flowering seasons, as appears to have happened in Costa Rican Heliconia species (Stiles, 1975). It is noteworthy that bromeliads in particular tend to have very restricted flowering seasons, so that observations conducted over a short period give a very incomplete picture of their importance for hummingbirds. The four species that were in flower in our three study areas at the time of our visit were only a fraction of the species present. The remaining hummingbird flowers belonged to 18 different families (omitting intro- duced plants), with only one or two species in each. Most of these families are well known to contain ornithophilous genera, the most interesting of the exceptions being the Melastoma- taceae. This very large family, with numerous neotropical species, is typically insect- pollinated; the flowers are open or have only short corolla tubes. The genus Huilaea, with a single species, is known only from a few temperate forest localities in Colombia. Its flowers are strikingly different from typical melastome flowers, showing several ornithophilous features, and it is reasonable to suppose that it evolved in association with forest humming- birds. Whether it is generally associated with Boissoneauaflavescens, which held territories at the Huilaea trees at Carare, must await further study. Acknowledgements We are most grateful to Dr Polidoro Pinto E., Director of the Instituto de Ciencias Naturales (I.C.N.), Universidad Nacional de Colombia, and to Dr Pedro M. Ruiz C, Head of the Zoological Section, I.C.N., for invaluable help in making arrangements for our expedition, and to the latter especially for putting his hacienda near Togui at our disposal; to Sr Enrique Zerda O. for help with the field work and in many other ways throughout the expedition; and to Sr Pablo Bernal, preparator at I.C.N., for much practical help with our journeys to and from our study areas. Dr Gustavo Lozano C. of the Botanical Section, I.C.N., kindly identi- fied most of our plant specimens. Dr James L. Luteyn of the New York Botanical Garden helped with identifications of the Ericaceae. We also acknowledge with thanks a grant from the Frank M. Chapman Memorial Fund of the American Museum of Natural History towards the travelling expenses of one of us (B.K.S.). References Baker, H. G. 1961. The adaptation of flowering plants to nocturnal and crepuscular pollinators. Q. Rev. Biol. 36: 64-73. 1975. Sugar concentrations in nectars from hummingbird flowers. Biotropica 7: 37-4 1 . Bolten, A. B. & Feinsinger, P. 1978. Why do hummingbird flowers secrete dilute nectar? Biotropica 10: 307-309. Carpenter, F. L. 1976. Ecology and evolution of an Andean hummingbird lOreotrochitus estella). Univ. Calif. Publs Zool. 106: 1-74. RELATIONSHIPS BETWEEN HUMMINGBIRDS & FLOWERS 1 39 Faegri, K. & van der Fiji, L. 1966. The principles of pollination ecology. Pergamon, Oxford. 248 pp. Feinsinger, P. & Chaplin, S. B. 1975. On the relationship between the wing disc loading and foraging strategy in hummingbirds. Am. Nat. 109: 2 1 7-224. & Colwell, R. K. 1978. Community organization among neotropical nectar-feeding birds. Am. Zool. 18: 779-795. -, Terborgh, J. & Chaplin, S. B. 1979. Elevation and the morphology, flight energetics, and foraging ecology of tropical hummingbirds. Am. Nat. 113: 481-497. Greenewalt, C. H. 1962. Dimensional relationships for flying animals. Smithson. misc. Collns 144 (2): 1-46. Hainsworth, F. R. & Wolf, L. L. 1976. Nectar characteristics and food selection by hummingbirds. Oecologia25: 101-113. Linhart, Y. B. 1973. Ecological and behavioral determinants of pollen dispersal in hummingbird- pollinated Heliconia. Am. Nat. 107: 5 1 1-523. McWilliams, E. L. 1974. Evolutionary ecology. Pp 40-64, in Flora Neotropica, Monograph no. 14, Pitcairnia (Bromeliaceae). Hafner Press, New York. Meyer de Schauensee, R. 1964. The birds of Colombia. Livingston, Narberth, Pa. xvi + 430 pp. Miller, A. H. 1963. Seasonal activity and ecology of the avifauna of an American equatorial cloud forest. Univ. Calif. Publs Zool. 66: 1-78. Skutch, A. F. 1964. Life histories of hermit hummingbirds. Auk 81: 5-25. Snow, B. K. 1973. The behaviour and ecology of hermit hummingbirds in the Kanaku Mountains, Guyana. Wilson Bull. 85: 163-177. 1974. Lek behaviour and breeding of Guy's Hermit Hummingbird Phaethornis guy. Ibis 116: 278-297. & Snow, D. W. 1972. Feeding niches of hummingbirds in a Trinidad valley. J. anim. Ecol. 41: 471-485. Stiles, F. G. 1975. Ecology, flowering phenology and hummingbird pollination of some Costa Rican Heliconia species. Ecology, Brooklyn 56: 285-30 1 . 1978. Temporal organization of flowering among the hummingbird foodplants of a tropical wet forest. Biotropica 10: 194-210. 1980. Ecological and evolutionary aspects of bird-flower coadaptations. Ada XVII Int. Orn. Congr. (in press). Willis, J. C. 1973. A dictionarv of the flowering plants and ferns. Cambridge Univ. Press. xxii + 1245pp. Wolf, L. L., Stiles, F. G. & Hainsworth, F. R. 1976. Ecological organization of a tropical, highland hummingbird community. J. anim. Ecol. 45: 349-379. Manuscript accepted for publication 22 July 1979 British Museum (Natural History) Monographs & Handbooks Feeding and the Feeding Apparatus in Waders A study of anatomy and adaptation in the Charadrii P. J. K. Burton 1974, 150pp, 41 line plates, 4to boards, £6.50 The Museum publishes some 10-12 new titles each year on subjects including zoology, botany, palaeontology and mineralogy. Besides being important reference works, many, particularly among the handbooks, are useful for courses and students' background reading. Lists are available free on request to : Publications Sales British Museum (Natural History) Cromwell Road London SW7 5BD Standing orders placed by educational institutions earn a discount of 10% off our published price. Titles to be published in Volume 38 Miscellanea Relationships between hummingbirds and flowers in the Andes of Columbia. By David W. Snow & Barbara K. Snow Miscellanea Mites of the subfamily Parasitinae (Mesostigmata: Parasitidae) in the British Isles. By K. H. Hyatt Miscellanea Printed by Henry Ling Ltd, Dorchester Gl Bulletin of the y British Museum (Natural Histdiyf Miscellanea Zoology series Vol 38 No 3 29 May 1980 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum's resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself, available separately, and individually priced. Volumes contain about 300 pages and are not necessarily completed within one calendar year. Subscriptions may be placed for one or more series Subscriptions vary according to the contents of the Volume and are based on a forecast list of titles As each Volume nears completion, subscribers are informed of the cost ol 1 next Volume and invited to renew their subscriptions. Orders and enquiries should be sent to: Publications Sales, British Museum (Natural History), Cromwell Road, London SW7 5BD, England. World List abbreviation: Bull. Br. Mus. nat. Hist. (Zool.) © Trustees of the British Museum (Natural History), 1980 ISSN 0007-1498 r 171 Vol 38 No 3 pp 141-173 British Museum (Natural History) Cromwell Road ,-,*. mo A London SW7 5BD Issued 29 May 1980 Miscellanea 9 JUN 1980 LIBRARY <£ , xOK '»AL Contents A new species of small Barbus (Pisces, Cyprinidae) from Tanzania, East Africa. ByR. G. Bailey A new species of Barbus (Pisces Cyprinidae) from Africa. By K. E. Banister . A new blind cyprinid fish from Iraq. By K. E. Banister & M. K. Bunni A new species of cichlid fish from the Malagarasi swamps and river Tanzania, E. Africa). By P. H. Greenwood A new catfish from Sierra Leone. By G. J. Howes A new genus of cheline cyprinid fishes. By G. J. Howes 141 145 151 159 165 171 A new species of small Barbus (Pisces, Cyprinidae) from Tanzania, East Africa. Roland G. Bailey Department of Zoology, Chelsea College, University of London, Hortensia Road, London SW100QX A collection of Barbus made by the author in northern Tanzania belongs to the group characterized by radiately striated scales and a thin, flexible last unbranched dorsal fin ray. It is described here as a new species. Barbus venustus sp.nov. (Fig. 1) HOLOTYPE. A male fish, 29-0 mm SL from Nyumba ya Mungu reservoir situated at 3°45'S, 37°25'E at an altitude of 670 m on the upper Pangani river, Tanzania, BM (NH) 1979.10.3:1. PARATYPES. Nineteen fishes from Nyumba ya Mungu reservoir and its affluent streams, BM(NH) 1979. 10.3:2-20. DESCRIPTION. Based on 20 fishes, 19-6-30-2 mm SL. Body moderately compressed, its depth equal to or, in mature females, a little greater than the length of the head. Predorsal profile convex with a slight nuchal hump. Snout rounded, shorter than the eye diameter which is slightly less than or equal to the interorbital width. Mouth moderate and subterminal. Anterior barbel short, posterior barbel longer, extending as far as the vertical to the mid-point of the pupil of the eye. Well developed sunken pit-lines are present on the cheek and operculum and, less distinctly, on the dorsal surface of the head. There may be as many as 12 lines between the preoperculum and the anterior orbital margin, some branching ventrally. In any line the pits are separated from each other and their mouths may rise above the level of the surrounding skin. Sharply pointed conical tubercles or spinules are developed on the snout and lower jaw of males. The caudal peduncle is relatively slender, its depth 1 -36-1 -76 (mean 1 -56) in its length. In alizarin stained preparations of two fishes, 26-0 and 27-8 mm SL, the post Weberian vertebral count is 27. The pharyngeal bones and teeth (Fig. 2) are similar in appearance in both specimens. The teeth number 2.3.5-5.3.2 and have recurved crowns. Fig. 1 A paratype of Barbus venustus, 27-6 mm SL. Bull. Br. Mus. nal. Hist. (Zool.)38 (3) : 141-144 Issued 29 May 1980 141 142 R. G. BAILEY Summary of morphometric data for 20 fishes, 19-6-30-2 mm SL; measurements are expressed as percentages of SL. range mean standard deviation Maximum depth 28-4-32-4 30-4 1-1 Length of head 27-9-30-4 29-0 0-7 Length of snout 6-3-8-3 7-4 0-6 Horizontal diameter of eye 9-2-11-0 10-2 0-5 Least bony interorbital width 8-3-10-2 9-3 0-5 Length of anterior barbel 0-9-3-0 1-9 0-7 Length of posterior barbel 3-3-5-9 4-3 0-7 Length of longest dorsal fin ray 24-8-28-2 26-9 1-0 Greatest length of pectoral fin 18-8-22-8 20-9 1-1 Distance from snout tip to dorsal fin origin 52-4-56-7 54-2 1-2 Length of caudal peduncle 18-2-21-8 20-1 1-0 Least depth of caudal peduncle 11-9-13-8 12-9 0-6 SCALES. Scales have radiate striae, (Fig. 3). Lateral line complete, with 22 (f. 1), 23 (f.7), 24 (f.6) or 25 (f.6) pore-bearing scales; its course dips ventrally on the flanks and returns to a midlateral position on the caudal peduncle. There are 4i (f.14) or 5 (f.6) scales between the dorsal fin origin and the lateral line and \\ (f.4) or 2 (f. 16) scales between the lateral line and pelvic insertion. Eight (f.3), 9 (f. 1 1) or 10 (f.6) scales encircle the least circumference of the caudal peduncle and 9 (f. 1 5) or 10 (f.5) are found in the pre-dorsal fin row. Fig. 2 A lateral view of the right pharyngeal bone from a fish 27-8 mm SL. Fig. 3 Striations on a flank scale from above the lateral line. FINS. The origin of the dorsal fin is on or slightly in advance of the perpendicular to the pelvic fin insertion. It has HI/7 (f.7) or HI/8 (f. 13) rays, and the third unbranched ray is NEW SPECIES OF SMALL BARBUS 143 unossified and flexible. The anal fin has III/5 rays. In both fins the last branched ray is typically divided to its base. The pectoral fin does not reach the base of the pelvic fin. COLORATION. In the preserved fishes, fixed in formalin, the dorsal surface of the head and body are brownish. Scales on the upper flanks and caudal peduncle are outlined in a reticulum of tiny spots and those of the lateral line series are intensely pigmented above and below the pores. A dark midlateral stripe of varying intensity runs from the tip of the snout to the base of the caudal fin. There is a small, rounded black spot at the base of the caudal fin and patches of pigment about the origin of the dorsal and, more distinctly, on the base of the anal fin. In life, the upper flanks, caudal peduncle and caudal fin are orange-red in colour. ETYMOLOGY. The specific name is derived from the latin for pretty, and alludes to the attractive appearance of this small fish in life. DIAGNOSIS AND AFFINITIES. Barbus venustus belongs to the group of African Barbus characterized by radiately striated scales and a thin and flexible last unbranched dorsal fin ray. It is separable from 12 other species in this group recorded in East Africa by its possession of a combination of the following features:- small adult size (20-30 mm SL), well developed pit-lines on the head, two pairs of short barbels, relatively large scales (22-25 in the lateral line, 8-10 around the caudal peduncle), a complete series of lateral line pore scales, and a dark midlateral stripe extending from the snout tip to the base of the caudal fin where it terminates in a prominent spot. Cephalic pit-lines have been reported in three other species present in East Africa. Two of these, Barbus radiatus Peters (a widely distributed polytypic species, Stewart 1977) and B. profundus Greenwood (Lake Victoria), are placed in the subgenus Enteromius (formerly Beirabarbus, see Greenwood 1970) on the basis of the form and arrangement of the pits. In this subgenus they are small, numerous and tightly packed in raised lines or ridges, (Greenwood, 1962). The pit-lines in B. venustus are well developed but from a comparison with those in B. radiatus from Tanzanian localities, they clearly differ from the Enteromius type being sunken, with rows of relatively large, discrete pits. This condition conforms more closely to that in B. cercops Whitehead (Lake Victoria basin), except that only 5 or 6 lines are visible on the cheek in this species compared with up to 12 in B, venustus. Barbus cercops also differs from the new species in coloration and the possession of two pairs of long barbels (equal to or longer than the eye) and smaller scales (29-32 in the lateral line series; 12 around the caudal peduncle). Whitehead (1960) considered that the nearest relatives of B. cercops were to be found in western Africa and Greenwood (1962) suggests the possibility of a phyletic significance in the fact that, of 14 species of small Barbus he examined with pit-lines of the non-Enteromius type, all except B. cercops were essentially West African in distribution. Subsequently Hopson A. J. & J. (1965) added a further 11 species with weakly developed pit-lines to Greenwood's list: all are from the Volta region. However, Poll & Lambert (1961), Poll & Gosse (1963) and Poll (1976) have found small Barbus with pit-lines of this sort in localities in the north-east, central and south-eastern parts of the Zaire basin. Moreover Tweddle collecting in Malawi (personal communication) has noted pit-lines in several as yet unidentified Barbus species and I have observed 5 or 6 weak pit-lines on the cheek in type material of B. barnardi Jubb, (AM.P1055, 1056-1062) from the Zambezi river system. It is possible therefore that pit-lines have been overlooked in small Barbus from other parts of Africa. As Greenwood (1962) comments, preservation undoubtedly affects the ease with which pits and pit-lines may be seen. Within the group defined above, the affinities of B. venustus are unknown. Although clearly separable in body shape, the depth of the lateral line scales and coloration, the 144 R. G. BAILEY distribution of the pit-lines in B. schoutedeni Poll & Lambert (River Dungu, Zaire) appears to be very close to their condition in B. venustus. Barbus pygmaeus Poll & Gosse (Central Zaire basin) shares with B. venustus a small adult size (<20 mm SL), large scales and a midlateral stripe, but there are only 4 pit-lines beneath the eye and it has a deeper body, longer barbels (about equal to or longer than the eye) and higher dorsal and anal fin ray counts (III/9 and III/6, respectively). DISTRIBUTION. Barbus venustus is known only from the type localities. BIOLOGY. Barbus venustus in the reservoir, was caught in shallow water among open stands of the emergent grass Paspalidium geminatum. The sample was obtained in July and August at the onset of the dry season but with the lake level still close to its maximum. Cladocerans, insect fragments (including mayfly nymphs and chironomid midge larvae), plant material and fine sand were recorded in the guts of dissected fish. Most specimens of both sexes in the sample were sexually mature. The number of eggs, <0-8 mm diameter, carried by two ripe females, (27-6 and 30-2 mm SL) were 190 and 1 70, respectively. Acknowledgement I am grateful to Dr Keith Banister for his comments on the draft of this paper. References Greenwood, P. H. 1962. A revision of certain Barbus (Pisces, Cyprinidae) from East, Central and South Africa. Bull. Br. Mm. nat. Hist. (Zool.) 8 : 153-208 — 1970. A revision of the cyprinid species Barbus (Enteromius) radiatus Peters, 1853, with a note on the synonymy of the subgenera Beirabarbus and Enteromius. Revue Zool. Bot. afr. 82 : 1-13. Hopson, A. J. & J. 1965. Barbus (Pisces, Cyprinidae) of the Volta region. Bull. Br. Mus. nat Hist. (Zool.) 13: 101-149. Poll, M. 1976. Poissons. Explor. Pare. natn. Upemba Miss. G. F. de Witte( 1946-49). 73 : 1-127. — & Gosse J. P. 1963. Contribution a 1'etude systematique de la faune ichthyologique du Congo Central. Ann. Mus. Roy. Afr. Centr. Sci. Zool. 116 : 42-1 11. — & Lambert, J. 196 1 . Barbus schoutedeni espece nouvelle du sous genre Clypeobarbus. Revue Zool. Bot. afr. 64: 265-271. Stewart D. J. 1977. Geographic variation of Barbus radiatus Peters, a widely distributed African cyprinid fish. Env. Biol. Fish. 1 : 1 13-125. Whitehead, P. J. P. 1960. Three new cyprinid fishes of the genus Barbus from the Lake Victoria basin. Revue Zool. Bot. afr. 62 : 106-1 19. Manuscript accepted for publication 1 5 October 1979 A new species of Barbus (Pisces, Cyprinidae) from Africa K. E. Banister Department of Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD Barbus amboseli sp. nov. TYPICAL SERIES. Holotype (BMNH 1979.9.4:1), 35-5 mm SL; paratypes 23-45 mm SL BMNH 1979.9.4:2-6. All these specimens were caught in the Amboseli National Park, Kenya by R. Horowitz. ETYMOLOGY. The specific name refers to the National Park, their only known locality, and is to be treated as a noun in apposition. DESCRIPTION. The description is based on the six known specimens listed above. The abbreviations in the table below are explained at the end of the paper. Unless stated otherwise all measurements are expressed as a percentage of the standard length. Table 1 SL(mm) 35-5 45 35 33 25 23 D 29-6 27-8 30-0 31-8 30-0 26-1 H 28-2 28-9 28-6 28-8 30-0 30-4 I 7-0 6-7 5-7 6-0 6-0 6-5 10 9-8 10-0 10-0 10-6 10-0 10-8 MW 5-6 5-5 4-9 6-0 6-0 5-4 Pet 19-7 21- 1 21-4 18-2 18-0 7 Cpl 19-7 25-6 21-4 22-7 24-0 21-7 Cpd 11-2 12-2 11-4 12-1 10-0 10-9 Snt 8-4 6-7 8-6 9-1 8-0 8-7 Ab + 6-7 + 7 - - Pb 5-6 7-8 7-2 6-6 6-0 5-4 Dsp 15-5 14-5 15-7 13-6 18-0 17-4 Dfm III-7 III-7 1 1 1-7 III-7 III-7 III-7 Pfm 1-15 1-15 1-15 1-15 7 7 Vfm 1-7 1-7 1-7 1-7 7 7 Afm III-5 1 1 1-5 1 1 1-5 III-5 in-5 1II-5 Cpsc 10 10 10 10 8 10 LI 20 19 20 20 20 20 Tvl 4Mi 4Mi 4Mi 4j-4i 7 7 The overall shape of this stocky species can be seen in Fig. 1 . The great majority of the meristic and morphometric characteristics do not differ significantly from those found in many other species of small Barbus. There are, however, two interesting exceptions to this conformity. Firstly, the lateral line scale count is unusually low (see above) and only the largest specimen has lateral line tubules; tubules are completely lacking in the others. Secondly, the anterior barbels appear to develop late. They are present in the largest specimen, incipient in the specimens of 35 and 35-5 mm SL but lacking in the three smallest specimens (Figs 2-4). A pharyngeal bone is shown in Fig. 5, the pharyngeal teeth number 5.3.1 (f.4) or 5.3.2. (f.l). There are 30 (f.2), 31 (f.3) or 32 (f.l) vertebrae excluding those comprising the Weberian mechanism but including PUi+U,. In all the specimens there are Bull. Br. Mus. nat. Hist. (Zool.)38(3) : 145-150 Issued 29 May 1980 145 146 K. E. BANISTER Fig. 1 Barbus amboseli Holotype Fig. 2 Head of the specimen of 23 mm SL. Note the absence of the anterior barbels. Fig. 3 Head of the fish of 35 mm SL. Note incipient anterior barbels. three unbranched dorsal fin rays, the posterior of which is strengthened and serrated on its posterior face. There are 3 (f.3) short, hooked gill rakers. The scales have widely-spaced radiating striae (Fig. 6). The coloration in life is unknown, apart from a verbal comment by the collector that it was 'unremarkable'. Formalin-fixed and alcohol-preserved specimens are brown ventrally and darker brown dorsally. Immediately behind the operculum there is a diffuse dark spot which, after three scales, forms a very narrow clearly defined mid-lateral stripe. One dark spot surrounds the insertion of the unbranched dorsal fin rays, another dark spot is present at the termination of the mid-lateral stripe. There is a slight concentration of pigment cells NEW SPECIES OF BARBUS 147 Fig. 4 Head of the largest fish, 45 mm SL. The anterior barbels are well developed. Fig. 5 Two views of the left pharyngeal bone of the largest fish. Fig. 6 The third scale from the 'lateral line series' of the holotype to show the striations. around the insertion of the unbranched anal fin rays. The edges of the flank scales are darker than the centres. The dorsal fin, anal fin and caudal peduncle spots are more clearly demarcated in the smaller specimens. In the largest specimens, the proximal parts of the dorsal fin rays are more pigmented, especially those of the last simple ray. DISTRIBUTION. This species is known only from the type locality, the east side of the Encongo swamp, south of the causeway, Amboseli National Park, Kenya. The waterway was described by the collector as a stream with alternating riffles and papyrus-lined pools, 2-5 feet wide and 1-3 feet deep. The Amboseli area is a region of internal drainage, the lowest part, in wet years, forming the shallow, saline Lake Amboseli. DIAGNOSIS AND AFFINITIES. The combination of a serrated dorsal spine, low lateral line count, complete absence or full complement of tubules on the lateral line scales, a 5.3.1. pharyngeal tooth pattern and the late development of the anterior barbels is unique among the small Barbus of east and central Africa. Throughout Africa, very few species have as few scales as Barbus amboseli. Barbus jae from Cameroons has 20-23 scales in the lateral line series (fide Boulenger, 191 1); B. pumilis 148 K. E. BANISTER Blgr from the White Nile has 19-21 (fide Boulenger, 191 1); 5. chiumbeensis Pellegrin from Angola has 17-22 (fide Poll, 1967). Although Barbus gribinguensis Pellegrin was described as possessing 18-22 scales (see e.g. Poll & Lambert, 1961) it has now been shown that Pellegrin erred in his counts (Blache, 1964:120). Barbus gribinguensis is now considered to be conspecific with Barbus pleuropholis and has 2 1-23 scales in the lateral line series. The complete absence of lateral line tubules in all the sample except the largest Barbus amboseli is a puzzling phenomenon that immediately invites two possibilities. Firstly, their development could be delayed. Should this be the valid explanation then it would be a most unusual event. Lateral line tubules normally develop along with scales formation in African Barbus species (personal observations.). For example, in Barbus neumayeri from east Africa scales become distinguishable between 10 and 20mm SL. Tubules are detectable in the better formed lateral line scales at less than 20 mm SL and the full complement of tubules is conspicuous at 22 mm SL (personal observations.). Another possibility is that Barbus amboseli is unusually variable in the degree of tubule development. Various other species of small Barbus from Africa have only a few lateral line tubules, the exact number of which varies. Of the species with a low lateral line count listed above, Barbus jae for example, most frequently has only two or three tubules but it is not rare to find specimens in which they are wanting. In Barbus pumilis only the first few scales of the lateral line are perforated. Although reduction in the number of lateral line tubules is a derived character, its presence in the species listed above does not imply that they form a monophyletic group. Rather, it seems that the character has been acquired independantly. Barbus pumilis, for example, has been considered by some authors (e.g. Poll & Lambert 1961) to belong to the sub-genus Clypeobarbus because of its conspicuously deep lateral line scales and its smooth, flexible dorsal fin 'spine'. Also, neither Barbus pumilis nor Barbus jae have barbels. Amongst others, a similar reduction in lateral line tubules occurs in the strikingly marked, large scaled and sexually dimorphic Barbus papilio and its relatives from Zaire (Banister & Bailey, 1979). Again, this group of species is not advocated as close relatives of Barbus amboseli. The only obvious common factor in all the species showing reduction in the lateral line tubules is that they are small species, and all have a low number of lateral line scales. However, the correlation is far from perfect because the dull, but sexually dimorphic, species Barbus haasianus from southern Africa has 35-38 'lateral line' scales yet lacks the tubules (Jubb, 1967); whereas Barbus chiumbeensis has all 17-22 lateral line scales perforated (Poll, 1967). A serrated dorsal fin spine occurs in approximately a third of the African Barbus species. It is impossible, at the moment, to know whether or not the type of dorsal fin spine can be used as an indicator of relationship. Most African Barbus species can be divided into three groups on the basis of their type of dorsal fin spine (Boulenger, 191 1). The spines can be ossified and smooth, ossified and serrated or unossified, smooth and flexible. The difficulties inherent in the use of these characters to establish phylogenetic relationships are a) our inability to determine which (if any) of the three states is derived (or, indeed, primitive), b) the fact that some apparently valid species have more than one state [e.g. Barbus miolepis fide Poll & Lambert 1964 and Barbus issenensis fide Almaca 1970 (now regarded as Barbus capita by Karaman 1971)] and c) the apparent incompatability of classification by spine type with classification by size i.e. into the 'big' and 'small' Barbus species, or by the type of scale striations (parallel or radiate). Which, if any, of these three commonly used schemes will result in a phylogenetic classification is unknown and great caution should be exercised in adopting one scheme in preference to another. The commonest configuration of pharyngeal teeth among African Barbus is 5.3.2 (inner to outer rows). Reduction from the 5.3.2 pattern (which, on the principle of commonality is the plesiomorph condition) is known in a few species; mostly the 'smaller' ones (Banister, 1973 and personal observation.). The commonest form of reduction in the smaller species is to lose the first tooth of the inner row resulting in a 4.3.2. pattern (Banister & Bailey, 1979 and personal observation.). As far as I know the 5.3.1 pattern present in most of the Barbus NEW SPECIES OF BARBUS 149 amboseli specimens is unique among the 'small' Barbus of Africa, although it reflects the general trend towards reduction in pharyngeal tooth number in the smallest species. The actual loss of the tooth is probably a physical response to the increasing diminution in size. The late development of the anterior barbel presents an intriguing problem. The number of barbels has achieved a great significance in the classification of Barbus species, some authors having given separate generic status to species with respectively 0, 2 or 4 barbels (e.g. Schultz, 1957). Increasing awareness of the considerable variability shown by many species has, therefore, inevitably caused doubts about the usefulness of barbels as characters above the species level. Even at the specific level, great care must be exercised in using barbel characteristics as diagnostic features. Barbus anoplus Weber, from South Africa, has populations with only the posterior pair of barbels, populations with four barbels and populations in which the females have only the posterior pair whilst the males develop the anterior pair as well (Jubb, 1967). In an apparently closely related species, B. motebensis Steindachner the anterior pair of barbels may or may not be present. Parenthetically, it should be noted that in both these species the number of lateral line tubules is also very variable (Jubb, 1967), although it is not known if this phenomenon can be linked with barbel development Barbus brevipinnis Jubb has populations with 2 or 4 barbels. Barbus brevipinnis, and to a lesser extent Barbus motebensis, have restricted ranges and Jubb (1967:91) has noted that there is often more variability in localised species. In contrast, however, Barbus anoplus is widespread. There was no particular significance in the selection of these examples, they were chosen at random from the literature to illustrate barbel variability. Barbus amboseli has an apparently limited distribution which, following Jubb's (1967) comment, might indicate that a great variability is likely. But the only possible significantly variable feature in the regrettably small sample is in barbel number. It has proved impossible to determine the sex of the specimens so the presence of the anterior barbels cannot be shown to be a sex-limited feature as in some populations of B. anoplus. Perhaps more usefully, though, a clinal explanation can be rejected because all the specimens came from the same site. Unless a larger sample becomes available I cannot elaborate on the statement that Barbus amboseli is unique in that the development of the anterior pair of barbels is a correlate of size and hence age. The relationships of Barbus amboseli are presently unknowable because, within a cladistic framework, its diagnostic features are of no significance. Acknowledgements Firstly, my thanks must go to the collector Mr R. Horowitz for bringing the specimens to London. The manuscript was improved by Dr P. H. Greenwood and typed by Ms Joan Ellis. References Almaca, C. 1970. Sur un cyprinide nord-Africain: Barbus issenensis ou Varicorhinus issenensis? Bull. Mm. Hist. not. Paris (2) 42 : 159-160. Banister, K. E. 1973. A revision of the large Barbus (Pisces, Cyprinidae) of east and central Africa. Bull. Br. Mus. nat Hist. (Zool.) 26 (1) : 1-148. & Bailey, R. G. 1979. Fishes collected by the Zaire River expedition, 1974-1975. Zool. J. Linn. Soc. 66 (3) : 205-249. Blache, J., 1964. Les poissons du Tchad et du bassin adjacent du Mayo Kebbi. Mem. O.R.S.T.O.M. Paris 4: 483 pp. Boulenger, G. A. 1911. Catalogue of the fresh- water fishes of Africa 2 London. Jubb, R. A. 1967. Freshwater fishes of southern Africa. A. A. Balkema, Cape Town. Karaman, M. L. 1971. Susswasserfische der Turkei. 8. Revision der Barben Europas, Vorderasiens und Nordafrikas. Mitt, hamburg zool. Mus. Inst. 67 : 175-254. Poll, M. 1967. Contribution a la faune ichthyologique d'Angola. Publcoes cult. Co. Diam. Angola 75: 1-381. 150 K. E. BANISTER — & Lambert, J. 1 96 1 . Barbus schoutedeni, espece nouvelle du sous-genre Clypeobarbus. Revue Zool. Bot.Afr. 64 (3^4): 265-271 1964. Au sujet de Barbus miolepis Blgr et des especes voisines ou synonymes (Pisces, Cyprinidae). Revue Zool. Bot. Afr. 70 (3-4) : 405^ 1 1 . Schultz, L. P. 1957. The generic names Barbus and Puntius. Trop. Fish Hobby 5: 14-15;29-31. Key to the abbreviations Ab Length of the anterior barbel A fin Number of rays in the anal fin BMNH Register number in the British Museum (Natural History) Cpd Least depth of the caudal peduncle Cpl Length of the caudal peduncle Cpsc Number of scales around the least circumference of the caudal peduncle D Greatest depth of the body D fin Number of rays in the dorsal fin Dsp Length of the last unbranched ray of the dorsal fin H Head length from the tip of the snout to the extremity of the opercular bone I Horizontal diameter of the eye LI Number of scales in the lateral line series (or the equivalent series in the specimens lacking the tubules) MW Greatest width of the mouth Pb Length of the posterior barbel Pet Greatest length of the pectoral fin P fin Number of rays in the pectoral fin SL Standard Length Snt Length of the snout from its tip to the anterior border of the orbit Tvl Number of scales in transverse series from the origin of the dorsal fin to the lateral line and thence to the mid-ventral line + Character present but not measurable Character absent ? Uncountable or unmeasurable because of size or damage All measurements were taken with dial calipers. Manuscript accepted for publication 1 5 October 1 979 A new blind cyprinid fish from Iraq K. E. Banister Zoology Department, British Museum (Natural History), Cromwell Road, London SW7 5BD M. K. Bunni Natural History Research Centre, University of Baghdad, Bab-al-Muadham, Baghdad, Iraq Introduction Trewavas (1955) described a cavernicolous cyprinid, Typhlogarra widdowsoni, from a sink-hole near Haditha, Iraq (±34°4'N, 42°24'E). Earlier, A. G. Widdowson had captured examples of Typhlogarra from a deeper sink-hole called the Pidgeon Hole (sic) (Widdowson 1954) but none of these reached England. On October 21, 1977, Basim M. Al-Azzawi, a fish collector in the Natural History Research Centre, University of Baghdad, visited a sink-hole at the Sheik Hadid shrine near Haditha to collect some specimens of Typhlogarra. The significance of this visit is that it revealed two species of cavernicolous cyprinids, Typhlogarra and the new species which we have pleasure in naming after Basim Al-Azzawi. Recently, Greenwood (1976) described a blind, depigmented loach, Noemacheilus smithi, which shares its subterranean waters in the Zagros mountains with Iranocypris typhlops. Here was the first Asiatic cave system known to support two fully cave-adapted species. The rarity of this event was emphasized by Greenwood (1976) who thought that the Zagros mountain situation was paralleled only by that in the Mammoth Cave, Kentucky. To that very short list we can add only the poorly known Yardee Creek Wells, Northwest Australia where the synbranchid Anommatophasma candidum Mees ( = Ophisternon candidumfide Rosen & Greenwood 1976) and the goby Milyeringia veritas Whitley occur (Mees 1962), and now the waters below Haditha. Caecocypris basimigen. et sp. nov. COMMENTS ON THE GENERIC STATUS OF THE TAXON. Until recently, it was accepted practice to consider a cavernicolous species as generically distinct from its nearest epigean relative. The absence of e.g. eyes and pigment were implicitly regarded to be of more 'taxonomic' importance in determining the rank of the taxon than any similarities that the hypogean species and its relatives may possess. The characin Anoptichthys jordani Hubbs & Innes, the cyprinid Caecobarbus geertsi Boulenger and the synbranchid Typhlosynbranchus boueti Pellegrin demonstrate, by their allocated generic names, the theory behind this approach. Recent reconsiderations of some cavernicolous species on a phylogenetic basis have shown the illogicality of a generic level separation for the hypogean member of an epigean genus. Indeed, Greenwood (1976) wrote in his description of the cavernicolous Noemacheilus smithi 'Certainly no useful or phyletic purpose would be served by creating for it (N. smithi} a new genus that could only be defined on the basis of such features as eyelessness, depigmentation and other regressional trends associated with cave life....' Typhlosyn- branchus boueti, mentioned above, has now, on phyletic grounds, been transferred to the epigean genus Monopterus by Rosen & Greenwood (1976). Anoptichthys jordani and its subterranean congeners, which have been the subjects of many biological studies, have been shown to be interfertile with the epigean Astyanax mexicanus and have frequently been called as such although (as far as we are aware) the genus 'Anoptichthys' has not been formally synonymised with Astyanax (see details and bibliography in Mitchell, Russell & Elliott 1977). Roberts and Stewart (1976:304) could find no reason for maintaining a separate genus for the blind, rheophilic spiny eel Caecomastacembelus brichardi; Bull. Br. Mus. nat. Hist, (Zool.) 38 (3) : 1 5 1-1 58 Issued 29 May 1980 151 152 K. E. BANISTER AND M. K. BUNNI consequently, they placed Caecomastacembelus in the synonymy with Mastacembelus. Although the phyletic reasonableness of these nomenclatural changes cannot be denied, they are only practicable when the closest epigean relative of the hypogean species has been detected. The reason behind our decision to place our new species in its own genus is that we have been unable to discover its relationships below the family level. HOLOTYPE. The holotype is a mature female, 49 mm SL, BMNH 1979.8.3:1 from a natural well in the Sheik Hadid sink hole, Haditha, Iraq. A further eight specimens 7-9 mm-36 mm SL from the same locality, BMNH 1979.8.3:2-9 are paratypes. DESCRIPTION. The description is based on nine specimens, 7-9-49 mm SL. SL 49mm 36mm 31mm 30-5mm30mm 27mm 24mm 22mm 7-9 mm D HW HL MW Cpl S-D LI Cpsc Dfm Cfm A fin Vfm Pfm Gr Unless stated otherwise all measurements are expressed as a percentage of the standard length. The key to the symbols will be found on p. 1 58. The body is nearly circular in cross section in the abdominal region but becomes progressively compressed caudad. The head is slightly depressed. The mouth is horse-shoe shaped in ventral view. The gape is oblique (Fig. 1) and it extends as far as the vertical from the middle of the nostrils. The lips are thin and covered with minute, hair-like villi. There are no barbels. At their symphysis, the dentaries are produced into a small knob, but there is no corresponding notch in the upper jaw. The gular region has a padded appearance and lacks the mental disc characteristic of the sympatric Typhlogarra widdowsoni. On the cheek there are five or six approximately vertical rows of inconspicuous pits which resemble the 'pit organs' found in some other cyprinids. The smallest specimen has a number of protuberant structures randomly distributed over the lower parts of the sides of the head. The relationship, if any, between these and the 'pit organs' on the larger specimens is uncertain. Excluding the smallest fish (see below), and that of 24 mm SL, all the specimens have a small slit marking the position of the eye. The orbit is filled by fat globules bound together with loose connective tissue. The 24 mm specimen lacks the slit. In the only specimen (36 mm SL) in which the orbital region was dissected there was no trace of the eye in the left orbit, but in the right orbit there was a small structure (Fig. 2) which we interpret as a very regressed eye. This consists of a small tough sphere with two strap-like bands running rostrad towards the anterior wall of the orbit. From the posterior part of the sphere there are two thin, thread-like bands which appear to have fastened onto the inter-orbital septum. The two sets of bands are tentatively interpreted as residual eye muscles. Histological examination will be necessary before we can have any confidence in our identification of these structures. The smallest fish has a vestigial eye visible through the translucent tissue (Fig. 3). There is 21-4 20-8 16-4 16-4 15-0 7 7 7 11-4 18-3 19-4 19-0 19-8 18-4 20-4 7 7 16-5 30-6 30-6 32-2 30-8 30-0 29-7 7 7 26-2 10-6 11-1 9-8 11-5 8-5 11-1 7 7 7-8 13-9 14-0 14-8 15-6 15-0 15-5 7 7 - 54-0 55-5 51-5 54-0 55-0 53-6 7 7 46-5 29 28 28 29 31 7 29 30 - 12 12 12 7 12 7 7 7 - III 7 III 9 III 7 III 8 III 8 III 8 III 8 III 8 7 I-9-9-I I-9-8-I 7 I-10-8-I 7 I-9-8-I I-9-9-I 7 19 115 115 115 115 115 115 115 115 4 116 117 18 117 117 117 7 117 - II 13 II 13 115 7 113 II 14 7 II 13 - 14 12 13 14 13 7 7 7 7 NEW BLIND CYPRINID FISH 153 10mm Fig. 1 Caecocypris basimi Lateral view (above), dorsal and ventral views of the head (below) a small spherical lens and faint traces of pigment manifest as pale brown stains forming a ring around, particularly, the dorsal and posterior parts of the lens. The cephalic lateral line canals are not enlarged. The nostrils are situated at the lateral margin of the flat top of the head almost at the level of the supraorbital bone. The margin of the anterior nostril is produced into a short tube, that of the posterior is flush with the surface of the head. In the specimen of 36 mm SL there are four thick sensory folds on each side of the central nasal lamella. 0-5 mm Fig. 2 Right orbit of the specimen of 36 mm SL. The isthmus is relatively short and broad, and is not joined to the branchiostegal membrane. This condition is rare in cyprinids and, to the best of our knowledge, does not occur in any other cyprinid in the Tigris and Euphrates system. The gill rakers are short and slightly curved. The number on the lower limb of the first gill arch is given on p. 1 52. The position of the fins can be seen in Fig. 1 and the number of rays is given on p. 1 52. All the simple rays are thin, flexible and smooth. The fins of the smallest example are in the process of differentiating. The pectoral fin is rayless and is carried on a 154 K. E. BANISTER AND M. K. BUNNI pedicel as is usual in the larval condition. The dorsal fin rays are formed and the fin is separate from the median fin fold. The anal fin is still linked to the median fold and the rays are just beginning to develop. 2mm Fig. 3 Caecocypris basimi 7-9 mm SL. In the five specimens radiographed there are 32 vertebrae excluding those comprising the Weberian mechanism and including PUi+U,. It is difficult to identify positively the first caudal vertebra from radiographs but there seems to be sixteen or seventeen abdominal vertebrae. Ten pairs of ribs are present. Dorsal and ventral intermuscular bones are poorly ossified and can only be detected posterior to the dorsal fin. The caudal fin skeleton does not appear to differ from the general cyprinid type. The alimentary canal is relatively short with only one major coil and is less than twice the body length. The holotype (a mature female) has over thirty large eggs visible as well as many more less mature ova. The largest eggs are over a millimetre in diameter. The pharyngeal bones are delicate and bear two rows of slender, conical teeth arranged 5. 1-1 .5 (f.3). The posterior edentulous process forms a right angle with the rest of the bone. Fig. 4 Left pharyngeal bone of the holotype of Caecocypris basimi. Occlusal and ventral aspects. SQUAMATION. Scales are present only on the larger specimens. A mid-lateral row (corresponding to the lateral line series in other cyprinids, but in this case distinguished by the absence of lateral line tubules) of scales are the best developed, but even they are thin and flexible. Several scales tore during their removal. The biggest fish has a complete mid-lateral row as well as scales over the whole of the caudal peduncle. The back, in front of the dorsal fin, and the abdominal region are scaleless. Above and below the mid-lateral series, in front NEW BLIND CYPRINID FISH 155 of the caudal peduncle, there are a few poorly-formed scales. The scales of the mid-lateral series and, to a lesser extent, those on the sides of the caudal peduncle are unusual in that they are deeper than long. The scale striations are more or less parallel (Fig. 5). Scale counts are given on p. 152. Fig. 5 Scale from the holotype to show the striations. COLORATION. In life the fish was dead-white suffused with pale pink. The branchial region was deep red, the colour of the blood showing through the translucent tissues. The fins were colourless and there did not appear to be any guanine on the body. Specimens fixed in formalin then preserved in alcohol are an opaque, flat white colour with translucent, pallid fins. THE LOCALITY. The specimens were collected in a sink-hole near the top of a small hill surmounted by the Sheik Hadid shrine. The hill is some 500 m to the west of the Euphrates on the northern outskirts of the town of Haditha. A few metres to the west of the shrine is a small quarry formerly a source of limestone for building. Access to the sink-hole is via an opening, 2-3 by 6 m, on the floor of the quarry. Below this opening is a cavern, 6 by 6-5 m, Details of the water analysis from the Sheik Hadid well Provided by the Department of Chemistry and Minerals, Iraq Foundation General of Minerals Date 16.6.1979 pH7-5 Non-Carbonate hardness 23-85 mgh1 Total dissolved solids® 22° C 3072 m Cations mgl-' Anions mgl-' Na+ 490 ci- 912-84 K+ 44-8 so4— 1020-64 Ca-H- 265-53 NO3- nil Mg++ 182-4 NOr nil co3- 18-0 HCO,— 231-84 its stone-covered floor some 5 m below the quarry floor. At the eastern end of this cavern i.e. the side next to the hill, is natural well with surface dimensions of 1-2 m by 3-4 m and a depth of about 2 m. The water enters the well from four sources, two narrow slits and two oval apertures, about 15-30 cms deep. At the time of collection the greatest depth of water 156 K. E. BANISTER AND M. K. BUNNI was 63 cms, but in spring the water table rises to fill the well. The well is shaded from direct sunlight by a rock shelf at the eastern edge of the sink-hole opening; only in late afternoon would direct sunlight reach the well. In all probability, the type locality of Typhlogarra (the Haditha hole of Trewavas, 1955) is the sink hole described above. There are no other sink-holes in the area that match the description given by Trewavas (1955). Widdowson's Pidgeon hole is a much deeper sink- hole some 12 km south of Haditha (Widdowson 1954). There are many sink-holes in the area descending to the water table but the Sheik Hadid site differs from all the others in that it has a welling of water to the surface whereas the others are all land collapses of one kind or another. However, the presence of Typhlogarra in Widdowson's Pidgeon hole and in the Sheik Hadid well, strongly suggests that the same underground water system is involved. It is not known whether the underground system has any contact with the nearby Euphrates system. At noon on 1 3 June 1979 the characteristics of the well water were: temperature 22° C; pH 7-5 and 02 content 1-5 mg 1~' and salinity 1'75 g h1. Although a detailed water analysis is provided later, the water may be summarised here as extemely hard, low in oxygen and salinity, and high in sulphates. NOTES ON THE BIOLOGY. In addition to the two fish species, the underground waters support a yet undescribed species of the blind amphipod Niphargus. The stomachs of the post larval specimens of Caecocypris were empty but that of the larva contained a small piece of what seemed to be a chitinous material that may have come from an insect larva or a copepod. Dragonfly larvae, copepods, diatoms and green algae occur in the well if not in the completely lightless parts of the system. The blind fish have been seen to enter the well from all four openings and there they swim close to the pebbly bottom of the well. If disturbed they hide among the stones or swim back through the openings. They avoid being in the well when the light is bright and are most abundant at night. It is therefore likely that they are light sensitive to a limited degree. On the basis of the collections made so far it seems that Typhlogarra is about twice as common as Caecocypris (68 cf 35). Mr Abdul Hameed Hasan has kindly informed us that the water temperature varies very little during the day or throughout the year, remaining between 21-5 and 22° C. He also observed that specimens less than a centimetre long are always present and concluded that the fishes breed throughout the year. Both species are remarkably hardy, being able to endure considerable changes in water temperature and chemistry. DIAGNOSIS. Caecocypris basimi can be distinguished from the sympatric Typhlogarra widdowsoni by the absence of barbels and mental disc. The former species has a large terminal mouth, the latter a small ventral mouth. Caecocypris is unique among cavernicolous cyprinids and middle eastern cyprinids in the possession of a free branchiostegal membrane. AFFINITIES. In justifying the erection of a new genus for this species we indicated that no close relative could be found. Although Caecocypris possesses some undoubted apomorphies they seem to be autapomorphies rather than synapomorphies. As far as we have been able to discover the only other cyprinids with a free branchiostegal membrane are species of the genus Hypophthalmichthys from China. Hypophthalmichthys is highly derived and specialised for filter feeding (Fang 1928). It is tempting to suppose that the free branchiostegal membrane in Hypophthalmichthys is related to its feeding habits. Should that be so, it is unlikely that the free branchiostegal membrane in Caecocypris would serve the same function because, a) the species lacks the other modifications associated with filter feeding and b) it is unlikely that a cave system would afford it an adequate planktonic diet. The similar form of the branchiostegal membrane in these two genera therefore can only be interpreted as a parallelism. The symphysial knob on the lower jaw is a character of mosaic distribution throughout the NEW BLIND CYPRINID FISH 1 57 cyprinids. It occurs in some rasborines, chelines and aspinines (Howes 1978) and, more interestingly, in the cavernicolous, microphthalmic Barbopsis devecchi from Somalia. Other resemblance between Barbopsis and Caecocypris are similar body shape, large mouth, depressed head, similar meristic counts, two rows of pharyngeal teeth and a right-angled bend in the pharyngeal bone. However, the regressional trends associated with a hypogean habitat are not as advanced in Barbopsis as they are in Caecocypris; Barbopsis is eyed, pigmented, and has four barbels. (It should be noted that the ancestor of Caecocypris probably had barbels because the maxillary foramen for the innervation of the anterior barbel is still present although reduced. As far as we are aware, the maxillary foramen is invariably present in the members of that large group of 'barbelled' cyprinids as well as in members of the same group displaying secondary loss of barbels.). The branchiostegal membrane in Barbopsis is fused to the short, broad isthmus. We doubt, however, whether these similarities can be used as evidence for a postulated relationship. The body shape, depressed head, large mouth and similar meristics are all of such frequent occurence among the cyprinids as to nullify their use for this purpose. The right-angled bend in the pharyngeal bone could just as well be an obligatory response to similar head shapes as a synapomorphy. We have no precedents, nor evidence, to enable us to argue either case. Two rows of pharyngeal teeth (albeit with different numbers of teeth in the inner row) could indicate: a) reduction from three rows, a regressional trend noticed in other cavernicolous cyprinids e.g. Caecobarbus (Thines, 1969) and Typhlogarra (personal observation), b) the character was acquired independently or c) a synapomorphy. In view of the fact that so little is known about the polarity of trends associated with pharyngeal teeth we can have no confidence in using the number of rows of teeth to define relationships. Overall, because we cannot find any undoubted synapomophies for Caecocypris and Barbopsis, it can only be assumed that the similarities are due to parallelism. No other Asiatic cave fish can be related to Caecocypris. Both Typhlogarra and Iranocypris typhlops are 'garrine' i.e. they have the mental disc, barbels and mouth structure characteristics of their epigean relatives in the genus Garra. A search among the fishes of the Tigris and Euphrates has proved equally fruitless. We cannot find a species that we could postulate as the epigean relative of Caecocypris. In a way, this is no surprise. The general conservatism, coupled with the occasional small saltatory morphological changes, in cyprinid characters has the effect of making it difficult, if not impossible, to distinguish between parallelism and the sharing of characters derived through common ancestry. When these problems are overlain by the regressional trends associated with cave life, the likelihood of detecting the relatives of Caecocypris diminishes further. Acknowledgements It is with great pleasure that we acknowledge the help given to us by the following people and institutions. Dr P. H. Greenwood for his encouragement and his criticism of the manuscript. Mr Anthony Smith for his advice on the manuscript, for arousing interest in the caves of Iraq and for acting as courier for the living specimens of Caecocypris. Mr Abdul Hameed Hasan for putting his collection of Hadid fish at the disposal of M.K.B. The department of Chemistry and Minerals, Foundation General of Minerals (Iraq) for the detailed water analysis. The drawings of the holotype are the work of Mr Gordon Howes. References Banister, K. E. (In press). The fishes of the Tigris and Euphrates rivers. In Rzoska, J. (ed.), Tigris and Euphrates, rivers of destiny. Monogr. biol. Fang, P. W. 1928. Notes on the gill-rakers and their related structures of Hypophthalmichthys nobilis and H. molitrix. Contr. biol. Lab. Sci. Soc. China 4 (5) : 1-30. Greenwood, P. H. 1976. A new and eyeless cobitid fish (Pisces, Cypriniformes) from the Zagros mountains, Iran. J. Zoo/. Lond. 180 : 129-137. 158 K. E. BANISTER AND M. K. BUNNI Howes, G. J. 1978. The anatomy and relationships of the cyprinid fish Luciobrama macrocephalus (Lacepede). Bull. Br. Mus. nat. Hist. (Zool.) 34 ( 1 ) : 1-64. Mees, G. F. 1962. The subterranean freshwater fauna of Yardie Creek station, North West Cape, Western Australia. Jl R. Soc. West. Aust. 45 : 24-32. Mitchell, R. W., Russell, W. H. & Elliott, W. R. 1977. Mexican eyeless characin fishes genus Astyanax : Environment, distribution and evolution. Spec. Publs Mus. Texas Tech. Univ. 12 : 1-89. Poll, M. 1961. Contribution a Tetude des poissons d'eau douce de Somali appartenant aux genres Gobius et Barbopsis. Boll. Musei 1st. biol. Univ. Genova 31 ( 1 83) : 1 5-35. Roberts, T. S. & Stewart, D. J. 1976. An ecological and systematic survey of fishes in the rapids of the lower Zaire or Congo river. Bull. Mus. comp. Zool. Harv. 147 (6) : 239-3 17. Rosen, D. E. & Greenwood, P. H. 1976. A fourth neotropical species of synbranchid eel and the phytogeny and systematics of synbranchiform fishes. Bull. Am. Mus. nat. Hist. 157 ( 1 ) : 1-69. Thines, G. 1969. L'evolution regressive des poissons cavernicoles et abyssaux. Masson et Cie, Paris. 394 pp. Trewavas, E. 1955. A blind fish from Iraq, related to Garra. Ann. Mag. nat. Hist. (12)8 : 551-555. Widdowson, A. G. 1954. Explorers of subterranean by-ways. Iraq Petroleum 3(12): 4-8. Key to the abbreviations A fin Number of rays in the anal fin C fin Number of rays in the caudal fin Cpl Length of the caudal peduncle Cpsc Number of scales round the least circumference of the caudal peduncle D Greatest depth of the body Dfin Number of rays in the dorsal fin GR Number of gill rakers on the lower limb of the first gill arch HL Head length, from the tip of the snout to the posterior extremity of the opercular bone HW Greatest width of the head LI Number of scales in the mid-lateral line MW Greatest width of the mouth P fin Number of rays in the pectoral fin S-D Horizontal between a line perpendicular to the tip of the snout and the origin of the dorsal fin SL Standard length V fin Number of rays in the ventral fin character absent ? character unmeasurable or uncountable because of small size or damage. All distances were measured with dial calipers. Manuscript accepted for publication 1 5 October 1 979 A new species of cichlid fish from the Malagarasi swamps and river (Tanzania, East Africa) Peter Humphry Greenwood Department of Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD Introduction The material described below was collected in the Malagarasi river and swamps by members of the East African Fisheries Research Organization (later the East African Freshwater Fisheries Research Organization) during a survey of the area carried out in August and September 1952. Details of the habitats from which specimens were obtained are given in Greenwood ( 1 954:40 1-402). The new taxon brings the total number of cichlid species recorded from the upper part of the Malagarasi system to six, of which five (the new species, Astatoreochromis vanderhorsti [Greenwood], Orthochromis malagaraziensis [David], Sarotherodon karomo [Poll], and Serranochromis janus Trewavas) are endemic; the sixth species, provisionally identified as Astatotilapia bloyeti (Sauv.), belongs to a poorly understood complex of taxa. It may well be specifically distinct from the type and other populations of A. bloyeti, and thus would also have to be considered an endemic of the Malagarasi system (see Greenwood, 1954, & 1979:283-286, 297-298 and 302-303). Astatotilapia paludinosa sp. nov. HOLOTYPE. An adult male 100 + 26 mm long, from the Malagarasi swamps at Katare (Tanzania); BMNH reg. no. 1956.7.9.266. The trivial name, from the Latin, refers to the nature of the type locality. DESCRIPTION (Figs 1 & 2). Based on 56 specimens (including the holotype) from the Malagarasi swamps at and near Katare, and the river below Uvinza. The material is divided into two groups, one of fishes 20-50 mm standard length, the other of fishes from 52-1 08 mm SL. Additional material, examined in the field, is incorporated in the colour descriptions and in the data on food and feeding habits. In the table of proportions, head length, body depth and caudal peduncle length are expressed as percentages of standard length; all other measurements (indicated thus f) are given as percentages of head length. M = mean. Standard Length 20-50 mm 52- 108 mm Depth of body 33-3-36-7 M = 34-8 33-4-40-1 M = 36-8 Length of head 33-3-38-3 M = 34-8 32-6-39-0 M = 35-0 Preorbital depth t 10-0-16-7 M = 13-2 14-0-19-0 M = 16-7 Interorbital width t 18-8-25-0 M = 21-3 18-2-24-0 M = 21-1 Snout length f 22-2-30-0 M = 26-4 28-0-33-3 M = 30-7 Eye diameter t 25-0-35-0 M = 30-2 23-5-30-9 M = 27-2 Cheek depth t 12-5-23-0 M = 19-4 22-8-28-5 M = 25-1 Caudal peduncle length 14-0-19-0 M = 15-8 1 1-5-16-4 M= 14-0 Dorsal head and snout profile straight, except for a slight interorbital concavity; sloping moderately steeply (c. 30°-35° to the horizontal). Snout as broad as or slightly broader than Bull. Br. Mm. nat. Hist. (Zool.) 38 (3) : 1 59-1 63 Issued 29 May 1 980 159 160 P. H. GREENWOOD 1cm Fig. 1 Astatotilapia paludinosa. Holotype. Drawn by Gordon Howes. long. Mouth slightly oblique; posterior maxillary tip extending to the vertical from the anterior orbital margin. Jaws equal anteriorly, the lower 33'3-46'0, M = 39'8 (fishes < 50 mm SL), and 38-0-44-5, M = 42-7 (fishes < 50 mm SL) percent of head length; length/breadth ratio of the lower jaw 1 -4-2- 1 (mode 2-0). Gill rakers. Short, 7-9 on the lower limb of the first arch, the lowermost one or two reduced. Scales. Ctenoid except on the nape and chest; lateral line interrupted, with 29 (f.4), 30 (f.14), 31 (f.24), 32 (f.l 1) or 33 (f.3) scales. The last 2-6 scales of the upper lateral line series are separated from the dorsal fin base by 1 to lj scales. Thoracic scales noticeably smaller than those occurring on the nape and belly. 7-10 (mode 9) scales between pectoral and pelvic fin bases; 5 or 6 (rarely 7) between dorsal fin origin and the upper lateral line. Cheek with 3 or 4 (mode), rarely 5 series of scales. Fins. Dorsal with 24 (f.l 4), 25 (f.31), or 26 (f.l 1) rays, anal with 10 (f.l), 1 1 (f.24), 12 (f.28) or 13 (f.3) comprising XIV-XVI, 9-11 and III, 7-10 spinous and soft rays for the fins respectively. First pelvic ray produced, reaching to the vent in small fishes and as far as the third anal spine in larger individuals. Pectoral fin shorter than the head, 22-7-29-2, M = 25-7 per cent of standard length. Caudal moderately rounded. Teeth. Outer series. In fishes between 2 1 and 70 mm SL, the majority of teeth in this series of both jaws are compressed and unequally bicuspid; individuals in the size range 77-95 mm SL have some unicuspid, caniniform teeth, chiefly in the lower jaw and posterolaterally in the upper. Above this size, moderately slender caniniform teeth predominate in both jaws. Tooth number is variable, but shows positive correlation with size; thus, fishes less than 45 mm SL have from 24-36 outer teeth in the upper jaw; in larger fishes the range is from 40-56 (one exceptional individual possessed only 32). In fishes less than 70 mm SL, the posterior few teeth in the premaxillary outer series are not noticeably enlarged, nor are they unicuspid (the usual condition in most specimens of Astatotilapia species in the size range above 40 mm SL). Larger specimens, however, have the posterior \-4 outer teeth larger than those preceding them, and invariably unicuspid. Inner series. In fishes less than 90 mm SL, inner teeth are mostly tricuspid, but in larger individuals unicuspid teeth predominate, although an admixture of both types is common. NEW SPECIES OF CICHLID FISH 161 The number of inner rows in either jaw is variable, from 1-3, with 2 rows as the modal frequency at all sizes. Lower pharyngeal bone (Fig. 2). Triangular, its dentigerous surface slightly broader than long (most fishes<90mm SL) or equilateral (fishes>90mm SL), its teeth arranged in 20-22 rows. Teeth of the median rows enlarged and stout, clearly cuspidate in small individuals, becoming conical and sub-molariform in the largest specimens. Fig. 2 Astatotilapia paludinosa. Lower pharyngeal bone; occlusal view (scale = 2 mm). Skeleton. The neurocranium and jaw skeleton are identical with those in other Astatotilapia species (see Greenwood, 1979:274, 282 and fig. 6). Vertebrae (excluding the fused PU, and U, elements) 28-30 (mode 29), comprising 12 or 13 (mode) precaudal and 15 or 16 (mode) caudal elements. The caudal fin skeleton in 21 of the 23 specimens examined radiographically has all the hypural elements free; in the two exceptional fishes hypurals 3 and 4 are fused. Coloration. Preserved material. Since coloration of the few females caught is not markedly different from that of males, the description is applicable to both sexes. Ground colour brownish; seven to nine dark transverse bars on the flank and caudal peduncle, with faint indications of an interrupted mid-lateral stripe. A broad and distinct sub-lachrymal stripe is usually present. Posterior spinous and entire soft part of the dorsal fin maculate, less so in females; caudal fin maculate, especially on its dorsal half. Pectoral and anal fins clear, the latter, in males, with a single row of 3 or 4 ocelli with a narrow clear surround. Pelvic fins clear, or the outer half black, the leading edge of the first ray whitish. Live coloration - breeding males. Ground colour grey-green; chest and belly dark silver-grey. Operculum golden; branchiostegal membrane blue-grey between the rami of the lower jaw, but yellow between the opercula. Lips iridescent blue. Dorsal fin yellow, lappets and margin red; dark red spots, often coalesced, on posterior spinous and entire soft part. Caudal fin with yellow lower and clear upper half on which deep-red maculae occur. Anal fin light grey, with a faint red flush; ocelli orange-yellow. Pelvics with outer half dusky, inner half yellowish; first ray blue. Transverse and longitudinal banding was observed in live fishes, especially in a male seen guarding an established territory. No information is available on the coloration of females. ECOLOGY. Distribution. Lower Malagarasi river and swamps (Tanzania). Habitat. The species was collected both in the swamp and the open river, at Katare and Uvinza respectively. At Katare, specimens were obtained from traps set amongst marginal vegetation, in open-water further off-shore and in nets set amongst dense off-shore stands of water-lilies. The greatest depth of water in any habitat was about 4 metres. A number of young A. paludinosa was found in a small and shallow pool partially isolated from the main swamp. In the open river at Uvinza specimens were caught within the marginal reed beds, in the river itself and in sheltered embayments with a dense surface cover of water-lily leaves. 162 P. H. GREENWOOD There appears to be no difference between the preferred habitats of this species and those of A. bloyeti inhabiting the same areas of the Malagarasi swamps and river. However, in the river at Uvinza A. bloyeti were relatively uncommon in areas away from the banks. FOOD AND FEEDING HABITS. Analyses of stomach and intestinal contents in forty fishes (size range 44-107 mm SL) from all habitats at both localities provided little concrete evidence on the food of A. paludinosa; they do, however, indicate that the species is a bottom feeder. In both the swamp area and in sheltered river embayments the bottom is largely covered by a thick deposit of semi-decomposed plant material. Detritus typical of this deposit was predominant in the stomach contents of twenty-five individuals. Since many of its constituents were partly decomposed before ingestion it is impossible to determine what material, if any, was digested by the fishes. That the feeding habits of A. paludinosa are facultative is suggested by the occurrence of fish-guts as well as bottom debris in the stomachs of eleven individuals caught near a site where commercial catches were cleaned. Analyses made in the field showed that over 80% of fishes from this station had fed on fish offal. Scales and fish ova found in these fishes should probably be attributed to the same source. Although some insect remains (Trichoptera, Ephemeroptera and Odonata larvae, and chironomid pupae) were recorded, insects, at least on the basis of this sample, do not contribute significantly to the diet of A. paludinosa. Apparently there is little difference between the feeding habits of A. paludinosa and syntopic individuals of A. bloyeti, except that 2 1 of the 47 A. bloyeti specimens examined had ingested (and partly digested) water-lily seeds. BREEDING. No data were obtained on breeding sites or habits, save for a single male seen guarding a 'nest' prepared in the sand substrate near emergent reeds. Sexually active fishes of both sexes were found in all habitats. Fishes less than 60 mm SL are sexually immature. DIAGNOSIS. Astatotilapia paludinosa is distinguished from most congeneric species by its small chest scales (which are markedly smaller than those on the ventrolateral aspects of the flanks and belly, but are not separated from them by an abrupt change in scale size; see Greenwood, 1979), and, where live coloration is known, by its coloration and colour patterns (i.e. from A. flaviijosephi [Lortet], A. nubila [Blgr.], the A. bloyeti complex, A. burtoni [Giinther], A. swynnertoni [Blgr.], A. calliptera [Gunther] and A. stappersi [Poll]). From the four Astatotilapia species with small chest scales, A. paludinosa is distinguished as follows: (i) From A. flaviijosephi by its finer lower pharyngeal bone and few, less extensively molarized lower pharyngeal teeth (for a description of male coloration in A. flaviijosephi see Werner, 1976). (ii) From A. desfontainesi (Lacep.) by its more slender head, especially its narrower snout (snout noticeably broader than long in A. desfontainesi, slightly longer than broad or, rarely, slightly broader than long in A. paludinosa}, straighter, less decurved dorsal head profile, finer lower pharyngeal bone, and by its coloration, especially that of adult males (see Kirchshoffer, 1953, for colour descriptions of A. desfontainesi). (iii) From A. calliptera by its smaller chest scales and in having outer row jaw teeth with acutely and unequally bicuspid crowns (as opposed to subequally bicuspid, obliquely truncate crowns in A. calliptera). Similar dental differences distinguish A. paludinosa from A. swynnertoni, which species also has relatively larger chest scales. Colour descriptions for A. calliptera and A. swynnertoni have been published by Bell-Cross (1976) and Jubb (1967) for the species respectively (the latter author treating swynnertoni as a synonym of calliptera; but see Greenwood, 1979:284). (iv) From A. dolorosa (Trewavas) by its narrower interorbital width (18-2-24-0, mean 21-1% head length, cf. 25-8%), shallower cheek (22-8-28-5, mean 25-1% head length, cf. 30-3%) and its shorter, broader, lower jaw(38'0-44'5, mean 42-7% head length, cf. 48-6%; length-breadth ratio 1-4-2-1, mode 2-0, cf. 2-5 in A. dolorosa). All these morphometric NEW SPECIES OF CICHLID FISH 1 63 comparisons are with A. paludinosa in the size range 52-108 mm SL; the holotype and unique specimen of A dolorosa is 93- 5 mm SL. For further comments on these two species see below. AFFINITIES. The small chest scales of A. paludinosa, apparently a derived character (see Greenwood, 1979:270-271), suggest that the species' closest living relatives are A. desfontainesi (Tunisia and Algeria), A. flaviijosephi (Israel and Syria) and A. dolorosa (Lake Edward basin, Uganda). In its overall morphology, particularly its more streamlined head shape, A. paludinosa resembles A. dolorosa more closely than it does the other two species, but the phylogenetic significance of that feature cannot be assessed; it would certainly seem to be the commonest form amongst the African Astatotilapia species, and thus may be the primitive condition. The extent to which the lower pharyngeal bone and its dentition are hypertrophied (presumably a derived feature) suggests that A. flaviijosephi and A. desfontainesi are more closely related to one another than to any other Astatotilapia species. As a corollary, the finer bone and dentition found in A. paludinosa and A. dolorosa is probably a feature of little phylogenetic significance. Unfortunately A. dolorosa is known only from the holotype, a specimen now rather flaccid and completely bleached of its body and fin markings. This fish differs from comparable-sized A. paludinosa in certain morphometric features (see under Diagnosis above), which, since they involve bony features are not likely to be affected by the relatively poor condition of the specimen. However, the importance of the features in helping to establish the affinities of A paludinosa are diminished by the absence of other A. dolorosa specimens. No further material of that species has been captured, despite recent collections made in the vicinity of the type locality. Clearly the relationships of A. paludinosa cannot yet be determined. It is, however, interesting to note that they do not appear to lie with congeners presently occurring in the same geographical area (Lake Tanganyika basin and drainage), that is A. burtoni, the A. bloyeti complex, or A. stappersi. STUDY MATERIAL AND DISTRIBUTION RECORDS Museum and Reg. no. Locality Collector BM(NH) 1956.7.9.266 (Holotype) BM(NH) 1956.7.9.267-285 (Paratypes) BM(NH) 1956 7 9 291-31 1 (Paratypes) Katare, Malagarasi swamps. Katare, Malagarasi swamps. Katare Malagarasi swamps. E.A.F.R.O. E.A.F.R.O. E.A.F.R.O. BM(NH) 1956 7 9 254-265 (Paratypes) Malagarasi river at Uvinza E.A.F.R.O. (below the rapids). References Bell-Cross, G. 1976. The fishes of Rhodesia. Salisbury. Greenwood, P. H. 1954. On two species of cichlid fishes from the Malagarazi river (Tanganyika), with notes on the pharyngeal apophysisofthe Haplochromis group. Ann. Mag. nat. Hist. (12)7 : 401-414. — 1979. Towards a phyletic classification of the 'genus' Haplochromis (Pisces, Cichlidae) and related taxa. Part \.Bull. Br. Mm. nat. Hist. (Zool.)35 (4) : 265-322. Jubb, R. A. 1967. Freshwater fishes of southern Africa. Cape Town. Kirchshoffer, R. 1953. Aktionssystem des Maulbriiters Haplochromis desfontainesii. Z. Tierpsvchol. 10 (2): 297-3 18. Werner, Y. L. 1976. Notes on reproduction in the mouth-brooding fish Haplochromis flaviijosephi (Teleostei: Cichlidae) in the aquarium. ./. nat. Hist. 10 : 669-680. Manuscript accepted for publication 1 5 October 1 979 A new catfish from Sierra Leone Gordon Jon Howes Zoology Department, British Museum (Natural History), Cromwell Road, London SW7 5BD A collection of freshwater fishes from Sierra Leone made by Dr A. I. Payne contained three specimens of a previously undescribed dwarf catfish. At first sight these small fishes (33-5- 36 mm SL) were thought to be the juveniles of a Synodontis species, but closer examination showed many significant differences. Furthermore, dissection of two specimens revealed that they were adult females, having ripe ovaries. In overall morphology the new catfish appears closer to members of the Mochokidae sensu lato than to those of any other siluroid family, (see p. 168). The possibility that these fishes belonged to the dwarf mochokid genus Microsynodontis was dispelled when comparisons were made with species of that genus. MOCHOKIELLA gen. nov. Fishes of a small size with a shallow body (depth 25-5-27-9 of SL). Head broad (almost equal to its length) with dorsal profile sloped, snout rounded, ethmoid with narrow rostral process. Nostrils widely separated, the posterior situated midway between the anterior nostril and the orbital rim. Anterior nostril tubular. Mouth small, lower lip developed only at the corners of the mouth. Premaxillary teeth short and conical, in a broad patch. Dentary teeth like those on the premaxilla, set in a broad semi-crescentic band. Vomerine teeth absent. One pair of maxillary and two pairs of mandibular barbels. The maxillary barbel long, reaching to the tip of the cleithral process. Outer mandibular barbel extending to halfway along the cleithral process, the inner reaching to the base of the pectoral spine. The outer barbel bears 3 long branches, or sub-barbels, and the inner has 4 (number constant in all specimens). The nuchal shield is well-developed, slightly rugose. There are 3 nuchal plates, the 3rd paired, divided by the 1st and 2nd dorsal fin spines. The cleithral process is narrowly triangular, extending to a line level with the edge of the 3rd nuchal plate. Eye supero-lateral without free orbital rim. Pectoral spine strong, almost reaching the origin of the pelvic fin, its anterior margin bears 23 serrations, the posterior margin 8-9 strong 'teeth'. The longest ray of the pelvic fin reaches to the origin of the anal fin. Anal fin base short, as long as the caudal peduncle. Adipose fin deep and long based, equal to its distance from the last dorsal fin ray. Caudal fin long and forked. Branchiostegal rays 6. Gill opening extending to the base of the pectoral fin spine; branchiostegal membrane forming a fold across the ventral surface of the head. Gill-rakers small, 6 on 1st ceratobranchial. Lateral line pores prominent, interrupted. Mochokiella is distinguished from all other African siluroids by a combination of characters including small adult size, uni-lateral branching of the mandibular barbels, supero-lateral position of the eye, bands of villiform jaw teeth, elongate humeral process, paired 3rd nuchal plate and distinctive colouration (see below). Type species: Mochokiella paynei sp. nov (Fig. 1) HOLOTYPE. An adult female, 33-5 mm SL BMNH 1979.8.22:1 collected from Kassawe Forest Reserve, Sierra Leone by Dr A. I. Payne, for whom the species is named. DESCRIPTION. Based on the holotype and two paratypes, BMNH 1979.8.22:2-3, 35-5 & 36 mm SL. Characters given here amplify the preceding generic description. The proportions given below are shown as percentages of the standard length except for nos Bull. Br. Mus. nat. Hist. (Zool.) 38 (3) : 165-1 70 Issued 29 May 1980 165 166 G. J. HOWES Holotype 1979.8.22: 1 Para types 1979.8.22:2-3 Standard length 33-5 mm 35-5 mm 36 mm Depth 27-0 25-5 27-9 Head length 31-5 31-5 29-2 Interorbital width 38-0 40-0 38-0 Snout -dorsal fin 39-0 39-5 39-0 Eyediam. 19-0 22-5 24-0 Mouth width 27-0 28-5 28-5 Head width 30-0 31-0 29-2 Snout length 38-0 40-0 43-0 Maxillary barbel 132-0 130-0 119-0 Mandibular barbel, outer 105-0 112-0 105-0 Mandibular barbel, inner 76-0 71-0 67-0 Cleithral process length 15-4 18-4 19-4 Cleithral process depth 28-2 31-0 23-0 Pectoral spine length 27-0 24-0 27-9 Pelvic spine length 12-0 12-6 12-5 Dorsal spine length 22-5 22-5 22-1 Dorsal - adipose distance 26-9 34-2 22-1 Adipose fin length 26-9 24-0 19-4 Caudal peduncle length 13-4 15-5 11-1 Caudal peduncle depth 12-0 11-3 9-7 Anal fin base length 14-0 12-7 16-6 Dorsal fin rays 116 116 116 Pectoral fin rays 16 16 16 Pelvic fin rays 16 16 16 Anal fin rays 110 19 110 Gill-rakers on lower arch 6 Unknown Unknown Vertebrae excluding fused anterior elements 27 27 27 3,5,6,8-1 1 , which are percentages of the head length; the depth of the cleithral process (13) is shown as a percentage of its length ( 1 2). COLORATION. The overall body colour pattern is mottled, the top of the head dark with indistinct transverse bands. A dark band links the nostrils to the eye and a diffuse band runs antero-ventrally from the eye to the mouth. A light stripe crosses the interorbital space. The cheeks and operculum bear dark patches. A distinct dark band runs across the supraoccipital, and is followed by a broad light patch and then another dark band on each 3rd nuchal plate, which are divided by the dorsal fin. From the base of the dorsal fin to the mid-lateral line is an area of dark brown pigment. Below the cleithral process this pigment is broken into a reticulated pattern. Just above the pelvic fin is a dark, irregular patch encircled by a light band. An extension of this band continues dorsad to meet its partner and form a light area anterior to the adipose fin. In the paratypes the mid-lateral blotch is much less distinct than in the holotype, the encircling band being broken and interrupted mid-dorsally. From the rear of the adipose fin a broad light stripe runs down to the posterior part of the anal fin. The dorsal part of the caudal peduncle is uniformly dark; below the lateral line is a lighter coloured spot just posterior to the anal fin, followed by a darker patch at the base of the caudal fin. Again, on the paratypes these markings are less distinct and more reticulated. In all specimens the ventral surface of the body is mottled with light brown pigment, a small darker patch occurs at the base of each pelvic fin. The adipose fin is mostly dark, but with a light posterior margin. In one of the paratypes the posterior tip has a distinct spot, and in the other there are two spots. All other fins are barred and blotched. All barbels are barred. NEW CATFISH FROM SIERRA LEONE 167 5mm Fig. 1 Mochokiella paynei; above, lateral view; below, dorsal and ventral views of the head. Life colours have been recorded from 6 imported aquarium specimens: The light areas are cream to beige and the dark areas Vandyke Brown. The markings on the head are more distinctive and appear as individual blotches and spots. The patterning differed little from that of the types, most noticeable was the variation in the markings of the adipose fin. However, the main features - the pale nuchal band, striped snout, mid-lateral reticulation and caudal peduncle band were well-developed. HABITAT AND HABITS. The type locality is a forest stream in Kassewe Forest Reserve with a pH of 6-6. The fish were caught overnight in a baited trap. Recent trade aquarium imports from Sierra Leone have included specimens of Mochokiella paynei, and six individuals have been observed in aquaria. The fishes readily seek shelter in cavities provided by rocks and wood; sometimes swimming vertically or upside-down against the cavity walls. They are most active at night, moving outside their hiding places. One fish was seen resting along an almost upright plant stem, supported by its caudal fin. All known aquarium specimens have been of the same or slightly larger size than the type series. Of the type specimens, two have ovaries each containing a total of c. 30 eggs. PHYLOGENETIC RELATIONSHIPS. Both the lack of osteological material of Mochokiella and knowledge of the distribution of certain anatomical characters throughout the Siluroidei makes any judgement of the taxon's relationships highly speculative. 168 G. J. HOWES B 5mm Fig. 2 Dorsal views of the heads and ethmoid regions of A. microsynodontis batesii, B. Mochokiella paynei, C. Synodontis melanopterus. All specimens are of the same size, 36 mm SL. N 1 ,2 & 3 = nuchal plates. Assignment to the Mochokidae is made solely on the grounds of superficial resemblance and must be viewed as a temporary expedient (see below). Character comparison of Mochokiella with Microsynodontis and Synodontis is given below and in Fig. 2. On the basis of more numerous similarities, Mochokiella more closely resembles Microsynodontis than any other mochokid. Present lack of data makes it impossible to say whether these similarities are to be regarded as based on apomorph or plesiomorph features. The ethmoid of both Mochokiella and Microsynodontis is long and narrow, in contrast to the broad, thick ethmoid of other mochokids (Figs 2A & B), the only exception being Mochokus. A similarly elongate ethmoid occurs in some bagrids, e.g. Chrysichthys and Auchenoglanis. Research on other otophysans - characoids and cyprinoids - has indicated that an elongate ethmoid is an apomorph character. However, it should be noted that an elongate ethmoid is present in Diplomystes, considered by some authors as a primitive siluroid (Regan, 1911; NEW CATFISH FROM SIERRA LEONE 169 Alexander, 1966). Mochokiella shares with Microsynodontis and Mochokus two other features. One is the elongate posterior cleithral process (humeral process) which is short and broadly triangular in other mochokids and bagrids. The other character is the number and form of the nuchal plates, there being 3 in all three genera, the 2nd partially divided by the dorsal fin spine and the 3rd completely divided. In other mochokids and amphiliids there are two nuchal plates (see Taverne & Aloulou-Triki, 1974). Mochokiella has the parapophyses of the 4th and 5th vertebrae well-developed, spine-like and extended laterally as in Synodontis. In Mochokus these parapophyses extend upward to contact the ventral surface of the nuchal plates and in Microsynodontis they contact each other laterally to form a continuous shelf- similar to that illustrated for Pimelodus by Alexander ( 1 966). The dentition of Mochokiella is distinctly different from that of other mochokids, and it more closely resembles that of bagrids. However, this is most likely to be a plesiomorph rather than an apomorph character. Mochokiella differs from other species of Microsynodontis in having greater body depth, longer head, greater interorbital width, longer snout to dorsal distance, larger eye, longer mandibular barbels, longer dorsal and pectoral spines and caudal fin forked instead of rounded. Nuchal plates Ethmoid Eyes Lower lip Dentary teeth Premaxillary teeth Mandibular barbels Mochokiella 3 elongate small, superolateral; no free orbital rim weak small, conical, all of equal size set in broad band small, conical, all of equal size set in broad patch c.45 on each pmx. with thin sub-barbels Microsynodontis 3 elongate Minute, superior; no free orbital rim well-developed slender, hook-like, the outer row longer than the inner, set in narrow band slender, elongate set in broad patch c. 45-50 with thick sub-barbels Svnodontis 3" short and broad large, lateral or supero-lateral; free orbital rim usually well- developed elongate, hook- like, outer row longer than the inner, set in small, rounded patch stout, elongate set in narrow band, c. 30 with numerous sub- barbels sometimes tuberculate Gill opening extending anterior to to pectoral base to pectoral base pectoral base Branchiostegal transverse ventral transverse fold fold lacking membrane fold Pectoral fin spine anterior serrations, 28 28-29 25-26 Lateral line interrupted interrupted complete Adipose fin long short variable Cleithral process elongate elongate broadly triangular Caudal fin forked rounded forked As currently recognized, the Mochokidae contains the genera Mochokus, Micro- synodontis, Atopocheilus, Acanthocleithrum, Chiloglanis, Euchilichthys, Synodontis, Hemi- synodontis and Brachysynodontis. Taverne & Aloulou-Triki (1974) recognized the diverse nature of this family by placing the latter three genera into the subfamily Simuldentinae. Although these authors listed a set of characters defining the subfamily, no comparison was made with the other non-included genera nor with other siluroids, Thus, the 170 G.J.HOWES distribution of these characters within the suborder is unknown. From preliminary observations of the osteology of the Mochokidae I would venture to suggest that the family is possibly polyphyletic. Acknowledgements My most sincere thanks are due to Dr A. I. Payne for collecting the catfishes and providing habitat information, and to Stephen Pritchard who first drew my attention to aquarium imports and provided not only data on the habits of live fishes but with the singular osteological preparation of the species. I am most grateful also to Mr & Mrs D. Lambourne for their notes on live specimens. I am indebted to Dr Max Poll, formerly of the Musee Royal de 1'Afrique Centrale, Tervuren for loaning the holotype of Microsynodontis polli and to Dr P. H. Greenwood for commenting on the manuscript. References Alexander, R. McN. 1966. Structure and function in the catfish. Proc. Zoo/. Soc. Lond., 148 (1): 88-148. Regan, C. T. 1911. The classification of the teleostean fishes of the order Ostariophysi. 2. Siluroidea. Ann. Mag. nat. Hist. (8)8 : 553-577. Taverne, L. & Aloulou-Triki, A. 1974. Etude anatomique, myologique et osteologique du genre SynodontisCuvier (Pisces: Siluriformes, Mochoch\dae)AnnlsMus. r. Afr. cent. 210 : 1-69. Manuscript accepted for publication 1 5 October 1 979 A new genus of cheline cyprinid fishes Gordon Jon Howes Zoology Department, British Museum (Natural History), Cromwell Road, London SW7 5BD In a recent review of the cyprinid genus Barilus (Howes, 1980), it was noted that one species, Barilius auropurpureus Annandale, 1919 was not a member of that genus but that it shared derived characters with the cheline group of cyprinids. This paper amplifies that statement and establishes a new genus to contain the species. Annandale (1918) described Barilius auropurpureus from Inle lake basin, south Shan States in Burma. He did not, however, compare the species with any other Barilius and gave no reason for including it in that genus. I have examined a series of specimens collected by Annandale and can find no characters that indicate affinity with Barilius or any member of the bariliine group (sensu Howes, 1980). On the contrary, auropurpureus shares apomorph features with the cheline group (sensu Howes, 1979). These characters are: marginal overlap of the supraethmoid by the frontals; triangular, lamellate kinethmoid; ethmo-vomerine bloc extended anteriorly and laterally as a shelf; wide midline separation of the medial anterior maxillary processes; anteriorly notched and hooked dentary; truncated supraoccipital; long, posteriorly curved lateral processes on the 2nd vertebra. Inclusion of auropurpureus in an existing cheline genus is precluded on the grounds of its autapomorphic features involving the ethmoid and jaw morphology (see below). I propose therefore that 'Barilius ' auropurpureus be assigned to a new genus: INLECYPRIS gen nov. TYPE SPECIES. Barilius auropurpureus Annandale, 1918 Slender fishes with large eye, pointed snout and obliquely angled mouth; ventral keel present. Supraethmoid produced laterally, the mesethmoid extending anteriorly beyond the supraethmoid as a shelf; preethmoid minute; kinethmoid lamellate, triangular; symphysial notch and process on dentary; large nasal; infraorbitals well developed, the 5th small and contacting the wide supraorbital; frontals narrow, overlapping the posterior margin of the supraethmoid; dilatator fossa small; parasphenoid horizontal, separated from the orbitosphenoid by a shallow septum; minute metapterygoid-quadrate fenestra; vertical cleithral limb short; postcleithrum well-developed, curved mesad; coracoids meeting along their ventral margins; neural complex low; 7 large lamellate supraneurals - 1st & 2nd fused; lateral processes of 1st and 2nd vertebrae elongate and caudally directed; total number of vertebrae 37 (4+10+22+1). A single species, Inlecypris auropurpureus (Annandale, 1918), known only from Inle lake basin, Burma. Annandale (1918; 1922) adequately described the species, its habitat and habits. RELATIONSHIPS. Within the cheline group Inlecypris shares derived characters with Chela. The ethmoid region of both taxa is very similar. The supraethmoid is overlapped posteriorly by the frontals, the mesethmoid is depressed and is produced anteriorly as a long shelf, the preethmoids are minute. Both genera have a flat, triangular kinethmoid (Figs 1 A & B). Other synapomorphic features are the enlarged nasal; the morphology of the upper jaw, which in Chela lacks a prominent mid-lateral maxillary ascending process and which in Inlecypris is reduced to a spine (cf. Figs 2A & B); the posterior curvature of the cranium, reduction of the supraoccipital process and absence of posttemporal fossae; morphology of the supraneurals; morphology and degree of development of the transverse processes on the 1st and 2nd vertebrae. Both taxa also possess an enlarged supracleithrum which covers the Bull. Br. Mus. nal. Hist. (Zool.) 38 (3) : 1 7 1-1 73 Issued 29 May 1 980 171 172 1mm B Fig. 1 Dorsal views of the ethmoid regions of A, Inlecypris auropurpureus and B, Chela laubuca. F, frontal; K.E, kinethmoid; ME, mesethmoid; MX, maxilla; N, nasal; PE, preethmoid; PMX, premaxilla; SE, supraethmoid. V2 VI MXP PC B Fig. 2 A. Inlecypris auropurpureus, lateral view of the neurocranium with jaws, part of suspensorium, pectoral girdle and vertebral column in articulation. B. Chela laubuca, lateral view of jaws. MQP, metapterygoid-quadrate fenestra; MXP, mid-lateral maxillary process; PC, postcleithrum; SC, supracleithrum; SY, symplectic; VI & 2, 1 st and 2nd vertebrae. NEW GENUS OF CHELINE CYPRINID FISHES 173 upper half of the cleithrum (cf. Fig 2 A with Fig 35 in Howes, 1979). The form of the lower jaw in Inlecypris more closely resembles that of other members of the cheline group viz Oxygaster and Salmostoma. Howes (1979) recognizes three lineages within the cheline group, the cheline lineage being represented solely by the genus Chela. Some modification to this classification is now required. Chela undoubtedly displays greater cranial and vertebral modifications than does Inlecypris (Howes, 1979). The absence in Chela of a symphysial notch and midlateral maxillary ascending process are possibly to be regarded as secondary loss characters - the reductive trend being witnessed in Inlecypris. Thus, now Chela and Inlecypris comprise the cheline lineage which forms the sister group to the combined assemblage of salmostomine and oxygastrine lineages (see cladogram in Howes, 1979). Writing of the ichthyofauna of the Inle basin, Annandale (1918) noted that of the 17 recorded genera 2 were endemic (Chaudhuria (Chaudhuriidae) and Sawbwa (Cyprinidae)). Sawbwa is a minute, scaleless cyprinid of unknown affinity. It does not possess cheline characters and may possibly belong to a subgroup of the rasborine complex. Annandale (1918) described another cyprinid genus, Microrasbora, represented by two species in the Inle basin, M. rubescens and M. erythromicron. He believed, however, that Microrasbora was widespread throughout the Malay Peninsula and that Rasbora heteromorpha Dunker, 1904 and R. maculata Dunker, 1904 should possibly be included in the genus. Paucity of specimens of Microrasbora prevents analysis of osteological characters but superficially its two contained species greatly resemble Sawbwa except that they are scaled. As in Sawbwa the ethmoid region is narrow and the lower jaw shallow; there is a patch of dark pigment surrounding the vent and the genital pore is enveloped in a thickened sheath. These characters are lacking in Rasbora heteromorpha and R. maculata. Thus, it would appear that there are four endemic genera in the Inle basin. Of the cyprinids, Inlecypris is a relatively plesiomorph member of its lineage. Its closest relative, the more derived Chela, is widespread throughout the South-East Asian archipelego. Acknowledgements I am most grateful to Drs P. H. Greenwood and K. E. Banister for commenting on the manuscript, and to Mr J. Chambers for preparing alizarin specimens. References Annandale, N. 1918. Fauna of Inle Lake. Fish and fisheries of the Inle Lake. Rec. Indian Mus. 14 : 33-64. — 1922. Fish and fishing in the Inle Lake. /. Bombay nat. Hist. Soc. 28 (4) : 1038-1044. Howes, G. J. 1979. Notes on the anatomy of Macrochirichthys macrochirus (Valenciennes, 1844), with comments on the Cultrinae (Pisces, Cyprinidae). Bull. Br. Mus. nat. Hist. (Zool.) 36 (3) : 147-201 . — 1980. The anatomy, phylogeny and classification of bariliine cyprinid fishes. Bull. Br. Mus. nat. Hist. (Zool.) 37 (3): 129-198. Manuscript accepted for publication 1 5 October 1 979 British Museum (Natural History) Associated publication on fishes The Cichlid fishes of Lake Victoria, East Africa. The biology and evolution of a species flock. P. H. Greenwood. Bull. British Museum (Natural History) Zool. Suppl. No. 6, 1974 vi+ 134 pp, 1 coloured plate, 77 text figures, 4to, (paper) £6.00 (boards) £8.25 Lists are available free on request to : Publications Sales British Museum (Natural History) Cromwell Road London SW7 5BD Standing orders placed by educational institutions earn a discount of 10% off our published price. Titles to be published in Volume 38 Miscellanea Relationships between hummingbirds and flowers in the Andes of Columbia. By David W. Snow & Barbara K. Snow Miscellanea Mites of the subfamily Parasitinae (Mesostigmata: Parasitidae) in the British Isles. By K. H. Hyatt Miscellanea Printed by Henry Ling Ltd, Dorchester Bulletin of the \ British Museum (Natural Historf^ Miscellanea Zoology series Vol 38 No 4 25 September 1980 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum's resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself, available separately, and individually priced. Volumes contain about 300 pages and are not necessarily completed within one calendar year. Subscriptions may be placed for one or more series. Subscriptions vary according to the contents of the Volume and are based on a forecast list of titles. As each Volume nears completion, subscribers are informed of the cost of the next Volume and invited to renew their subscriptions. Orders and enquiries should be sent to: Publications Sales, British Museum (Natural History), Cromwell Road, London SW7 5BD, England World List abbreviation: Bull. Br. Mus. nat. Hist. (Zool.) Trustees of the British Museum (Natural History), 1980 ISSN 0068-1498 Zoology series Vol 38 No 4 pp 175-236 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 25 September 1980 Miscellanea Contents 3\%H A redescription of Trochammina nana (Brady) (Protozoa: Foraminiferida), with observations on several other Recent Trochamminidae in the Collections of the British Museum (Natural History). By P. Bronnimann & J. E. Whittaker . Crinoidea collected by the Meteor and Discovery in the NE Atlantic. By A. M. Clark A new abyssal genus of the family Ophiuridae (Echinodermata: Ophiuroidea). By G. L. J. Paterson A revision of the spider genus Macopaeus (Araneae: Salticidae). By F. R. Wanless A revision of the spider genus Orthrus (Araneae: Salticidae). By F. R. Wanless . A note on Lonchophylla (Chiroptera: Phyllostomatidae) from Ecuador and Peru, with the description of a new species. By J. E. Hill Page 175 187 211 219 225 233 A redescription of Trochammina nana (Brady) (Protozoa: Foraminiferida), with observations on several other Recent Trochamminidae in the Collections of the British Museum (Natural History) P. Bronnimann Laboratoire de Paleontologie, Universite de Geneve, 13, rue des Maraichers, nil- Geneve-^ Switzerland J. E. Whittaker Department of Palaeontology, British Museum (Natural History), Cromwell Road, London SW7 5BD Introduction As part of a major revision of the Trochamminidae from several parts of the world, in particular the waters of the South Atlantic (Bronnimann, 1976, 1979; Bronnimann & Beurlan, 1977, 19770; Bronnimann & Whittaker, in preparation), the present authors have found in necessary to examine the important Recent foraminiferal Collections of H. B. Brady and Heron-Allen & Earland, housed in the British Museum (Natural History) (B.M.N.H.). In the Brady Collection, several of the trochamminids have been found to be in need of taxo- nomic revisison, and this paper deals with the confusion surrounding Haplophragmium nanum. Brady (1881 : 50) proposed for a small trochospiral agglutinated foraminifer collected by the Challenger Expedition, the name Haplophragmium nanum. At the time he did not specify the station(s) at which the new species occurred, and he described it, without illustration, as follows: Test minute, rotaliform, depressed; superior face somewhat convex; inferior, plane, more or less excavated at the umbilicus; margin rounded, lobulate. Consists of about two revolutions, each composed of about six inflated segments, often irregular in shape and disposition. Shell texture thin, resembling that of//, canariense. Diameter. 1/75 inch(0'34 mm).' H. nanum Brady is therefore a valid name according to Article 12 of the International Code of Zoological Nomenclature. The lack of illustration is however unfortunate in the light of the ensuing taxonomic disorder. Subsequently, Brady (188 la) reported H. nanum from Arctic waters. This was material from the Austro-Hungarian North Polar Expedition and had been sent to Brady for identification by his friend Dr F. Karrer of Vienna. Brady's text was translated into German and apart from one or two mis-translations and an extra plate (pi. 1) showing the locations of the soundings in the vicinity of Franz-Josef Land and Novaya Zemlya, is identical to the English text; the German version was not however published until 1 882. Haplophragmium nanum is said by Brady (188 la: 406) to be. . . 'the commonest of all the smaller arenaceous forms, over the whole area reported by these soundings . . .'.He also states that the Arctic specimens are . . . 'rather smaller than those described from the Challenger dredgings [Brady, 1881] and seldom exceed 0'25 millim. in diameter; they are generally less convex on the superior surface and altogether somewhat thinner . . .'. The Bull. Br. Mus. nat. Hist. (Zool.) 38 (4): 175-185 Issued 25 September 1 980 175 1 76 P. BRONNIMANN & J. E. WHITTAKER accompanying illustration (Brady, 1 88 1 a, pi. 2 1 , figs 1 a-c) is apparently of a single specimen in spiral, edge and umbilical views, and is the first published figure. It must however be emphasized that this specimen was not from the Challenger material from which the original description was made. In the Challenger Report, Brady (1884: 311, pi. 35, figs 6a-c, 7, 8) described H. nanum again in more detail, giving the locations of his figured specimens. That shown in plate 35, figs 6a-c appears at first glance to be identical to the specimen in plate 21, figs la-c of his 1 88 1 a paper, but on close examination can be seen to differ in several respects particularly in edge and umbilical views. Perhaps Brady had merely instructed his engraver, Mr Hollick, to make the previous illustration a little more accurate, or may have had drawn a completely different specimen. According to our records, the specimen shown in Brady, 1 884, pi. 35, figs 6a-c, comes from a slide of about 40 individuals from the Austro-Hungarian North Polar Expedition station 518, off Franz-Josef Land (B.M.N.H. slide no. ZF 1546), at 1 13 fathoms (207 m), but it is not possible to ascertain with certainty which particular individual was figured. Brady did not himself designate holotypes nor in general give any indication as to which syntype(s) he illustrated. Nevertheless, this figured specimen is typical of the suite in ZF 1546. Figures 7, 8, of Brady's plate 35 were also attributed to H. nanum and came from Challenger station 346, north of Ascension Island in the South Atlantic (B.M.N.H. slide no. ZF 1547), at 2,350 fathoms (4,300 m). Two specimens were used for the engraving; only 8 syntypes now remain. Brady (1884) was obviously of the opinion that the Arctic and South Atlantic forms were conspecific, but probably because he had seen many more specimens from the Austro- Hungarian North Polar Expedition dredgings than from the Challenger, he wrote that ... (p. 312) ... 'fig. 6 represents the typical form, while figs 7 and 8 are examples of irregularly grown tests'. Nevertheless, he felt it necessary to modify his original description to accommodate the Arctic forms . . . 'Test minute, Rotaliform, depressed; consisting of about two convolutions, the outermost of which is composed of six or seven somewhat inflated segments, often irregular in shape and disposition. Superior face flat or slightly convex; inferior face convex, more or less excavated at the umbilicus. Walls thin; texture resembling that of Haplophragmium canariense; colour light-brown, the final segment somewhat lighter than the rest. Diameter 1/75 inch (0*34 mm) or less.' Subsequent workers (e.g. Barker, 1960) have always followed Brady (1884) and apart from referring it to Trochammina, have accepted this wide concept of the species without question. After studying the early literature on H. nanum outlined above, our growing apprehension that Brady had in fact confused two distinct species was confirmed on examining the respective syntypic slides in the Museum. It was also possible to produce scanning electron micrographs of several specimens from each slide and the results are illustrated in this paper. The two forms are shown not only to be separate species, but also to warrant separate generic status. Haplophragmium nanum Brady must be based on the Challenger South Atlantic material from north of Ascension Island (Stn 346) as the first valid description (Brady, 1881) was made without reference to the Arctic collections (R. V. Melville, pers. comm.). The species is now redescribed and illustrated (Figs 1-9) as a Trochammina and a lectotype chosen. Specimens from the Austro-Hungarian North Polar Expedition stations, even though later stated by Brady (1884) to ... 'represent the typical form . . .', must be given a new name; we propose Portatrochammina bipolaris sp. nov., as it has also been found to occur in Antarctic as well as Arctic waters (see Figs 15, 16, 1 8-3 1 ). We have examined every one of the very many slides in the Collections labelled Haplophragmium nanum (or Trochammina nand), ranging in age from Cretaceous to Recent and from present day polar, temperate and tropical regions, in order to ascertain the true geographical and stratigraphical range of the species under review. This information is given in the Systematic part below (pp. 178-183). Of the slides in the Brady Collection attributed to H. nanum, several contain a further new species, not previously illustrated by REDESCRIPTION OF TROCHAMMINA NAN A (BRADY) 177 Figs 1-9 Trochammina nana (Brady). Figs 1-3, 6, Lectotype (ZF 3971); spiral, umbilical, oblique-umbilical and edge views. (Believed to be specimen illustrated by Brady, 1884, pi. 35, fig. 8), x 118. Figs 4, 5, Paralectotype (ZF 3972); umbilical and spiral views in immersion oil, x 164. Figs 7-9, Paralectotype (ZF 3973); spiral, umbilical and oblique-umbilical views, x 168. All from Challenger station 346, South Atlantic. 178 P. BRONNIMANN&J. E. WHITTAKER Brady, which is thought worthy of formal description. This is given the name Trochammina pintoi sp. nov. and is described from Challenger station 323, at 1 ,900 fathoms (3,480 m), in the S.W. Atlantic (B.M.N.H. slide no. 1958.3.13.801-808). It is illustrated in Figs 10-14, 17. Systematics The material illustrated in this paper is housed in the collections of the Protozoa Section, Department of Palaeontology, registration numbers ZF 397 1 -ZF 3984 inclusive. Family: TROCHAMMINIDAE Schwager, 1 877 Subfamily: TROCHAMMININAE Schwager, 1877 Genus: TROCHAMMINA Parker & Jones, 1859 Type species: Nautilus inflatus Montagu, 1 808. Trochammina nana (Brady) 1881 Haplophragmium nanum Brady: 50. 1 884 Haplophragmium nanum Brady; Brady: 3 1 1 (pars), pi. 35, figs 7, 8 only. 1960 Trochammina nana (Brady) Barker: 72 (pars), pi. 35, figs 7, 8 only [after Brady, 1 884]. MATERIAL. 6 syntypes, B.M.N.H. slide no. ZF 1547, plus 2 further specimens from the type locality in the Heron-Allen & Earland Collection, ex Sidebottom 'Challenger type-slide Collection', C 12, square 6. LECTOTYPE. B.M.N.H. ZF 3971, ex slide no. ZF 1547, Challenger station 346, north of Ascension Island, South Atlantic, lat. 2°42'S, long. 14°41'W, depth 2,350 fathoms (4,300 m). Believed to be the specimen figured by Brady, 1 884, pi. 35, fig. 8. Refigured in Figs 1-3,6, by scanning electron microscopy. DESCRIPTION (LECTOTYPE). Test a loosely coiled trochospire, subcircular in umbilical and spiral views and strongly lobulate. In edge view, convex spirally, concave umbilically, well rounded at the periphery. Proloculus very large, followed by 9 subglobular chambers arranged in almost two whorls and increasing only slowly in size; 6 chambers in final whorl. Radial sutures deeply incised, straight to slightly curved. Umbilicus deep and open. Aperture an interiomarginal extra-umbilical arch with lip, without direct communication to the umbilical depression; because of loose coiling, apertures of the penultimate and antepenultimate chambers incompletely covered by the subsequent chambers and partially open. Wall agglutinated, single layered, imperforate and thin (5 urn). Agglutinant, very fine, consisting of quartz flakes and spicules of siliceous sponges. Colour light yellow-brown, except for the colourless final chamber. DIMENSIONS (LECTOTYPE). Maximum diameter 410um, minimum diameter 340 |im, maximum width 200 urn. Maximum external proloculus diameter 120um. Diameter of umbilical depression 100 urn Length of apertural opening 70 urn. VARIATION (PARALECTOTYPES). The 5 paralectotypes from slide ZF 1547 are morpho- logically similar to the lectotype. The megalospheric specimen illustrated in Figs 7-9 (ZF 3973) is smaller with a test diameter of between 230 and 290 urn. The single microspheric individual (unfigured) has 1 7 chambers arranged in just over 2 whorls with 8 chambers in the final whorl; the test diameter ranges between 320 and 440 urn and the proloculus is 20 urn across at its widest point. Specimen ZF 3972 (Figs 4, 5) has a test diameter of between 240 and 290 um and a large megalospheric proloculus of 90 um maximum diameter. A fragment of a further paralectotype, placed in dilute HC1 did not break down but turned colourless; the test is very slightly calcareous. The 'Challenger type-slide Collection' was originated by H. Sidebottom to illustrate examples of the species figured by Brady (1884) and was picked from topotypic material. REDESCRIPTION OF TROCHAMM1NA NAN A (BRADY) 1 79 Sidebottom's specimens of Trochammina nana comprise a microspheric individual with a maximum diameter of 470 um and 9 chambers in the last whorl and a megalospheric individual almost identical in size and morphology to the lectotype. REMARKS. As all the specimens of T. nana s.s. examined exhibit the same overall test morphology, it is difficult to see why Brady thought them to represent aberrant specimens. In fact they seem to form a well defined taxon without close relatives within Trochammina. The other specimen figured by Brady (1884, pi. 35, fig. 7) is dextrally coiled and cannot be recognized amongst the remaining syntypes nor in Sidebottom's Challenger type-slide. DISTRIBUTION. Known only from the type locality. Recent. Trochammina pintoi sp. nov. Figs 10-14, 17. ?1 88 1 Haplophragmium nanum Brady: 50 (tpars). 1 884 Haplophragmium nanum Brady; Brady: 3 1 1 (pars), not illustrated. DIAGNOSIS. A tightly coiled, 6-7 chambered, moderately inflated Trochammina, with a flat spiral side and broadly rounded at the periphery. NAME. In honour of Dr Jraja Damiani Pinto, of Porto Alegre, Brazil, for his contribution to the study of the microfauna, in particularly the Ostracoda, of the Brazilian Shelf. MATERIAL. 8 specimens, B.M.N.H. slide no. 1958.3.13.801-808 plus 3 further specimens from the type-locality in the Heron-Allen & Earland 'Students' Collection'. HOLOTYPE. B.M.N.H. ZF 3974, ex slide no. 1958.3.13.801-808, Challenger station 323, S.W. Atlantic (Argentine Basin), lat. 35°39'S, long. 50°47'W, depth 1,900 fathoms (3,480 m). Figured by scanning electron microscopy in Figs 12, 14, 17. DESCRIPTION (HOLOTYPE). Test a trochospire of just over two whorls, moderately inflated but tightly coiled. Subcircular and virtually non-lobulate in umbilical and spiral views. In edge view, spirally flat, umbilically convex, broadly rounded at the periphery. Chambers inflated, closely embracing, 6 in final whorl, increasing only slowly in height and width; seen umbilically and spirally chambers sub-triangular in shape. On spiral side, early whorls form a slightly sunken area due to the overlap along the spiral suture of the chambers in the final whorl. Radial sutures straight, flush on spiral side, slightly depressed on umbilical side. Umbilicus tight. Aperture, a narrow umbilical peripheral slit without a lip, apparently separated from the umbilical depression and thus of Trochammina-type. Wall rather thick, agglutinated, single layered and imperforate, made up of quartz grains. Colour uniformly light brown-yellow. DIMENSIONS (HOLOTYPE). Maximum diameter 350um, minimum diameter 310um, maxi- mum width 220 Jim. VARIATION (PARATYPES). All the paratypes have a morphology very close to that of the holotype. The paratype (ZF 3975) illustrated in Figs 10, 11, 13, has 7 chambers in the final whorl, and about 21 chambers, including the proloculus, in all (the exact number of chambers and whorls cannot be determined). The maximum and minimum diameter of the test is 350 and 300 um respectively, the greatest width 200 um. The maximum diameter of the remaining paratypes ranges from 300 to 450 um. REMARKS. Trochammina rossensis Warthin from the Ross Sea, Antarctica (Warthin, 1934: 3, text-figs 1-3), shows similarities to our new species. In edge view, although the spiral side is also very flat the periphery of Warthin's species is sub-acute; moreover, it has the well developed umbilical flaps and aperture-type of a Portatrochammina (see below). We have included, somewhat tentatively, the original reference to H. nanum (i.e. Brady, 1881) at least in part, in our synonymy of T. pintoi. This is because our new species comes from a South Atlantic Challenger station, as does the true T. nana, even though the 180 P. BRONNIMANN & J. E. WHITTAKER Figs 10-14, 17 Trochammina pintoi Bronnimann & Whittaker sp. nov. Figs 12, 14, 17, Holotype (ZF 3974); edge, umbilical and spiral views: figs 12, 1 7, x 132; fig. 14, x 135. Figs 10, 11, 13, Paratype (ZF 3975); umbilical, edge and oblique-umbilical views, x 1 32. Both from Challenger station 323, S. Atlantic. Figs 15, 16, 18, 19 Portatrochammina bipolaris Bronnimann & Whittaker sp. nov. Figs 15, 16, Paratype (ZF 3976); spiral and umbilical views in immersion oil,x 178. Figs 18, 19, Paratype (ZF 3977); spiral and umbilical views in immersion oil, x 200. Both from the Beaufort Sea (N.W. Canada), Arctic Ocean. REDESCRIPTION OF TROCHAMMINA NAN A (BRADY) 1 8 1 type-description of the latter does not seem to allow for its morphology, nor is it illustrated by Brady in any of his subsequent papers. It is more likely, however, that only later, in collating the data for his Challenger Report (1884), did Brady decide to include this form (the type-locality of T. pintoi is listed under the distribution of//, nanum: Brady, 1884, p. 3 12) within his wide concept of that taxon. DISTRIBUTION. Known only from the type-locality. Recent. Genus: PORTATROCHAMMINA Echols, 1971 Type species. Portatrochammina eltaninae Echols, 1971 Portatrochammina bipolaris sp. nov. Figs 15, 16, 18, 19-31. 1 88 1 a Haplophragmium nanum Brady; Brady: 406, pi. 2 1 , figs 1 a-c (non H. nanum Brady, 1881). 1 882 Haplophragmium nanum Brady; Brady: 99, pi. 2, figs la-c. 1 884 Haplophragmium nanum Brady; Brady: 3 1 1 (pars), pi. 35, figs 6a-c only. 1894 Haplophragmium nanum Brady; Goes: 22 (pars), pi. 5, figs 124-126 only. 1920 Trochammina nana (Brady) Cushman: 7, pi. 1, fig. 4. 19200 Trochammina nana (Brady); Cushman: 80, pi. 17, fig. 1 [after Brady, 1884, pi. 35, figs 6a-c]. 71931 Trochammina nana (Brady); Wiesner: 11 2, pi. 17, fig. 202. 71948 Trochammina nana (Brady); Cushman: 42, pi. 5, figs la-c. 1952 Trochammina nana (Brady); Phleger: 86, pi. 13, figs 31, 32. 1953 Trochammina nana (Brady); Loeblich & Tappan: 50, pi. 8, fig. 5. 1 960 Trochammina nana (Brady); Jarke: 6 1 9, pi. 4, figs 1 Oa, b. 1 960 Trochammina nana (Brady); Barker: 72 (pars), pi. 35, figs 6a-c only [after Brady, 1 884]. 1969 Trochammina nana (Brady); Vilks: 45, pi. I,figs23a, b. 1971 Portatrochammina wiesneri (Parr); Echols: 166 (pars), pi. 7, figs 2a-c (7 la-c) only (non Trochammina wiesneri Parr, 1950). DIAGNOSIS. A small compressed Portatrochammina with 6-8 chambers in the final whorl. Edge view with diagnostic flat or only slightly convex spiral side. Last formed chambers increasing rapidly in size and quite lobulate. Umbilical area wide but shallow and covered by a series of well developed overlapping flaps. NAME. Alludes to the bi-Polar distribution of this species. MATERIAL. Over one hundred specimens from various Austro-Hungarian North Polar Expedition stations to the south of Franz- Josef Land, and about 40 specimens from the vicinity of the South Shetlands and Graham Land, Antarctica (material from Discovery and Terra Nova Expeditions), both in the Museum Collections; several specimens from the Beaufort Sea, N. W. Canada, donated by Drs Schafer & Vilks. HOLOTYPE. B.M.N.H. ZF 3979, selected from about 40 specimens in slide no. ZF 1546, Austro-Hungarian North Polar Expedition station 518, off Franz- Josef Land, Arctic Ocean, depth 1 1 3 fathoms (207 m). The slide is labelled as containing specimen(s) figured by Brady, 188 la, pi. 21, figs la-c and Brady 1884, p. 35, figs 6a-c, but actual specimen(s) cannot be recognized. Figured by scanning electron microscopy in Figs 20, 2 1 , 26. DESCRIPTION (HOLOTYPE). Test a low trochospire; elongate-ovate in outline in umbilical and spiral views, final portion somewhat lobulate. In edge-view, compressed, almost flat spirally, concave umbilically and rounded to sub-acute peripherally; final chamber characteristically flat on the spiral side and strongly convex on the umbilical side. Chambers about 14 in number, radially elongate and tangentially short; 7 in the final whorl, overlapping at the umbilical ends along the radial sutures; later formed chambers increasing rapidly in size. Radial sutures straight to sinuous on the umbilical side and slightly curved on the spiral side. Shallow umbilical depression covered by the partially overlapping free umbilical flaps of the chambers of the final whorl, typical of a Portatrochammina. Aperture begins as a low interiomarginal opening near the periphery extending as a slit around the umbilical flap; last 182 P. BRONNIMANN & J. E. WHITTAKER Figs 20-31 Portatrochammina bipolaris Bronnimann & Whittaker sp. nov. Figs 20, 21, 26, Holotype (ZF 3979); umbilical, spiral and oblique-umbilical views, x!76. Figs 22, 25, Paratype (ZF 3980); edge and umbilical views, x 200. Figs 23, 29, 31, Paratype (ZF 3981); edge, umbilical and spiral views: fig. 23, x 172; figs 19, 21 x 180. Figs 24, 27, Paratype (ZF 3982); umbilical and edge views, x 168 and x 182 respectively. Fig. 28, Paratype (ZF 3983); edge view, x 1 52. Fig. 30, Paratype (ZF 3984); oblique-umbilical view, x 250. Figs 20, 21, 24, 26, 27, from Austro-Hungarian North Polar Expedition, station 518, off Franz-Josef Land (U.S.S.R.), Arctic Ocean (between lats. 79° and 80°N). Fig. 28, from Austro- Hungarian North Polar Expedition, station 516, off Franz-Josef Land, depth 130 fathoms (238m), ex Brit. Mus. Nat. Hist, slide no. 1955.10.1.312-345. Figs 22, 25, from Discovery station 177, off South Shetland Islands, Antarctica (lat. 63°18'S, long. 60°20'W), depth 1,080 m, ex Heron-Allen & Earland 'type-slide' no. 697, square 54. Figs 23, 29-31, from the Beaufort Sea (N.W. Canada) Arctic Ocean (north of lat. 70°N). REDESCRIPTION OF TROCHAMMINA NAN A (BRADY) 1 83 formed aperture, as well as preceeding apertures being in direct communication with the umbilical depression and presumably all functional. Wall agglutinated, single layed and imperforate, consisting of very fine quartz elements interspersed locally with irregular shaped, larger grains. Colour of early portion of test dark brown, grading into lighter brown, final chamber virtually colourless. DIMENSIONS (HOLOTYPE). Maximum diameter 270 urn, minimum diameter 200 um, width 70 um. Proloculus diameter 10 um. VARIATION (PARATYPE). Figs 22-25, 27-31, show both six and seven chambered forms and illustrates the variation in morphology in spiral, umbilical and edge views. Particularly variable is the chamber shape and lobulation of the final portion of the test and the length of the free portion of the umbilical flap of the last-formed chamber; in some forms this flap can almost cover the earlier flaps (e.g. Fig. 24), in others it is short, so that the flaps of all the chambers in the final whorl are exposed (e.g. Fig. 29). The maximum diameter of the specimens examined in our study varied between 150 and 320 urn. The number of whorls varies between about 2\ and 3. Two specimens donated by Drs Schafer & Vilks (ZF 3976, ZF 3977) are illustrated in a clearing medium (Figs 15, 16, 18, 19). The proloculus can be clearly seen and in both cases has a diameter of about 20 um. The wall of the final chamber is 2-3 jj.ni thick. In dilute HC1, the tests do not disintegrate nor does the colour change; the wall is very slightly calcareous. REMARKS. Unambiguous references to our new species, with good illustrations, are not common in the literature. In particular, no mention is made of the umbilical flaps in earlier papers nor can they be readily seen in many published figures; the most surprising example of this omission being Brady himself (188 la; 1884), whereas we know, from the type- material that the flaps exist. The vast majority of citations of Haplophragmium nanum or Trochammina nana auctt., nevertheless, do not belong to any of the three species revised in this paper. That the species should occur in both Polar seas and not elsewhere, is perhaps surprising, but the Antarctic form (see Figs 22, 25) would appear identical to that from the Arctic (Figs 20, 21, 23, 24, 26-31). The Antarctic specimens of P. bipolaris have been checked against the types of the species of Trochamminidae described by Parr (1950), and preserved in the South Australian Museum, Adelaide, and those of Heron-Allen & Earland (1922; 1932) and Earland (1933-36) in this institute, all from that region. None would appear to be synonymous with our new species. A specimen of the type-species of Portatrochammina, P. eltaninae Echols, from the Terra Nova Expedition, station 194 (26), collected off Gates Land, Antarctica (lat. 69°43'S, long. 163°24'E), is illustrated in Figs 32-34, for comparison with P. bipolaris. It has been checked against the holotype of Echol's species (U.S. National Museum reg. no. 687210; figured in Echols, 197 1 , pi. 8, figs la-d). P. eltaninae is a five-chambered species and is less compressed and has more globose chambers than P. bipolaris; furthermore, the umbilical flap of the last- formed chamber almost completely covers the umbilicus. DISTRIBUTION. Arctic Ocean - widespread in the Canadian Arctic (Cushman, 1920; Phleger, 1952; Vilks, 1969; Schafer & Vilks's material, herein); around Greenland (Phleger, 1952; Loeblich & Tappan, 1953); the Russian (Brady, 188 la; 1882; Jarke, 1960, and material herein) and Scandinavian Arctic (Goes, 1894; Jarke, 1960). Its most southerly substantiated record in the N. Hemisphere is from the Faroes Channel (slide marked 'cold area', Porcupine Expedition material, B.M.N.H.). Antarctic - Scotia Sea (as P. wiesneri (Parr); Echols, 1971); off the South Shetlands and Graham Land (Heron-Allen & Earland, 1922; 1932, unfigured, but verified from their collections, see herein). A Recent species only, as far as we are aware. Records exist back to the Cretaceous (e.g. Chapman, 1892: 6, pi. 5, figs 15a-c), but these are not conspecific. A shelf species, living down to about 1 ,000 m, but in Arctic waters commonly found between 1 50 and 300 m. 184 P. BRONNIMANN& J. E. WHITTAKER Figs 32-34 Portatrochammina eltaninae Echols. ZF 3978. Spiral, edge and umbilical views of same specimen, for comparison with P. bipolaris, x 142. From British Antarctic (Terra Nova) Expedition station 194 (26), off Gates Land, Antarctica, depth 180-200 fathoms (329-366 m) ex Heron- Allen & Earland 'type slide' no. 316, square 16. Acknowledgements The writers are indebted to: Drs C. T. Schafer and G. Vilks, Atlantic Geoscience Centre, Bedford Institute of Oceanography, Dartmouth, Canada, for donating specimens of P. bipolaris from the Beaufort Sea; Dr R. Cifelli, U.S. National Museum, Washington, D.C., for loaning material in his care; and to Mr J. Cann, Salisbury College of Advanced Education, Salisbury East, South Australia, for scanning electron micrographs of Parr's types. Mr R. V. Melville, International Commission of Zoological Nomenclature, kindly advised us on the taxonomic problem concerning T. nana. Mrs C. A. Whittaker took the scanning electron micrographs for this paper and these were printed by the Electron Microscope Unit of the B.M.N.H. The work of one of us (P.B.) was in part funded by the Fonds National Suisse. References Barker, R. W. 1960. Taxonomic notes on the species figured by H. B. Brady in his report on the Foraminifera dredged by H.M.S. Challenger during the years 1873-1876. Accompanied by a reproduction of Brady's plates, xxiv + 238 pp., 1 15 pis. Tulsa (American Association of Petroleum Geologists, Special publication 9). Brady, H. B. 1881. Notes on some of the Reticularian Rhizopoda of the Challenger Expedition. Part 3. 1. Classification. 2. Further Notes on New Species. 3. Note on Biloculina mud. Q. Jl microsc. Sci., London, (n.s.) 21:31-71. (January, 1881). 188 la. On some Arctic Foraminifera from Soundings obtained on the Austro-Hungarian North-Polar Expedition of 1872-1874. Ann. Mag. nat. Hist., London, (ser. 5), 8 : 393^18, pi. 21. (December, 1881). 1882. Uber einige Arktische Tiefsee-Foraminiferen gesammelt wahrend der osterreichisch- ungarischen Nordpol-Expedition in der Jahren 1872-1874. Denkschr. Akad. Wiss. Wien, 43 (2): 91-110, pis 1,2. 1884. Report of the Foraminifera dredged by H.M.S. Challenger during the years 1873-1876. Rep. scient. Results Voy. Challenger (Zool.), London, 9, xxi + 8 1 4 pp., 115 pis. Bronnimann, P. 1976. Two new genera of Recent Trochamminidae (Foraminiferida). Archs Sci. Geneve, 29:215-218. 1979. Recent benthonic foraminifera from Brasil. Morphology and ecology. Part 4. Trochamminids from the Campos shelf with description of Paratrochammina n. gen. Paldont. Z., Stuttgart, 53 : 5-25, Figs 1-10. & Beurlan, G. 1977. Recent benthonic Foraminifera from Brasil. Morphology and Ecology. Part 1. 1. Polystomammininae, new subfamily of the Trochamminidae, and description of Poly- stomammina planulata (Mikhalevitch), 1971, from the Campos Shelf. 2. Spiroplectamminoides camposi Bronnimann and Beurlan, n. gen., n. sp., from the Campos Shelf. Archs Sci. Geneve, REDESCRIPTION OF TROCHAMM1NA NAN A (BRADY) 1 85 30: 77-90, pis 1-3, (1977). Part 2. 3. Cribrostomoides Cushman and Haplophragmoides Cushman from the Campos Shelf. 4. Trochammina brasiliensis Bronnimann and Beurlan, n. sp., from the Campos Shelf, he. cit.: 243-262, pis 1-4, (1977a). & Whittaker, J. E. (in preparation). A revision of the Trochamminidae (Protozoa: Foramini- ferida) of the Discovery Reports, described by Heron-Allen & Earland, 1932-36. Bull. Br. Mus. nat. Hist., London (Zool.). Chapman, F. 1892. The Foraminifera of the Gault of Folkestone -2. Jl R. misrosc. Soc., London, 1892: 3 19-330, pis 5, 6. Cushman, J. A. 1920. The Foraminifera of the Canadian Arctic Expedition, 1913-18. Rep. Can. arct. Expd., Ottawa, 9(M): 1-13, pi. 1. 1920<2. The Foraminifera of the Atlantic Ocean. Part 2. Lituolidae. Bull. U.S. natn. Mus., Washington, 104 (2) : 1-1 1 1 , pis 1-18. 1948. Arctic Foraminifera. Spec. Publs Cushman Lab., Sharon, Mass., 23 : 1-79, pis 1-8. Earland, A. 1933-36. Foraminifera. Part 2. South Georgia. Discovery Rep., Cambridge, 6 : 27-138, pis 1-7. (1933). Part 3. The Falklands sector of the Antarctic (excluding South Georgia), loc. cit., 10 : 1-208, pis 1-10, (1934). Part 4. Additional records from the Weddell Sea sector from material obtained by the S.Y. Scotia loc. cit., 13 : 1-76, pis 1, 2, 2 A (1936). (see also Heron-Allen & Earland, 1932). Echols, R. J. 1971. Distribution of Foraminifera in sediments of the Scotia Sea area, Antarctic waters. In Reid, J. L. (ed.) Antarctic oceanology 1. Antarctic Res, Ser. Washington, 15 : 93-1 68, pis 1-16. Goes, A. 1894. A synopsis of the Arctic and Scandinavian Recent marine Foraminifera hitherto discovered. K. Svenska VtenskAkad. Handl, Stockholm, 25 (9) : 1-127, pis 1-25. Heron-Allen, E. & Earland, A. 1922. Protozoa. Part 2, Foraminifera. Nat. Hist. Rep. Br. antarct. Terra Nova Exped. (Zool.)., London, 6 : 25-268, pis 1-8. & 1932. Foraminifera. Part 1. The ice-free area of the Falkland Islands and adjacent seas. Discovery Rep., Cambridge, 4, 29 1-460, pis 6-1 7. Jarke, J. 1960. Beitrag zur Kenntnis der Foraminiferenfauna der mittleren und westlichen Barents-See. Int. Revue ges. Hydrobiol., Berlin, 45 : 581-654, pis 1-13. Loeblich, A. R. & Tappan, H. 1953. Studies of Arctic Foraminifera. Smithson. misc. Collns, Washington, 121 (7) : 1-150, pis 1-24. Parr, W. J. 1950. Foraminifera. Rep. B.A.N.Z. antarctic Res. Exped., Adelaide (ser. B, Zool. & Bot.) 5: 233-392, pis 3-1 5. Phleger, F. B. 1952. Foraminifera distribution in some sediment samples from the Canadian and Greenland Arctic. Contr. Cushman Fdnforamin. Res., Bridgewater, Mass., 3 : 80-89, pis 13, 14. Vilks, G. 1969. Recent foraminifera in the Canadian Arctic. Micropaleontology, New York, 15: 35-60, pis 1-3. Warthin, A. S. 1934. Foraminifera from the Ross Sea. Am. Mus. Novit., New York, 721 : 1-4, text-figs 1-5. Wiesner, H. 1931. Die Foraminiferen der Deutschen Siid polar-Expedition 1901-1903. Dt. Sudpol. Expd. (Zool.)., Berlin, 20 : 49-165, pis 1-24. Manuscript accepted for publication 13 November 1979 Crinoidea collected by the Meteor and Discovery in the NE Atlantic A. M. Clark Department of Zoology, British Museum (Natural History), Cromwell Rd., London SW7 5BD Introduction The paper includes new records for mostly little-known Crinoidea collected by the Meteor and Discovery in the Atlantic from off NW Spain to the vicinity of the Canary Islands and from the Great Meteor Bank, S from the Azores, the source of a new genus and species, Meteorometra monticola. Tables of numerical data for some of the less well-known species are provided. The ranges of several species are extended and affinities of some discussed. A distribution map is included. Systematic descriptions Family COMASTERIDAE Neocomatella europaea A. H. Clark Actinometra pulchella (part) P. H. Carpenter, 1888: 304-306, pi. 4, pi. 52, fig. 1; Koehler & Vaney, 1910: 32. [Non A. pulchella Pourtales, 1878.] Neocomatella europaea A. H. Clark, 1913a : 4; 1931 : 150-153; Cherbonnier, 1970a : 343, 345, 346, 347; 19706 : 1266, 1268, 1269, 1270, 1271; Sibuet, 1974 : 791. MATERIAL. Meteor cruise 9c, st. lOlb AT 33, 36°36'3'N, 14'14'W (Josephine Bank, W of Portugal); 1 70-300 metres; 5 specimens. Discovery st. 8966, 31°21-22'N, 10040'5-39-2'W (off Morocco); 686-742 metres; 1 specimen. Discovery st. 9015, 28°46'8^6-0'N, 12°47'4^6-8'W (off S Morocco); 610-637 metres; arms only. The description of this species given by A. H. Clark is limited mainly to the form of the cirri. Unfortunately all the present specimens are badly broken but some descriptive remarks can be made from them. All probably had c. 20 arms, not exceeding 40-50 mm in length. One of the larger Meteor specimens and the Discovery one both have the arm breadth at the first syzygy (which is usually 1+2) 1-1 mm. The strong, incurled cirri are one or two deep around the discoidal centrodorsal, numbering only XVIII plus ii immature ones in one specimen but XXXIII in another. There are up to 15 cirrals, laterally compressed, the distal ones broadened dorso- ventrally. The longest segment, the fourth or fifth, measures up to 1'3 mm, with length : median breadth 3'2-3'5 : 1; the total cirrus length being up to 12 mm. There are small dorsal spines on the distal segments. The division series are rounded laterally. The ossicles are not at all flared at the joints but are ornamented along their distal edges by a series of very fine rugosities. All the IIBr series are of two ossicles. On the brachials the ornamentation becomes stronger, the outer part is fluted, each small ridge leading to a distally-directed spine, though the ossicle is slightly Bull. Br. Mm. nut. Hist. (Zool.). 38 (4): 1 87-210 Issued 25 September 1 980 187 188 A.M.CLARK flared and the profile is irregular. Combs are present on the first three pinnules, which on one specimen have c. 30, c. 25 and c. 24 segments respectively, P, measures 7'0 mm. DISTRIBUTION. Northern Bay of Biscay to Western Sahara, NW Africa; 337-1 700 metres. AFFINITIES. Messing (MSc thesis, 1975) has studied a large collection of Neocomatella pulchella (Pourtales) from the West Indian area, ranging to St. Paul's Rocks, NE from Brazil. He noted that a relatively small specimen with arm length only c. 50 mm (compared with the usual 100+ length of most examples of this species collected) has XVI cirri, with 14-17 segments, measuring 10mm in length, not markedly different from the specimens of N. CRINOIDEA COLLECTED IN NE ATLANTIC 1 89 europaea mentioned above. However, already at this size, the longest cirrals in the West Indian specimen have length : median breadth only 2.1 : 1 and the cirri must be much stouter, otherwise it might be thought that the eastern Atlantic Neocomatella is only a stunted paedomorphic form of the West Indian species. Family THALASSOMETRIDAE Koehlermet ra porrecta (P. H. Carpenter) Antedon porrecta P. H. Carpenter, 1888:250-252, pi. 52, figs 3-5; Koehler & Vaney, 1910:32; Hartlaub, 1912:285. Antedon flava Koehler, 1895:475; Koehler & Vaney, 1910 : 31. Crotalometra porrecta: A. H. Clark, 19136 : 46; Mortensen, 1927 : 25. Koehlermetra porrecta: A. H. Clark, 1950 : 101-105. Koehlermetra flava: A. H. Clark, 1950 : 105-107. MATERIAL. Meteor cruise 36, st. 98 AT 149, 25'31'5'N, 16'02-2'W (off Western Sahara); 658-888 metres; 12 specimens. Discovery st. 7984, 25°27'N, 16°10'W; 811-890 metres; 2 adult specimens and 1 pentacrinoid. All the specimens are somewhat broken, often from the first syzygy, which is at 2 + 3 on arms following IIBr series of four ossicles - the usual number, or at 3 + 4 on arms based on IIBr series of two, occasionally at 1 + 2 or even not until 10 + 11. Where more than one IIBr series occur, these are usually adjacent and posterior. The centrodorsal is domed in one specimen, otherwise more or less broadly truncated, even to the extent of being discoidal. Measurements are given in table 1 . The cirri are arranged one or two deep around the sides. The sixth or seventh cirral is usually a transition segment and is the longest; it has a triangular raised patch dorsally. The division series are quite smooth except sometimes for a few small knobs at the sides. They have a flattened shoulder-like flange each side continuing distally into the base of the first pinnule (PD) on the IIBr2. The larger specimens have c. 20 arms. The arm length in one of them (the third in table 1 , taken by the Discovery) was probably 120-130 mm. This specimen has c. 12 segments in Fig. 1 Map of part of the NE Atlantic, with contours for 200, 2000 and 4500 metres, showing records of the following 1 8 species of bathyal and abyssal Crinoidea: V Neocomatella europaea A. H. Clark • Thalassometra lusitanica(P.H.C.) + Meteorometra monticola n.sp. • Trichometra cubensis (Pourtales) x Pentametrocrinus atlanticus (Perrier) O Bathycrinus gracilis Wyville Thomson H Rhizocrinus magnus Gislen CD Zeuctocrinus gisleni A. M. Clark Li Gephyrocrinus grimaldii Koehler & Bather D Koehlermetra porrecta (P. H. Carpenter) & T. omissa (Koehler) A Orthometra hibernica (A.H.C.) ES Atelecrinus balanoides P.H.C. A Thaumatocrinus jungerseni A.H.C. + Democrinus par/aid Perrier B Porphyrocrinus thalassae Roux O Anachalypsicrinus nefertiti A.M.C. Cyathidiumforesti Cherbonnier & Guille (Note: A few of the older records, e.g. from Porcupine and Travailleur have been slightly modified to conform with the depth contours.) An additional seven stalked species have been recorded from the area, namely: Annacrinus wvvillethomsoni (Jeffreys in W. Thomson) (from the Bay of Biscay to the Canary Is in 1330-2000 metres); Porphvrocrinus incrassatus Gislen (Bay of Biscay, also St. Helena, 2400-2780 metres); Conocrinus cabiochi Roux (Bay of Biscay, 1975-2070 metres); C. cherbonnieri Roux (Bay of Biscay, 330-510 metres); Monachocrinus perrieri (Koehler & Vaney) (off Morocco and the Azores, also South Africa, 1620^t600 metres); M. recuperatus (Perrier) (off Morocco and Azores, 2300-4260 metres); also IDemocrinus conifer (A. H. Clark) but the Josephine Bank and Portuguese records of A. H. Clark (1923) at least are probably conspecific with D. parfaiti. as already noted (A. M. Clark, 1977 : 177). The abbreviations for the sea mounts are as follows: AB- Ampere Bank; CB Cromer Bank; CsB Cruiser Bank; DB Dacia Bank; GB Gettysburg Bank; GMB Great Meteor Bank; JB Josephine Bank; OB Ormonde Bank; SB Seine Bank. (Gettysburg and Ormonde Banks together make up the Gorringe Ridge). 190 A. M.CLARK co .5 CJ 2 c ro 10 CO '"O C — .£> 3 *-. o J fN fN ON ON ON 1 — o") £ ^-C CO + 2 > *-" OO O — ' — ^t^|oo«r> TJ- O oo £ ^ " ^'3 00^ r-os^ ON>OVOONOVO c3 -C O0\- ^^^^ — 00 E c 'co oouoo ONONONmvot^r- .« i-J m^l-1^- O^ 0s* I* — ^^ OO | ^^ */^ ^^ rn r*"1) -^T r^ ro (N , ^ > si '+> ^'7> >~ — S"3 — v1 "* 6 _^ |~"j + T"! ^ ^x* *y* ^* r~^ ^ CO 1) z xxggxgxxxxx>^O^fN^ "§ + G Mrs) OONOOONONONONOOONOO s2^ = p W CO CJ !"1 1 a* >" — 1 + t-^^)^^r-rNfNOOO 55 N C ^ "cO (NO I >/^ I I Tf I ro ro ro fN I fO I fN — I fN I IfNIOOOOIfNIOO • I I I • fN I I I fN O — fN I — I fN I OO fN fN I fN sO v,^ • ^O */~> CO C^\ f*^ CO I ro ro >/"> Tf f"> Tf I Tf O do — I o r- m o S C f ) c « fi <^ — fN p •- rOrOf^r^rnlfNlfNfNfNfNfN1 !fNfNfN — 00 + J 2 iO»nir)mmT»-T*rKirvirNifsirornfNfNfNOOOOOOr~- — — — — — — — oooo o Z _ ^-.fslfs, CRINOIDEA COLLECTED IN NE ATLANTIC 1 97 arm width 1-2-1-5 mm the ratio is 1-6-2-7/1, on the few specimens where it can be measured, the cirri are so frequently all lost. Larger specimens have the division series and arm bases approximating and distinctly flattened laterally, as described for West Indian specimens of Trichometra cubensis by A. H. Clark (in Clark & Clark, 1967), whereas T. delicata A. H. Clark from the Bay of Biscay and Portugal, in which the arm length is not known to exceed 40 mm, is similar to smaller specimens in having these ossicles rounded laterally. The division series, brachials and pinnulars mostly have flared and spinose distal edges. The distal intersyzygial interval is usually two muscular joints. Most of the pinnules are broken; the longest intact P, found measures 7*7 mm and has 19 segments, whereas P2 and P3 in all but the smallest specimens have 10-14 and 10-12 segments respectively. AFFINITIES. The related species Orthometra hibernica, also small and with spinose distal ends to the brachials and pinnulars and recorded from W of Ireland to the Josephine Bank, W of Portugal, differs in having a much more flattened centrodorsal, without papillae, and appreciably shorter cirrals, even the longest less than half again as long as their median width. Owing to the delicacy of this species and damage to the specimens collected, the fact that West Indian specimens have been recorded as having up to only 35 rather than 42 cirrals but as many as 25 rather than 19 pinnulars in P, is probably not significant. Nevertheless a direct comparison between specimens from the two sides of the ocean is desirable. The minimum depth recorded from the western tropical Atlantic is only 210 metres (Meyer, Messing & Macurda, 1978), compared with 1300 metres in the NE Atlantic, though SW of Iceland one Ingolf station was only 31 1 metres. A. H. Clark (in Clark & Clark, 1967) recorded as T. cubensis specimens from the area S of Newfoundland, W of Greenland and off Iceland, as well as in the West Indies. Meyer et al. have suggested that there may be two species-group taxa, a northern and a southern. Zoogeographically this idea has considerable justification and accordingly the references to the //tgo//"specimens (either as T. cubensis or T. delicata} in the synonymy of T. cubensis given above are queried. The reference to the Helga specimen from west of Ireland as T. delicata by A. H. Clark (19130) is additionally queried, having been referred to Orthometra hibernica by A. H. Clark (in Clark & Clark, 1967). However, since the cirri were described as having only 1 5-20 segments, the longest as much as three times as long as wide, I think this very unlikely and the specimen could after all have been a Trichometra. One specimen from Discovery st. 7711 had its cirri clasped around the stalk of the hyocrinid Anachalypsicrinus nefertiti. Antedonid sp. cf. 'Antedon' omissa Koehler Fig. 3 ? 'Antedon omissa Koehler, 1909 : 268-269, pi. 32, fig. 10. IThalassometra omissa: A. H.Clark, 1950: 184-185. MATERIAL. Discovery st. 7984, 25°27'N, 16'10'W (off Western Sahara); 810-890 metres; 2 specimens. DESCRIPTION. The larger specimen has ten arms 45 + ?c. 10mm long, arm width at the first syzygy (Br3 + 4) 1'5 mm, length from the proximal edge of the IBr, to this syzygy 3'5 mm and to the second syzygy at 9 + 10 8'4 mm. The centrodorsal is hemispherical, 3'0 mm in diameter and 1-8 mm high, a ratio of l'7/l. The cirrus sockets number XXXVII and are spaced and irregularly arranged with long pointed papillae among them as well as on the apical area. The single remaining peripheral cirrus has 29 segments and is 1 5 mm long. The fifth and longest segment has length/median width 2-75/1; most of the cirrals are constricted medially and flared distally but the shorter distal ones have a dorsal crest, highest medially. 198 A. M.CLARK Fig. 3 Antedonid sp. cf. 'A. ' omissa Koehler. Discovery st. 7984. [The tip of the terminal claw is broken.] The scale equals 2 mm. The ossicles of the division series have a small lateral flange each side; this and the arm bases are very sharply flattened laterally and the distal edges of the ossicles, even at the syzygies, bear fine pointed spines. The more distal brachials have spinose frills along their distal edges. The second syzygy is at 9 + 10 in all but two cases, where it is at 7 + 8 or 12 + 13; the third syzygy is usually at 14+15 or 15 + 16; the distal intersyzygial interval is usually two muscular joints. P,, P2 and P3 have 19, 13 and 13 segments and measure 7'6, 5*8 and 6'0 mm. The first two pinnulars are short, the rest progressively more elongated. If breakage occurs it is usually after the second pinnular. P4, sometimes P3, is the first genital pinnule. The dorsal end of the distal edge of each pinnular is flared and bears a cluster of relatively elongated spines. The side plates in the pinnule ambulacra are partially rod-like with one or two proximal branches and a distal expansion into an irregular lattice. The covering plates are also inconspicuous and form an irregular lattice. The smaller specimen is similar, though less spinose; it has lost all the cirri. AFFINITIES. These two specimens were taken at the same station with some much larger specimens of Koehlermetra porrecta from which they were distinguished by having only ten arms, a hemispherical rather than truncated centrodorsal with numerous pointed papillae, CRINOIDEA COLLECTED IN NE ATLANTIC 1 99 no transition segment in the single remaining cirrus, distinctly spinose division series as well as brachials and pinnulars and inconspicuous side and covering plates in the pinnules. At first they were thought to be Trichometra cubensis, collected from as close as Morocco (32°14'N), though the minimum depth for that species in the NE Atlantic appears to be 1300 metres, from the present records. T. cubensis too has ten arms, division series and arm bases somewhat flattened laterally in larger specimens (arm width at the first syzygy > 1 -2 mm) and P, much the longest pinnule. However, all the specimens in the present collection referred to T. cubensis have blunt conical centrodorsals with papillae only on the apical area, XLV or more cirri (with at least 35 segments in the longest ones at arm width > 1-2 mm) arranged in fairly regular vertical lines, relatively longer division series and brachials (length to 3 + 4 from the proximal edge of IBr, 4-0-4-5 mm at width 1 -5 mm, rather than only 3' 5 mm), and only very fine bent rods in the pinnule ambulacra. There is a possibility of identity with Antedon omissa Koehler, 1909, collected nearby SE of Tenerife, Canary Is in 1330-1340 metres, although Koehler referred this species to the Basicurva-group - the Thalassometrida - prompting A. H. Clark (1950) to include the species in Thalassometra itself. The single, poorly preserved specimen had the centrodorsal hemispherical, the division series slightly flattened laterally, the brachials flared and after about the seventh with spinose distal edges and P, much larger than P2 (from the figure). However, nothing is said about papillae on the centrodorsal and the cirri are given as including only ten principal ones with others arranged above them, the number of cirrals being up to only 21; also P, and P2 have only 12 and 6 segments and the distal intersyzygial interval is given as three to five muscular joints. The ambulacral plates are not described so affinity with the Thalassometrida cannot be certain. The only clue as to the size of the holotype of A. omissa is a disc diameter of 4 mm. The centrodorsal diameter would presumably be appreciably smaller than this, perhaps 2-3 mm. In spite of the irregular position of the second syzygy occasionally, the two Discovery specimens appear to be referable to the Antedonidae and run down to the vicinity of Trichometra in the key to the subfamily Bathymetrinae by A. H. Clark (in Clark & Clark, 1967). This key is unfortunately far from satisfactory since it depends strongly on the relative cirral length and pinnular numbers, which vary to some extent with size. Hopefully more material will be forthcoming from the further collecting going on off NW Africa which should help towards a better assessment of the taxonomic position of these specimens. METEOROMETRA gen. nov. DIAGNOSIS. A genus of Antedonidae, subfamily Bathymetrinae* having a low hemispheri- cal centrodorsal on which the cirri are more or less irregularly arranged and the dorsal pole is papillose; peripheral cirri with 17-25 segments, the longest ones flared distally and about twice as long as their median widths, the distal ones about as wide as long; the ossicles of the division series and arm bases are more or less constricted medially and flared and finely spinose distally; the pinnules all consist of attenuated segments after the first two short ones of each, the distal ends of the pinnulars being markedly flared and spinose; P, is markedly and consistently shorter and more slender than P2 and the following pinnules. TYPE SPECIES. Meteorometra monticola sp. nov.f AFFINITIES. The hemispherical centrodorsal with cirrus sockets not regularly arranged, peripheral cirri mostly with more than 20 dorsally-flared segments and P, with usually 1 5-20 elongated segments ally Meteorometra with the genera included in the Bathymetrinae, of which Orthometra, Hathrometra, Phrixometra and Trichometra are recorded from certain parts of the Atlantic. * Since the taxa which have been included in the Bathymetrinae and Zenometrinae include all intermediate arrangements of the cirrus sockets and shapes of the centrodorsal from completely irregular on a low hemispherical ossicle in Bathymetra to well-marked vertical columns on a conical or columnar centrodorsal in Zenometra, I consider that these two subfamilies should be merged. i" Monticola - dweller on a mountain. 200 A. M. CLARK Orthometra with O. hibernica recorded below (p. 203) is easily distinguished by the low, flattened centrodorsal, lacking apical papillae, and the relatively numerous and short cirrals (25-32 in peripheral ones, the longest with length/median width less than 1-5/1). Hathrometra with two species, H. tenella (Retzius) from NE America and H. sarsi (Diiben & Koren) from Norway to Greenland (but not W or S of the British Isles), is a much more delicate genus with very attenuated cirri and P, much longer than P2. Phrixometra - predominantly a Southern Ocean genus though the type-species, P. longipinna (P. H. Carpenter), ranges N to the River Plate (c. 37°S, 54°W) - also has P, usually the longer pinnule, or at least it is similar in proportions to P2, unlike that of Meteorometra; it is remarkable in being viviparous, females having brood pouches alongside the gonads. Also the division series and arm bases in Phrixometra are constricted and spaced laterally, not closely aligned. Trichometra with a Hawaiian species as well as T. cubensis above, (p. 195), is probably the closest to Meteorometra, also having the centrodorsal papillose apically and most of the ossicles flared and spinose or rugose but again it differs in having P, stouter and more elongated than P2, while the cirri and cirrals are more numerous (XL or more, the longest rarely with less than 30 segments). The other bathymetrin genera with cirrals of comparable proportions include Boleometra (from Japan), Nepiometra (from the Philippines) and Tonrometra (from the East Indies and Southern Ocean). All these have P, larger than P2 and indeed the reversal of proportions of these two pinnules in Meteorometra appears to provide a very basic distinction. Meteorometra monticola sp. nov. Fig. 4 MATERIAL. Meteor cruise 9a, st.-AT 88, 29°55'N, 28°20'W (Great Meteor Bank); 300 metres; 2 specimens. Cruise 9c, st. 148a AT 55, 30°06'7'N, 28°27'W; 323-3 14 metres; 19 specimens. Cruise 9c, st. 1 59a AT 60, 29°50'2'N, 28°29-8'W; 308-3 10 metres; 14 specimens. Cruise 9c, st. 172 BSN 24, 29°47-9'N,28°23-3'W; 300-310 metres; 1 specimen. Cruise 9c, st. 1 80a AT 77, 30°0 1 • 1'N, 28°24'W; 3 1 5-320 metres; 6 specimens. Cruise 9c, st. 186 KT 82, 29°55'9'N, 28°35'W; 308-330 metres; 1 specimen. Cruise 9c, st. 1 89 KT 84, 30°05- 1 'N, 28°38-4'W; 340-305 metres; 2 specimens. Cruise 19, st. 126 DD 90/91, 30°09-5'N,28°29-5'W; 342-357 metres; 1 specimen. Cruise 19, st. 129 AT 96, 30'01-5'N, 28°30'W; 292 metres; 5 specimens. Cruise 19, st. 131 AT 97, 30°08'N,28°38-5'W; 330-473 metres; 8 specimens. Cruise 19, st. 133 DD 100, 29°49-5'N,28°19-5'W; 315-326 metres; 1 specimen. Cruise 19, st. 133 AT 101, 29°48'5'N, 28°22'5'W; 306-308 metres; 3 specimens. Cruise 19, st. 134 AT 104, 29°46'N, 28°29'W: 320-620 metres; 3 specimens. Cruise 19, st. 136KT 106, 29°59-5/N,28°23/W: 307-337 metres; 2 specimens. Cruise 19, st. 137 KT 107, 30°03'N, 28°23'W; 307-337 metres; 2 specimens. Cruise 19, st. 140 AT 109, 30°03'N,28033-5'W; 295-3 10 metres; 2 specimens. Cruise 19, st. 157 KT 1 15, 29°55-5'N,28°30'W; 292-297 metres; 5 specimens. Cruise 19, st. 1 57 KT 1 16, 29°58'N, 28'20'W; 322-422 metres; 4 specimens. Cruise 19, st. 163 KT 123, 30°09'5'N, 28°32-2'W; 307-337 metres; 1 specimen. Cruise 1 9, st. 1 64 KT 1 24, 30°06'N, 28°28'4'W; 307-3 1 7 metres; 2 specimens. TYPE MATERIAL. All the specimens are small and more or less broken. The holotype is the best specimen from cruise 19, st. 1 57 KT 1 1 5; the paratypes are the others from this sample and the specimens from cruise 9c, sts 148a, 180a and 189, most of which are included in table 4. DESCRIPTION. The holotype is specimen 7 in table 4. The arms are all broken by 27 mm and may not have measured more than 40 mm when complete. The centrodorsal is hemispherical, diameter/height 1-9/1-1 mm = 1-7/1; the cirri are CRINOIDEA COLLECTED IN NE ATLANTIC 201 Fig. 4 Meteorometra monticola gen. & sp. nov. Centrodorsal from specimen 1 3 in table 4, Meteor st. 148a, otherwise from the holotype, st. 157, with enlargements of a brachial and a pinnular to show the ornamentation. The scale for the main figure equals 2 mm. irregularly arranged around the sides and the apical area is papillose. There appear to be only c. XXV cirri, the minimum observed in the type series. The peripheral ones have 19-23 segments of which the fourth and longest has length/median width up to 2-0/1 . The ossicles of the division series and arm bases are markedly constricted medially and flared distally, where they are armed with very fine close serrations. Beyond the arm bases these serrations become coarser, fewer and more spaced, elongating into spines. The second syzygy is usually at Br9 + 10 and the distal intersyzygial interval is two, occasionally three, muscular joints. The first pinnule is markedly more slender and somewhat shorter than P2 and the following pinnules; P2 is the first genital pinnule. After the first two short segments, the remaining pinnulars become progressively more attenuated; the distal ends of all but the terminal one are markedly flared and spinose. The frequent breaking point for the distal pinnules is after the second pinnular. There appear to be no rudiments of side or covering 202 A. M.CLARK - •C — ON — "SoCu r~ •/•> CO r*^ ^o o^ co i */^ */^ ^f •— 8 '2 fft. .— — ^ ~^ ---- --•*• 1 o _c ^ •o 00 *O fO CN fN ^ r~- ^^ r** rs ^^ ^> oo i^^ i ro r^ os ^^ ON C c ON ON >/-> SO 0< 3 "/^ \o ^o */^ t^^ t*** w*i ^o * ^^ ^^ f**i ^f ON ON ^I-J d d CO 00 cu" ^™ m O •o et a u C^l O> CO so C N fN cOTt0o — rf § O ON OO t^ r^ do oo — oo dor~-t~-sbTrTt i """ i—i d d d " d d ^ oT 0 o" O SO CO CO VD 1 J SO SO so CJ so d d CO CO g cu 0 ^ c^« fN _i_ tfc C/5 .j_ *-^ * "i 00 r^ — oo + "™ so t-~ O i/~> so co 1 o fN r~ o r-~ oo O C/5 c ON Q OO ON do r~ r-- ON do do so >o 1 O ,3 oo g t/2 £ °? in t/5 co 10 — Tf co co co TJ- co fN co oo — < ON t — i 2 1— '— X rN rs rs CN fN fN fN (NfNfSfN — >S xx§Sxxx> vOxXv-v-^r, !x"XO xg xxx 1 x X X X X -,: X o c X — CO d. t/3 13 "e e c 0-5 oooNOoooor^r^ONONO so — en ^1 ^ 1 rsi - fN - fN fN C "o ^ u oo g fN .2 g e OfNOOOCNONt^l^OO ONOOOOI^ >O I CO — • 7 o CJ ^ 0 W-) — Tj- - 3- r- oo t- OfN *O ^' ^ so ^* ^ -J o -4t 5 Q. g g >, c ON i "1 i<£ 10 f^ f^N ^^ ( >£££ */~l f*^ */~l 10 10 I/") >/") 10 10 f^l {**) f^ 5 Wj JS •^ <-• oooooooooooooooooooo o """" QJ ^^ w so .£ O •^ -— ' (A J-** (U ^ */"} 00 US — c d ^^ ^^^ ^^ — ^^ ^^ . — ^^ ^^ — 05 L^ *-; Z — " fN CO ^T o so r^- oc ai CRINOIDEA COLLECTED IN NE ATLANTIC 203 plates in the pinnule ambulacra, but on the disc the ambulacra are bordered by opaque plates. PARATYPES. The maximum arm length was probably c. 50 or perhaps 55 mm. The mean diameter/height of the centrodorsal in 15 specimens is 1-8/1. In the largest specimens the longest cirrals are relatively slightly shorter than in the holotype, with length/median width c. 1-8/1. Some of the cirrus sockets in most specimens are arranged in vertical lines but this is much less regular than in the specimens referred to Trichometra cubensis. P, is always markedly smaller than P2, the mean inverse length ratio in ten specimens being 1- 5/1. The two largest specimens (1 and 2 in the table) give an idea that P3 increases allometrically with respect to P2 since the lengths of these two pinnules are almost the same, where the smaller specimens tend to have a P2/P3 length ratio of 1 -2-1 -5/1 . Several samples included also material of the hydroid Polyplumaria flabellata G. O. Sars, which range from the Congo (Zaire) N to Iceland in 30-800 metres. I am indebted to Dr. P. F. S. Cornelius for this information. Some specimens have their cirri clasping branches of a gorgonian. DISTRIBUTION. On present evidence Meteorometra monticola appears to be endemic to the Meteor Sea-mount but it is also likely to occur on the nearby Cruiser Sea-mount (c. 32°N, 28°W) and, like Cyathidiumforesti (see p. 209), found on the slopes of the Meteor Bank, may even be found around the islands of the Azores at a similar depth of c. 300 metres, though the greater part of the Azores Plateau exceeds 1 500 metres. Orthometra hibernica (A. H. Clark) Trichometra hibernica A. H. Clark, 1913a : 2-3. Orthometra hibernica: Clark & Clark, 1967 : 680-682; A. M. Clark, 1970 : 34-36, fig. 11. [Non O. hibernica: A. H. Clark, 1923 : 57 = Heliometra glacialis]. MATERIAL. Meteor cruise 9c, st. Ill KD 39; 36°53'3'N, 14°24-7'W (Josephine Bank, W of Portugal); 960-980 metres; 1 specimen. Although the arms are all detached from the first syzygy, some of the proximal pinnules remain intact and allow a description of these appendages for the first time. The arm width at 3 +4 is 0*75 mm and the length to 9 + 10 5'75 mm. The total arm length probably did not exceed c. 20 mm. The centrodorsal is very low, as in the paratype figured in 1970. The longer cirri have c. 30 short segments, most with a rounded transverse dorsal 'ridge' which is approximately median on the very short distal segments. The division series and proximal brachials have finely serrated distal edges; on the remaining brachials the fine spines become fewer but longer. Some proximal pinnules are as follows: PI with 13 segments is 3*2 mm long P2 with 12 segments is 2*7 mm long; another with 1 1 segments is 2- 5 mm P3 with 10 segments is 2*7 mm and is slightly thicker, likewise P4 with 10 segments 2-5 mm. The first two pinnulars are short but the following ones become attenuated and are flared and spinose at their distal ends. The more distal pinnules are probably relatively longer with more expanded joints. DISTRIBUTION. This record provides an extension of range to the south from the Bay of Biscay NW from Brittany and confirms a maximum depth of more than 750 metres; it is otherwise known only from W of Ireland (53-54°N, 1 IfW, 698-749, ?9 14, metres). Family PENTAMETROCRINIDAE Pentametrocrinus atlanticus (Perrier) Eudiocrinus atlanticus Perrier, 18836 : 725; Koehler, 1909 : 271-274, pi. 32, figs 204 A. M. CLARK Eudiocrinus (Pentametrocrinus) atlanticus: Koehler & Vaney, 1910:31. Pentametrocrinm atlanticus: A. H. Clark, 1913fl : 4; 1923 : 44; Mortensen, 1927 : 24, fig. 1 1; Clark & Clark, 1967 : 790-794; Messing, 1978: 700-708, figs 1-5, 8-11, 13-16; Meyer, Messing & Macurda, 1978:423. Pentometacrinus atlanticus: Cherbonnier, 1970a : 343, 348. Pentametacrinus atlanticus: Cherbonnier, 19706: 1266, 1268. MATERIAL. Discovery st. 7967, 29°20'N, 12°15'W (NE of Lanzarote, Canary Is); 1576-1539 metres; 1 specimen. Discovery st. 8519/7, 24002'N, 16°59'W (off Western Sahara); 1037-994 metres; 15 specimens. Discoverysi. 8931;24°58-8'-25°05-4/N, 16°3H-32-2'W; 955-1012 metres; 1 specimen. Discovery st. 9042, 42°15-0-17-8'N, H°22-0-19-7'W(offNW Spain); 1662-1541 metres; 3 specimens. / As usual all the specimens collected of this distinctive five-armed species are more or less badly broken. One of the largest (from st. 9042) has the centrodorsal diameter 4-0 mm and height 2-0 mm, the shape being low hemispherical. All the LXVIII cirri are broken. In contrast, a specimen from st. 7967 with centrodorsal diameter/height 3*2/1 '3 mm and almost discoidal in shape, has only XIX cirrus sockets in one irregular peripheral ring. Messing gives a range of the diameter/height ratio from 1-4-3-6/1 and a maximum cirrus number of c. XC so clearly there is a great range in the centrodorsal shape and cirrus number. A detached cirrus 20 mm long from st. 8519 has 18+ segments, the distalmost one broken so at least one more when intact; length/width of the fourth and longest segment is 2-3/0-4 mm or nearly 6/1. There is also a more slender complete detached cirrus 21-5 mm long with 19 segments. A. H. Clark (in Clark & Clark, 1967) gave the maximum cirral number as 17 and maximum length as 35 mm but Messing (1978) queried the length as an error of interpretation. Three specimens from st. 8519 are unusual in having the first pinnule sometimes before the first syzygy (brachials 4 + 5); two of them have a P, on the right side of Br2 and a Pa on the left of Br3 on one arm only, while the third has a Pa on the left of Br3 on two arms. The centrodorsal diameter in these three is 3- 1 , 2-75 and 2-6 mm and their arm widths at 4 + 5 are 1-75, 1-5 and 1-5 mm respectively. Otherwise the first pinnule is P2 on Br5, as usual in P. atlanticus. Only P. varians (P. H. Carpenter) from the Maldive area to Japan regularly has P, and Pa present. These localities are within the known range of the species which extends in the NE Atlantic from SW of Ireland to NW of Africa and in the W Atlantic from off Martinique (Lesser Antilles) to the Blake Plateau N of the Bahamas. Thaumatocrinus jungerseni A. H. Clark Thaumatocrinus jungerseni A. H. Clark, 1915: 149, 150; 1923: 13-17, figs 2-4; Clark & Clark, 1967:782-785. MATERIAL. Discovery st. 7711/57, 52°53'3'N, 19°52'W (far W from southern Ireland); 2742-2734 metres; 6 specimens. Discovery si. 771 1/62, 52°50'0'N, 20°02'8'W; 2727-2720 metres; 18 specimens. Discovery st. 771 1/85, 53°05'8'N, 19055'7'W; 2652-2626 metres; 2 specimens. The specimens collected of this distinctive species with its ten unbranched arms arising from ten radials are more or less badly broken, particularly with regard to the cirri. However, some of the Discovery specimens exceed in size the type-material from the Ingolfcol lections from W of Iceland for which the maximum arm length was estimated by A. H. Clark (1923) as 1 10 mm and the centrodorsal diameter 3*0 mm. A specimen from Discovery st. 771 1/85 has several arms almost intact, measuring c. 130 + ?c. 5 mm; since this has the centrodorsal diameter 3'0 mm and the arm width at the first syzygy (4 + 5) 1-4 mm, a larger broken specimen with these measurements 4-4 and 1-7 mm may have had arms as much as 200 mm long. CRINOIDEA COLLECTED IN NE ATLANTIC Table 5 Thaumatocrinus jungerseni A. H. Clark. Numerical data from ten specimens from Discovery st. 77 1 1/62 and one (no. 7) from st. 77 1 1 785. 205 Arms Width Length Centrodorsal Cirrus at4 + 5 to4 + 5 diameter number No. 1-7 4-2 4-4 XVII 1 1-3 4-25 3-0-3-3 XIII 2 1-5 3-8 3-3-3-6 XVI 3 •3 4-2 3-1 IX 4 •6 3-75 3-1 XIII 5 •4 3-9 2-9-3-3 XIV 6 •4 3-8 3-0 XVIII 7 •2 3-75 3-7 XIV 8 •25 3-75 3-2 XIV 9 1-2 3-75 2-5-2-9 X 10 1-0 3-3 2-25-2-4 XI 11 The size of the peripheral cirri is very variable and some of the sockets are markedly projecting so that the outline of the centrodorsal is often rather irregular, resulting in a range of diameter according to the aspect, as given for some specimens in table 5. All the cirri are broken at or before the fifth cirral. The larger cirri of larger specimens have length/median width of the first three cirrals less than 1/1. In the tenth specimen in table 5 this ratio in a stout cirrus is 1-0/1 , while the smallest (last) specimen in the table has even its stoutest cirrus with a third cirral ratio of 1- 5/1 and a slender one with as much as 3-2/1. As I noted in 1967 (in Clark & Clark), the relative length of the proximal cirrals, as well as the number of cirri, vary too much for a practicable distinction between T. jungerseni and T. renovatus P. H. Carpenter from the Southern Ocean (near the Crozet Is, S of Australia and off the Antarctic continent). However, the position of the first genital pinnule will probably serve to distinguish betwen them. In T. renovatus this pinnule is Pa, P2 or Pb, whereas in the Discovery specimens of T. jungerseni it is usually P3 or Pc, sometimes P4 and only rarely Pb. A few specimens have the proximal pinnules intact; the largest in table 5 has the segment number and length of P,, P2andP3 : 34, lO'O mm; 22, 9*1 mm and 32, 12*1 mm, P3 being the first genital pinnule. A specimen with width at the first syzygy 1-4 mm gives counts of 22, 7- 1 mm, 24, 7'9 mm and 23, 8'3 mm. It should be noted that the position of the first arm syzygy is rather variable in this material. In c. 50% of cases it is at Br4 + 5 but may also be at 3 + 4 or 5 + 6. One specimen was seen to lack P, on seven of the ten arms. DISTRIBUTION. These new records provide an extension of range somewhat to the southward in the North Atlantic, the type material being from the Denmark Strait and from SW of Iceland (c. 62°N, 30°W); the depth range is slightly extended beyond 2075 metres to 2734 metres. Family BATHYCRINIDAE Democrinus parfaiti Perrier Democrinus parfaiti Perrier, 1883a : 450-451; Gislen, 1938:28-29; 1947:7-9, fig. 3; Roux, 1977 : 39-40, figs 4, 9, 10, 1 1 , 16, pi. 2, figs 6-8, pi. 5, figs 1-6; A. M. Clark, 1977 : 1 72-1 77, fig. 3. MATERIAL. Meteor cruise 36, st. 127 FD 181; 33°40'N, 08°55'W (off Morocco); 988 metres; 1 specimen lacking arms beyond the first brachial. 206 A.M.CLARK Table 6 Bathycrinus gracilis Wyville Thomson. Numerical data from the 7 Meteor specimens for comparison with table 1 of A. M. Clark (1977); no. 5 is from st. 24, no. 2 from st. 30 and the rest from st. 38. The height of the radials cited is the mid-radial side view (i.e. slightly foreshortened). CALYX: 1 2 3 4 5 6 7 Height (basals and radials) — 1-7 2-0 — — — — Basals height — 0-50 0-58 — — — — Basals lower width — 0-70 0-75 — — — — Radials height 1-08 1-10 1-46 1-50 1-75 1-90 2-42 Radials lower width (A) 0-75 0-83 0-92 1-42 1-67 1-42 1-50 Radials upper width (B) 1-83 1-90 2-25 3-25 3-33 3-25 3-83 A as % B 41 44 41 44 50 44 39 POST-RADIAL SERIES: Crown length (incl. radials) c. 18 18+ 20+ — 20++ c.40 c.65 IBr, and 2 length 2-42 2-60 3-00 3-42 3-58 3-75 4-33 IBr,_2 width 1-33 1-42 1-50 — 2-58 2-33 2-92 IBr,-Br4 length 5-08 5-58 6-20 7-10 7-70 7-70 8-60 Br4+5 width 0-58 0-67 0-71 0-92 1-17 0-87 1-21 Bathycrinus gracilis Wyville Thomson Fig. 5 Bathycrinus gracilis Wyville Thomson, 1872 : 772-773; P. H. Carpenter, 18846 : 243-245, fig. 16, pi. 8a, figs 1-3; Koehler, 1909 : 254; A. M. Clark, 1977 : 164-167, fig. la-f. MATERIAL. Meteor cruise 3, st. 24, 42°17'2'N, 14°46'3'W (W of Cape Finisterre); 5270 metres; 1 crown. St. 30 AT 3, 42°55-4'N, 14°07'9'W; 5260 metres; 1 specimen, missing most of the stalk. St. 38 AT 5, 42°10-7'N, 14°20-8'W; 5275 metres; 4 crowns and one specimen missing most of the stalk. This collection has at last yielded well-developed specimens (unfortunately only two) of Bathycrinus gracjlis in which the basal ring is still attached to the radials and the crown is not in process of regeneration. Sadly no more than 1 1 proximal columnals remain, all of them discoidal. In both these specimens the junction between radial and basal rings tends to form an angle of c. 1 50°, the basal ring itself tapering downwards slightly but to a much less marked degree than the radials. The degree of splay of the radial ring can be expressed by the angle off the vertical derived from the vertical height and top and bottom diameters of the ring. It ranges from 24-3 1 ° in the seven specimens from the Meteor compared with 30 and 3 1° in the two Chain specimens included in table 1 of A.M.C. (1977) and 30-37° in the three specimens of B. aldrichianus in the same table. Angles of 26^0-320 are obtained for the three specimens of B. aldrichianus for which Macurda & Meyer (1976) cite measurements of the top diameters of both basal and radial rings, although the degree of splay appears to be greater in the crown shown in their fig. 1A. (Unfortunately they only cite measurements of the lower widths of individual radials of their detached crowns, some of which were of larger size - maximum radial ring diameter 2'41 mm compared with 1*56 mm for the largest one with basal ring attached.) Macurda & Meyer note that the basal ring in B. aldrichianus is only partially fused, sutures being distinguishable in the lower part. The same is true in the specimen of B. gracilis from Meteor st. 30. Absence of fusion of the basal ring in the adult has been supposed by Gislen (1938) to be characteristic of Monachocrinus A. H. Clark, 1917, which is also supposed to differ from Bathycrinus in having a curved rather than angular junction between radial and basal rings, as well as in having non-muscular and muscular joints regularly alternating throughout the arms. These characters are discussed and doubt thrown on their generic weight in a paper on stalked crinoids now in the press (A. M. Clark, Rec. Aust. Mus.). CRINOIDEA COLLECTED IN NE ATLANTIC 207 Fig. 5 Bathycrinus gracilis Wyville Thomson. Meteor st. 30. The scale equals 2 mm. The Meteor specimens agree closely with the two crowns taken SW from Ireland by the Chain and described in 1977 (A. M. Clark), having low but sharp keels and lateral flanges on the division series and first brachials, giving way to fluting on the following ossicles, which are flared at their distal ends, making a slightly serrated profile to the arm (fig. 5). Though indistinct, the non-muscular joints are usually at 1+2,4 + 5, 7 + 8, 10+11 and then alternate regularly with muscular joints but one arm of the third specimen in table 6 appears to have 1+2,6 + 7 + 8,9 + 10+11. Because of variations in the positions of these joints, the first pinnule may be on brachial 10 or 12, though it is usually on Brl 1. Out of 38 arms intact to the first pinnule, in 24 it is at Brl 1, in 9 at Brl2 and in 5 at BrlO. Since Macurda & Meyer record their specimens of B. aldrichianus as having the first pinnule from Br8 to Brl 1, the mean position of this pinnule may justify maintaining distinction of these two taxa at the specific level. No other differences in the calyx or crown are evident but it is possible that material of B. gracilis with intact stalks will reveal differential characters. DISTRIBUTION. Apart from the record from off the U.S.A. of Bathycrinus serratus A. H. Clark, 1908, reduced to a synonym of B. aldrichianus by A. H. Clark (1949), all the records of crowns or intact specimens which can confidently be referred to B. aldrichianus are from the equatorial or South Atlantic in 3300-5600 metres. B. gracilis is recorded from SW of Ireland (c. 50°N), between the Azores and Portugal (c. 39°N, 21°W) and now from W of northern Spain, in 4430-5275 metres, which appears to show a distribution limited to the West European Basin. However, the absence of intermediate records may be an artefact of collection, considering the great depths involved. Family PHRYNOCRINIDAE Zeuctocrinus gisleni A. M. Clark Zeuctocrinus gisleni A. M. Clark, 1973 : 277-281, fig. 5, pi. 2; Roux, 1977 : 34, figs 6, 7, pi. 1, figs 6-8, pi. 3, figs 1-7. MATERIAL. Discovery st. 8511/2, 41°49'N, 1 1'06'W (off NW Spain); 2574-2584 metres; 1 specimen. DISTRIBUTION. This record provides a small extension of range SW from Roux's Thalassa 208 A. M. CLARK sample in the southern Bay of Biscay (c. 44°N, 8°W) and takes the depth range beyond the 2380-2432 metres of the type (and only other known) locality, W of Ireland (c. 53°N, 20°W). Porphyrocrinus thalassae Roux Porphyrocrinus thalassae Roux, 1977 : 34-38, figs IB, 2, 3, 5, 1 1 A, 12, 13, 14, 16B, pi. 1, figs 1-5, pi. 7, figs 1-6, pi. 8, figs 1-6, pi. 9, figs 1-6; A. M. Clark, (in press). MATERIAL. Discovery st. 8511/2, 41°49'N, 11°06'W (offNW Spain); 2574-2584 metres; 1 specimen. The contemporaneous discovery by the Thalassa and Discovery of this species with its secondary arm branching comparable with that of Phrynocrinus nudus A. H. Clark from Japan, helps to justify the synonymizing with Phrynocrinidae by Roux (1977) of the nominal family Porphyrocrinidae, which I split off from the Bathycrinidae in 1973 to accommodate Porphyrocrinus Gislen and Naumachocrinus A. H. Clark on the basis of their fused plate rather than dendritic stalk attachment. In 1977 (p. 168) I noted that this character alone is not of sufficient weight to justify a family distinction, prompted by new observations on stalk attachments in the Bathycrinidae. Partly on the basis of the fine structure of joints, Roux also cast doubt on my grouping of Zeuctocrinus, in which the second post-radial ossicle is a primary axillary, with Phrynocrinus, where any axillaries are secondary, irregular and asymmetrical, rather than with Bathycrinus. This may well be important but at the same time the absence of discoidal proximal columnals in fully grown specimens of Zeuctocrinus forms a strong resemblance to Phrynocrinus. Possibly even Phrynocrinidae should be merged with Bathycrinidae, as Roux (1977) implied, though in his summary classification of the order Millericrinida (1978) he listed Phrynocrinidae including the four genera mentioned in this discussion. Clearly further study is needed on the affinities of these taxa to settle the best disposition. Further remarks on Porphyrocrinus thalassae will be published in the proceedings of the 1978 Conference on Echinoderms in Sydney (A. M. Clark, Rec. Aust. Mus. [in press]). DISTRIBUTION. This record extends the range slightly SW from the type-locality in the Bay of Biscay (c. 44°N, 07°W). Family HYOCRINIDAE Gephyrocrinus grimaldii Koehler & Bather Gephyrocrinus grimaldii Koehler & Bather, 1902 : 68-79; Koehler, 1909 : 256-264, pi. 1, fig. 12, pi. 32, figs 1-9; Roux, 1977:31. MATERIAL. Discovery st. 9042, 42°1 5'0-1 7'8'N, 1 1°22-0-19-7'W (off NW Spain); 1541-1662 metres; 3 specimens. A transparency by Dr P. J. Herring verifies Koehler's observation of the colour (1909) as bright yellow in life. DISTRIBUTION. This locality is intermediate between those of the type material, near the Canaries and Madeira and the Thalassa record from the Bay of Biscay (c. 49°N, 1 1°W) and is also within the known depth range of 1470 (71420)- 1970 metres. Appendix For the sake of complete coverage of the Meteor crinoid collections, the following record of a species of the relict genus Cyathidium from the Great Meteor Bank, admirably described by Fechter (1973), should be mentioned. Sadly the going to press of this paper overlapped with that of a prior description of the same species by Cherbonnier & Guille (1972) based on material from the Azores in 380-900 metres. CRINOIDEA COLLECTED IN NE ATLANTIC 209 Family HOLOPODIDAE Cyathidium foresti Cherbonnier & Guille Cyathidium foresti CherbonmeT&Guilte, 1972 : 2 193-2 196, pi. 1. Cyathidium meteorensis Fechter, 1973 : 162-169, figs 1,2. MATERIAL. Meteor, 1967, st. 173, 29°42'2'N, 28'20'5'W (Great Meteor Bank); 850-580 metres. St. 1 78, 29°57-9'N, 28° 1 5'7'W; 690-6 10 metres. A total of 20 specimens. Acknowledgements I am indebted to Dr H. Fechter of the Zoologische Sammlung des Bayerischen Staates, Munich and Dr A. L. Rice of the Institute of Oceanographic Sciences, Wormley, for access to the material. References Carpenter, P. H. 1884a. On the Crinoidea of the North Atlantic between Gibraltar and the Faroe Islands. Proc. R. Soc. Edinb. 1883-84 : 353-378. 1884ft. Crinoidea. 1. The stalked crinoids. Rep. sclent. Results Voy. Challenger. Zool. 11(32): 1-440, 62 pis. 1886. On the variations in the form of the cirri in certain Comatulae. Trans. Linn. Soc. Lond. 2 : 475-480, 1 pi. 1888. Crinoidea. 2. Comatulae. Rep. scient. Results Voy. Challenger. Zool. 26 (60) : 1-399, 70 pis. 1 89 1 . Notes on some crinoids from the neighbourhood of Madeira. J. Linn. Soc. Zool. 24 : 64-69. Cherbonnier, G. 1970a. Echinodermes recoltes par la Thalassa au large des cotes ouest de Bretagne et du Golfe de Gascogne (3-12 Aout, 1967). Bull. Mus. natn. Hist. nat. Paris 41 [1969] : 343-36 1 , 6 figs. 1970ft. Echinodermes recoltes par la Thalassa au large des cotes d'Espagne et du Golfe de Gascogne (18-2 5 Octobre, 1968). Bull. Mus. natn. Hist. nat. Paris 41 : 1266-1277, 2 figs. & Guille, A. 1972. Sur une espece actuelle de crinoide cretacique de la famille Holopodidae: Cyathidium foresti nov. sp. C. r. hebd. Seanc. Acad. Sci., Paris (Ser. D) 274 ( 1 5) : 2 1 93-2 1 96, 1 pi. Clark, A. H. 191 la. The recent crinoids of the coasts of Africa. Proc. U.S. natn. Mus. 40(1808) : 1-51. 191 1ft. Note sur les crinoides actuels du Museum d'Histoire naturelle de Paris. Bull. Mus. natn. Hist. nat. Paris 17 : 243-260, 2 figs. 19130. On a collection of recent crinoids from the waters about Ireland. Scient. Invest. Fish Brch Ire. 1912(4): 1-5, 1 9 1 3ft. Notes on the recent crinoids in the British Museum. Smithson. misc. Collns 61(15): 1-89. 1915. Die Crinoiden der Antarktis. Dt. SudpolExped. 16, Zool. 8 (2) : 101-209, 9 pis. 1923. The Danish Ingolf-Expedition. 5. Crinoidea. Dan. Ingolf-Exped. 4 (5) : 1-58, 58 figs. 193 1 . A monograph of the existing crinoids. 1 . The comatulids. 3. Superfamily Comasterida. Bull. U.S. natn. Mus. 82 (1)3 : 1-816, 82 pis. 1949. Echinoderms from the mid-Atlantic dredged by the Atlantis in the summer of 1948. J. Wash. Acad. Sci. 39 : 37 1-377. 1950. A monograph of the existing crinoids. 1. The comatulids. 4c. Superfamily Tropiometrida. Bull. U.S. natn. Mus. 82 (l)4c : 1-383, 32 pis. & Clark, A. M. 1967. A monograph of the existing crinoids. 1. The comatulids. 5. Suborders Oligophreata (concluded) and Macrophreata. Bull. U.S. natn. Mus. 82 (1)5 : 1-860, 53 figs. Clark, A. M. 1970. Echinodermata: Crinoidea. Marine Invertebrates of Scandinavia. No. 3. 55 pp., 19 figs. Oslo. 1973. Some new taxa of recent stalked Crinoidea. Bull. Br. Mus. nat. Hist. (Zool.) 25 : 267-288, 6 figs, 2 pis. 1977. Notes on deep-water Atlantic Crinoidea. Bull. Br. Mus. nat. Hist. (Zool.) 31 : 157-186, 5 figs. [In press] Inter-relationships of recent stalked non-Isocrinid Crinoidea. Rec. Aust. Mus. Fechter, H. 1973. Cyathidium meteorensis spec, nov., ein neuer Crinoide aus der Familie Holopodidae. Helgolander wiss. Meeresunters. 25 (1) : 162-169, 2 pis. 210 A.M.CLARK Gislen, T. 1938. A revision of the recent Bathycrinidae with a study of their phylogeny and geographical distribution. Ada Univ. lund. N.F. 34 (10) : 1-30, 18 figs. 1947. The echinoderms collected by the Skagerak Expedition in the eastern Atlantic, 1946. 2. Crinoidea. Goteborgs K. Vetensk.-o. vitterhSamh. Handl. B5 (10) : 1-9, 3 figs. 1955. West african crinoids. Atlantide Rep. 3 : 83-92, 3 pis. Hartlaub, C. 1 890. Beitrag zur Kenntniss der Comatuliden-Fauna des Indischen Archipels. Nachr. Ges. Wiss. Gottingen. 1890 : 168-187. 1912. Die Comatuliden. Reports on the results of dredging, . . . , by the U.S. coast survey steamer Blake. Mem. Mus. comp. Zool. Harv. 27 : 277^91, 18 pis. Koehler, R. 1895. Dragages profonds executes a bord du Caudan dans le Golfe de Gascogne. Revue biol. N. Fr. 7 : 439-496. 1909. Echinodermes provenant des campagnes du yacht Princesse- Alice. Result. Camp, sclent. Prince Albert 34 : 1-3 1 7, 32 pis. & Bather, F. A. 1902. Gephyrocrinus grimaldii crinoide nouveau provenant des campagnes de la Princesse- Alice. Mem. Soc. zool. Fr. 1902 : 68-79, 4 figs. & Vaney, C. 1910. Note preliminaire sur les crinoides du Travailleur et du Talisman. Bull. Mus. natn. Hist. nat. Paris 1910 : 23-32, 3 figs. M'Andrew, R. & Barrett, L. 1857. List of the Echinodermata dredged between Drontheim and the North Cape. Ann. Mag. nat. Hist. (2) 20 : 43^16. Macurda, D. B. & Meyer, D. L. 1976. The morphology and life habits of the abyssal crinoid Bathycrinus aldrichianus Wyville Thomson and its paleontological implications. J. Paleont. 50 : 647-667, 5 figs, 5 pis. Messing, C. G. 1978. Pentametrocrinus atlanticus (Perrier) (Crinoidea: Echinodermata): a review. J. nat. Hist. 12 : 699-708, 18 figs. Meyer, D. L., Messing, C. G. & Macurda, D. B. 1978. Zoogeography of tropical western Atlantic Crinoidea (Echinodermata). Bull. mar. Sci. 28 : 4 1 2-44 1,12 figs. Mortensen, T. 1927. Handbook of the echinoderms of the British Isles, ix + 47 1 , 269 figs. London. Pennant, T. 1 777. British Zoology. Ed. 4. 4 : 1-1 36, 93 pis. London. Perrier, E. 1883a. Sur un nouveau crinoide fixe, le Democrinus parfaiti, provenant des dragages du Travailleur. C. r. hebd. Seanc. Acad. Sci., Paris 96 : 450-452. 18836. Sur des Eudiocrinus de 1'Atlantique, et su la nature de la faune des grandes profondeurs. C. r. hebd. Seanc. Acad. Sci., Paris 96 : 725-728. Pourtales, L. F. de. 1869. List of the crinoids obtained on the coasts of Florida and Cuba, by the U.S. Coast Survey Gulf Stream Expedition in 1867, 1868, 1869. Bull. Mus. comp. Zool. Harv. 1 : 355-358. Roux, M. 1977. Les Bourgueticrinina (Crinoidea) recueillis par la Thalassa dans le golfe de Gascogne: anatomic comparee des pedoncules et systematique. Bull. Mus. natn. Hist. nat. Paris. Zool. 296:25-83, 17 figs, 10 pis. 1978. Ontogenese, variabilite et evolution morphofonctionnelle du pedoncule et du calice chez les Millericrinida (Echinodermes, Crinoides). Geobios No. 1 1 (2) : 2 1 3-24 1,15 figs, 2 pis. Sibuet, M. 1974. Echinodermes de la mer d'Alboran. Bull. Mus. natn. Hist. nat. Paris. Zool. 155 : 789-798. Thomson, C. Wyville, 1872. On the crinoids of the Porcupine Deep Sea Dredging Expedition. Proc. R. Soc. Edinb. 7 : 764-773. Manuscript accepted for publication 28 November 1979 A new abyssal genus of the family Ophiuridae (Echinodermata : Ophiuroidea) Gordon L. J. Paterson Department of Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD Introduction A new genus of abyssal ophiuroid from the NE Atlantic belonging to the subfamily Ophiurinae of the family Ophiuridae is described. Its relationship to the genera Ophiotrochus and Bathylepta is discussed and justification given for the transfer of these two genera from the Ophioleucidae to the Ophiurinae,. both of which are rediagnosed. Uriopha is the fourth purely abyssal genus belonging to the family Ophiuridae to have been discovered in recent years, Perlophiura Belyaev & Litvinova, 1972, Abyssura Belyaev & Litvinova, 1976 and Bathylepta Belyaev & Litvinova, 1972 are the other three. It is perhaps slightly surprising that a new genus should be found in what could be considered a well investigated region. This might be due to the small size of the specimens resulting in their being overlooked by past researchers. The specimens of Uriopha form part of a collection of deep sea ophiuroids made by the Institute of Oceanographic Sciences' Discovery Investigations off North West Africa. Systematic descriptions URIOPHA gen. nov. DIAGNOSIS. A genus of the subfamily Ophiurinae, family Ophiuridae, with the disk completely covered by granules obscuring the plates beneath; the disk is convex or even domed; no arm combs or genital papillae are developed; the jaw has a spiniform apical papilla flanked on each side by oral papillae which becomes more blocklike distally; the adoral plates are long and thin; the oral shields are small and triangular; the second oral tentacle pore emerges superficially, away from the mouth slit; the tentacle pores are round, distinct and occur along the whole arm length, each has one large rounded scale and sometimes a smaller indistinct scale underlying it; there are two conical appressed arm spines on each lateral arm plate beyond the disk. TYPE-SPECIES. Uriopha ios sp. nov. DERIVATION OF NAME. Uriopha is an anagram ofOphiura, a genus of the Ophiuridae, and ios is derived from the initials of the Institute of Oceanographic Sciences. Uriopha ios sp. nov. Fig. 1 The eleven individuals of the type series have disk diameters of 3-4 mm. Viewed from above the disk is round to subpentagonal; in side view it appears high and in many specimens domed. A dense coating of granules, each about 0*04 mm in diameter, covers the dorsal surface and ventral interradial areas, obscuring the plates, including the oral and adoral shields but not the oral plates. The underlying dorsal plates are numerous, small and slightly imbricated, more so in the domed specimens. Except for the centrodorsal plate in some specimens, the primary rosette is usually indistinct. The radial shields are irregular though Bull. Br. Mm. not Hist. (Zool.) 38 (4): 211-218 Issued 25 September 1 980 211 212 G. L. J. PATERSON Fig. 1 Uriopha ios sp. nov. (a) Holotype, (b-e) Paratypes. (a) Ventral view of part of disk, (b) Dorsal view of part of denuded disk, (c) Lateral view of arm fragment, (d) Adradial and (e) abradial view of oral plate. Bar scales = 1 mm.. approximately rectangular in shape, length: maximum breadth is 1- 5-2-0 : 1. They may be wholely contiguous or only distally so. Interradially the radial shields are separated by several plates. Arm combs are absent. Ventral disk scales are slightly larger and not imbricated like those of the dorsal surface. The apical papilla is spiniform; on each side of it are 3 or 4 oral papillae, broadening progressively from peglike to blocklike. The oral plates are long and rectangular and are raised up in the interradial midline where they meet. The adoral plates meet on the midline although this suture is partially obscured by the oral shield; they are approximately four times longer than wide, flaring distally where they meet the first lateral arm plates. The oral shield is triangular with a straight distal side; it partially overlies the adorals and is totally proximal to the two genital slits, not spanning them. The second oral tentacle pore arises superficially and away from the mouth slit and is surrounded on three sides by 4-6 tentacle scales, arising partly from the first ventral arm plate and partly on the bordering edge of the adoral plate. These scales are distinct from and do not form a continuous series with, the mouth papillae; in some cases they may appear continuous but there is a distinct gap between the last oral papilla and the nearest tentacle scale (see Fig. la). The genital slits are edged by granules not differentiated from those of the disk. The fragile arms are almost cylindrical, slightly higher than wide. They are often broken off short. The first dorsal arm plate is overlaid by the disk; it is contiguous with the next plate but subsequent plates are separated becoming progressively smaller down the arm. The first complete dorsal arm plate is trapezoidal with a convex distal edge, subsequent plates are NEW ABYSSAL GENUS OF OPHIURIDAE 213 triangular with a slight convex edge. In side view the dorsal arm plates appear swollen above the lateral arm plates (see Fig. Ic). The two arm spines are conical situated towards the ventral end of the lateral arm plates. The first ventral arm plate is pentagonal sometimes with a cleft or furrow running proximally into the mouth slit. The following plates are pentagonal with the lateral edges indented by the tentacle pores, the distal edge is slightly convex. The tentacle pores are round, distinct and present on all the arm segments. The tentacle pore of the first arm segment has two, sometimes three tentacle scales, the remaining pores have one distinct rounded scale sometimes with one small indistinct scale. The holotype (fig. 3) has a disk diameter of 3*4 mm. The arms of the holotype and para- types are all broken but from one specimen which has an almost intact arm, measuring 5'5 mm, it can be estimated that the arms were two to three times the disk diameter. Type locality: Stn 8524 1, Discovery Investigations, 20°46'N : 22°42'W (west of Cap Blanc and north of Cape Verde Islands), 4412 m. DISCUSSION. The superficial position of the second oral tentacle pore prompts inclusion of this genus in the subfamily Ophiurinae, the second subfamily, Ophiolepidinae differs in having this opening within the oral slits. Uriopha shows a combination of characters which readily distinguish it from the other genera of the Ophiurinae. Certainly some individual characters unusual in this subfamily, such as the low number of arm spines or the granulation of the disk, are shared by particular genera, suggesting possible relationships, but these are unsupported by the other characters. For example, the genera Ophiomastus Lyman, 1878, Ophiopyrgus Lyman, 1878 and Ophiuraster H. L. Clark, 1939 also have a low number of arm spines, two or less, and a simple form of tentacle scale on the arm pores but they have large thick and sometimes swollen disk and arm plates and in the case of Ophiomastus and Ophiopyrgus the disk is dominated by the centrodorsal, primary rosette and radial shields. Uriopha by contrast has small thin disk and arm plates and the primary rosette is indistinct. The genital slits also differ, being long and well developed in Uriopha but short or absent altogether in these three genera, while the shapes of the adoral and oral shields are very different and only Uriopha possesses a granulated disk. However, Uriopha does appear to have a genuine affinity with the genus Ophiotrochus Lyman, 1878, referred by Matsumoto (1915) to the Ophioleucidae when he differentiated that family from the Ophiuridae (Ophiolepidae in Matsumoto's terminology) (see Fig. 2). Both have a similar arrangement and number of arm spines, with two arm spines set on the ventral end of the lateral arm plate, though Ophiotrochus has more than two arm spines on the first arm segments; both have similar tentacle pore and scale form on the free arms, though Uriopha has two to three scales on the first arm pore; the genital slits are developed to a similar degree; the adoral and oral shields are of similar shape and arrangement and both have granulated disks though in Ophiotrochus this is limited to the ventral side. This affinity does not imply inclusion of the new genus in the Ophioleucidae since several factors support a proposition that not only Ophiotrochus but also Bathylepta Belyaev & Litvinova, 1972, should be transfered from the Ophioleucidae to the Ophiurinae. First and most importantly, the second oral tentacle pore arises superficially outside the mouth slit whereas the other ophioleucid genera have the second oral tentacle pore opening entirely within the mouth slits. The alignment of this pore is the primary character of the Ophiurinae and alone warrants their distinction from the Ophioleucidae. Within the order Chilo- phiurina, comprising the families Ophiodermatidae, Ophionereididae, Ophiocomidae, Ophiuridae and Ophioleucidae, only the Ophiurinae possesses this character. Koehler (1904) first noted its importance and suggested that it might be used to separate groups within the genus Ophioglypha (Ophiura, sensu H. L. Clark, 1911). In his revised classification of the Ophiuroidea, Matsumoto (1915, 1917) regarded the alignment of the second oral pore as an important character at the family-subfamily level. In his parallel keys to the family groups of the Chilophiurina (Matsumoto, 1915, pp. 74-75) this character separates his Ophiomastinae, now Ophiurinae, from all the other family-group units. Although the second oral tentacle of the Ophiurinae emerges ventrally in basically two ways, 214 G. L. J. PATERSON Fig. 2 Ophiotrochus longispinus H. L. Clark, (a) Ventral and (b) dorsal view of part of disk, (d) Adradial and (c) abradial view of oral plate. Bar scales = 1 mm. either opening via a slit or directly on to the ventral surface, both Koehler (1922) and Matsumoto (1915, 1917) consider them to be variations of one character. Such a distinct character suggests that the relationships between the subfamilies Ophiurinae and Ophiolepidinae should be investigated and further study may warrant raising them to family status. In addition Ophiotrochus and Bathylepta lack the large contiguous successive dorsal and ventral arm plates, the well developed peg-like oral papillae and the long flattened arms characteristic of the Ophioleucidae. These two genera currently form an anomalous group within the Ophioleucidae not closely related to Ophioleuce Koehler, 1904 and the other genera included. One character thought by Fell (1960) and Matsumoto to be important but now set aside is that of the insertion and degree of integration of the arms with the disk. This character seems to be affected by both the extent of the genital slit and the state of preservation of the specimen and may be misleading. Matsumoto (1917) considered Ophiotrochus to be a neotenous ophioleucid probably because of the ventral insertion of the arms even though he noted that the second oral tentacle pore emerges superficially, unlike that of the other genera of the Ophioleucidae. Bathylepta is related to both Uriopha and Ophiotrochus by the reduced number of arm spines and the development and shape of the adoral and oral shields. It differs mainly in the juvenile form of the oral papillae, which could be interpreted as a neotenous condition, the absence of genital slits and of tentacle scales except on the second oral tentacle pore (see Belyaev & Litvinova, 1972; Madsen, 1977). Due mainly to the work of Schoener (1967, 1969) on post larval development of several species of deep water ophiuroids, we are better able to recognize neotenous characters prolonged from early ontogenetic stages. Her observations indicate that such features as the lack of tentacle scales, the rudimentary form of the oral papillae and the lack of genital slits in Baihylepta; the fused oral papillae in Ophiotrochus and the small non-contiguous dorsal arm plates in Bathylepta and Uriopha and probably their absence in Ophiotrochus, are all characters found in the early stages of most species. As ontogeny progresses the tentacle scales appear. Also the superficial adradial edge of the oral plate differentiates into a NEW ABYSSAL GENUS OF OPHIURIDAE 215 Fig. 3 Uriopha ios sp. nov. (a) Dorsal and (b) ventral view of holotype. 216 G. L. J. PATERSON flange which subsequently enlarges and becomes indented to form individual oral papillae. Further development varies with the species, in some, such as Ophiura ljungmani, the papillae separate from one another and become peglike, whereas in Ophiomusium lymani they remain contiguous and block-like. As the dorsal arm plates enlarge, in many cases they become contiguous while in others they remain small and separate. By a disk diameter of c. 2-3 mm the adult features are recognisable and further growth involves only minor changes in proportions. Schoener showed that in Ophiura sarsi the juvenile stages up to 0'9 mm disk diameter have only two arm spines, rather than the adult three, suggesting that this character in the genera under discussion may be a neotenous condition; however, I have been unable to confirm this for the juvenile stages of any other ophiuroid species. The removal of Bathylepta and Ophiotrochus from the Ophioleucidae leaves the family more homogenous and allows the following more restricted definition: OPHIOLEUCIDAE Matsumoto, 1915 A family of the Chilophiurina with the disk comprised of scales completely or partially covered by a coat of granules sometimes also covering the radial shields, adoral and oral shields; the adoral shields are usually long and flared distally; the oral shields are small, triangular or arrow shaped; the teeth are papilliform and arranged in a single series; the second oral tentacle pores open within the mouth and are not visible externally; the arms are long, flat and wide and taper slowly to the end of the arm; the dorsal and ventral arm plates are large and successive ones are contiguous, at least proximally; the tentacle pores are variable in size, sometimes enlarged but becoming smaller distally, armed with one to several scales; the disk may appear superficial, overlying the arm bases. The familly as restricted includes the genera Ophioleuce, Koehler 1904, Ophiernus Lyman 1878, Ophiocirce Koehler, 1904, Ophiopyren Lyman 1878 and Ophiopallas Koehler 1904 (Ophioperla Koehler, 1912 was inexplicably and erroneously referred to the Ophioleucidae from the Ophiuridae by Fell, 1960 and followed by Spencer & Wright, 1966. Later Fell, 196 1 , moved Ophioperla back to the Ophiurinae). The Ophiurinae can be defined as follows: OPHIURINAE Lyman, 1865 A subfamily of the Chilophiurina with a disk covered by scales, which are usually distinct but may be obscured by a covering of opaque skin or granules; the primary rosette is often conspicuous, sometimes dominating the disk; the radial shields are usually conspicuous; opposite each arm base a comb of papillae or spinelets is often present and may be continuous ventrally with papillae bordering the genital slits; the oral and adoral shields are variously developed (though ill-defined in the specialized genus Astrophiurd); the teeth are narrow, pointed or papilliform; at the apex of the jaw no tooth papillae are present but usually a single infradental papilla bordered on each side by a continuous series of pointed, rounded or blocklike papillae; the second oral tentacle pore opens either directly or via a slit onto the ventral surface; the arms are of moderate or short length, widest proximally, tapering, flat or cylindrical, usually not moniliform, the exceptions being Ophiotrochus and Perlophiura which possess moniliform arms; the lateral arm plates are very evident, the dorsal and ventral arm plates being usually small so that successive ones are contiguous only on the basal segments, in some cases the dorsal arm plates may be undeveloped altogether as in Anthophiura and Ophiotrochus; the arm spines are mostly small and appressed; the tentacle pores are of varying sizes and may be found throughout the arm length or only proximally, usually armed with rounded or spine-like scales, which are absent in Haplophiura, the tentacle scales are variable in number and arrangement; the disk may appear to surround the arm bases. The genera belonging to this subfamily are as listed by Fell (1960) with the following additions: Anophiura H. L. Clark, 1939, Ophioperla, Perlophiura, Abyssura, Ophiotrochus, Bathylepta, and Uriopha. NEW ABYSSAL GENUS OF OPHIURIDAE 217 Uriopha, Ophiotrochus and Bathylepta are compared and contrasted in table 1. While these three genera form a distinct closely-related group within the Ophiurinae, further study is needed before their relationships to the remaining genera in this subfamily can be clarified. Table 1 Comparing the genera Bathylepta, Ophiotrochus & Uriopha. Character Bathylepta Ophiotrochus (see fig. 2a-d) Uriopha (see fig. 1 a-e) Area of granulation Total Ventral only Total on disk Arm spine number 2 4 on first arm 2 joint past the disk then 2 Adoral shields Well developed Well developed Well developed Oral shield Small triangular Small triangular Small triangular Genital slits Absent Present Present Dorsal armplate Present, triangular Absent Present, triangular Radial shields Long thin not Wider than long Wider than long contiguous contiguous or contiguous or distally so distally so Oral papillae Very neotenous Wide, blocklike, Peglike fused becoming block-like distally 2nd oral tentacle With 2 small With one large With 4-6 scales pore scales fused opercular plate Tentacle scales Absent One large round One large round of arms scale scale Acknowledgements I would like to thank Miss A. M. Clark, British Museum (Natural History), for her encouragement and for critically reading the manuscript; Dr A. L. Rice, Institute of Oceanographic Sciences, for providing the material; Dr R. P. S. Jefferies, British Museum (Natural History), for linguistic help and Mr P. A. Richens, British Museum (Natural History), who took the photographs. References Belyaev, G. M. & Litvinova, N. M. 1972. New genera and species of deep-sea Ophiuroidea. Byull. Mosk. Obshch. Ispyt. Prir. 77 (3) : 5-20. (In Russian with English summary). & 1976. The new and rare ophiuroid species from the Pacific and Indian Oceans. Trudy Inst. Okeanol. 99 : 126-139. (In Russian with English summary). Clark, H. L. 1911. North Pacific ophiurans in the collection of the United States National Museum. Bull. U.S. natn. Mus. 75 : 1-302. 1939. Ophiuroidea. Scient. Rep. John Murray Exped. 1933-34. 6 (2) : 1-136. Fell, H. B. 1960. Synoptic keys to the genera of the Ophiuroidea. Zoo/. Publs Viet. Univ. Wellington 26 : 1-44. 1961. Ophiuroidea of the Ross Sea. Fauna of the Ross Sea. Part 1 . N.Z. Dep. sci. industr. Res. Bull. 142: 1-79. Koehler, R. 1 904. Ophiures de mer profonde. Siboga Exped. 45a : 1-1 76. 1912. Echinodermes. Deux. Exped. antarct. franc. Sci. nat. 1-272. 1922. Ophiurans of the Philippine Seas. Bull. U.S. natn. Mus. 100 : 1-486. Lyman, T. 1865. Ophiuridae and Astrophytidae. ///. Cat. Mus. comp. Zoo/. Harv. 1 : 1-200. 1878. Ophiuridae and Astrophytidae of the exploring voyage of H. M.S. 'Challenger', 1. Bull. Mus. comp. Zoo/. Harv. 5 (7) : 1-168. 218 G. L. J. PATERSON Madsen, F. J. 1977. The Ophioleucidae (Ophiuroidea). Galathea Rep. 14: 109-122. Matsumoto, H. 191 5. A new classification of the Ophiuroidea. Proc. Acad. nat. Sci. Philad. 47 : 43-92. - 1917. A monograph of Japanese Ophiuroidea, arranged according to a new classification. J. Coll. Sci. imp. Univ. Tokyo** (2) : 1-408. Schoener, A. 1967. Post-larval development of five deep-sea ophiuroids. Deep-Sea Res. 14 (6) : 645-660. — 1969. Atlantic ophiuroids: some post larval forms. Deep- Sea Res. 16 (1) : 127-140. Spencer, W. K. & Wright, C. W. 1966. Asterozoans. In R. C. Moore (editor), Treatise on invertebrate paleontology pt U 3 ( 1 ) : U4-U 1 07. Manuscript accepted for publication 29 November 1979 A revision of the spider genus Macopaeus (Araneae : Salticidae) F. R. Wanless Department of Zoology, British Museum (Natural History) Cromwell Road, London SW7 5BD Introduction The genus Macopaeus Simon, 1900 formerly included three species. The type species M. spinosus Simon belongs in the subfamily Lyssomaninae and is clearly related to Asemonea O. P. -Cambridge and Pandisus Simon. The original description of M. spinosus was based on a female, but Simon (1900) did not indicate the number of specimens examined. A female type specimen was examined by Roewer (1965) who diagnosed the genus but did not provide a description of the species. Despite the fact that this specimen has been subsequently lost and that the epigyne has never been figured, the spines on leg I are arranged in a distinctive manner and it is possible to identify the species with reasonable certainty. The other two species M. madagascarensis Peckham & Peckham from Madagascar, and M. celebensis Merian from Celebes are not related to M. spinosus, but belong in the genus Brettus Simon which has recently been revised by the author (Wanless, 1979). In the present paper Macopaeus is redefined; M. spinosus, Brettus celebensis comb, n., and B. madagascarensis comb, n., are described and figured; and a neotype is designated for M. spinosus. The measurements were made in the manner described by Wanless (1978). Genus MACOPAEUS Simon Macopaeus Simon, 1900 : 381. Type species Macopaeus spinosus Simon, by original designation and monotypy. Simon 1901 : 394, 397, 399. Merian, 1911 : 303. Petrunkevitch, 1928: 181. Roewer, 1954 : 932; 1965 : 6. Bonnet, 1957 : 2684. DEFINITION. Spiders of medium size (4'0-8'0 mm). Carapace: longer than broad, moderately high with elevated eye region (Fig. 1 A, C); fovea long, sulciform, positioned midway between posterior lateral eyes and posterior thoracic margin. Eyes: with black surrounds except anterior medians; disposed as in figure 1A; all relatively large; posterior ocular quadrangle longer than broad, entire quadrangle about 50% of carapace length. Clypeus: moderately high, slightly concave. Chelicerae: of medium size, more or less vertical; promargin with 4 teeth, retromargin with 8. Maxillae: of medium length, blades rounded, outer lateral margins slightly excavated; slightly divergent. Labium: subtriangular, about half maxillae length, lateral margins slightly excised basally. Sternum: cordiform. Pedicel: short. Abdomen: elongate ovoid; spinnerets of medium length, anteriors robust, medians and posteriors slender; tracheal system not examined. Legs: long and slender; legs I with robust ventral spines on metatarsi and distal half of tibiae; other leg spines less robust, more scattered on legs II and less numerous on legs III and IV; claws pectinate, tufts present. Female palp: long and slender. Epigyne: openings distinct, spermathecae large; vulva not examined. A fuller definition can only be given when the male is discovered and additional material becomes available for study. AFFINITIES AND DIAGNOSIS. Macopaeus evidently belongs in the subfamily Lyssomaninae and seems to lie near Asemonea and Pandisus, two closely related genera both of which are known from Madagascar. It is separated from Pandisus by the large posterior median eyes, and from Asemonea by having the posterior ocular quadrangle longer than broad. Bull. Br. Mus. nat. Hist. (Zool.) 38 (4): 2 19-223 Issued 25 September 1980 219 220 F. R. WANLESS Macopaeus spinosus Simon (Fig. 1A-F) Macopaeus spinosus Simon, 1900:381, 9, Madagascar, Antongil (MNHN, Paris) [not examined, presumed lost]. Simon, 1901 : 396, 399. Petrunkevitch, 1928: 181. Roewer, 1954:932; 1965:6. Bonnet, 1957:2684.. REMARKS. Roewer (1965) examined, but did not describe the type specimen of this species which has been subsequently lost. The specimen described below agrees more or less with the original description (Simon, 1900) and with subsequent remarks made by Simon (1901) in respect of leg I spination, and is here formally designated neotype. DIAGNOSIS. M. spinosus, known only from the female is separated from all other known lyssomanids by the disposition of the eyes and structure of the epigyne (Fig. 1 A, E). MALE. Unknown. FEMALE NEOTYPE. Carapace (Fig. 1 A, C): pale amber with brownish marginal bands and faint bands extending from posterior lateral eyes to near posterior thoracic margin; thinly clothed in short amber hairs with iridescent violet hairs in eye and foveal regions. Eyes: fringed by whitish hairs that are iridescent under some angles of illumination. Clypeus: sparsely covered in pale amber hairs. Chelicerae: yellow-brown, shiny; thinly clothed in amber hairs; promargin with 4 teeth, retromargin with 8. Maxillae and labium: whitish yellow to pale yellow. Sternum (Fig. IF): whitish yellow; glossy, with scattered pale amber hairs. Coxae: subequal in size; pale yellow. Abdomen: whitish yellow with black markings; thinly clothed in fine iridescent pale amber hairs; spinnerets pale yellow. Legs: pale yellow to light amber with sooty lateral markings; legs I (Fig. 1 B) with strong spines, those on the tibiae restricted to distal half of segment. Epigyne (Fig. 1 E): vulva not examined. Dimensions (mm): total length 5'6; carapace length 2-32, breadth 1*84, height 1'36; abdomen length 3'2; eye row widths, anterior medians 1*2, anterior laterals 1'48, posterior medians 0*8, posterior laterals 1-0; quadrangle length 1-2. Ratios: AM : AL : PM:PL: 15 : 9'5 : 6 : 8; AMM : CL(clypeus) : 15:6. DISTRIBUTION. Madagascar. MATERIAL EXAMINED. Neotype 9, Madagascar, Beanana, ii. 1970 (A. Lambillon, MT. 142.599 part) (MRAC, Tervuren). Brettus celebensis (Merian) comb. n. (Fig. 2A-D) Macopaeus celebensis Merian, 1911: 304, 9. Holotype 9, Celebes, Wald bei Duluduo (NM, Basel, 88a) [Examined]. Roewer, 1954 : 932. Bonnet, 1957 : 2684. AFFINITIES AND DIAGNOSIS. B. celebensis is closely related to B. anchorum Wanless from India, but may be distinguished by the relatively short and wide seminal ducts (Fig. 2C). The presence of dense dorsal fringes on tibiae I (evidently lacking in B. anchorum) may also be a useful diagnostic character. MALE. Unknown. FEMALE HOLOTYPE. Carapace (Fig. 2A, D): dark brownish orange; clothed in recumbent short white hairs. Eyes: with black surrounds except AM; anteriors fringed in whitish hairs. Clypeus: clothed in whitish hairs. Chelicerae (Fig. 2B): brownish orange; clothed in pale brown hairs; promargin with 3 teeth, retromargin with 6 or 7. Maxillae: orange to yellow- Orange with median brown markings. Labium: brown tipped yellow-orange. Sternum: brownish orange with darker margins, shiny; clothed in fine whitish hairs. Coxae: unequal in size, I and IV slightly largest; brownish orange. Abdomen (detached, shrunken and partly rubbed): brown-black, clothed in recumbent pale yellowish hairs, iridescent under some angles of illumination. Legs: brown to yellow-brown; clothed in fine whitish iridescent hairs REVISION OF THE GENUS MACOPAEUS 221 Fig. 1 Macopaeus spinosus Simon, neotype 9: A, dorsal view, B, leg I; C, carapace, lateral view; D, cheliceral teeth, inner view; E, epigyne; F, sternum. with dense grey-black ventral fringes on femora I, patellae I and tibiae I, also dense dorsal fringes on tibiae I; spines strong and numerous. Palps: long and slender; pale brown to whitish yellow. Epigyne (Fig. 2C): very similar to that of B. anchorum, but the ducts are wider and do not extend to the posterior margin of the somewhat rectangular opening; vulva not examined. Dimensions (mm): total length about 6'6; carapace length 2*96, breadth 2*28, height 1'6; abdomen length 3'6; eyes, anterior row 1*48, middle row 1'04, posterior row 1'28; quadrangle length 1-2. Ratios: AM : AL : PM : PL : 14 : 7-5 : 5 : 6; AL- PM - PL : 7-5 - 1 1; AM : CL (clypeus) : 14:6. DISTRIBUTION. Celebes. MATERIAL EXAMINED. Holotype 9, data given in synonymy. Brett us madagascarensis (Peckham & Peckham) comb. n. (Fig. 3A-D) Macopaeus madagascarensis Peckham & Peckham, 1903 : 182, 9. Holotype 9, Madagascar (MCZ, Harvard) [Examined]. 222 F. R. WANLESS Fig. 2 Brettus celebensis (Merian), holotype 9: A, dorsal view; B, chelicera; C, epigyne; D, carapace, lateral view. M. madagascariensis: Roewer, 1954 : 932; 1965 : 6. Bonnet, 1957 : 2684. [Unjustified emendation]. AFFINITIES AND DIAGNOSIS. B. madagascarensis is a distinctive species readily separated from other known species of Brettus by the structure of the epigyne (Fig. 3C) and distribution. MALE. Unknown. FEMALE HOLOTYPE. Carapace (Fig. 3A, D): brown with margins and eye region clothed in recumbent short white hairs (long bright red hairs in eye area, described in the original description are no longer evident). Eyes: with black surrounds except AM; anteriors fringed in whitish hairs. Clypeus: clothed in whitish hairs. Chelicerae (Fig. 3B): yellow-brown with sooty markings; irregularly clothed in short whitish and long yellow-brown hairs; promargin with 3 teeth, retromargin with 5. Maxillae and labium: yellow-brown to light brown. Sternum: pale brown, shiny. Coxae: unequal in size, I and II largest; pale brown to brown. Abdomen (detatched, shrunken and partly rubbed): brown-black; irregularly covered in recumbent short whitish hairs, iridescent under some angles of illumination. Legs: mostly detached and broken, I and II missing; generally dark brown to yellow-brown; spines moderately strong and numerous. Palps: missing. Epigyne (Fig. 3C); vulva not examined. Dimensions (mm): total length about 4-4; carapace length 2-2, breadth 1*76, height 1'27; abdomen length 2*04; eyes, anterior row 1'16, middle row 0'8, posterior row 1*0; quadrangle length 0-96. Ratios: AM : AL : PM : PL : 10'5 : 6 : 4-5 : 5; AL-PM- PL : 6-8; AM: CL: 10-5:4. DISTRIBUTION. Madagascar. MATERIAL EXAMINED. Holotype 9, data given in synonymy. REVISION OF THE GENUS MACOPAEUS 223 Fig. 3 Brettus madagascarensis (Peckham & Peckham), holotype $: A, dorsal view; B, chelicera; C, epigyne; D, carapace, lateral view. Acknowledgements I wish to thank the following colleagues for providing specimens for study. Professor P. L. G. Benoit, Musee Royal d'Afrique Centrale, Tervuren, Belgium (MRAC, Tervuren) Professor H. W. Levi, Museum of Comparative Zoology, Harvard, U.S.A. (MCZ, Harvard) and Dr. E. Sutter, Naturhistorisches Museum, Basal, Switzerland (NM, Basel). References Bonnet, P. 1957. Bibliographia Araneorum, 2(3) : 1927-3026. Imprimerie Douladoure, Toulouse. Merian, P. 1911. Die Spinnenfauna von Celebes. Beitrage zur Tiergeographie im Indoaustralischen Archipel. Zoo/. Jb. 31 : 165-354. Peckham, G. W. & Peckham, E. G. 1903. New species of the family Attidae from South Africa, with notes on the distribution of the genera found in the Ethiopian region. Trans. Wis. Acad. Sci. Arts Lett. 14: 173-278. Petrunkevitch, A. 1928. Systema Aranearum. Trans. Conn. Acad. Arts Sci. 29 : 270 pp. Roewer, C. F. 1954. Katalogder Araneae. 2 Abt. B: 924-1290. Institut Royal des Sciences Naturelle de Belgique, Bruxelles. 1965. Die Lyssomanidae und Salticidae - Pluridentati der Athiopischen Region (Araneae). Annls Mus. r. Afr. cent. No. 139 : 86 pp. Simon, E. 1900. Descriptions d'arachnides nouveaux de la famille des Attidae. Annls Soc. ent. Belg. 44:381-407. 1901. Histoire Naturelle des Araignees, 2(3) : 381-668. Paris, Roret: Libraire Encyclopedique.. Wanless, F. R. 1978. A revision of the spider genera Belippo and Myrmarachne (Araneae: Salticidae) in the Ethiopian region. Bull. Br. Mus. nat. Hist. (Zool.) 33 (1) : 139 pp. 1979. A revision of the spider genus Brettus (Araneae: Salticidae). Bull. Br. Mus. nat. Hist. (Zool.) 35(2): 127-200. Manuscript accepted for publication 25 July 1979 A revision of the spider genus Orthrus (Araneae : Salticidae) F. R. Wanless Department of Zoology, British Museum (Natural History) Cromwell Road, London SW7 5BD Introduction The genus Orthrus Simon 1900, known only from Borneo and the Philippines, is comprised of three known species, two of which are described here as new. The genus was formerly placed in the salticid subfamily Lyssomaninae on account of somatic similarities with Asemonea Simon and Onomastus Simon. Furthermore, Orthrus and Onomastus were considered to form a link between lyssomanines and ordinary salticids "Les genres Onomastus et Orthrus font le passage des Asemonea aux Salticides ordinaires" (Simon, 1901). Onomastus and Asemonea, which are at present being revised by the author, evidently belong in the Lyssomaninae, but Orthrus shows no close affinities with lyssomanine spiders. Galiano (1976) in a review of lyssomanine genera also expressed doubts about the systematic position of Orthrus and, unfortunately, its affinities with other salticid subfamilies is uncertain as the genitalia are relatively simple and of the kind often found within the family. Although the biology of Orthrus species is poorly known, they are considered to be active spiders which hunt their prey over vegetation. In some respects they resemble lyssomanines: females and juveniles of O. muluensis sp. n., are green in life and have been collected from the underside of leaves, a habitat often frequented by New World Lyssomanines. Curiously, nothing appears to have been published on the habitat or behaviour of Old World species. The measurements were made in the manner described by Wanless (1978), but for the leg spination the system adopted is that used by Platnick & Shadab (1975). Genus ORTHRUS Simon Orthrus Simon, 1900 : 28. Type species Orthrus bicolor Simon, by original designation and monotypy. Simon, 1901 : 395, 398, 400. Petrunkevitch, 1928: 181. Roewer, 1938:78, 1954:932. Bonnet, 1958 : 321 5. Galiano, 1976 : 60, 66, 67. DEFINITION. Small to medium spiders ranging from about 3'8 to 4-8 mm in length. Species sexually dimorphic, males with porrect, elongate chelicerae and sometimes with distinct colour markings; not hirsute, but fringes often present on male palps and legs I. Carapace: moderately high, longer than broad, fovea moderately long, positioned just behind posterior lateral eyes; microsculpturing not marked; cuticle weakly irridescent (under some angles of illumination). Eyes: with black surrounds except anterior medians, surrounds sometimes edged in whitish or pale yellow guanin; anterior row more or less contiguous, apices recurved in frontal view; posterior median eyes relatively small, slightly nearer to anterior laterals than to posterior laterals; posterior row narrower than anterior row; quadrangle length between 49 and 53% of carapace length. Clypeus: low. Male chelicerae: of different shapes, porrect, elongate; fang long; promargin with 3 or 4 teeth, retromargin with 6 or 8. Female chelicerae: of medium size, slightly inclined anteriorly; promargin with 3 teeth, retromargin with 7. Maxillae: elongate, divergent. Labium: tongue-shaped, slightly longer than broad, slightly less than half maxillae length. Sternum: scutiform. Coxae: subequal in females, coxae I largest in males. Pedicel: short. Abdomen: elongate, sometimes tapered; scuta lacking; spinnerets moderately long and slender, subequal in length; tracheal system not Bull. Br. Mus. nat. Hist. (Zool.) 38 (4): 225-232 Issued 25 September 1 980 225 226 F. R. WANLESS examined (insufficient material). Legs: long and slender; fringes of short stiff hairs sometimes present on male legs I; spines moderately strong to weak, less numerous on posterior legs; claws pectinate; tufts present, scopulae lacking. Female palp: long and slender, terminal claw apparently lacking. Male palp: simple; retrolateral tibial apophysis distally blunt or pointed; embolus originating subbasally to apically, shape relatively short and stout to moderately long and slender; tegulum unmodified, seminal duct peripheral. Epigyne: relatively simple; openings indistinct; posteriorly a semicircular depression (presumably for retaining the retrolateral tibial apophysis during copulation); primary spermathecae small, secondaries large with leaf-like fertilization ducts. AFFINITIES AND DIAGNOSIS. The simple structure of the secondary genitalia would appear to exclude this genus from the Lyssomaninae, unfortunately affinities with other salticine subfamilies are unknown and a useful diagnosis cannot be given at the present time. Hopefully the position will become clear when other Oriental genera have been revised. REMARKS. It may be of interest to note that a survey of Oriental salticids in the collections of the British Museum (Natural History) produced several pale yellow females from the Malay Peninsular which were initially considered to be congeneric with Orthrus, but later proved otherwise. Similarly, the New Guinea collections produced several males with elongate chelicerae that closely resembled Orthrus, but these too belonged in another 'unknown' and unrelated genus. List of species in the genus Orthrus Simon 1900 Orthrus bicolor Simon, 1 900 O. muluensis sp. n. O. palawanensis sp. n. An identification key is hardly necessary as the species can be readily distinguished from one another by the structure of the genitalia. Orthrus bicolor Simon (Fig. 1A-E) Orthrus bicolor Simon, 1900 : 28, O>/~> — ~O •/"> OO m in >n r~- r— "/~> ^t1 to v> in m Tt Tt in >o 6 ~~ Tf MD ?N 6 oo oo 6 r^* ^^ ON ON ON ^^ ro ^f Tj* Tf — ^^ ^^ n -*t ^^ ^^ ^^ * oo m ON 9^6^-; ^^96 ON 66 r- o — ON o I llll ll 1 O— • O l/~> ^^> (^^ f<^ ^^ T^* l^6rMON9^— 6 — — ^TTT^^T 7 77 ll — O — ' ON ONO OO OO OO a ^ ^ o 2 "O -C> as O 3 *- O II mo m i T" ^ "T1 6 6 "* ^ T "P r- do 6 T "? *? 7 7 T ^ 7 °T 7 77 suoiuods ON ON ON ON ON ON ON ON ON — ON 1 — ON ON ON ON ON ON ON ON ON ON OO ON ON GO 1 0 . 1 \ association GO > '5 c/i DC « .S ~& t! S o S o O 0> (U Cd S "a I 1" E •s .1 tu >^ Parasitine \ <5 2