S6M logs Zoology Series S\2 NATURAL HISTORY MUSEUM VOLUME 65 NUMBER1 24 JUNE 1999 The Bulletin of The Natural History Museum (formerly: Bulletin of the British Museum (Natural History) ), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology. The Zoology Series is edited in the Museum’s Department of Zoology Keeper of Zoology Prof P.S. Rainbow Editor of Bulletin: Dr B.T. Clarke 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 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. All papers submitted for publication are subjected to external peer review for acceptance. A volume contains about 160 pages, made up by two numbers, published in the Spring and Autumn. Subscriptions may be placed for one or more of the series on an annual basis. Individual numbers and back numbers can be purchased and a Bulletin catalogue, by series, is available. Orders and enquiries should be sent to: Intercept Ltd. P.O. Box 716 Andover Hampshire SP10 1YG Telephone: (01264) 334748 Fax: (01264) 334058 Email: intercept @ andover.co.uk Internet: http://www.intercept.co.uk Claims for non-receipt of issues of the Bulletin will be met free of charge if received by the Publisher within 6 months for the UK, and 9 months for the rest of the world. World List abbreviation: Bull. nat. Hist. Mus. Lond. (Zool.) © The Natural History Museum, 1999 Zoology Series ISSN 0968-0470 Vol. 65, No. 1, pp. 1-72 The Natural History Museum Cromwell Road London SW7 5BD Issued 24 June 1999 Typeset by Ann Buchan (Typesetters), Middlesex Printed in Great Britain by Henry Ling Ltd., at the Dorset Press, Dorchester, Dorset Bull. nat. Hist. Mus. Lond. (Zool.) 65(1): 1-13 Issued 24 June 1999 Phylogenetic relationships of Toad-headed lizards . Agamidae) based on morphology _ E.N. ARNOLD 9 JL 1999 — } Department of Zoology, The Natural History Museum, Cromwell Road, London SW7 SBD, UK ae i CONTENTS | GENERAL MAE CVA TV GUT OOM ics runs owns dos Sakae canon ak geo esa ase eas ends Cane wa as totnns cause ounoncnw ana casbnassevunck'usessUchpiiiansnnaaUsdeyssdudsveasanedesennuns dv sacnsessiwe 1 ERE VAM CAS INA TIS OME zID SILO CET) LILLIES acacia Wes sg ae eae asta ns cha sak hc hand ch ann cp aca we nna aia sdemwddennnstceccnedannatens 1 MOM OLOcirall ChALACLELS USE (0) ESUIMNALE PINVIO SCD se cnc. ceatten noe anascucncccevnuacrapaneasvencceupuascnosnhsscWnascsrasaadeceicusvstvasae in cesadecngadeaxutastes 2 FBI O MEMEUIC AMALV SIS) sosec.cessasqcs-cevaeveconaendevsveviesscnacent<vestaqsssaunasaccacayodssvassveaneedeauesudvasaveterasssachaiscdvivsaevsseveszesudebsacasucvsdatasccscesteaastases 7 DDPTSCSTES SLO rpete eee ties cece Pace Ps eet nape Sn scents cain Mean sae aac say ebse ranean ee es ae Mean cc enc ofaectn assis eravvacstesSvatich saasuvoeass¥ebeves steevenee 9 PSCMMOWLE CS CMENES! 4-.nsstonuananottne ce rete wwe ateecaskicsssceworvencasesedsanesuceacive deecstaevsusvssivdacsaue th veveuacsocesuecsctvaiesbcedes sxackoseencvenessviesesendievaes 11 FELT HCE Secs ccaa aster raed eas ueOe sesacee ce fae cee eae exc er ate a ne was nes ps nome camera Sasthy conc ehvasancsdsctatenvate means Gciuswesesdasenavedneues 11 SyNopsIS. Phrynocephalus together with its sister-group, Bufoniceps is most closely related to other advanced Palaearctic and African agamids. They have been regarded as the sister-group of all these species or derived from African Agama (Moody, 1980, morphological data) or as the sister of Laudakia (Joger, 1991, albumin immunology) but reassessment of morphology suggests a relationship to Trapelus. Parsimony analysis of 46 morphological characters, involving 54 derived states, of 25 species of Phrynocephalus indicates that successive branches arising from the main lineage of the genus are as follows: P. mystaceus; P. maculatus; P. arabicus; the P. interscapularis group — (((P. clarkorum, P. ornatus) ((P. euptilopus, P. luteoguttatus) (P. interscapularis, P. sogdianus))); P. scutellatus; P. golubevi; P. reticulatus; P. raddei. There is then a group of 11 species in which relationships are generally poorly resolved, although within this P. theobaldi, P. roborowskii and P. vlangalii are clearly closely related to each other and perhaps to P. forsythii, and the tuberculated species, P. helioscopus, P. persicus, P. rossikowi and P. strauchi may also form a clade. There is no clear morphological evidence that the northeastern species, P. axillaris, P. versicolor, P. przewalskii. and P. guttatus (which also extends far to the west) form a holophyletic group. Phrynocephalus does not appear to share its general phylogeographic pattern with other Asian reptiles and this may consequently result from dispersal rather than vicariance events. The phylogeny suggests the ancestor of Phrynocephalus occurred inArabia-NW India area whence there were three independent invasions of Central Asia: by the ancestors of PR. mystaceus, of P. interscapularis + P. sogdianus, and of P. golubevi and its sister group, the latter later extending north and eastwards into Mongolia, China and Tibet. Phrynocephalus appears to have primitively occupied aeolian sand habitats but to have spread to harder substrates from which sandy habitats were sometimes reinvaded. Degeneration of the outer and middle ear occurred in the early history of Phrynocephalus but was partly reversed in P. axillaris and the P. theobaldi group. Arnold 1992 which was created for Phrynocephalus laungwalaensis INTRODUCTION Sharma, 1978. Moody (1980) placed Phrynocephalus (including P. laungwalaensis) with what at the time was usually called Agama Daudin 1802, in his group 6 of the Agamidae. WithinA gama, as then understood, this author recognised several separate genera: Agama s. str., Xenagama Boulenger 1895, Pseudotrapelus Fitzinger 1843, Trapelus Cuvier 1817 and Stellio Laurenti 1768. However, the name Stellio is unavailable (Stejneger, 1933) and the assemblage it was used to denote by Moody is paraphyletic, comprising distinct Palaearctic and mainly African assemblages (Joger, 1991; Baig & Bohme, 1997) of which the former is probably a clade and the members of the latter more closely related to such taxa as Agama, Pseudotrapelus and Trapelus (personal observations). Leviton, Anderson, Adler & Minton (1992) argue for the use of Laudakia Gray, 1845 for the Palaearctic forms, a course followed here. The more recent suggestion (Henle, 1995), that Laudakia should be ee confined to some members of this assemblage and the rest placed in RELATIONSHIPS OF PHRYNOCEPHALUS Placoderma Blyth, 1854, requires more thorough assessment of the relationships of these lizards before it is adopted. The name Phrynocephalus is the sister group of the monotypic Bufoniceps Acanthocercus Fitzinger, 1843 is available for the remainder of the Toad-headed agamids, Phrynocephalus Kaup 1825, are a found in the mainly Palearctic desert regions of Asia, from Eastern Turkey and Russia to Mongolia, and southwards to southern Arabia and Pakistan. Species in the south and centre of the range of the genus are, in the main, well defined but, in the northeast, boundaries between them are often less clear and numerous nominal taxa have been described (see e.g. Zhao & Adler, 1993). This makes the total number of species in the genus uncertain but it is likely to be in excess of 30. In this paper, an estimate of phylogeny is made for 25 of the better defined species using morphological characters, includ- ing external features and some internal ones derived from the skeleton, middle ear, shoulder muscles and abdominal arteries. © The Natural History Museum, 1999 E.N. ARNOLD Fig. 1 Anterior views of right nasal area of skulls showing differences in contribution of the maxilla (m) to the posterior wall of the narial opening. a. Small, does not contact septomaxilla (s) (Bufoniceps laugwalaensis). b. More extensive contribution, especially dorsally, and broad contact with septomaxilla (Phrynocephalus mystaceus). c. More extensive still, both dorsally and ventrally, broad contact with septomaxilla maintained (P. euptilopus). Fig. 2 Anterior views of right nasal area of skulls. a. Dorsal process of maxilla (m) tapering upwards, maxilla extending outwards below lateral process of prefrontal bone (pf) which is large (P. euptilopus). b. Dorsal process of maxilla blunt above, maxilla not extending markedly outwards below lateral process of premaxilla which is relatively small (P. persicus). forms that Moody allocated to Stellio (Schatti & Gasperetti, 1994; Henle, 1995; Baig & Bohme, 1997). Unweighted Wagner tree analysis of the morphological data presented by Moody (1980) indicated that Phrynocephalus was derived from a paraphyletic Agama s. str., while compatability analysis, and Wagner tree analysis where characters were weighted according to their consistency index in an initial run, suggested that Phrynocephalus was sister to all other members of Moody’s Group 6 (Moody, 1980). Results of isozyme analysis have been interpreted as indicating that Phrynocephalus is the sister group of Laudakia (Ananjeva & Sokolova, 1990), a result in agreement with immunological studies (Joger, 1991). In contrast, areassessment of morphology (pers. obs.) suggests that the sister group of Phrynocephalus + Bufoniceps is Trapelus. Shared features that appear derived within Moody’s Group 6 include the following: maxillae in contact beneath premaxilla, lateral prefrontal processes very large, palatine roof of interorbital canal narrow or absent, vomers fused, squamosal spatulate with no hook-shaped projection on its lateral margin, presacral vertebrae usually 22 or fewer; nostrils directed forwards rather than sideways, no enlarged subocular scales (reversed in some Phrynocephalus), external ear opening reduced in size, no spinous scales on dorsum of neck (reversed in some Phryncocephalus), no caudal autotomy, scales on tail not in regular whorls; nasal passage long and flexed, depressor mandibulae muscle extends partly over tympanum. MORPHOLOGICAL CHARACTERS USED TO ESTIMATE PHYLOGENY Skull 1. Contribution of the maxilla to the posterior wall of the narial opening of the skull (Figure 1). Small, does not contact septomaxilla (0); more extensive especially dorsally, broad con- tact with septomaxilla (1); more extensive still both dorsally and ventrally, broad contact with septomaxilla maintained (2). 2. Dorsal process of maxilla (Figure 2). Tapering upwards (0); broad and ending bluntly above (1). 3. Maxilla extends clearly outwards below the anterior surface of the lateral process of the prefrontal bone (Figure 2). No (0); yes (1). 4. Relationship of maxillary and nasal bones below the lateral process of the nasal (Figure 3). Widely separated (0); more narrowly separated (1); in contact (2). PHRYNOCEPHALUS PHYLOGENY 3 Fig.3 Anterior views of right nasal area of skulls showing differences in arrangement of maxilla (m), septomaxilla (stippled)) and nasal bones (n). a. maxilla and nasal widely separated below lateral process (Ip) of nasal (P. scutellatus). b. Maxilla and nasal more narrowly separated below lateral process of nasal and the space filled by the septomaxilla (P. versicolor). c. Maxilla and nasal in contact below lateral process of nasal (P. persicus). pa Fig. 4 View into anterior right orbit, showing width of posterior face of prefrontal bone (pf) relative to that of the posterior platine (pa) and variation in lateral extension of the prefrontal relative to the infraorbital canal (io). a. Posterior face of prefrontal broad, extends laterally across infraorbital canal. b. Posterior face of prefrontal narrow, does not extend across infrarorbital canal CEES 2 epi ssoe Fig. 5 Difference in size of the parietal foramen (black) in adults. a. Small, diameter less than distance from the lateral edge of the parietal bone (p) (P scutellatus). b. Large, diameter more than distance from lateral edge of parietal bone (P. persicus). 5. Maxilla in contact with septomaxilla on surface of skull below (0); narrowed (1). lateral process of nasal(Figure 3). No (0); yes but does not reach 9. Prefrontal bone extends laterally across infraorbital canal (Fig- nasal (1); yes and reaches nasal (2). ure 4). No (0); yes (1). 6. Nasal bone projects laterally over maxilla anteriorly. No (0); 10. Size of parietal foramen in adults (Figure 5). Relatively small, yes (1). its lateral diameter less than its distance from the lateral edge of 7. Size of lateral process of prefrontal (Figure 2). Relatively small the parietal bone (0); large, its lateral diameter more than its (0); large and extended laterally (1). distance from the lateral edge of the parietal bone (1). 8. Width of posterior face of prefrontal bone in orbit relative to 11. Body of parietal bone relative to its supratemporal processes width of posterior part of palatine (Figure 4). Relatively broad (Figure 6). Upper surface of body of parietal bone relatively flat Fig. 6 Lateral profiles of supratemporal process (left) and body of parietal bone (right), arrow indicates border between the two regions. a. Upper surface of body of parietal bone running more or less smoothly into upper margin of supratemporal process (P. mystaceus). b. Upper suface of body of parietal abruptly raised relative to upper margin of the supratemporal process (P. persicus). c. Similar, but upper surface of body of parietal tuberculated (P. scutellatus). and running more or less smoothly into upper margin of supra- temporal processes which is relatively flat (0); upper surface of body of parietal bone abruptly raised relative to upper margin of supratemporal processes (1). Other skeletal features 12. Number of scleral ossicles. Twelve (0); eleven (1); ten in some individuals (2). ERSTE Eo KG FO EES S HRT Vg i) JM Wek Peers Lt a SI ee SPE IO TETRIS | mh S) AT PED eT og, {} on C's 13% 14. E.N. ARNOLD Acrodonta including Agamidae usually have 12 scleral ossicles in each eye instead of the usual lizard number of 14. While Bufoniceps possesses 12 there is further reduction in Phrynocephalus: most species and individuals have 11 ossicles but some members of at least a proportion of species in the P interscapularis group have 10. This is true of P. interscapularis, P. luteoguttatus, P. sogdianus and P. ornatus. Occurrence of 10 ossicles may in fact be wider, but the other two members of the P. interscapularis group (P. eutilopus and P. clarkorum) are known from relatively few specimens, so checks on ossicle number have been very limited in these. Number of presacral vertebrae. Usually 22, occasionally 23 in some species (0); usually 21, occasionally 20 (1). Substantial data on presacral vertebral number are given by Whiteman (1978) and my own observations confirm his. Excep- tions to the usual numbers occur in some species but nearly always constitute a small minority of not more than about 15% of individuals. Number of caudal vertebrae. Usually 40—50 or more (0); usually less than 40 (1). Again, my own observations confirm data given by Whiteman (1978). External features Se 16. 7. 18. Largest individuals exceed 60mm from snout to vent. Yes (0); no (1). Outline of body viewed from above. Robust and rounded (0); more slender (1). Position of nostrils relative to line joining anterior corners of eyes when head viewed from in front (Figure 7). Nostrils clearly below line (0); nostrils intersecting line or above it (2); interme- diate (1). Differences in position of the nostril are associated with differ- ences in the conformation of the distal limb of the tubular nasal vestibule. The proximal limb of the vestibule is more or less vertical in all cases, running downwards from its connection with the primary nasal chamber. Where the nostril is low, the distal limb of the vestibule is relatively short and runs obliquely upwards and outwards from the base of the proximal limb to the nostril. In animals where the nostril is high the distal limb runs more or less vertically upwards parallel to the proximal limb and is about as long as this. Number of internasal scales between the nasal scales (Figure 7). Usually two or more — 0; usually one or nil — 1. Fig.7 Anterior views of heads showing differences in nostril position and in number of scales between nasal scales. a. Nostrils lower and separated by two or more internasal scales (P. theobaldi). b. Nostrils high and nasal scales in contact or separated by a single internasal scale (P. arabicus). PHRYNOCEPHALUS PHYLOGENY eee ey | ey J ee SS Fig. 8 Left side of head showing differences in number of horizontal rows of scales immediately above the supralabials counted below the anterior eye, and in the size of the subocular and anterior temporal scales. a. 2 or 3 rows above supralabials, subocular and one or more anterior temporals enalarged and elongate (P. clarkorum); b. 4-5 rows above supralabials, suboculars and anterior temporals not enlarged (P. golubevi). Paann N suena seeeeeesti oe ae ae oe sud cine LR pea A Fig.9 Ventral views of underside of head showing differences in scalation. a. Scaling more or less uniform (P. arabicus). b. Scaling heterogeneous, with curved lateral row of enlarged scales, a large central patch of enlarged pointed scales, and scales at sides of throat, behind level of angle of mouth, very small and granular (P. euptilopus). 19. Single internasal scale with a vertical keel. No (0); yes (1). 20. Number of horizontal rows of scales immediately above the 7AM 22. 23 24. supralabials, counted below anterior part of eye (Figure 8). Usually three rows, occasionally two (0); usually four or even five rows (1). Enlarged subocular scales (Figure 8). Not or only weakly differentiated (0); one or more enlarged, keeled, antero- posteriorly elongated scales (1). One or more enlarged, diagonally keeled and elongated scales on anterior temporal region (Figure 8). No (0); yes (0). External ear opening. Present (0); absent (1). A lateral row of enlarged throat scales beginning in mental area and curving backwards and outwards usually to the vicinity of the angle of mouth, separated from lower labial scales anteriorly by one to three rows of scales (Figure 9). No (0); yes (1). DE: 26. Pile 28. 2). Enlarged scales in curved lateral row on throat keeled. No (0); yes (1). A large central patch of enlarged pointed scales on throat, the more postero-lateral ones directed outwards and backwards (Figure 9). No (0); yes (1). Scales at sides of throat, behind level of angle of mouth very small and granular (Figure 9). No (0); yes (1). Some scales on posterior temporal region and on sides of anterior neck enlarged, elongate and pointed, and directed outwards and upwards. No (0); yes (1). Distinct enlarged, raised, often pointed tubercles on dorsum of body. No (0); yes (1). Tubercles are enlarged scales that project markedly above the general level of the dorsal body skin. They may be sex- ually dimorphic in Phrynocephalus, often being more strongly we: A aprons 4 '2e52 Paes Fig. 10 Left side of tail base in P. roborowskii, showing enlarged spinose scales. 30. 31. 32% 33). 34. 315), developed in males than females. Tubercles are frequently clumped, especially anteriorly, and there is considerable varia- tion in the number associated in such groupings. Tubercle form is also variable and is especially narrow, pointed and elongate in P. forsythii, which also shows particularly strong sexual dimor- phism. Although the presence of tubercles is usually a clear-cut condition, their development is sometimes sporadic and weak. For instance, many P. theobaldi lack them but a few animals have somewhat enlarged scales that are raised and form weak tubercles posteriorly. Scales at sides of tail base distinctly enlarged and often spinose (Figure 10). No (0); yes (1) Horizontal fringe of pointed upturned scales on posterior sur- face of proximal thigh. No (0); yes (1). Subdigital lamellae on distal part of fourth toe of pes (Figure 11). With two or more keels or at least projections from the free edges of the lamellae (0); with a single keel or none (1; Narrow light longitudinal stripes often present on flanks. No (0); yes (1). Dark pigment frequent in mid-line area of belly in adults. No (0); yes (1). Distal tail often with substantial dark pigment at least ventrally, where it may form transverse bars. No (0); yes (1). Soft parts 36. Palatal flaps. Large (0); reduced or absent (0) 37 . Tympanum. Well developed and robust(0); reduced to a delicate membrane (1); absent (2). This and other ear features of acromiotrapezius occiput episterno- scapulodeltoideus cleidomastoideus E.N. ARNOLD Fig. 11 Underside of fourth toes of pes (anterior edge above), showing extent of lateral fringes of pointed scales and number of keels on subdigital lamellae. a. Fringes small, especially anteriorly, two keels distally (P. theobaldi). b. Large fringes, single keels (P. mystaceus). Bufoniceps and Phrynocephalus are discussed further elsewhere (Armold, submitted). 38. Pars inferior of extracolumella. Large (0); small or absent (1) 39. Pharyngeal opening of middle ear. Large, length 15-25% of head length (0); distinctly reduced, length about 10-14% of head length (1); minute or absent (2). 40. Episterno-cleidomastoideus muscle present (Figure 12). Yes (0); very reduced (1); absent (2). 41. Episterno-cleidomastoideus muscle a single strap (Figure 12). Yes (0); with two branches (1). 42. Episterno-cleidmasoideus muscle extends anteriorly to occiput (Figure 12). No (0); yes (1). 43. Scapulodeltoideus muscle extends upwards immediately ante- rior to insertion of acromiotrapezius muscle on scapula. No (0); yes (1). 44. Origin of caecal artery on doral aorta (Figure 13). Anterior and close to mesenterica cranialis artery and well posterior to coeliac artery (0); close to and usually in front of coeliac artery, occasion- ally behind (1). The caecal artery, which arises from the dorsal aorta and sup- plies the intestine, exhibits interspecific variation in the position of its origin on the aorta, relative to the origins of the coeliac artery, which runs to the stomach, and the mesenterica cranialis artery, which like the caecal artery supplies the intestine (Henke, 1974). In at least some Sitana and Draco, and inAcanthocercus, Fig. 12 Diagramatic representations of superficial muscles of the right shoulder and neck. a. Episterno-cleidomastoideus muscle a single strap not extending to the occiput; no dorsal extension of scapulodeltoideus muscle anterior to insertion of acromotrapezius muscle. b. Episterno-cleidomastoideus muscle divided, the anterior branch reaching the occiput; a dorsal extension of scapulodeltoideus muscle anterior to insertion of acromiotrapezius muscle present. PHRYNOCEPHALUS PHYLOGENY a b c caecal coeliac coeliac coeliac caecal caecal m. cranialis m. cranialis m. cranialis Fig. 13 Variation in position of the origin of the caecal artery on the dorsal aorta in Phrynocephalus. a. close to origin of mesenterica cranialis artery. b, c. Close to origin of coeliac artery. Xenagama, Agama s. str., Pseudotrapelus and Trapelus, the caecal artery originates well posterior to the coeliac artery and close to and anterior to the mesenterica cranialis artery (Figure 13a). Ina wide range of agamids, including Laudakia, the caecal and coeliac arteries originate close together, with the former usually, although not always, anterior (Figure 13b, c). (Informa- tion from Henke, 1974 and personal observations). Within Phrynocephalus some species exhibit an anterior origin of the caecal artery, either a short distance in front of that of the coeliac artery or, much less commonly, just posterior to it. In contrast, the remaining members of the genus and Bufoniceps 2 Ro t= oO = 2 o o o ny ag wm 2 ES gr 8. io 8 2 Deo as OFS Hoe Oo oS US GU Ol! CUM US CUCU 3 o fF 5 as <~< «& ° 2 ‘3s 2 2 = = o oO =I a oO oO D Ss =) = 5 > © © £ & =F = = ° 5 Oo s s EP ce f=) (0 oO, t0° 2 = % “Se [= (0) Efe No > et: Cn «nn > eo = ae « Sa» 2 88 98 96 P. reticulatus show a posterior origin close to the mesenterica cranialis artery. Other characters 45. Viviparous, giving birth to fully-formed young. No (0); yes (1). 46. Tail used frequently in intraspecific signalling. No (0); yes (1). Hemipenial features It has been suggested that features of the hemipenis delineate species groups within Phrynocephalus (Semenov & Danayey, 1989). These authors illustrate apparent differences in lobe length and in whether calyces are present on the lobes. However, personal observations of a wide range of species, including P. mystaceus, P. maculatus, P. arabicus, P. euptilopus, P. interscapularis, P. helioscopus, P. theobaldi, P. vlangalii, P. guttatus, P. versicolor and P. przewalskii, suggest that the hemipenis in these forms is consistently deeply lobed with a honeycomb structure on the outer lobe surfaces. Possibly the differences described by Semyonovy and Danayev result from examining hemipenes preserved in different stages of eversion. PHYLOGENETIC ANALYSIS The data set (Appendix 1) consists of 46 characters most of which are binary but eight include three states. Trapelus and Laudakia were used as alternative outgroups. Analysis was initially carried out using the Hennig86 program (Farris, 1988) with the options ie- and bb*, which apply branch swopping to a single tree certain to be of minimum length. When characters were ordered and Trapelus used as the outgroup, two trees of 110 steps were produced with a consistency index of 0.49 and a retention index of 0.79. With Laudakia as the outgroup two trees were again produced, with a length of 112 steps, consistency index 0.48 and retention index 0.79. P. raddei P. przewalskii P. guttatus P. versicolor P. persicus P. rossikowi P. strauchi P. helioscopus P. axillaris P. forsythii P. viangali P. roborowskii P. theobaldi o N 95 76 69 67 58 67 95 Fig. 14 Estimate of phylogeny of Phrynocephalus and Bufoniceps using Trapelus as an outgroup. Tree produced by parsimony analysis using branch and bound on a tree guaranteed to be of shortest length. Figures indicate degree of bootstrap support, only that of 57% or above being shown. oo tm 2 > eg e) s ie) Trapelus Bufoniceps P. mystaceus P. maculatus P. arabicus P. ornatus P. clarkorum P. euptilopus P. luteoguttatus P. interscapulari P. sogdianus P. scutellatus P. golubevi P. reticulatus P. raddei P. przewalskii P. guttatus P. versicolor P. persicus P. rossikowi P. strauchi P. helioscopus P. axillaris P. forsythi P. viangali P. roborowskii P. theobaldi Fig. 15 Tree in Figure 14 after being subjected to successive approximations character weighting using Hennig86 program (Farris, 1988), resulting in P scutellatus and P. golubevi being resolved as successive branches. Characters that define lettered nodes are as follows (brackets indicate some degree of parallelism; R indicates reversal). A 17, 18, (32); B 1.1, 12.1, 23, 35, 37.1, 46; C 1.2, 44; D 15; E 37.2, 38, 39.2; F 3, 12.2?, (21), 24, (36), 42; G (14), 28, 43; H 16, 22, 33; 125, 26, 27; J (8), 19, 31; K 13, 32R; L 10, 20; M (4), 17.2R; N 17.1R; O 18R, 44R; R (29); S 29; T 30; U 6, (8), (34), 38R, (39.2R), 45. Trapelus Bufoniceps P. mystaceus P. arabicus P. maculatus P. ornatus P. clarkorum P. euptilopus P. luteoguttatus P. interscapularis P. sogdianus P. scutellatus P. golubevi P. reticulatus P. raddei P. przewalskii P. guttatus P. versicolor P. persicus P. rossikowi P. strauchi P. helioscopus P. axillaris P. forsythii P. viangali P. roborowskii P. theobaldi Fig. 16 Conservative estimate of phylogeny for Phrynocephalus and Bufoniceps. Only nodes supported by two or more characters of low homoplasy are shown. PHRYNOCEPHALUS PHYLOGENY In both cases the consensus has the same topology (Figure 14). When all characters were unordered, trees of 102 steps were pro- duced which are congruent with those where characters were ordered, but with less resolution in the clade consisting of P. przewalskii and its nearest relatives (the topology of this region of the tree is the same as that shown in Figure 16.). When all these analyses were repeated using the ‘heuristic search’ option of the PAUP 3.1.1 programme (Swafford, 1993), results were identical. Bootstrapping (100 replicates), using this programme, was also applied to the ordered tree rooted on Trapelus and nodes with bootstrap support over 50% are indicated in Figure 14. Use of the successive approximations character weighting option in Hennig86 produced little change in the original tree based on unordered characters and rooted on Trapelus, merely resolving the trichotomy in the consensus tree involving P. scutellatus and P. golubevi, by making them successive branches on the main lineage of Phrynocephalus. Principal states supporting nodes are shown in Figure 15. It will be seen that some 13 nodes are supported by two or more conserva- tive characters that show little or no homoplasy. The other nodes are defined by single or noisy characters. A conservative tree recognis- ing the nodes based on the former features, or with bootstrap support above 50% (and in many cases both) is shown in Figure 16. Several nodes on the main lineage of Phrynocephalus are quite well supported and a number of other subclades can be recognised. Thus six species constituting a holophyletic group with marked internal structure form the Phrynocephalus interscapularis group consisting of P. interscapularis, P. sogdianus, P. euptilopus, P. luteoguttatus, P. clarkorum and P. ornatus. The clade has geographi- cal coherence, occurring in western Pakistan, Afghanistan, eastern Iran and adjoining central Asia. Another well defined clade, the P. theobaldi group, includes P. theobaldi, P. roborowskii and the rather more different P vlangalii. The similar tuberculated species, P helioscopus, P. persicus, P. strauchi and P. rossikowi may form another unit, although it lacks marked bootstrap support. DISCUSSION Biogeography Ananjeva and Tuniyev (1992) speculate about the history and bioge- ography of Phrynocephalus in the former USSR. Their complex hypothesis is difficult to assess as it is not based on an estimate of phylogeny for the species concerned and does not include other members of the Phrynocephalus clade. Phrynocephalus is a characteristic element of the deserts of Palaearctic Asia, like the lacertid genus Eremias and the gecko assemblage including Cyrtopedion, Agamura, Bunopus, and Crosso- bamon etc. Its area cladogram is not shared with these other taxa and there is substantial sympatry between species and species groups. It therefore seems likely that parts of the genus dispersed into at least some areas of its huge range. The estimate of phylogeny suggests that the ancestor of the present species occurred in the south of the present distribution of Phrynocephalus, possibly within the area running from western Arabia to northwestern India. This region appears to contain the primary range of Trapelus, which may be the sister of the Bufoniceps + Phrynocephalus clade, and Bufoniceps itself occurs in northwest India. Many of the basal branches of mainPhrynocephalus lineage are found wholly or partly in this area, including P. maculatus (Arabia to Pakistan), P- arabicus (Arabia), some members of the P. interscapularis group (S.Afghanistan, SW. Pakistan) andP. scutellatus (central and eastern Iran, S. Afghanistan and SW. Pakistan). 9 From this putative source area, there may have been at least a triple invasion of the presently warm and arid lowland regions of central Asia (Turkmenistan, Uzbekistan, Tadzhikistan, Kirgizstan, southern Khazakstan): by the P. mystaceus and P. interscapularis- sogdianus lineages and by the ancestor of P golubevi and the members of its sister group (shown in Figure 15, 16). The latter invasion has given rise to a series of taxa in the area ( including P golubevi, P. reticulatus, P. raddei and the P. helioscopus group). There was then apparently eastward spread: into the Tibetan region, by the ancestor of the P. theobaldi group and perhaps P. forsythii, and further north into Northwest China and southern Mongolia. On the basis of morphology, it is not clear whether extension into the latter region represents a single invasion and radiation or independent invasion by several lineages. A variety of additional movements by particular lineages has also occurred. For instance, although within the P. helioscopus group P. strauchi and P. rossikowi have relatively small allopatric ranges, P. helioscopus is widespread in former Soviet Central Asia and the very similar P. persicus on the soutwestern periphery of the range of this species extends into eastern Turkey and Iran. P. guttatus now has a broad distribution from northwest China westwards as far as the north Caspian area. Unfortunately, there is little or no fossil record of Phrynocephalus and its immediate relatives. Material assigned to Phrynocephalus has been reported from the Pliocene of eastern Turkey (Zerova & Chkhikvadze, 1984), but the precise relationships of these fossils are unknown and it is not even certain whether they represent a member of the clade made up of all present species of Phrynocephalus or if they fall outside this grouping. This arrangement of branches on the main lineage of the Phrynocephalus-Bufoniceps clade correlates with species distinct- ness. As noted, the older southern side-branches comprise very well differentiated taxa, whereas later ones in central Asia often involve more similar species and this trend is especially marked among the relatively recent, more terminal branches in the Northwest China- Southern Mongolia region, where species are very variable, their boundaries poorly defined and their taxonomy often confused. Structural niche Most members of the majority of genera in Moody’s Group 6 (Moody, 1980) climb to some extent. This is true of Laudakia, most Acanthocercus andA gama s. stt., Pseudotrapelus and mostTrapelus. Members of the latter genus, the likely sister-group of Bufoniceps + Phrynocephalus, spend a lot of time on the ground but many of them also climb in bushes. In contrast to these, Bufoniceps and Phrynocephalus themselves are strictly ground-dwelling, a derived condition. There has been dispute as to whether the ancestral spatial niche of Phrynocephalus is soft, wind-blown sand. This is suggested by Chernov (1948), Whiteman (1978) and Semenov (1987), but Golubev (1989) and Ananjeva & Tuniyev (1992) consider the group arose in gravel and sandstone deserts. The estimate of phylogeny presented here supports the former hypothesis, with Bufoniceps and three of the four basal external branches of the main Phrynocephalus lineage being found in loose-sand habitats. (References to use of soft-sand habitats: P. mystaceus — Ananjeva & Tuniyev, 1992; P. arabicus — Arnold, 1984, Gallagher & Arnold, 1988; P. clarkorum and P. orna- tus — Clark, 1992; P. luteoguttatus — Minton, 1966; P. euptilopus — Smith, 1935; P. interscapularis — Ananjeva & Tuniyev, 1992; P. sogdianus — Bannikoy et al., 1979). Shifts to firmer ground occurred in P. maculatus and independently in the ancestor of the clade containing P. scutellatus and its sister group. There was some 10 subsequent shift back to looser substrates in P. guttatus (Ananjeva & Tuniyev, 1992) and P. przewalskii. Another indication that aeolian sand habitats are primitive is that a number of features conferring performance advantage in such environments first appear on the internal branch of the phylogeny on which these habitats are entered, that is the ancestral lineage of the Bufoniceps + Phrynocephalus clade. These are discussed below. Changes in morphological features Principal changes in morphology in the history of the Bufoniceps- Phrynocephalus clade are listed in the caption of Figure 15. A high proportion of the characters in the data set (Appendix 1) show a single change on the phylogeny. Overt reversals occur in such features as size (in P. euptilopus) and the pattern of arteries arising from the aorta. Simple parallelisms are quite frequent in the remain- ing characters, but few of these are really noisy. Body size decreases early in the history of the main lineage of Phrynocephalus. Many features that appear likely to confer per- formance advantage in aeolian sand habitats develop at the base of the Bufoniceps + Phrynocephalus clade and, as noted, are concur- rent with entry into such habitats. These features include: lateral fringes of elongate scales on the digits that prevent the feet sinking into soft surfaces (Carothers, 1986); reduction of the keeling on the digital lamellae, which may be less necessary to reduce heat intake in soft-sand environments (Arnold, 1998); fringes of elongate scales along the edges of the eyelids, countersunk jaws, valvular nostrils, and a U-shaped nasal vestibule consisting of vertically parallel and subequal proximal and distal limbs, all of which features appear to exclude sand (Stebbins, 1943, 1944, 1948), although very long nasal passages may also protect the main nasal cavity from desiccation; skin covering the tympanum that may protect it from damage during burial activity, and lateral prefrontal processes that possibly protect the eyes during the same process. Some of these features initially associated with aeolian sand habitats persist in less basal forms that occur on firmer substrata. Thus, toe and eyelid fringes and countersunk jaws occur in all Phrynocephalus, although they are less marked in species that are not found on loose sand. The outer limb of the nasal vestibule is shortened in most firm-ground forms, a shift associated with the changed position of the nostril (p. 5). This feature represents a reversion towards the primitive condition found in other Group 6 agamids. It is also associated with increased contact between the maxillary and nasal bones, either directly or via the septomaxilla. These nasal features occur in more terminal Phrynocephalus species on the main lineage of the genus and have developed in parallel in P. maculatus. Other changes loosely associated with shift to firmer substrates include reduction in size of the lateral processes of the prefrontal bones, reduction in number of presacral vertebrae, increase in number of scale rows above the upper labial scales, increase in size of the parietal foramen of the skull and reversal in the pattern of the arteries arising from the aorta. The high altitude P. theobaldi group is characterised by a number of features, including viviparity, something that often develops in cold conditions (Shine, 1985). Within this group, P. vlangalii devel- ops a nostril structure that is even more reversed than in other firm-ground forms. The external and middle ear is heavily modified in the early history of the main Phrynocephalus lineage, the tympanum disap- pearing, the extracolumella decreasing in size and the pharyngeal opening becoming very reduced or absent. These changes may be associated with greater use of subterranean rather than aerial vibra- E.N. ARNOLD tion in hearing when lying under the sand. They partly reverse in the P. theobaldi group and perhaps independently in P. axillaris. Cer- tainly the former species do not usually bury directly in the substratum and use permanent burrows instead (K. Autumn, pers. comm.) Members of the P interscapularis group possess a range of features that are rare or absent in other Phrynocephalus (see caption of Figure 15); their functional significance is uncertain. Behaviour Phrynocephalus has a number of distinctive behaviour patterns. The appearance of burial by fast lateral oscillation of the flattened body (discussed by Arnold, 1995) is concurrent with entry into aeolian sand habitats at the base of the Bufoniceps-Phrynocephalus clade and, like some morphological features already discussed, is likely to be an adaptation to this environment. In line with this, such shimmy burial is best developed in more basal species (e.g. Bufoniceps — Sharma (1978), P. mystaceus, P. interscapularis — Ananjeva & Tuniyev (1992), P. arabicus, P. scutellatus, P. reticulatus (pers. obs.)). Lateral oscillation often persists in species secondarily occurring on harder substrata, for instance in P. maculatus (pers. obs) and P. helioscopus (Ananjeva & Tuniyev, 1992). In such cases this behaviour may be modified and not necessarily always used for burial. When sprayed with water, P. helioscopus adopts a distinctive posture in which the hindquarters are raised and the head lowered. Any liquid on the back then moves forward by capillary action in the channels between the scales (and probably by gravity when enough water is present) towards the mouth where it is ingested (Schwenk & Greene, 1987). Presumably, such behaviour permits advantage to be taken of even minor precipitation and condensation, something likely to be a significant benefit in the arid regions where P. helioscopus lives. P. arabicus from the United Arab Emirates re- sponds to spraying very similarly (pers. obs.). As these two species are widely separated on the estimate of phylogeny for Phrynocephalus, this stereotyped behaviour may well be more widespread than presently known. It could not be demonstrated in Trapelus flavimaculatus, also from the United Arab Emirates, so it may be confined to Phrynocephalus and possibly Bufoniceps. Phrynocephalus species are also distinctive in using the tail for intraspecific signalling (e.g. Arnold, 1984; Ross, 1989, 1995). For instance, it may be raised, curled upwards in the sagittal plane and wagged laterally. Movements usually expose conspicuous markings on the underside of the tail, such as a dark tip and transverse bars and sometimes areas of bright pigment as well. Tail signalling has been investigated for a number of Central Asian species by Dunayev (1996), who recognises seven distinct ways in which the tail may be used (Dunayev, Figure 3). Of the species considered in the present paper, the following are listed as investigated: P mystaceus, P. maculatus, P. interscapularis, P. sogdianus, P. reticulatus (as P. ocellatus), P. raddei, P. strauchi, P. helioscopus, P. versicolor and P. guttatus. When data for P. arabicus (Ross, 1995) is incorporated, it is apparent that more basal forms on the main Phrynocephalus lineage have less complex tail displays than the others. When the seven display features are treated as two-state characters (absent or present) and subjected to parsimony analysis on their own, they produce the following consensus tree which is congruent with the estimate of phylogeny based on morphology: (P mystaceus, P. maculatus (P. arabicus (all other species))). However, the supposed P. maculatus on which Dunayev’s observations were based are from the small area of Tadjikistan where P. golubevi occurs, a species which was previously not separated from P maculatus. If the animals concerned are in fact P. golubevi, the tree based on tail signalling is no longer congruent with that from morphology. PHRYNOCEPHALUS PHYLOGENY Ecological analogues of Phrynocephalus Small diurnal lizards, that are sit-and-wait foragers, have high body temperatures when active and in many cases signal with their tails, are found in several desert systems. Apart from Phrynocephalus, they include the agamids Ctenophorus and Tympanocryptis in Aus- tralia, the phrynosomatid sand lizards in North America (Uma, Callisaurus, Holbrookia and Cophosaurus), tropidurines in south America (Leiolaemus), geckoes in southern Arabia and Somalia (Pristurus) and lacertids in Southwest Africa (Meroles anchietae). However, although they show significant parallels in morphology and behaviour, these derived features are not necessarily assembled in the same order (Arnold, 1994). Nomenclature As presently understood, Phrynocephalus is a well-defined clade defined by six synapomorphies not found in closely related agamids (numbers 1.1, 12.1, 23, 35, 37.1 and 46 in the present data set). Besides lacking these, Bufoniceps, the sister taxon of Phrynocephalus, possesses at least one apomorphy not found in the latter genus, namely a very short tail. Golubev & Dunayev (1997) suggested that Bufoniceps should be expanded to include P. mystaceus, P. maculatus, P. arabicus, P. ornatus, P. clarkorum, P. luteoguttatus, P. euptilopus, P. interscapularis and P. sogdianus. These are all basal members of Phrynocephalus and their inclusion in Bufoniceps would create a new grouping that is clearly paraphyletic and reduce Phrynocephalus to a smaller and less well defined clade. The suggestion should consequently be rejected. ACKNOWLEDGEMENTS. I am grateful to Jens Vindum (California Acad- emy of Sciences for the loan of material of Phynocephalus roborowskiii, P. rossikowi and P. strauchi. Ed Wade drew Figures 7-11. REFERENCES Ananjeva, N.B. & Sokolova, T.M. 1990. The position of the genus Phrynocephalus Kaup 1825 in agamids system. Trudy Zoologicheskogo Institituta, Akademiya Nauk SSSR, Leningrad, 207: 12-21. (In Russian, English Summary) Ananjeva, N.B. & Tuniyey, B.S. 1992. Historical biogeography of the Phrynocephalus species of the USSR. Asiatic Herpetological Research, 4: 76-98. Arnold, E.N. 1984. Ecology of lowland lizards in the eastern United Arab Emirates. Journal of Zoology, London, 204: 329-354. 1992. The Rajasthan toad-headed lizard, Phrynocephalus laungwalaensis (Rep- tilia: Agamidae), represents a new genus. Journal of Herpetology, 26: 467-472. — 1994. Do ecological analogues assemble their common features in the same order? An investigation of regularities in evolution, using sand-dwelling lizards as examples. Philosophical Transactions of the Royal Society London B, 244: 277-290. 1995. Identifying the effects of history on adaptation: origins of different sand diving techniques in lizards. Journal of Zoology, 235: 351-388. 1998. Structural niche, limb morphology and locomotion in lacertid lizards (Squamata, Lacertidae); a preliminary survey. Bulletin of the Natural History Mu- seum, London (Zoology), 64: 63-90 Submitted. Modes of ear reduction in iguanian lizards (Reptilis: Iguania), different paths to similar ends. Baig, K. J. & Bohme, W. 1997. Partition of the ‘Stellio’ group of Agama into two distinct genera: Acanthocercus Fitzinger, 1843, and Laudakia Gray, 1845 (Sauria: Agamidae). Herpetologia Bonnensis, 1997; 21-25. Bannikoy, A.G., Darevskii, I.S., Ischenko, V.G., Rustamov, A.K. & Shcherbak, N.N. 1977. Opredelitel Zemnovodnykh i Presmykayushchikii Fauny SSR. Moscow: Prosveshenie Carothers, J.H. 1986. An experimental confirmation of morphological adaptation: toe Bt fringes in the sand-dwelling lizard Uma scoparia. Evolution, 40: 871-874. Chernoy, S.A. 1948. Reptilia. In E. N. Pawlovsky and B. S. Vinogradov (eds) The Animal World of the USSR, Vol.2 The Zone of Desert. Pp. 127-161. Moscow: USSR Academy of Science Press. (In Russian) Clark, R. 1992. Notes on the distribution and ecology of Phrynocephalus clarkorum Anderson & Leviton 1967 and Phrynocephalus ornatus Boulenger 1887 in Afghani- stan. Herpetological Journal, 2: 140-142. Dunayey, E.A. 1996. On the possible use of the ethological features in the taxonomy and phylogeny of toad agamas, Phrynocephalus (Reptilia, Agamidae). Russian Journal of Herpetology, 3: 32-38. Farris, J.S. 1988. Hennig86 program and documentation. Port Jefferson, New York: published by the author. Gallagher, M.D. & Arnold, E.N. 1988. Reptiles and amphibians of the Wahiba Sands, Sultanate of Oman. Journal of Oman Studies Special Report No.3: 405-413. Golubey, M.L. 1989. Phrynocephalus guttatus (Gmel) or Ph. versicolor Str. (Reptilia, Agamidae): what Phrynocephalus species occurs in Kazakhstan? Vestnik Zoologii, 5: 38-46. — & Dunayev, E.A. 1997. On the relationships between agamid genera Phrynocephalus Kaup 1825 and Bufoniceps Arnold 1992. Abstracts of the Third World Congress of Herpetology, 2-10 August 1997, Prague, Czech Republic. p. 86. Henke, J. 1974. Vergleichend-morphologische Untersuchungen am Magen-Darm- Trakt der Agamidae und Iguanidae (Reptilia: Lacertilia). Zoologische Jahrbiicher, Anatomie, 94: 505-569. Henle, K. 1995. A brief review of the origin and use of ‘Srellio’ in herpetology and a comment on the nomenclature and taxonomy of agamids of the genus Agama (sensu lato) (Squamata: Sauria: Agamidae). Herpetozoa, 8: 3-9. Joger, U. 1991. A molecular phylogeny of the agamid lizards. Copeia, 1991: 616-622. Leviton, A.E., Anderson, S.C., Adler, K. & Minton, S.A. 1992. Handbook to Middle East Amphibians and Reptiles. Contributions to Herpetology, 8:\—252. Minton, S.A. 1966. A contribution to the herpetology of West Pakistan. Bulletin of the American Museum of Natural History, 134: 27-184. Moody, S. 1980. Phylogenetic and Historical Biogeographical Relationships of the Genera in the Family Agamidae (Reptilia: Lacertilia). Ph.D. Thesis, University of Michigan. Ross, W. 1989. Notes on ecology and behaviour with special reference to tail signalling in Phrynocephalus maculatus (Reptilia: Agamidae). Fauna of Saudi Arabia, 10: 417-422. — 1995. Tail signalling in populations of Phrynocephalus arabicus Anderson 1894 (Reptilia: Agamidae). Zoology in the Middle East, 11: 63-71. Schatti, B. & Gasperetti, J. 1994. A contribution to the herpetofauna of southwest Arabia. Fauna of Saudi Arabia, 14: 348-423. Schwenk, K. & Greene, H.W. 1987. Water collection and drinking in Phrynocephalus helioscopus: a possible condensation mechanism. Journal of Herpetology, 21: 134-139. Semenoy, D.V. 1987. Systematics, ecology and behaviour of the Phrynocephalus of the guttatus-group (Reptilia, Agamidae, Phrynocephalus). Thesis. Severtsov Insititute of Evolutionary Morphology and Animal Ecology, Moscow. & Danayey, Y.A. 1989. Morphology of the hemipenis and classification of lizards of the genus Phrynocephalus (Reptilia, Agamidae). Zoologicheskii Zhurnal, 68: 59— 64. Sharma, R.C. 1978. A new species of Phrynocephalus Kaup (Reptilia: Agamidae) from the Rajasthan Desert, India with notes on its ecology. Bulletin of the Zoological Survey of India, 1: 291-294. Shine, R. 1985. The evolution of viviparity in reptiles: an ecological analysis. Biology of the Reptilia, 15B: 606-694. Smith, M.A. 1935. The Fauna of British India, Reptilia and Amphibia, 2 — Sauria. London: Taylor & Francis. Stebbins, R.C. 1943. Adaptations in the nasal passages for burrowing in the saurian genus Uma. American Naturalist, 77: 38-S2. Stebbins, R.C. 1944. Some aspects of the ecology of the iguanid genus Uma. Ecologi- cal Monographs 14: 311-332. Stebbins, R.C. 1948. Nasal structure in lizards with reference to olfaction and conditioning of the inspired air. American Journal of Anatomy, 83: 183-221. Stejneger, L. 1933. In Smith, M.A. Some notes on the monitors. Journal of the Bombay Natural History Society, 35: 615-619. Swofford, D.L. 1993. Phylogenetic Analysis Using Parsimony, version 3.1.1. Washing- ton: Smithsonian Institution. Whiteman, R.S. 1978. Evolutionary History of the Lizard Genus Phrynocephalus (Lacertilia, Agamidae). M.A. Thesis: California State University, Fullerton. Zerova, G.A. & Chkhikvadze, V.M. 1984. Review of Cenozoic lizards and snakes in the USSR. News. Georgian SSRAcademy of Sciences, Biological Series, 10: 319-325 (In Russian). Zhao, E-M & Adler, K. 1993. Herpetology of China. Contributions to Herpetology,10: 1-521. 12 Appendix 1 Data set for Phrynocephalus and its relatives. Figures above columns refer to characters listed on pp. 2—7. — indicates no data or character uncheckable or intermediate; v indicates character variable. 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 10 Laudakia Trapelus i Bufoniceps P. mystaceus P. maculatus P. arabicus 6 P. ornatus P. clarkorum 8 9 10 11 P. euptilopus P. luteoguttatus 2 P. interscapularis P. sogdianus 12 13 P. scutellatus P. golubevi 14 15 16 17 18 19 20 21 P. reticulatus P. raddei P. rossikowi P. strauchi P. persicus il? P. helioscopus P. forsythii P. roborowskii P. theobaldi P. viangalii 22 23 24 25 P. axillaris P. guttatus 26 P. versicolor Pf P. przewalskii 28 E.N. ARNOLD PHRYNOCEPHALUS PHYLOGENY Appendix 1 continued 27 28 29 300) sil 82 33 34 35 365 sy 38 39) 40) 41 42 43 44 45 46 26 Laudakia Trapelus 1 Bufoniceps 3 4 5 6 7 8 P. mystaceus P. maculatus P. arabicus P. ornatus P. clarkorum P. euptilopus P. luteoguttatus 10 11 12 13 0 P. interscapularis P. sogdianus P. scutellatus 0 P. golubevi 14 15 16 17 18 0 P. reticulatus P. raddei P. rossikowi N P. strauchi P. persicus 19 P. helioscopus P. forsythii 20 21 P. roborowskii P. theobaldi P. vlangalii 22 23 24 25 P. axillaris P. guttatus 26 P. versicolor 27 P. przewalskii 28 13 Bull. nat. Hist. Mus. Lond. (Zool.) 65(1): 15-21 Issued 24 June 1999 Rita sacerdotum, a valid species of catfish from Myanmar (Pisces, Bagridae) CARL J. FERRARIS, JR. Department of Tehthyology, California Academy of Sciences, Golden Gate Park, San Francisco, CA 94118, SYNOPSIS. Rita sacerdotum Anderson, 1879, is the valid name for the only species of the south Asian bagrid catfish genus Rita that resides in Myanmar. This species is distinguished from other species of Rita by a comparatively short dorsal-fin spine that never extends to the adipose fin base; palatal tooth patches, composed primarily of uniformly sized molariform teeth, that are in broad contact across the midline anteriorly, but diverge posteriorly and terminate in an acute point; and a small eye that is only about one-eighth to one-tenth the length of the head. Rita sacerdotum resides in the Sittoung and Ayeyarwaddy rivers, at least as far north as Myitkyina, Kachin State. This species is redescribed and a new key to the species of Rita is provided. INTRODUCTION The south Asian catfish genus Rita is broadly distributed in India, Bangladesh, Nepal, the Indus plain of Pakistan, and the Ayeyarwaddy system of Myanmar. Jayaram (1966) provided a systematic account of the species of the Rita, and subsequent reviews have been provided by Misra (1976), Jayaram (1977, 1981), and Talwar & Jhingran (1991). In each of these treatments, four species of Rita are recognized, although the specific names used for some of the species differ. The identity of the single species of Rita that inhabits the Ayeyarwaddy River, the focus of this paper, also varies among these studies. This uncertainty has existed since the earliest accounts of | the presence of Rita in Myanmar waters, its persistence due in part | to the dearth of material available for study and an enigmatic | proposal of a name for the Ayeyarwaddy species. I recently obtained | additional specimens of this species, and this prompted me to reexamine the question of their identity. The discovery of the holotype of Rita sacerdotum Anderson, 1879, a specimen that has, for all intents and purposes, been lost since the species was named, | revealed that there is only one species of Rita in the Ayeyarwaddy system, and that Rita sacerdotum is the valid name for that species. METHODS AND MATERIALS Measurements are all straight line distances. Specimen lengths are all reported as standard length. Institutional abbreviations follow | Levitonetal. (1985). Other abbreviations used herein are: HL — head | length; SL — standard length. When refering to previously published accounts of the region and in the list of the specimens examined, I repeat the name Burma (or Burmah), for the country now known as Union of Myanmar. In all | other places, use Myanmar. Throughout the text, I use the officially | accepted spellings: Yangon for Rangoon, Ayeyarwaddy for Irrawaddy, Bago for Pegu, and Sittoung for Sittang or Sitang. | MATERIAL EXAMINED | Rita chrysea, 54 specimens, 61—205 mm. INDIA: Orissa, Mahanadi River at Cuttack, K. Jayaram, CAS 54540 (6:107—130 mm, | cleared & stained). Mahanadi River basin, Sonepur fish market, T. Roberts, 19-22 Feb 1985, CAS 61855 © The Natural History Museum, 1999 (43:61—205 mm). Mahanadi River at Amicut, Cuttack, K. Jayaram, 23 Oct 1954, SU 48799 (2:110—-118 mm). Bihar, Sheonath River at Bisrampur, A. Herre, 13 Dec 1940, SU 41043 (1:106 mm). Rita gogra, 6 specimens, 112—205 mm. INDIA: Andhra Pradesh, Poona, Bombay Pres., A. Herre, 1940, SU 41044 (1:123 mm). Maharashtra, Godavari River, Nanden market, K. Jayaram, 10 Feb 1955, USNM 11494 (1:112 mm). Karnataka, Krishna River basin, Tungabahdra River or reservoir at Hospet, Hampi, or Kampli, T. Roberts, 28 Jan—3 Feb 1985, CAS 62088 (5:138-205 mm). Rita kuturnee, 8 specimens, 57-97 mm. INDIA: Andhra Pradesh, Tungabhadra River, K. Jayaram, 10 Feb 1955, USNM 114950 (3:61—88 mm);Tungabhadra River, at Kurnool, K. Jayaram, 10 Feb 1955, SU 48798 (2:57-59 mm). Karnataka, Krishna River basin, Tungabahdra River or reservoir at Hospet, Hampi, or Kampli, T. Roberts, 28 Jan—3 Feb 1985, CAS 62077 (1:59 mm). Maharashtra, Poona, A. Herre, 9 Apr, 1937 SU 34868 (2:83- 97 mm). Rita rita, 16 specimens, 24-258 mm. INDIA: Bihar, Ganges River at Patna, T. Roberts, Apr-May 1996, CAS 92501 (1:61 mm). Uttar Pradesh, Allahabad, Ganga River, 8— 12-1974 [sic], USNM 317823 (1:214 mm). West Bengal, Hugli River at Pulta, A. Herre, 23 Oct 1954, SU 34866 (10:70—94 mm, one additional specimen in lot, not examined). Calcutta, A. Herre, 9 Apr 1937, SU 34867 (2:187-194 mm). Calcutta, A. Herre, 6 Apr 1937, SU 14132 (1:258 mm). BANGLADESH: North Central Region, Tangail District, Ganges River basin, 5 Nov 1992, CAS 92411 (1:24 mm). Rita sacerdotum, 23 specimens, 22-690 mm. MYANMAR: ‘3rd Defile of Irrawaddy River, Upper Burmah, Dr. Anderson’, BMNH 1875.8.4.7 (1:690 mm, holotype of Rita sacerdotum). Sittoung River, E. Oates, BMHN 1891.11.30:242 (1:285 mm), BMNH 1891.11.30.343 (1, disarticulated dry skeleton, not measured). Ayeyarwaddy Division, Wa-ke-ma town market, 17 Sep 1996, Myint Pe, NRM 40631 (4:107—135 mm). Bago Division, Bago market, 25 Oct 1997, C. J. Ferraris, Myint Pe, Mya Than Tun, BMNH 1998.3.11.1 (1:150 mm), CAS 99210 (1:126 mm). Yangon Division, Hlaing River, 31 Oct 1997, C. J. Ferraris, Mya Than Tun, CAS 99309 (10:22—70 mm). Insein market (northern Yangon), July 1996, Myint Pe, AMNH 224490 (1:195 mm). Insein market, Nov 1997, Pe, C. J. Ferraris, Mya Than Tun, USNM 348211 (2:191—203 16 mm). Rangoon Market, A. Herre, 14 Nov 1940, SU 39869 (2:172— 184 mm). HISTORY OF THE IDENTIFICATION OF THE AYEYARWADDY RITA Day (1873) provided the first mention of Rita from the Ayeyarwaddy River in his account of the fishes of India and British Burma, under the name Rita ritoides (Valenciennes, 1840), aname now considered a junior synonym of Rita rita (Hamilton, 1822) (Jayaram, 1966). Several years later Day (1877) included the Ayeyarwaddy within the range of Rita buchanani Bleeker 1853, another junior synonym of Rita rita. Day acknowledged that Rita ritoides might have been the appropriate name for the species, but departed from his earlier use of that name and, without explanation, used R. buchanani instead. The name Rita sacerdotum was proposed in Anderson (1878 [1879]) for a species from the middle reaches of the Ayeyarwaddy River. As noted by Jayaram (1966), several authors have attributed the description of this species to Francis Day, presumably on the basis of a statement in the book’s acknowledgements (Anderson, 1878 [1879]: xxiv) which states that Day ‘favored me with a list of fishes collected on the First Expedition, and undertook the descrip- tion of certain species’. However, the species described by Day are those published elsewhere (Day 1870a, 1870b, 1871) and not the ones that first appeared in Anderson (1878 [1879]). The style of writing and the choice of anatomical characters are significantly different from that of Day’s other published species descriptions. It is important to note that the actual publication date for the species description, and for the volume as a whole, differs from that on the title page. A statement in the published corrigenda that follows the title page clearly indicates that publication was unexpectedly de- layed past 1878, the date on the title page, and was issued, instead, in 1879. Anderson’s (1878 [1879]) description of Rita sacerdotum was based on his field observations of living examples of the species that were treated as pets by the residents of a Buddist temple as well as a single specimen that was secured and illustrated. The account was published in a summary of an expedition to western Yunnan, along with accounts of other species from Yunnan and ‘upper Burmah’. Because of the title of the publication, some accounts have mistak- enly cited the type locality of this species as Yunnan. In neither Day’s (1888) Supplement to the fishes of India, nor his modified and updated version of his earlier book (Day, 1889) is Anderson’s Rita sacerdotum (or the other two fish species described by Anderson) mentioned. The reason for this curious omission is unknown. Itis possible, but highly unlikely, that Day was unaware of Anderson’s book with its included species accounts. Day and Anderson must have known each other, as evidenced by the above mentioned acknowledgement of Day’s assistance by Anderson. Day may have considered the species to lie outside the scope of his own book, as it was described from Upper, rather than British, Burma. For whatever reason, Day’s failure to include mention of Rita sacerdotum in either of the two accounts he published on fishes of southern Asia appears to have been a major factor in the subsequent oversight of Anderson’s name. Vinciguerra (1890) reported on a specimen of Rita from the vicinity of Yangon, under the name Rita ritoides. He noted that his specimen differed from the typical form of R. ritoides in the relative length of the dorsal spine and the shape of the humeral process. Vinciguerra compared his specimen with the description of Rita sacerdotum, and decided that it too differed from his specimen on C.J. FERRARIS, JR. several features, but that the two specimens shared a comparatively short dorsal-fin spine. On that basis, he concluded that two distinct forms of Rita ritoides existed, one in Myanmar and one in India. After a period of more than a half century without any mention of Rita from the Ayeyarwaddy, Jayaram (1966) revised the genus Rita and concluded that two species were found in Myanmar: Rita rita and R. kuturnee (Sykes, 1839). Inclusion by Jayaram of Rita rita in the fauna of Myanmar appears to be based solely on the literature accounts of Day (1873) and Vinciguerra (1890). Jayaram tentatively placed Rita sacerdotum into the synonymy of that species. All of the specimens from the Ayeyarwaddy River, or elsewhere in Myanmar, that were cited as having been examined by Jayaram were listed in the account of Rita kuturnee. However, Rita kuturnee, and its widely used junior synonym Rita hastata (Valenciennes, 1840), is a species otherwise known only from the rivers of peninsular India. Talwar & Jhingran (1991) doubted that R. kuturnee actually occurs in Myanmar, even though Jayaram (1977, 1981) had continued to list it in subsequent accounts of the distribution of that species. Misra (1976) included Myanmar in the distribution of Rita rita, but not that of R. kuturnee. In his abbreviated synonymy for R. rita, there is no mention ofR. sacerdotum, and the publication ofAnderson (1878 [1879]) is likewise missing from the literature cited. Talwar & Jhingran (1991) similarly listed Rita rita as the only species of Rita from Myanmar, but they tentatively included Rita sacerdotum in the synonymy of that species. IDENTITY OF THE AYEYARWADDY SPECIES OF RITA Although much of the recent literature suggests that the Rita species inhabiting the Ayeyarwaddy River is Rita rita, the species in that basin is, in fact, clearly distinct from R. rita. During this study, specimens of R. rita, from various parts of the Ganges basin, the type locality of the species, were found to exhibit characters lacking in specimens from the Ayeyarwaddy. As first noted in Jayaram (1966), the palatal teeth of theAyeyarwaddy Rita specimens are not arranged in the broad, elliptical patches characteristic of R. rita but, instead, in ‘pear-shaped’ patches that tapered posteriorly nearly to a point (Figure la). In addition, the dorsal-fin spine of R. rita is long and stout with its length at least equal to the head length. The adpressed spine usually extends well past the adipose-fin origin, at least in large individuals. Day (1877) noted that the relative size of the spine was apparently allometric, and that in small individuals it may only equal the head length, but that in larger individuals it may exceed 1.3 times HL. In Ayeyarwaddy specimens, in contrast, the dorsal-fin spine is never as long as the head and, more typically, it is shorter than the head minus the snout, even in the largest specimens. In contrast to the prevailing view, Jayaram (1966) identified the Ayeyarwaddy specimens as R. kuturnee. It appears that his conclu- sion is based primarily, but erroneously, on the similarity of the palatal tooth patches in the two species. In placing the Ayeyarwaddy Rita into R. kuturnee, he also looked beyond several striking differ- ences between the two species. For example, the eye size of R. kuturnee is significantly larger than that of the Ayeyarwaddy Rita. In his diagnosis of R. kuturnee, Jayaram (1966) lists the eye size as ‘Eye 3.07 (2.70 to 4.70 of up to 8.80 in specimens from Burma) in head length; 1.35 (1.00 to 1.50 or 3.90) in interorbital space width; 1.39 (1.00 to 1.50 or 3.00) in snout length.’ It is possible that Jay- aram interpreted the consistant disparity in eye proportions between the Indian and Ayeyarwaddy specimens as a result of allometric growth in R. kuturnee. All of the specimens he examined from ae Ra Set, ee a a MYANMAR CATFISH Fig. 1 Diagrammatic representation of tooth patches on jaws and palate of Rita species. A. Rita sacerdotum Anderson, 184 mm, SU 39869. Scale bar = 1 mm.; B. Rita kuturnee (Sykes), 97 mm, SU 34868. peninsular India are substantially smaller (36 to 103 mm) than any of the specimens listed from Myanmar (184 to 318 mm). In the Ayeyarwaddy specimens that I examined (22 to 285 mm), the eye length is always 8 to 10 times in the head length. Curiously, despite the observation that the Ayeyarwaddy specimens has small eyes, Jayaram used the eye size of R. kuturnee to help distinguish it from R. chrysea, a species for which he lists the eye diameter as ‘3.76 (2.83 to 5.22)’ in head length. Clearly he did not take into account the Ayeyarwaddy specimens in this diagnosis of R. kuturnee. The shape of the palatal tooth patches, a characteristic on which Jayaram placed heavy emphasis, also differs between Rita kuturnee and the Ayeyarwaddy form. The ‘pear-shaped’ tooth patches that Jayaram (1966, Figure 1b) described and illustrated as characteristic of R. kuturnee appear, in fact, to be those of the Ayeyarwaddy river species, and not R. kuturnee. In all of the specimens of R. kuturnee that I examined, the palatal tooth patches are slender, crescent- shaped arches that are either separated at the midline (Figure 1b), or meet only for the width of a single row of teeth. The palatal tooth- patches in Rita kuturnee have stout conical teeth, larger in size than those of the premaxilla, rather than the broadly rounded, or molari- form ones that predominate in the palate of the Ayeyarwaddy species (Figure la). As with the size of the eye, it is possible that Jayaram assumed that his specimens of R. kuturnee from peninsular India were juveniles, with incompletely developed palatal tooth patches, and that the adult condition in the peninsular population is like that in the Ayeyarwaddy specimens. Even in the smallest examined specimens from the Ayeyarwaddy basin the palatal tooth patches are broadly in contact across the midline and are composed primarily of molariform teeth. Thus, I conclude that the Ayeyarwaddy form is not conspecific with Rita kuturnee. The Ayeyarwaddy River Rita population has never been consid- ered conspecific with either of the two other Indian Rita species, R. chrysea Day, 1877 and R. gogra (Sykes, 1839), and I have found no reason to assign either name to theAyeyarwaddy fishes. Rita chrysea, restricted to the Mahanadi River and nearby tributaries in Orissa and considered to be the smallest species of Rita (Talwar & Jhingran, 1991), is characterized by a large eye (2.8 to 5.2 in HL) and by having a broad, nearly rectangular, patch of molariform teeth that extends across the midline of the palate (Jayaram, 1966). Rita gogra, which is sometimes listed as Rita pavimentata (Valenciennes, 1840) (e.g., Misra, 1976; Talwar & Jhingran, 1991), is known only from rivers of the Deccan region of peninsular India, including the Krishna, Harda, Godavari, Tungabhadra, Manjra, Bhima, and Mutha- Mula (Jayaram, 1966). Although similar in overall appearance with the Ayeyarwaddy Rita, R. gogra can be distinguished immediately by the unusual shape of its head. The dorsal surface of the head, posterior to the orbits, is dominated by a bilaterally symmetrical swelling formed by massive extensions of the adductor mandibulae muscle that cover the cranial roofing bones. All other species of Rita, including the Ayeyarwaddy species, have the dorsal surface of the cranium covered only with skin, through which the cranial roofing bones can easily be palpated. In addition, the Ayeyarwaddy Rita can be distinguished from R. gogra by the color of the mental barbel (black in R. gogra, white in the Ayeyarwaddy species). The palatal tooth-patch in R. gogra has finely conical teeth anteriorly and increasingly large molariform teeth posteriorly (Jayaram, 1966). Thus, it must be concluded that the Ayeyarwaddy Rita is not conspecific with any of its Indian congeners. The only remaining name that might apply is Rita sacerdotum, which was described from the Ayeyarwaddy. The description and published illustration of that species, however, only vaguely resembles a Rita, and character- istics of the Ayeyarwaddy species are either absent from the 18 description, or in variance with the illustration. Although Jayaram (1966) followed Vinciguerra (1890) in placing Rita sacerdotum into the synonymy of Rita rita, he place the specimens he examined from the Ayeyarwaddy into a second species of Rita. I have been unable, so far, to find any specimens that represented a second species from Myanmar. Clearly, an examination of the holotype of Rita sacerdotum was necessary to determine whether it indeed represented a species of Rita different from the one that I, and others, have observed. THE HOLOTYPE OF RITA SACERDOTUM ANDERSON Anderson (1878 [1879]) did not indicate where the holotype of Rita sacerdotum, or any of the other species described in the same paper, were deposited. Although I expected to find the specimen in The Natural History Museum, London, the holotype was not listed in its type catalog, and there was no entry for R. sacerdotum in their species catalog. In fact, no specimen of Rita collected by Anderson was listed in the catalog. An exhaustive search through the registers did uncover a Rita sacerdotum collected by Anderson, without any indication that it was a holotype. With the assistance of the staff of the Fish Section of the Zoology Department, the specimen was C.J. FERRARIS, JR. found among the collection of stuffed, dried fish specimens. Its identity as the holotype was promptly made by comparsion of the stated locality information and by direct comparision with the published illustration. It is puzzling that the specimen was never recognized as the holotype of Rita sacerdotum. Although the specimen was registered in 1875, prior to Anderson’s publication, the register entry (BMNH 1875.8.4.7) lists the name and is surrounded by entries for the other species named by Anderson. It is even more suprising that although the specimen was registered with the new name during Albert Giinther’s tenure, he did not include the specimen in his personal annotated copy of his catalog (Giinther, 1868) or annotate the register entry to indicate that the specimen was a holotype. Nonethe- less, with the discovery of the holotype, it is now possible to clarify some peculiar features in the illustration of Rita sacerdotum and, with that information, finally resolve the identity of theAyeyarwaddy Rita. The holotype of Rita sacerdotum is a dried, stuffed specimen, 69 cm in standard length (Figure 2). The specimen appears to have been placed on display at two different times, based on the fact that the stuffed skin has two forms of wire attachments. One set of mounts, extending from the ventral surface of the body, indicate that the specimen was at one time mounted freestanding, probably on a Fig. 2 Rita sacerdotum Anderson, holotype, 69 cm, BMNH 1875.8.4.7. Fig. 3 Published illustration of holotype of Rita sacerdotum, reproduced from Anderson (1878 [1879], pl. 79, Fig. 3). MYANMAR CATFISH Fig.4 Rita sacerdotum Anderson, 150 mm, BMNH 1998.3.11.1. Fig.5 Rita sacerdotum Anderson, 126 mm, CAS 99210. wooden stand. A second pair of wires protrudes from the left side of the body, suggesting that the specimen was mounted on a wall, with the right side of the body on display. The published illustration of the holotype (Figure 3) resembles the mounted specimen quite closely, except for some damage to the fins. Most importantly, the elongated caudal region of the body, which is identical in proportion to that in the illustration, suggests that the illustration was probably prepared from the dried mount rather than the freshly collected specimen. The body of the specimen is disproportionally long and the caudal region is far more slender and cylindrical than other specimens of Rita from the Ayeyarwaddy (Figures 4, 5). This unusual body form, and the illustration that resulted from drawing the dried specimen, have made comparison between the illustration and fresh specimens of the species problem- atic. On close inspection, it appears that the body of the mounted specimen must have been stretched well beyond the normal propor- tions of the species when it was stuffed. As Anderson collected only a single specimen, it is reasonable to assume that he or his taxider- mist had no model to use to shape his specimen, once it was skinned and the vertebral column removed. Although the general shape of the body does not closely resemble the other specimens from the Ayeyarwaddy, other features of the body are, in fact, quite similar and clearly indicate that the holotype and the other available speci- mens are conspecific. The shape of the palatal tooth patches, the unusually short dorsal spine, and small eye combine to distinguish this species from its congeners. All of the specimens that I have examined exhibit this same suite of characters, albeit with some ee Uf ontogenetic variation. It appears therefore that there is only one species of Rita in the Ayeyarwaddy system, and that the oldest available name for that species is Rita sacerdotum. DIAGNOSIS AND REDESCRIPTION OF RITA SACERDOTUM As noted above, the inaccurate taxidermic preparation of the holotype of Rita sacerdotum made the specimen longer than it would have been in life, and this precludes using the specimen for any propor- tional measurements standardized against the body length. Therefore, any statement in the description that relates a body measurement to the standard length does not include the holotype. Diagnosis Rita sacerdotum is readily distinguished from all congeners by the following combination of characteristics: eye small, its diameter 10— 13% head length; dorsal-fin spine length no greater than the length of the head posterior to the snout; adpressed dorsal-fin spine does not extend to adipose-fin origin; and palate with a single crescent- shaped patch of primarily large, bluntly conical, teeth of approximately uniform size. Description Body elongate, slightly compressed anteriorly, progressively more 20 compressed toward caudal fin. Body deepest at dorsal-fin origin, its depth at that point approximately equal to distance from nasal barbel to opercular margin; body depth decreases gradually to adipose-fin origin, more rapidly thereafter. Least depth at caudal peduncle approximately equals snout length. Skin of body and head covered with thick coat of mucous, anchored by fine filamentous projections from skin surface; filaments largest and most dense on chin and opercular margin of head, and, especially, on lateral surface of body ventral to dorsal fin. Vent slightly anterior to anal-fin origin. Lateral line midlateral and straight from past tympanum to hypural plate; anterior portion of lateral line more dorsally situated; lateral line bent sharply in the dorsal direction onto base of upper caudal-fin lobe posterior of hypural plate margin. Lateral line pores extend laterally from canal, through thick mucous coat. Anterior canal pores ramify and spread in asymmetric pattern over pectoral-girdle elements and tympanum. Cephalic canal pores similarly branch over dorsum of head and onto opercle. Head large, its length approximately 3% times in SL; head slightly depressed, at pectoral-fin origin its depth approximately 80% its width; head depth at orbit approximately 2/3 its width. Dorsal profile of head straight from orbit to snout, slightly convex posteriorly; ventral profile nearly straight. Mouth nearly terminal; upper jaw slightly overhangs lower. Teeth in upper jaw conical and sharply pointed, in 6 to 8 irregular rows. Tooth-bearing surface of premaxilla long and nearly transverse, its long axis four to five times its short axis. Tooth-bearing surface of mandible elongated, tapering posteriorly. Teeth in lower jaw pointed and conical along anterior margin of jaw, approximately equal in size to those of upper jaw; two rows of bluntly rounded teeth, much larger in size than conical teeth, present mesially; only blunt teeth present along posterior part of mandible. Palate with coalesced tooth patch extending across mid- line. Tooth patch convex anteriorly, concave posteriorly, with nearly parallel lateral margins. Teeth on palate nearly all in form of bluntly rounded pegs, slightly larger in diameter posteriorly, except for one or two rows of somewhat smaller teeth along lateral and anteriolateral margins of toothplate. Gill rakers 24 to 29; anterior 8 to 10 rakers on lower arch rudimentary, shorter than intervening spaces; posterior rakers moderately long and thick. Eye small, ovoid, with long axis parallel to body length; long diameter of orbit approximately 1/3 snout length, 1/5 interorbital width, and equal to or slightly greater than 1/10 head length. Orbital margin free. Anterior naris situated along anterior margin of snout, its opening a short tube, flared at margin, directed anteriorly. Posterior naris remote from anterior naris, and slightly more laterally situated; its anterior margin located midway between snout tip and anterior margin of orbit. Naris surrounded by short rim, connected to nasal barbel anteriorly. Head with three pairs of barbels. Maxillary barbel extends from fold between upper lip and skin of snout; barbel filamentous, without fleshy attachment to snout. Maxillary barbel short, not extending to margin of bony opercle. Nasal barbel short, its length approximately equal to orbital diameter; adpressed barbel reaches only to anterior margin of orbit. Ventral surface of head with single pair of mandibular barbels; barbel originates at vertical through anterior orbital margin; barbel filamentous, extending to, or nearly to, vertical through pectoral spine origin. Dorsal surface of supraoccipital, posttemporal and pterotic bones granular, remainder of head covered with smooth skin. Adductor mandibulae does not extend onto dorsal surface of cranium. Upper lip with several rows of short papillae along margin; papillae often multifurcated at tip. Lower lip broadly connected to C.J. FERRARIS, JR. skin of chin, separate laterally. Lip margin with papillae comparable to those of upper lip, at least medially. Opercular membrane free from isthmus at margin, but attached more basally; membranes broadly connected across midline, but separated posterior to isthmus connection. Branchiostegal rays 7 or 8. Dorsal-fin origin at approximately 40% of SL. Fin quadrangular, first ray longest and approximately two times that of last ray; last ray without membranous extension to body; fin margin straight. Fin base approximately 1/2 of HL and shorter than interspace between dorsal fin and adipose fin. Dorsal-fin spine stout, with sharply pointed tip. Spine length equals head length minus snout, or approxi- mately 15% SL. Anterior margin of spine produced into sharp keel, without serrations; lateral and posterior surfaces smooth. Dorsal spine preceded by fully formed spinelet. Dorsal fin preceded by coarsely granular predorsal bone; lateral extent of predorsal bone approximately equals that of supraoccipital spine. Dorsal fin rays II,7; posterior two rays appear as one, split at base. Adipose fin large; anterior fin margin straight, convex distally. Fin extends posteriorly well past its posterior insertion. Caudal fin deeply forked, lobes with acutely pointed tips; lobes slightly asymmetrical, dorsal lobe longer and sometimes with filamentous extension. Length of dorsal most primary ray approxi- mately three times length of middle rays. Procurrant rays few, short, not extending anteriorly onto caudal peduncle. Caudal fin rays iL /eeshail Anal fin quadrangular, anterior rays longest; posterior rays pro- gressively shorter, fin margin straight. Last ray not connected to caudal peduncle by membrane. Fin base short, approximately equal to that of adipose fin. Anal-fin origin slightly posterior to vertical through adipose fin origin. Anal-fin rays iv, 9-10. Pelvic fin abdominal, its origin posterior to vertical through posterior insertion of dorsal fin. First branched ray longest, follow- ing rays only slightly shorter. Adpressed fin just reaches anal-fin origin. Pelvic-fin rays 1,6. Pectoral fin acutely pointed; first branched ray longest, its length approximately three times posterior-most ray. Pectoral-fin spine stout, sharply pointed at tip. Spine with short filament at tip, length of filamentous extension approximately equals snout length. Outer margin of spine produced into acute keel; keel very finely serrated for basal quarter, smooth for remainder of its length; in small specimens, most of spine margin covered with tiny transverse serrations. Inner spine margin with densely packed, pointed, retrorse serrations; serration height greater than length of space between successive serrations. Humeral process acutely pointed posteriorly, with a slightly rounded tip. In larger specimens, process becomes more rounded posteriorly, as in holotype (Figures 2, 3). Surface of humeral process granular, with granulations less coarse than those of cranial surface. Pectoral-fin rays I,10 or I,11. Coloration in preservative Body gray, darker dorsally, gradually becoming lighter ventrally; abdomen nearly white. Head dark gray dorsally, white ventrally; transition between gray and white regions fairly abrupt, occuring ventral to eye and approximately in line with maxillary barbel origin. Operculum gray with white margin. Orbit surrounded by distinct white ring. Maxillary barbel dark grey, mental barbel nearly white. Dorsal, anal, and pectoral fins pale, with broad black margin. Pelvic fin uniformly pale or with some indication of dark margin. Caudal fin with fine dark margin on middle rays; darkened margin progressively larger toward lobe tips. MYANMAR CATFISH Distribution Rita sacerdotum appears to be distributed widely through the Ayeyarwaddy River basin of Myanmar. Specimens examined during this study were all from the lower portions of the basin, but I observed a few large individuals in markets as far upriver as Manda- lay. In addition, Rita is seen occasionally in the Myitkyina market (U Tun Shwe, pers. comm.). Outside of the Ayeyarwaddy basin and its extensive delta, there is one record of specimens from the Sittoung River (BMNH 1891.11.30.242—243), and two from the Bago River (BMNH 1998.3.11.1 and CAS 99210). Natural History Little is known about Rita sacerdotum. While small specimens, up to about 25 cm, are routinely found in markets of Yangon and smaller delta villages, at least during the rainy season (April to September; U Myint Pe, pers. comm.), large specimens are only rarely seen in markets. All of the specimens examined during this study appear to be juveniles and there is no published indication of the size of maturity for this species. Individuals as small as 22 mm were obtained from the tidal rivers in the vicinity of Yangon in November, 1997. The presence of these tiny individuals in the lower course of the river suggests that Rita may reproduce in the estuarine part of the river and disperse more widely throughout the river at a larger size. This idea is supported by anecdotal reports that large numbers of large Rita appear at an island pagoda, in the middle reaches of the Ayeyarwaddy River, for a short period of time during monsoon season (U Nyi Nyi Lwin, pers. comm.). This may be indicative of a spawning migration. Examination of the gut contents of a few specimens revealed that Rita sacerdotum feeds on a variety of aquatic and terrestrial inverte- brates. Several specimens contained fragments of small glass prawns, and others contained pieces of winged insects. A comprehensive study of the food habits is not possible at this time due to the relatively small number of specimens available and the fact that the specimens in collections represent only juveniles. KEY TO SPECIES OF RITA 1. Dorsal surface of head, between eyes and occipital spine, covered with tckalayeriot muscle; pelvic fin BACK ........cerecncescnssecceseorseccessrsuceenses Bees sate ses Rita gogra (rivers of the Deccan region of peninsular India) Dorsal surface of head covered only with skin; pelvic fin pale ........ 2, PS CAS HUTA Men NS GUE nsec avacnceesscseevaveceneseecceescoseeascsteesascisieunstsatorseces 3 een O40) Je EMM a auc, sn tevnseunannct teesuaies Mhousaanecisnstaenctanscessnedeientaess 4+ 3. Dorsal-fin spine as long, or longer than length of head, adpressed spine extends to or beyond adipose-fin origin (in specimens greater than 100 mm); palatal teeth in two elliptical patches, not meeting at midline; teeth on posterior extent of lower jaw and palate molariform, much ADOC Ra ATID Lett OIMUC EL Dest tee are oe cece ce fs sant car cmoSew sess de Slee ceaust sesso Dorsal-fin spine no longer than head minus snout, adpressed spine not reaching adipose-fin origin; palatal teeth in a single crescent patch that extends across midline of palate; teeth on palate more or less uniform in SUZ et OUICe IVAN) eeccceer ens Nearer tee ES os aca cas cbeca caveat sdenestintscendsscocess Rita sacerdotum (Myanmar: Ayeyarwaddy and Sittoung River basins) 4. Palatal teeth in slender patches along lateral margin of palate, no larger than teeth in upper jaw and not meeting at midline (Figure 1B), dorsal- fin spine smooth anteriorly, except for few serrae basally ...............0.. ee tS Rita kuturnee (rivers of Deccan region of peninsular India) 21 Palatal teeth in large quadrangular patch that covers most of palate; teeth large and molariform in middle of patch, smaller laterally; dorsal-fin spine with single row of antrorse serrae, for at least basal 2/3 of spine ACKNOWLEGEMENTS. Examination of specimens for this study was facili- tated by David Catania and William Eschmeyer (CAS), Darrell Siebert and Oliver Crimmon (BMNH), and Richard Vari (USNM). Tyson Roberts pro- vided valuable insights during early stages of this study. The photograph of the holotype of Rita sacerdotum was arranged by Darrell Siebert and taken by Phil Hurst (BMNH). Al Leviton provided the photographic reproduction of the published illustration of Rita sacerdotum, from his personal copy of Anderson (1878 [1879]). Molly Brown drew figure | and Alison Schroeer drew figure 5. Travel to Myanmar was undertaken as part of a series of consultancies for the United Nations Food and Agriculture Organization. Kent Carpenter and Dora Blessich of that organization were instrumental in making these trips possible. Collection of specimens in Myanmar was made possible by the Myanmar Department of Fisheries; several individuals, including U Hla Win, U Nyi Nyi Lwin, U Myint Pe and U Mya Than Tun, provided me with assistance and valuable information. Financial support to travel to London and to examine specimens at The Natural History Museum was provided by TWA and the Inhouse Research Fund of the California Academy of Sciences. Without the help of all of these persons and organiza- tions, this study could not have been undertaken. The manuscript was improved by comments from James Atz, Nigel Merrett, Darrell Siebert and Richard Vari. REFERENCES Anderson, J. 1878 [1879]. Anatomical and zoological researches: comprising an account of the zoological results of the two expeditions to Western Yunnan in 1868 and 1875; and a monograph of the two cetacean genera, Platanista and Orcella. 984 pp., 84 pls. Bernard Quaritch, London. Day, F. 1870a. Remarks on some of the Fishes in the Calcutta Museum. — Part I. Proceedings of the Zoological Society of London 1869 (3): 511-527. 1870b. Remarks on some fishes in the Calcutta Museum — Part II. Proceedings of the Zoological Society of London 1869 (3): 548-560. 1871. Monograph of Indian Cyprinidae. Parts 1-3. Journal of the Asiatic Society of Bengal 40 (pt 2, no. 1-4): 95-142, 277-367, 337-367, Pls. 9, 21-23. 1873. Report on the fresh water fish and fisheries of India and Burma. 307 pp. Office of the Superintendent of Government Printing, Calcutta. 1877. The fishes of India, being a natural history of the fishes known to inhabit the seas and fresh waters of India, Burma, and Ceylon. Part 3:369-552, pls. 79-138. Bernard Quaritch, London. 1888. Fishes of India, supplement, October 1888. pp. 779-816. Bernard Quaritch, London. 1889. The fauna of British India, including Ceylon and Burma. Fishes, 1. xx +548 pp. Taylor and Francis, London. Jayaram, K.C. 1966. Contributions to the study of bagrid fishes (Siluroidea: Bagridae). 1. A systematic account of the genera Rita Bleeker, Rama Bleeker, Mystus Scopoli, and Horabagrus Jayaram. Internationale Revue der Gesamten Hydrobiologie 51 (3): 433-450. — 1977. Aid to the identification of the siluroid fishes of India, Burma, Sri lanka, Pakistan, and Bangladesh. I. Bagridae. Records of the Zoological Survey of India, Miscellaneous Publications, Occasional Paper (8): 1-41. 1981. The freshwater fishes of India, Pakistan, Bangladesh, Burma and Sri Lanka —a handbook. viii + 475 pp, 13 pls. Zoological Survey of India, Calcutta. Leviton, A.E., Gibbs Jr, R.H., Heal, E. & Dawson, C.E. 1985. Standards in herpetology and ichthyology: Part 1. Standard symbolic codes for institutional resource collec- tions in herpetology and ichthyology. Copeia 1985 (3): 802-834. Misra, K.S. 1976. The fauna of India and the adjacent countries. Pisces (second edition), Vol. 3, Teleostomi: Cypriniformes: Siluri. xxi + 367 pp., 15 pls. Controller of Publications, Delhi. Talwar, P.K. & Jhingran, A.G. 1991. Inland fishes of India and adjacent countries. xx + 1158 pp. Oxford & IBH Publishing Co., Pvt. Ltd. New Delhi, Bombay and Calcutta. Vinciguerra, D. 1890. Viaggio di Leonardo Fea in Birmania e regione vicine. XXIV. Pisci. Annali del Museo Civico de Storia Nationale de Genova, Milano, (serie 2a) 9: 129-362, pls. 7-11. Bull. nat. Hist. Mus. Lond. (Zool.) 65(1): 23-29 Issued 24 June 1999 Indian Ocean echinoderms collected during the Sindbad Voyage (1980-81): 4. Crinoidea JANET I. MARSHALL CROSSLAND KX (S| Foe. | Museum of Tropical Queensland, 70-84 Flinders St, Townsville, Queensland 4810 Australia ANDREW R.G. PRICE Ecology and Epidemiology Group, Department of Biological Sciences, University of Warwick, Coventry CV4 7AL, UK CONTENTS A EMCNCA ULE Ck area tany Meant ee cvcetsecas caso ace MU ae Pee EEO c casa int Sac taal cs dexdacisaavucysvedersusessausucescsseseseaeepeinseenaeecucsiciessaaqeseauasesours 23 Meare tiled hg cit AMIE REL OGS be cane Savi casheat scenes die pvawceseen-eu-t sa ssaxasceuauwannanvatttansdnsnedssdacacsnueedacanceictvaccsudssicecesaedvancssetbacsdeace coacstenscctangt vasesebe 23 Nea Screens tae ee Re ah A a oe ect maaan nec Bie cay cam te obo ve dexGael duesuaedomencs ates tenons cota yevaes vnsinbucn oye MRasesveaugdevvesanvadevestcuas 23 MEI CVI GSI OTM ec cos co sean cane cnia auc pean en Seo denned d Sencans znsaps GuaSaunctevadacse detudch saacabia ss soehtoesnes asec nana vaauesach cuiiaus dicoeSbauh dewtaa dosussesasdvepecdnes 29 PS TOC LE LSE IMC INS sen oye cdeaa nears aie el td ncn cae wanes encca nasigy eater cee cab sccccnnt ener cpncass ad uosisndede conse athersueesoxeseseWacecuunarasuassteeSevaaaseseseunses 29 INCE R CM COSI eens ce atu see we ncaa Saudaxosiusaiwanseamasionatanasde tedavesaadenesesvasduscacsarctabantsncdsaesanchdesseuttovsasscobreccsdelsacesadsaccndcceassocasccasadtevexateents 29 SYNOPSIS. Thirty species of shallow-water Crinoidea, representing eighteen genera in six families, are recorded from collections made during the Sindbad Voyage (Oman to China) from the Lakshadweep (Laccadive) Islands, Sri Lanka and Pula We Sumatra). Following the zoogeographic subdivisions of Clark and Rowe (1971), extensions of range are recorded for at least six of the species: Clarkcomanthus albinotus (Indonesia/East Indies); Comanthus briareus (Sri Lanka area); Comanthus gisleni (Sri Lanka area & Indonesia/East Indies); Comanthus suavia (Sri Lanka area & Indonesia/East Indies); Comanthus wahlbergii (Maldive area, Sri Lanka area & Indonesia/East Indies), and Oxycomanthus bennetti (Indonesia/East Indies); and possibly also Comaster parvicirrus (Sri Lanka area — doubt about earlier record) and Comaster multifidus (Maldive area — specimens poorly preserved). In addition to the taxonomic treatment, ecological information for each crinoid species (habitat types, depth range) is provided and broadly analysed. INTRODUCTION | Systematically, crinoid taxonomy has undergone relatively few | changes since the monumental works of A.H. Clark (1915-1967), the major exception to this being the recent revisions to the family Comasteridae by Hoggett and Rowe (1986) and Rowe, Hoggett, _ Birtles and Vail (1986). | This paper is the fourth in a series reporting the collection of | echinoderms made during a cruise by one of us (ARGP) across the _ Indian Ocean from Oman to China. The expedition was undertaken in a replica of an ancient Arab sailing vessel, “Sohar’. Systematic accounts of the other echinoderm classes have already been published (Price & Reid, 1985; Marsh & Price, 1991; Price & Rowe, 1996). Thirty species of shallow-water crinoids from six families are listed, including nine new distribution records. Generally, comments | are made where the record extends or modifies a range of distribu- | tion, or to clarify the identification. Where no comment is offered, | the species was already known from the region and is widespread in | the Indo-West Pacific. | MATERIALS AND METHODS Specimens were collected by one of us (ARGP) and other expedition members atlocalities from Chetlat Island, Lakshadweep (Laccadives), Sri Lanka and Pula Wé, Sumatra (Indonesia). Details of sampling localities are shown in Figure 1. Sampling was undertaken principally © The Natural History Museum, 1999 on coral reefs using scuba. At each locality, details of habitat type and depth range were recorded, along with the number of individuals of each species. The number of specimens collected is placed in parenthesis after each sample number in the Material lists for each species. Material was fixed and preserved using standard methods (Lincoln & Sheals, 1979). Conditions on board and for specimen storage on ‘Sohar’ were not as sophisticated as on modern research vessels. Hence not all specimens returned were in good condition. Specimens were identified by JIMC. Where the identification was uncertain, due to the changes to crinoid taxonomy by Rowe, Hoggett, Birtles and Vail (1986) and Hoggett and Rowe (1986), confirmation was sought from one of the authors of those papers. In some cases, subsequent re-examination of specimens has engendered doubt, and this doubt is expressed in the text of this paper. . Where three or more specimens of a species were collected, representative specimens of that species were sent to the Singapore Museum (SM), as the regional museum; otherwise material was divided between the Natural History Museum (NHM), London, and the Western Australian Museum (WAM), Perth. Species are listed in families, and within each family, alphabeti- cally by genus and species. RESULTS Throughout this account synonymy has been confined, where possi- ble, to a single reference from which the original reference can be traced. J.I.M. CROSSLAND AND A.R.G. PRICE . 24 oO SO E 60 70 80 90 100 Muscat OMAN Chetlate ANDAMANS$ ° LACCADIVES «,° oe SRI LANKA NICOBARS# Tangalla MALDIVES ™ Ug Bau Fig. 1 (a) Map of northern Indian Ocean showing sampling areas (@) during Sindbad Voyage, with insert (b) for Pula Wé Sumatra. INDIAN OCEAN ECHINODERMS Class Crinoidea Family COMASTERIDAE 1. Alloeocomatella pectinifera (A.H. Clark, 1911) SEE. Clark and Rowe, 1971:6—7; Hoggett and Rowe, 1986:122; Messing, 1995: 644. MATERIAL. NHM — 810501C/2 (2), 810505C/1 (1); WAM — 810425F/2 (1); SM — 810421 A/1 (1). COLLECTION SITES. Pula Wé, Sumatra. Sabang Bay, Seukundo, and Ug Seukundo, HABITATS AND DEPTH RANGE. Subtidal rock/coral, coral reef, on gorgonian; 2—30m. COMMENTS. The species was described by A.H. Clark (1911) and placed (with reservation) in the genus Comissia, later to be included in anew genus Alloeocomatella by Messing (1995). The species has been found in the Maldives, Indonesia, the Great Barrier Reef (GBR) of Australia, Papua New Guinea, New Caledonia and the Marshall Islands. 2. Capillaster multiradiatus (Linnaeus, 1758) SEE. Clark and Rowe, 1971:6—7. MATERIAL. NHM — 810422E/3 (1), 810423C/4 (2), 810425F/2 (4), 810426B/1 (1), 810427D/3 (2), 810428D/2 (1), 810430A/1 (1), 810430A/7 (1), 810430A/23 (1), 810430A/30 (1), 810430A/31 (1), 810504A/2 (1); WAM — 810422C/3 (1), 810422D/9 (1), 810424D/5 (1), 810425B/1 (2), 810525E/1 (2), 810425E/2 (2), 810427D/1 (1), 810428D/1 (1), 810430A/11 (1), 810430A/22a (1), 810430A/25 (1 of 2), 810430A/24b (1), 810430A/29 (1); SM — 810422C/4 (1), 810422D/7 (1), 810424B/4 (1), 810425F/1 (4), 810425F/4 (2), 810425F/5 (1), 810426A/1 (1), 810430A/10 (1), 810430A/21b (1), 810430A/21d (2), 810430A/25 (1 of 2). COLLECTION SITES. Klah, Seulakoe, Sabang Bay, Ug Murung, Ug Tapa Gadja, Ug Seukundo and Rubiah, Pula Wé, Sumatra. HABITATS AND DEPTH RANGE. Subtidal rock, coral reef, coral rubble, fire coral, soft coral and gorgonian; 2—30m. COMMENT. This species is well-known across the Indo-Pacific region. However, its habitat varies; in some regions it inhabits exposed situations, in others it is cryptic. Specimens have been recorded from 0.5—1 m in Madagascar, to 77 m in the Bay of Bengal (Clark, 1972). In this collection, habitat and depth also varied. 3. Capillaster sentosus (Carpenter, 1888) SEE. Clark and Rowe, 1971: 6-7. WAM = 810501C/4 (1). Ug Seukundo, Pula Wé, Sumatra. Subtidal rock; 20m. MATERIAL. COLLECTION SITE. | HABITAT AND DEPTH. _ 4. Clarkcomanthus albinotus Rowe, Hoggett, Birtles and Vail, 1986. SEE. Rowe, Hoggett, Birtles and Vail, 1986: 238. MATERIAL. WAM -— 810428E/1(1). COLLECTION SITES. Ug Tapa Gadja, Pula Wé, Sumatra. 25 HABITAT AND DEPTH. Soft coral, 2—10m. COMMENT. This is a marked extension of range for this species, previously only recorded from the Great Barrier Reef, Papua New Guinea (Messing, 1994) and Japan. 5. Clarkcomanthus littoralis (Carpenter, 1888) SEE. Rowe, Hoggett, Birtles and Vail, 1986: 236. MATERIAL. NHM — 810423D/1 (broken), 810425C/3 (1), 810428C/4 (2); WAM — 810420A/2 (fragmented), 810430A/12 (1), 810501A/3 (1), 810504C/4 (1); SM —810421C/2 (1), 810428E/2 (1), 810430A/19 (1). COLLECTION SITES. Klah, Seukundo, Ug Bau, Subang Bay, Ug Tapa Gadja and Ug Seukundo, Pula Wé, Sumatra. HABITATS AND DEPTH RANGE. 10m. Subtidal rock/coral, coral reef; 2— 6. Clarkcomanthus luteofuscum (H.L. Clark, 1915) SEE. Rowe, Hoggett, Birtles and Vail, 1986: 233. MATERIAL. WAM - 810427D/1 (1). COLLECTION SITES. Ug Murung, Pula Wé, Sumatra. HABITAT AND DEPTH. Soft coral, 2—10m. 7. Comanthina nobilis (Carpenter, 1884) SEE. Rowe, Hoggett, Birtles and Vail, 1986: 243. MATERIAL. NHM — 810425A/18d (2), 810425D/1 (1), 810430A/ 34 (1); WAM — 810425A/18e (2), 810430A/16 (1); SM — 810425A/ 18 (1), 810430A/2 (1), 810501A/5 (2). COLLECTION SITES. Ug Murung and Ug Seukundo, Pula Wé, Sumatra. HABITATS AND DEPTH RANGE. 30m. Subtidal rock/coral, coral reef; 6— 8. Comanthina schlegelii (Carpenter, 1881) SEE. Clark and Rowe, 1971:6—7; Rowe, Hoggett, Birtles and Vail, 1986: 244. MATERIAL. NHM — 810422D/4 (1), 810423A/2 (1), 810430A/13 (1), 810504C/3 (1); WAM — 810423C/2a (1), 810430A/6 (1), 810504C/1 (1); SM — 810421C/1 (1), 810421C/4 (1), 810424D/3 @: COLLECTION SITES. Pula Wé, Sumatra. Seukundo, Klah, Ug Bau and Ug Seukundo, HABITAT AND DEPTH RANGE. Subtidal rock and coral reef; S—20m. 9. Comanthus briareus (Bell, 1882) SEE. Rowe, Hoggett, Birtles and Vail, 1986:218 MATERIAL. NHM — 810424B/5 (1); WAM — 810204A/8 (1). COLLECTION SITES. Kalpitiya, Sri Lanka; Seulakoe, Pula Wé, Sumatra. HABITAT AND DEPTH RANGE. Coral reef; 3—5m, 20—30m. COMMENT. The Sri Lankan specimen is a new locality record for 26 this species, and extends its range west into the Indian Ocean from Indonesia. 10. Comanthus gisleni Rowe, Hoggett, Birtles and Vail, 1986. SEE. Rowe, Hoggett, Birtles and Vail, 1986: 219. MATERIAL. NHM -— 810124A/1 (1), 810204A/3 (1),810430A/14 (1), 810430A/22a (3), 810504B/1 (1); WAM — 810124A/1 (1), 810124A/6 (1), 810204A/2 (1), 810425F/1 (1), 810430A/15 (1); SM — 810422D/5 (1), 810430A/4 (1), 810430A/14 (1), 810504A/1 (1). COLLECTION SITES. Galle and Kalpitiya, Sri Lanka; Klah, Sabang Bay, Ug Bau, Ug Seukundo and Rubiah, Pula Wé, Sumatra. HABITATS AND DEPTH RANGE. reef, soft coral; 2—20m. Subtidal rock, coral rubble, coral COMMENT. All these specimens represent extension of the range of this species into the northern Indian Ocean. It has been recorded from the coast of Western Australia, but otherwise only from the Pacific Ocean coasts and islands, Thailand, Papua New Guinea and Japan (Rowe et al., 1986: 221; Messing, 1994). 11. Comanthus mirabilis Rowe, Hoggett, Birtles and Vail, 1986. SEE. Rowe, Hoggett, Birtles and Vail, 1986: 226. MATERIAL. NHM —810501F/2 (1); WAM — 810427C/1 (1); SM — 810430A/5 (1). COLLECTION SITES. Ug Bau and Ug Seukundo, Pula Wé, Sumatra. HABITATS AND DEPTH RANGE. 30m. Subtidal rock/coral, coral reef; 5— COMMENT. The WAM specimen has 45 arms. The IIIBr series are mostly 2; where an arm has broken off and regenerated there is usually another devision series and extra arms, otherwise division series beyond IIIBr are randomly distributed. Most pinnules beyond P, are broken, so comb distribution further out cannot be ascer- tained. 12. Comanthus parvicirrus (Miller, 1841) SEE. Rowe, Hoggett, Birtles and Vail, 1986:211; Hoggett and Rowe, 1986: 125. MATERIAL. NHM-—810123A/1 (1),810123B/2 (1), 810203A/1 (1 of 3), 810206A/1 (1), 810212A/2 (1), 810420A/1 (1), 810425F/1 (1), 810427C/2 (1 of 3), 810428C/2 (1); WAM — 810126B/5 (1),810124A/6 (1);810203A/1 (1 of 3), 810204A/2 (1), 810204A/6 (1), 810425F/3 (1), 810427C/2 (1 of 3), 810428C/5 (1),810430A20a (1); SM — 810125A/2 (1), 810126B/6 (1), 810203A/1 (1 of 3), 810204A/5 (1), 810204A/7 (1), 810425A/18d (1), 810427C/2 (1 of 3), 810430A/20c (1), 810430A/21a (1), 810501E/10 (1). COLLECTION SITES. Galle, Hikkaduwa, Kandakkuliya, Kalpitiya, Negombo and Unawatuna, Sri Lanka; Klah, Sabang Bay, Ug Murung, Ug Bau, and Ug Seukundo, Pula Wé, Sumatra. HABITATS AND DEPTH RANGE. 20m. COMMENT. ‘The specimens from Sri Lanka may constitute a new record for C. parvicirrus as it is now defined (Rowe et al., 1986), Subtidal rock/coral, coral reef; 2— J.1.M. CROSSLAND AND A.R.G. PRICE depending on the correctness of H.L. Clark’s (1915) identification of a specimen from the region. 13. Comanthus suavia Rowe, Hoggett, Birtles and Vail, 1986. SEE. Rowe, Hoggett, Birtles and Vail 1986: 222. MATERIAL. NHM -— 810501B/1 (1); WAM — 810123A/2(1), 810124A/6, 810501B/4 (1). COLLECTION SITES. Galle, Sri Lanka; Rubiah, Pula Wé, Sumatra. HABITATS AND DEPTH RANGE. Subtidal rock, coral reef; 5—20m. COMMENTS. This is a major extension of range, as the species was originally thought to be restricted to the northern Great Barrier Reef and New Guinea. Two specimens, whose identity was originally in doubt, have now been confirmed as this species. One, 810124A/6, has 7—9 triangular comb teeth, recurved but with bases not in contact; terminal tooth small, proximal tooth usually saucer-shaped. Combs appear irregularly, e.g. on P,, P,, P, ork), PPPs. The centrodorsal is stellate with cirrus buds and cirrus scars, and subradial clefts are present. Specimen 810501B/4 has short combs with 4+2 teeth, triangular but not in lateral contact, a saucer-shaped proximal tooth, and a smaller secondary tooth on some pinnules. 14. Comanthus wahlbergii (Miller, 1843) SEE. Rowe, Hoggett, Birtles and Vail, 1986: 228. MATERIAL. NHM — 810123B/1 (1), 810421A/5(1); WAM — 810204A/2 (1); SM — 8101423B/3 (1), 810204A/4 (1). COLLECTION SITES. Chetlat I., Laccadive Islands; Galle and Kalpitiya, Sri Lanka; Seukundo, Pula Wé, Sumatra. HABITATS AND DEPTH RANGE. Subtidal rock, coral reef; 3—30m. COMMENT. The collection of C. wahlbergii from the Laccadives, Sri Lanka and Sumatra, Indonesia fills the gaps in the distribution of this species around the Indian Ocean. 15. Comaster multifidus (Miller, 1841) SEE. Clark and Rowe 1971: 6. MATERIAL. NHM-—810421A/4 (1); WAM — 810425E/1 (1); SM — 801210B/5 (1, fragmented), 801212A/2 (1, fragmented). HABITATS AND DEPTH RANGE. Subtidal rock and coral; 10—30m. COLLECTION SITES. Chetlat I., Laccadive Is; Sabang Bay and Seukundo, Pula Wé, Sumatra. COMMENT. This species is well known from Indonesia and north- ern Australia, and from the South Pacific. The record from the Laccadives is a marked extension of range, but identification is not positive because of the condition of the specimens. 16. Oxycomanthus bennetti (Miiller, 1841) SEE. Clark and Rowe, 1971:6-7; Rowe, Hoggett, Birtles and Vail, 1986:259. MATERIAL. NHM — 810421B/3 (1), 810422D/9 (1),810423C/5 (1), 810423C/5 (1),810424D/2 (1),810424D/4 (1), 810425A/18b (1), 810425A/18 (1), 810425E/1 (1), 810425F/2 (1), 810425F/3 (3), 810428B/2 (1); WAM — 810422D/6 (1), 810423C/2b (1), 810423C/ 3 (2), 810423C/6 (1), 810424A/3 (1), 810424D/6 (1),810424D/7 INDIAN OCEAN ECHINODERMS (1), 810425A/18c (1), 810427D/1 (2), 810428A/5 (1), 810428B/3 (1), 810504C/2 (1); SM—810421C/5 (1), 810422D/8 (1), 810423A/ 1 (1), 810423C/1 (1), 810424D/1 (1), 810427D/1 (1), 810427D/2 (1), 810428A/6 (1), 810430A/8 (2),810430A/35 (1), 810430A/36 (1), 810501C/4 (2). COLLECTION SITES. Seukundo, Klah, Ug Bau, Seulakoe, Sabang Bay, Ug Murung and Ug Seukundo, Pula Wé, Sumatra. HABITATS AND DEPTH RANGE. Subtidal rock/coral, coral reef, soft coral; 2—30m. COMMENT. Rowe ef al. (1986) do not record this species from Indonesia, although it was recorded from Papua New Guinea by Messing (1994); therefore this collection fills in the gap between the Andaman Islands and the Philippines. Family HIMEROMETRIDAE 17. Amphimetra tessellata (Miiller, 1841) SEE. Clark and Rowe, 1971:6-7. MATERIAL. WAM — 810425E/2 (1). COLLECTION SITE. Sabang Bay, Pula Wé, Sumatra. HABITAT AND DEPTH. On gorgonian; 10—20m. 18. Himerometra robustipinna (Carpenter, 1881) SEE. Clark and Rowe, 1971:8—9. MATERIAL. NHM — 810423E/1 (1). COLLECTION SITES. Ug Bau, Pula Wé, Sumatra. HABITAT AND DEPTH. Ona wreck, 5m. Family MARIAMETRIDAE 19. Lamprometra palmata (Miiller, 1841) SEE. Clark and Rowe, 1971: 8-9. MATERIAL. NHM — 810425E/1 (1), 810504D/2 (1 of 2); WAM — 810124A/7 (1), 810212A/4 (1); SM —810425F/3 (1), 810504D/2 (1 of 2). COLLECTION SITES. Galle and Unawatuna, Sri Lanka. HABITATS AND DEPTH RANGE. 20m. Subtidal rock, coral, coral reef; 2— 20. Oxymetra finschi (Hartlaub, 1890) SEE. Clark and Rowe, 1971: 8-9. MATERIAL. WAM — 810430A/33 (1). Ug Seukundo, Pula Wé, Sumatra. Subtidal rock; 12—13m. COLLECTION SITE. HABITAT AND DEPTH RANGE. 21. Stephanometra indica (Smith, 1876) SEE. Clark and Rowe, 1971:8—9. MATERIAL. NHM — 81020A/3 (1); WAM — 801212A/1 (frag- mented); SM — 810421C/3 (fragmented). 27 COLLECTION SITES. Chetlat I., Laccadive Islands; Klah and Seukundo, Pula Wé, Sumatra. HABITAT AND DEPTH RANGE. Coral reef; 4-8m. COMMENT. Even though two of the three specimens are frag- mented, they are easily identifiable as this widely-distributed Indo-Pacific species. 22. Stephanometra spinipinna (Hartlaub, 1890) SEE. Clark and Rowe, 1971:8-9. MATERIAL. NHM — 810423D/1 (1). COLLECTION SITE. Ug Bau, Pula Wé, Sumatra. HABITAT AND DEPTH. Coral reef; 2—8m. Family COLOBOMETRIDAE 23. Cenometra bella (Hartlaub, 1880) SEE. Clark and Rowe, 1971:10—11; Meyer and Macurda, 1980:88— 89. MATERIAL. NHM — 810423C/7 (1). COLLECTION SITE. Ug Bau, Pula Wé, Sumatra. HABITAT AND DEPTH. On gorgonian; 10—20m. 24. Colobometra perspinosa (Carpenter, 1881) SEE. Clark and Rowe, 1971:10-11. MATERIAL. NHM — 810421B/5 (1), 810425F/2 (2); WAM — 810427C/1 (2), 810501D/3 (1); SM — 810424A/2 (1), 810501 A/1 (1), 810501B/2 (1). COLLECTION SITES. Seukundo, Klah, Sabang Bay, Ug Murung and Ug Seukundo, Pula Wé, Sumatra. HABITATS AND DEPTH RANGE. Subtidal rock/coral, coral reef on gorgonian; 2—30m. 25. Decametra brevicirra (A.H. Clark, 1912) SEE. Clark and Rowe, 1971:10-11. MATERIAL. WAM — 810425D/2 (1). COLLECTION SITE. Sabang Bay, Pula Wé, Sumatra. HABITAT AND DEPTH. Subtidal rock and sand; 25m. COMMENT. Clark & Rowe (1971) implied that the main key char- acteristic of D. brevicirra, the similarity in segment numbers in P, and P,, distinguishing this species from its congeners D. mylitta (A.H. Clark, 1912) and D. chadwicki (A.H. Clark, 1911), would not ‘hold good’ when more specimens from the type locality, the Bay of Bengal, had been collected. This specimen, from Sumatra, clearly has 10 segments on both proximal pinnules. It differs from the other specimen of this genus collected in the same area, which clearly keys out to D. parva (below) on the basis of having a higher cirrus segment number. It may be time for a thorough re-examination of the genus, as there are doubtless more records than there were in 1971. 26. Decametra parva (A.H. Clark, 1912) SEE. Clark and Rowe, 1971:10-11. MATERIAL. NHM — 810428A/11 (1). 28 COLLECTION SITE. Ug Bau, Pula Wé, Sumatra. HABITAT AND DEPTH. Subtidal rock and coral, 20—30m. 27. Oligometra carpenteri (Bell, 1884) SEE. Clark and Rowe, 1971:10-11. WAM -— 810124A/7 (1). Galle, Sri Lanka. Subtidal rock; 10—15m. MATERIAL. COLLECTION SITE. HABITAT AND DEPTH. COMMENT. This is an extension of range for the species, which is well known along much of the Great Barrier Reef and has been recorded in Indonesia. This specimen has much less well-developed keels on the proximal pinnules than in specimens from the GBR, where the two species of the genus are quite distinct. However, the pinnules are wider than long, and lack flaring of their distal ends of segments. Only O. serripinna has been previously recorded from the Sri Lanka area. 28. Oligometra serripinna (Carpenter, 1881) SEE. Clark and Rowe, 1971:10-11. MATERIAL. NHM -—810425D/2 (1); WAM — 810425E/2 (1); SM— 810422C/5 (1). COLLECTION SITES. Klah and Sabang Bay, Pula Wé, Sumatra. HABITATS AND DEPTH RANGE. Subtidal rock/sand, coral reef, on gorgonian; 10—30m. J.I.M. CROSSLAND AND A.R.G. PRICE COMMENT. See above. Family TROPIOMETRIDAE 29. Tropiometra carinata (Lamarck, 1816) SEE. Clark and Rowe, 1971:10-11. MATERIAL. NHM — 810126B/5 (4), 810428D/3 (1); WAM — 810123B/3 (1), 810213A/2 (4), 810428D/6; SM — 810123A/1 (1), 810212A/4 (3). COLLECTION SITES. Galle, Hikkaduwa, Unawatuna and Tangalla, Sri Lanka; Ug Tapa Gadja, Pula Wé, Sumatra. HABITATS AND DEPTH RANGE. Subtidal rock/coral, coral reef; 3— 15m. COMMENT. J. carinata is well known from Indian Ocean reefal areas. Family ANTEDONIDAE 30. Antedon parviflora (A.H. Clark, 1912) SEE. Clark and Rowe, 1971: 10-11. MATERIAL. NHM — 810425D/2 (1). COLLECTION SITE. Sabang Bay, Pula Wé, Sumatra. HABITAT AND DEPTH. Subtidal rock and sand, 25m. Table 1 Regional distribution of crinoids from the Sindbad Voyage (names in parenthesis are equivalent zoogeographic subdivision of sampling area, following Clark & Rowe, 1971) Laccadives (Maldive area) Comanthus wahlbergii Comaster multifidus Stephanometra spinipinna Sri Lanka (Sri Lanka area) Comanthus briareus Comanthus gisleni Comanthus parvicirrus Comanthus suavia Comanthus wahlbergii Lamprometra palmata Oligometra carpenteri Tropiometra carinata Pula Wé, Sumatra (Indonesia/East Indies) Alloeocomatella pectinifera Capillaster multiradiatus Capillaster sentosus Clarkcomanthus albinotus Clarkcomanthus littoralis Clarkcomanthus luteofuscum Comanthina nobilis Comanthina schlegelii Comanthus briareus Comanthus gisleni Comanthus mirabilis Comanthus parvicirrus Comanthus suavia Comanthus wahlbergii Comaster multifidus Oxycomanthus bennetti Amphimetra tessellata Himerometra robustipinna Oxymetra finschii Stephanometra indica Stephanometra spinipinna Cenometra bella Colobometra perspinosa Decametra brevicirra Decametra parva Oligometra serripinna Tropiometra carinata Antedon parviflora ee INDIAN OCEAN ECHINODERMS DISCUSSION The crinoids of the tropical Indo-west Pacific region (Africa, Indonesia, Philippines, tropical Australia and the South Pacific) are relatively well-documented (Clark & Rowe, 1971). The region between the Red Sea and Indonesia has to date produced a rela- tively depauperate crinoid record, but the reasons for this are not clear. Unfortunately, the Sindbad collection does not resolve the problem. The low number of crinoids in this collection from the Laccadives and from Sri Lanka is probably due to a combination of two factors: lower abundance and diversity of this group in the localities collected, and limited sampling time available in those regions. This situation is unfortunate, as the areas of the northern Indian Ocean, except for the western fringe of Indonesia, are not well-represented in any collections of echinoderms, so that spe- cies and even generic distributions within the region are not well-known. In fact, the majority of specimens collected during the Sindbad Voyage are from around the small island of Pula Wé, at the western tip of Sumatra, Indonesia. SE Asia is the region of the Indo-West Pacific associated with greatest echinoderm species richness (Clark & Rowe, 1971), and Indonesia in particular is commonly regarded as the centre of distribution for coral reefs, other invertebrate groups and marine tropical diversity in general (Veron, 1995; Gray, 1997). Crinoid records from this voyage’s collection are divided into the different regions sampled in Table 1, which also shows the equiva- lent zoogeographic subdivisions adopted by Clark & Rowe (1971). The observed distribution is highly skewed, with all but two of the 30 species collected in Sumatra, eight in Sri Lanka and only three in the Laccadives. Regional comparison based on more comprehensive records, including Indian Ocean data of Clark and Rowe (1971) and the results of James (1989) for the Laccadives (13 additional spe- cies) and Sri Lanka (14 additional species), shows species distributions to be less uneven. Nevertheless, the resulting pattern reveals a progressive increase in species richness from the Maldive area to Sri Lanka to East Indies/Indonesia, as suggested in the Sindbad data (Table 1). However, the Laccadives, in particular, probably remain undersampled. These islands are a prohibited area under the control of India, and access will probably continue to be restricted. Range extensions, to the western fringe of Indonesia (Pula Wé) and into the Indian Ocean, are recorded for at least six of the 30 crinoid species collected during the Sindbad Voyage, as follows: Clarkcomanthus albinotus (Indonesia/East Indies); Comanthus briareus (Sri Lanka area); Comanthus gisleni (Sri Lanka area & Indonesia/East Indies); Comanthus suavia (Sri Lanka area & In- donesia/East Indies); Comanthus wahlbergii (Maldive area, Sri Lanka area & Indonesia/East Indies); Oxycomanthus bennetti (In- donesia/East Indies); and possibly also Comaster parvicirrus (Sri Lanka area — depending on validity of an earlier record) and Comaster multifidus (Maldive area?— specimens poorly pre- served). Of the crinoids represented, Capillaster multiradiatus and Oxycommanthus bennetti were the most common, each occurring in 19% of the samples, followed by Comanthus parvicirrus which occurred in 9% of the samples. The first two species also occupied a relatively wide range of depths (2-30 m) and habitats compared to most other species collected. A more comprehensive ecological and biogeographic assessment of echinoderms of Pula Wé, Sumatra is currently in progress. BY) ACKNOWLEDGEMENTS. We are grateful to Dr R. Dalley, P. Hunnam, P. Dobbs and D. Tattle for their considerable assistance during field work. One of us (A.R.G.P.) would also like to thank T. Severin, leader of the Sindbad Voyage, for the kind invitation to participate in the expedition which was made possible by generous support from the Ministry of Natural Heritage and Culture, Sultanate of Oman. Financial assistance to A.R.G.P. from the Leverhulme Trust is gratefully acknowledged. Thanks are also due to Dr Anne Hoggett (Lizard Island Research Station, GBR, Australia) for assist- ance with identification of specimens in doubt; and to Dr Frank Rowe for confirming several identifications. REFERENCES Clark, A.H. 1911.A new unstalked crinoid from Christmas Island. Annals & Magazine of Natural History 7: 644-645. Clark, A.H. 1915. A monograph of the existing crinoids. 1(1). Bulletin of the United States National Museum, 82(1): 1-406. Clark, A.H. 1921. A monograph of the existing crinoids. 1(2). Bulletin of the United States National Museum, 82(2): xxv + 797, 57 pls. Clark, A.H. 1931.A monograph of the existing crinoids. 1(3). Superfamily Comasterida. Bulletin of the United States National Museum, 82(3): vii + 816, 82 pls. Clark, A.H. 1941.A monograph of the existing crinoids. 1(4). Superfamily Mariametrida (except the family Colobometridae). Bulletin of the United States National Museum, 82(4a):vii + 603, 61 pls. Clark, A.H. 1947. A monograph of the existing crinoids. 1(4b). Superfamily Mariametrida (concluded — the family Colobometridae) and Superfamily Tropiometrida (except the families Thalassometridae and Charitometridae). Bulletin of the United States National Museum, 82(4b):vii + 473, 43 pls. Clark, A.H. 1950. A monograph of the existing crinoids. 1(3). Superfamily Tropiometrida (the families Thalassometridae and Charitometridae). Bulletin of the United States National Museum, 82(4c): vii + 383, 32 pls. Clark, A.H. & Clark, A.M. 1967. A monograph of the existing crinoids. 1(5). Suborders Oligophreata (concluded) and Macrophreata. Bulletin of the United States National Museum, 82(5):xiv + 860, 53 figs. Clark, A.M. & Rowe, F.W.E. 1971. Shallow-water Indo-West Pacific Echinoderms. 238p. British Museum (Natural History), London. Clark, A.M. 1972. Some crinoids from the Indian Ocean. Bulletin of the British Museum (Natural History) 24(2):73-156, 17 text figs. Clark, H.L. 1915. The echinoderms of Ceylon (other than Holothurians). Spolia zeylanica 10(37): 83-102. Gray, J.S. 1997. Marine biodiversity: patterns, threats and conservation needs. Biodiversity & Conservation 6: 153-175. Hoggett, A.K. & Rowe, F.W.E. 1986. A reappraisal of the family Comasteridae A.H. Clark, 1908 (Echinodermata: Crinoidea) with the description of a new subfamily and a new genus. Zoological Journal of the Linnean Society 88:103—142, 3 figs. James, D.B. 1989. Echinoderms of Lakshadweep and their zoogeography. Bulletin of the Centre for Marine Fisheries Research Institute 43: 97-144. Lincoln, R.J. & Sheals, J.G. 1979. Invertebrate Animals: Collection and Preservation. 150p. British Museum (Natural History) and Cambridge University Press. Marsh, L.M. & Price A.R.G. 1991. Indian Ocean echinoderms collected during the Sindbad Voyage (1980-81): 2. Asteroidea. Bulletin of the British Museum (Natural History ), Zoology 57(1): 61-70. Messing, C.G. 1994. Comatulid crinoids (Echinodermata) of Madang, Papua New Guinea, and environs: Diversity and ecology, pp. 237—243. Jn David, B., Guille, A., Feral, J-P & Roux, M. (eds). Echinoderms through time. Balkema. Rotterdam. Messing, C.G. 1995. Alloeocomatella, a new genus of reef-dwelling feather star from the tropical Indo-West Pacific (Echinodermata: Crinoidea: Comasteridae). Proceed- ings of the Biological Society of Washington 108(3):436-450. Meyer, D.L. & Macurda, D.B. 1980. Ecology and distribution of the shallow-water crinoids (Echinodermata) of the Palau Islands and Guam (Western Pacific). Micronesica 16: 59-99. Price A.R.G. & Reid, C.E. 1985. Indian Ocean echinoderms collected during the Sindbad Voyage (1980-81): 1. Holothurioidea. Bulletin of the British Museum (Natural History ), Zoology 48(1): 1-9. Price A.R.G. & Rowe, F.W.E. 1996. Indian Ocean echinoderms collected during the Sindbad Voyage (1980-81): 3. Ophiuroidea and Echinoidea. Bulletin of the British Museum (Natural History ), Zoology 62(2): 71-82. Rowe, F.W.E., Hoggett, A.K., Birtles, R.A. & Vail, L.L. 1986. Revision of some comasterid genera from Australia (Echinodermata: Crinoidea) with descriptions of two new genera and nine new species. Zoological Journal of the Linnean Society 86:197-277, 10 figs. Veron, J.E.N. 1995. Corals in Time and Space. University of New South Wales Press, Sydney. Bull. nat. Hist. Mus. Lond. (Zool.) 65(1): 31—50 On the hybrid status of Rothschild’s Parakeet Psittacula intermedia (Aves, Psittacidae) PAMELA C. RASMUSSEN KX (Si CTobE.1) NHB 336 MRC 114, Smithsonian Institution, Washington, D.C. 20560-0131, USA NIGEL J. COLLAR BirdLife International, Wellbrook Court, Girton Road, Cambridge CB3 ONA, UK SYNOPSIS. The name Psittacula intermedia was attached to seven dataless specimens sent from India to England between 1895 and 1907, six of which are now at the American Museum of Natural History, the other being at The Natural History Museum, Tring, U.K.Their origins and taxonomic standing have long puzzled authorities, since they look intermediate between male Plum- headed Parakeet P. cyanocephala and Slaty-headed Parakeet P. himalayana, and no definite field records exist. Although a hybrid origin has been suggested, intermedia has recently been considered a valid species on the bases of: (a) uniformity of characters; (b) a single origin; (c) a non-captive origin; (d) an old description of hybrid cyanocephala x himalayana which does not match intermedia; (e) reports of captive intermedia in the 1990s; and (f) biochemical analysis of captive birds. For this study, we examined all published intermedia specimens. For hybrid diagnoses we compared morphology of adult males qualitatively and mensurally with the putative parental species, including also Grey-headed Parakeet P. finschii and Blossom-headed Parakeet P roseata. We examined six live adult hybrid cyanocephala x himalayana bred by two different aviculturists, as well as one live bird in India claimed to be intermedia, and we considered published avicultural evidence. Our analyses showed all the defences of the specific status of intermedia to be wanting, as follows: (a) considerable variation exists in the original material; (b) the specimens could not all have had a single origin; (c) six of the seven specimens showed signs of captivity; (d) the 65-year-old account of cyanocephala x himalayana only furnishes passing descriptions of juveniles, and is therefore not comparable with the adult intermedia specimens; (e) all the specimens examined in Bombay are hybrid cyanocephala x krameri, while other captive intermedia in Austria and India are of uncertain provenance (but the former appear to becyanocephala x finschii); and (f) the biochemical analysis was seriously flawed, most importantly in that the specimens used were not intermedia but hybrid cyanocephala x krameri. Neither cyanocephala nor himalayana shows any morphological characters incompatible with being parent to intermedia, and all features of the latter are explained by a combination of the two former species. Moreover, mensurally the AMNH intermedia fall midway between cyanocephala and himalayana. All known male cyanocephala x himalayana possess plumage features and measurements matching AMNH’s five adult male intermedia, while the previously undescribed female hybrid has the head paler than himalayana and drabber than female cyanocephala. This evidence leaves no doubt that intermedia is a hybrid of Issued 24 June 1999 cyanocephala and himalayana. INTRODUCTION Rothschild’s or the Intermediate Parakeet Psittacula intermedia was described from a single dataless specimen (Rothschild 1895) backed up by a later series which also lacked data but had been exported from Bombay (Rothschild 1907, Hartert 1924). However, apart from being listed in Peters (1937), this parakeet was overlooked until mentioned by Ripley (1953) as a species of Indian origin. Of subsequent authors who have considered the status of intermedia and treated it as taxonomically valid, only Biswas (1959) had examined Rothschild’s entire series. Walters (1985) had access only to a single specimen in the bird collections of The Natural History Museum, Tring, U.K. (BMNH); in Bombay, Sane (1975, 1977, Sane et al. 1987) had only his own captive birds; while Inskipp and Inskipp (1995) simply reviewed the literature on intermedia; and Bhargava (1998) had only his own specimens in India. Psittacula intermedia has been accepted uncritically as a species by several authors (e.g. Howard and Moore 1991, Monroe and Sibley 1993). However, Salvadori (1907), in reference to the type, had stated that intermedia was ‘. . . not improbably established on a hybrid!’. Immediately after Salvadori’s comment, Rothschild (1907) men- tioned having obtained six more specimens, which he maintained *.. Should certainly dispose of any doubt regarding the distinctness of intermedia’ . Conversely, Husain (1959), on the basis of the single © The Natural History Museum, 1999 skin at the BMNH, considered that intermedia was a hybrid between Plum-headed Parakeet Psittacula cyanocephala and Slaty-headed Parakeet P. himalayana, while Forshaw (1973) concluded the same after examination of the series at the American Museum of Natural History, New York (AMNH); Wolters (1975) also treated it as a probable hybrid. Still other authors have remained undecided as to its status (Peters 1937, Ali and Ripley 1969, 1981, Wirth 1990); Juniper and Parr (1998) and Collar (1997) tentatively gave it a species account pending publication of the present study. Psittacula intermedia is currently listed as a globally threatened species with IUCN status Vulnerable, but for which no specific threats have been identified (Collaretal. 1994), although its possible extermination by collectors has been suggested (Walters 1985). The fact that the phenotype of Psittacula intermedia places it midway between cyanocephala and himalayana was acknowledged in both the original description and in the specific epithet chosen (Rothschild 1895). Since then, no characters of intermedia have been identified that either differ from those of cyanocephala or himalayana or are not manifestly intermediate between these two (contra Inskipp et al. 1996, whose cited references nowhere demonstrate non-intermediacy). Moreover, the area of origin of intermedia has never been accurately pinpointed, despite fairly extensive subsequent ornithological work in many presumably likely areas; even Rothschild (1907) admitted that “speculations as to its exact locality were useless, as these collections contained forms 32 exclusively found in the Eastern Himalaya as well as others occur- ring only in the north-western portions of India.’ A statement by Ali and Ripley (1969) concerning intermedia — ‘never consciously seen alive in the wild state by any ornithologist’ — remains true today. Recent reports within India (R. Bhargava pers. comm. 1996, in litt. 1997, 1998; Ahmed et al. 1996, Anon. 1997, 1998a, b; Mookerjee 1997; Bhargava 1998; Them 1998), and three birds identified as intermedia (Sane 1975, 1977, Sane et al. 1987, S. R. Sane pers. comm. 1997), do not clear up the mystery, as all refer to captive birds of uncertain provenance, and some of these are of problematic identification as well (Rasmussen and Collar 1998, Bhargava 1998). Appeals for information (Rothschild 1907, Sane 1977, Wirth 1990, Inskipp and Inskipp 1995) have not led to the discovery of a wild population. If intermedia were a typical diurnal, noisy Psittacula, it would be a most unusual bird, not only for having escaped the attentions of field ornithologists for over a century in one of the best-known parts of tropical Asia, but also for showing complete intermediacy in numerous characters between two clearly differentiated congeners. There are two possible explanations for this double circumstance: the first is that it is an extremely rare species and therefore requires concerted conservation attention; the second is that it is not a species at all, but a hybrid. Only the second explanation accounts for both of its unusual traits. In this paper we reexamine the evidence for specific status vs. hybrid origin of intermedia, based on plumage and mensural analyses both of museum specimens and of newly located captive birds of known parentage. METHODS SPECIMENS EXAMINED Eight museum specimens have been published in the primary litera- ture as Psittacula intermedia, although one (AMNH 621545) has been considered an immature himalayana (Biswas 1959, 1990, Forshaw 1973). Each was thoroughly examined, photographed, and measured for this study: AMNH 621539 (holotype), 621540-621542, 621544-621545; BMNH 1980.3.1; and BNHS (Bombay Natural History Society) 26758. In addition, we examined another uncatalogued specimen belonging to Mr. Sane, as well as a photo- graph of three unaccessioned specimens in the possession of R. Bhargava. In the early 1930s, Rothschild’s entire series of intermedia went to AMNH along with most of the rest of his collection. Subse- quently, one (BMNH 1980.3.1, formerly AMNH 621543) was exchanged to the then British Museum (Natural History) (M.P. Walters, pers. comm. 1997), where it had already resided since 1959 on long-term loan (Knox and Walters 1994). The BMNH specimen was lent to AMNH so that we could compare it there with the remainder of the series. A colour transparency of AMNH 621540 (placed with the speci- mens; date and photographer unknown) taken sometime after 1973 — based on an accompanying note: “Psittacula ‘intermedia’ believed to be a hybrid himalayana x cyanocephala see Forshaw (1973: 336)’ — shows that it had long central rectrices when the photo was taken, but these were lacking when the specimen was first photographed by PCR in 1993, and it cannot now be determined if the rectrices were fully grown. Estimates of lengths of the tail and of the yellow tip of the central rectrix of AMNH 621540 were made from the transpar- ency, in which the subject is 1/3 natural size and photographed from the side. BNHS 26758 is essentially dataless (label data: male, aviary bird, P.C. RASMUSSEN AND N.J. COLLAR S. R. Sane, Bombay, 12/90), as is Sane’s second uncatalogued specimen; these are two of the three birds examined from his collection. The first may or may not be the specimen described in Sane (1975, 1977), referred to as having died in 1978, and as being in the BNHS collection (Sane et al. 1987), but if ‘12/90’ refers either to date of death or to date of accession it can hardly be the same individual. The only other specimens reputed to be intermedia of which we are aware are those preserved by R. Bhargava, and we have seen photos of three of those. However, adult female and especially immature intermedia (see below) would readily escape notice among series of similar congeners, and may well exist undetected in mu- seum collections. We assessed variability among the eight published putative intermedia specimens in the plumage and mensural characters listed in Tables 1-4 and the Appendix. PHOTOGRAPHIC EVIDENCE Most of the published information on intermedia was recently summarized by Inskipp and Inskipp (1995). Through perusal of the avicultural literature we located an additional, previously unrecog- nized, published photograph of an intermedia-like bird, and correspondence with aviculturists and researchers (after the main Statistical analyses for this paper were complete) resulted in addi- tional unpublished information, including the location of several more captive birds, some of documented parentage. HYBRID DIAGNOSES For hybrid diagnoses (Graves 1990), plumage, other external char- acteristics, and measurements of adult males were compared among the species of Psittacula that either had been suggested previously as possible parental taxa (Husain 1959, Forshaw 1973, Wolters 1975) or for which the phenotype of the presumptive hybrids indicated the likelihood of those species being involved. Sane’s birds were com- pared indirectly with the other intermedia specimens (Table 1) and with series of adult male cyanocephalaand Rose-ringed Parakeets P. krameri (Table 2), while other intermedia were compared with series assembled at the National Museum of Natural History (USNM) of each of the potential parental species: cyanocephala (n = 21), himalayana (n = 26), Grey-headed Parakeet P. finschii (n = 17), krameri (n= 10), and Blossom-headed Parakeet P. roseata (n= 11). Additionally, all adult males of these species in the collections of AMNH, the Academy of Natural Sciences of Philadelphia (ANSP), the Museum of Comparative Zoology (MCZ), and the University of Michigan Museum of Zoology (UMMZ), as well as several smaller collections, were examined and measured; their plumage characters (which did not differ materially from those of the series assembled at USNM) were not included in the analyses, but their measurements are included in the Appendix and in the statistical analyses. A few unsexed specimens with plumage characteristics diagnostic of males were included. All other species of Psittacula were ruled out as potential parental species as they have plumage characters strongly incompatible with the phenotype of specimens reputed to be intermedia. Mensural characters of adult males (listed in Appendix) were used to evaluate which (if any) of the species listed above could potentially be parental species of the intermedia specimens. Measure- ments taken as far as possible for each specimen were: culmen length (from distal edge of cere); height and width of maxilla (upper mandible, at distal edge of cere); minimum distance between nares; width of (lower) mandible; wing length (straightened and flattened); shortfalls of each primary (P1—P10, with P1 outermost) from wingpoint; for P1, distance from notch on inner web to feather tip, maximum width, and width at notch; widths of P2—5, each taken at 38 TAXONOMIC STATUS OF PSITTACULA INTERMEDIA ajeymeds Apysiys diy 0} ysinyq wuoyrun anjq-ystusess oped feondiyyja ‘ayerapour SUISSTU SUISSTUL SUISSTUL [eondiyja ‘ayesopour MoleU ay1YM-YsIMmoy[od anyjq-ystusel3 [eondiyya ‘ayerapour MOLIeU MOTJ9A ound uda13 [OOS pajurod ‘ayeopour MOLIeU JOyeI MO][aX asosuLd anyq sep [eondiyja ‘aresapour a1e1apour MO]J9A ayed anyjq [eAor arenbs ‘proiq peoiqg MO]JA oped aniq yep aienbs ‘ayerapoul oye19pour MOT[OA anyjq [eo orenbs ‘a\e1apour SUISSTU SUISSTUL SUISSTUI orenbs ‘prog dn [x Jo pearg Ta Jo dn yo inojop Ty JO uonsesprut Jo mojoD TIVE €d jo adeyg YSIS ISIS auou ouou quasaid Suryor| quasoid juasoid quasaid == sy19A09 BuIMIoddn 0} a8uy ont g osu Ajsni 1YsI]s ouou auou ouou quasaid juasaid quasaid quasoid quasaid yoyed Jopynoys Jo souasa1g ONIM YSIS YSIS Suryory] suons suons ayeJapoul suons aye1opou a}e1Opoul duini uo a8un anyq Jo yua}xq SAT[O-MOTI[AA IYS1IG SAT[O-MOT[OA JYSUG AIO-[9A JOyIEp “|S UdAIB [OOD = IATO-"]9A TYP "|S 9ATJO-"J9A TYP ‘IS SAT[O-MOT]OA SAT[O-MOTJAA —- DAT[O-MOT[OA yorqg Jo 10[oD aye1apoul ayerapoul aye1apoul auou Moe AIDA MoLIeU AIDA peoig A100 aye1apoul peroiq Je]JOO uo anjq Jo JuaX| UPIPLITA UPIPLITA uaeis-anjq ayed =usess-anjq Aysnp = uder8-anjq ajed = uaaas-anjqayed §=—s uaauS-anyq ayed ~=—uaa3-anyq aed uaais-anjq aed Ie][oo adeu jo 10joD AGO anyjq uRe[NIa9 YystIq «= aN{q URayNIIO 1YsLUqQ Aois-any]q Aois Aysnp anjq-Aois anjq-Aais anjq-Aois anjq-Aois anjq-Aois adevu/uMmoio IeaI JO INojOD ystXois oeyt]-ystAois yurd-ystmoy[aA ude13-MO][ah yuid-ysrjdind yuid doap yuid-ystjdind yuid-ysyjdind = yurd-ystjdind SaI0] JO INO[OD ajayduio9 aa[duios auou auou A ouou auou ouou auou a[qipueul uo yep Jo Juax_ por 1ysuq por 1y311q aeJopou ouou suons ayed A190 aed suons suons PI[IXew UO pos/asuvIO Jo yJdaq pol [je por [je aSurIO ouou asurIO 7% jewruru asur10 Y aSUPIO % asuvl0 % —«R| [Xe UO pol/asuPIO Jo JUaIX_ avaH ynpe ynpe qnpe onyeUUUT y[Npe {inpeqns ynpe ynpe ynpe ssejo os pansojeyeoun 8SL9T Ve 086l SPSIc9 brsic9 CrsSIc9 IvS1c9 OvSIc9 6€S 179 Jojourey) oues SHN4@ HNW4 HNWYV HNNV HNNAV HNWV HNWV HNWNV “DIPAWAAJUI DINIVDIJIS_ O} paynginye susuttoads [enprlArpur Jo ssojoeseyo sayneyyenb [euslayxq | I[Qey, 34 P.C. RASMUSSEN AND N.J. COLLAR Table 2 Qualitative characters used in hybrid diagnosis between Psittacula cyanocephala, P. krameri, and S. R. Sane’s specimens and living bird. Character P. cyanocephala Sane’s birds Maxilla all yellowish all red Mandible black black Cere shape moderately wide, rounded narrow, nearly straight Cere colour medium grey pale greyish-horn Orbital skin dark grey fleshy whitish Lores pattern rather wide, not prominent no line Forehead reddish-purple Auriculars and central face reddish-purple Lower face purplish Rear crown shining mauve Lower border neck collar as nape Nape bright bluish-green Upper wing coverts variably bluish-green Shoulder patch maroon Underwing coverts pale turquoise-blue Rump variably bluish-green Uppertail rich dark blue Tail tip white, moderate width, spatulate Foot dark pinkish-grey rather narrow, very prominent slight line greenish tinge dull purplish-blue cerulean blue cerulean blue broken orange-chestnut Viridian slightly bluish slight tinge on one slightly bluish slightly bluish pale blue-green concolorous with rest or narrow whitish; slightly spatulate pale pinkish-grey P. krameri all red black narrow, straight whitish orange rather narrow, very prominent strong line green bluish-green lime green powder blue nearly complete rose-orange slightly bluish-green lacking bluish absent yellowish-green yellowish-green lime-green concolorous with rest, not spatulate whitish tip of next shortest feather; lengths of longest (R1, central) and next more lateral (R2) rectrices (both taken from insertion of central rectrices); maximum width of yellow or white tip and approximate maximum width at distal end of blue or green area of R1 (with feathers flattened out); distances between tips of each rectrix (except R1) of one side and the next shortest (next more lateral) rectrix; widths of each rectrix of one side at the tip of the next shortest one; approximate distance from tips of R1 and R2 to definite blue or green part of feather (= length of pale tip); tarsus length; minimum width of tarsus; length of claw of middle toe (from distal edge of scute); length of hindclaw (from distal edge of scute). Feathers in sheath or in a damaged or heavily worn state were not measured, and if there was a difference in length between rectrices of a pair, the longer one was measured. Maximum skull width was measured over skin and compressed feathers for specimens in which palpation and/ or x-rays indicated that the widest portion of the skull was intact and not padded with stuffing. Two specimens from the Rothschild series (BMNH 1980.3.1 and AMNH 621545) showed very different plumage and mensural characters from each other and from the remainder of the specimens in this series, and so were treated as unknowns in the analyses. AMNH 621545, although thought a female intermedia by Rothschild (1907), was considered by Biswas (1959, 1990) and Forshaw (1973) to be an immature himalayana, the latter opinion being shared by us after examination. We therefore compared its plumage characters with known immature himalayana and finschii, and its measure- ments with nine juveniles (sexes combined) of the former. Univariate statistics and principal components analyses (PCAs) using correlation matrices were run on external and skeletal meas- urements using SYSTAT for Windows (Version 5.0) on an IBM-compatible PC. Variables for PCAs were chosen to allow the inclusion of selected individual intermedia specimens without esti- mation of missing data, which would be inadvisable owing to the small sample size of intermedia. EVALUATION OF ORIGIN OF SPECIMENS To test the idea that the Rothschild Collection series of intermedia had the same origin — an argument first put forward long ago by Hartert (1924) — we compared preparation styles and materials used among these specimens by external examination and study of x-rays. We also compared them with native-prepared (e.g., ‘Bombay prepa- ration’, ‘India’, and ‘Madras’) skins of other Psittacula species (cyanocephala, himalayana, roseata, finschii) at AMNH, MCZ, and USNM. To permit analysis of certain aspects of preparation styles and materials used, radiographs (x-rays) were taken of the Rothschild intermedia specimens, and of native skin specimens of himalayana, finschii, and cyanocephala for comparison. X-rays (ventral and lateral views) were made of the AMNH and BMNH intermedia (including the putative immature himalayana) by M. N. Feinberg, Department of Ichthyology, AMNH (30 kV and 3 mA for 2 min, using Kodak Industrex-M Ready-pack film), and for the other specimens by PCR in the Fish Division, National Museum of Natural History (USNM; 25 kV and 5 mA for 30 sec, using Kodak Industrex SR film). To evaluate whether the Rothschild Collection series originated from wild, not captive birds — an argument used by Biswas (1959) to support species status — we examined the Rothschild specimens for presence of: overgrown bill and claws; broken remiges and rectrices; overly worn feathers due to delayed moult; abrasion damage to feathers of the type resulting from repeated contact with cage bars; and dirt on plumage, bill, and feet consistent with a confined environment. EXAMINATION OF CAPTIVE BIRDS We examined five living adult hybrids belonging to Mr M. Sedgemore that are the progeny of an experimental pairing of a male cyanocephala and a female himalayana. The female parent, which died in the nest shortly after producing the second of two hybrid broods in successive years, was considered unsalvageable as a specimen; the male parent died more recently and the skin is preserved as BMNH 1998.33.2.We took hand-held photographs and aviary videotape of all five hybrids, as well as several measurements (taken by PCR while the birds were held by Sedgemore) of bill, wing, and tail. All the hybrids were in some stage of moult, so certain measurements could not be taken. The recently moulted central rectrices of the single female hybrid are now at the BMNH, and Sedgemore also gave us several photographs of the hybrids, both as juveniles with their parents and as adults. PCR also examined and videotaped the single live bird claimed to be intermedia remaining in Sane’s collection in December 1997. We also studied photos sent by TAXONOMIC STATUS OF PSITTACULA INTERMEDIA MrL. Critchley of yet another adult male hybrid and its parents. This bird was one of five hybrids in a single brood that Critchley incidentally produced by housing a male cyanocephala in a mixed aviary with a female himalayana, but he sold the other four hybrids to a pet shop in the U.K. before they attained adult plumage. Photos of all captive birds discussed in this paper are on file both with BirdLife International HQ and the senior author, and selected photos showing each specimen will appear on a colour plate accompanying a short article on Psittacula intermedia (Rasmussen and Collar in press). RESULTS SANE’S CAPTIVES BNHS 26758 and Sane’s other two birds, one stuffed and one alive as of December 1997, are all captive, dataless males identified as intermedia. However, all proved to be considerably different from the Rothschild specimens, possessing characters of both cyano- cephala and krameri, but none inconsistent with their being hybrids between the latter two species (Tables 1 and 2; Appendix). Sane’s birds have much bluer, paier sides to the head and a greener mid- and hindcrown than do any of the Rothschild Collection intermedia; like the latter they have mainly lilac cheeks, but many feathers of the head are multicoloured, at least on the two skin specimens. On the sides of the crown and edge of the black moustache, most of the individual feathers have peach-coloured bases and blue tips; on the cheek the bases tend to be peach and the tips lilac; while the feathers of the centre of crown and nape have green centres and blue tips. The specimens have entirely black lower mandibles and all-red upper mandibles except for the paler tips. Their soft-part colours and the Table 3 Component loadings for PCAs of (A) bill width, wing, and tail measures for a model including juvenile Psittacula himalayana; (B) bill, wing length, and rectrix 5 measures, including Sane’s mounted specimen and three of Sedgemore’s living birds; and (C) head and wing measures, including BNHS 26758 (Sane’s study skin). Factor A B (G Measurement 1 2 1 2 1 2 3 HEAD Culmen length - - 0.93 0.03 0.82 -0.49 0.06 Maxilla height - - 0.96 0.06 0.87 -—0.36 0.10 Bill width 0.67 0.06 0.95 0.05 0.85 —0.38 —0.20 Skull width - - - = 0.84 -—0.24 —0.02 WING Wing length 0.95 0.11 0.93 -0.12 0.92 -0.20 0.12 P3 shortfall - - - — -0.09 0.81 0.06 P4 shortfall ~ - - - 0:67 . 0:57 0:24 PS shortfall - - = - 0.86 0.41 0.09 P6 shortfall 0.93 0.16 — - 0.92 0.29 0.10 P7 shortfall 0.93 0.24 —- - O95 0n16" 10115 P8 shortfall 0.97 0.14 —- - O95 O21) 10M P9 shortfall 0.96 0.19 —- = OMOGTy HOMs One P10 shortfall 0.96 0.19 — - OO5— OS OkkG P1 notch length - - - - 0:68) —0'55, 10512 Pl maximum width 0.68 —0.43 - P1 notch width - - - - 0.36 —0.57 —0.45 P2 width 0.62 -0.59 — - 0.58 0.56 —0.42 P3 width - - — _ 0.66 0.31 —0.49 P4 width = - - - 0.81 0.11 —0.34 PS width - - - - 0.86 -—0.16 —0.27 TAIL R1 width 0.83 0.02 —- - - - - R2 width 0.75 -0.50 0.56 0.76- —- - - 35 feathering at the bill base are all unlike AMNH intermedia. The maxillae of Sane’s birds are smoothly rounded on lateral view and not particularly robust proximally, being very similar in shape to krameri, not himalayana. Two of the three individuals bear no indication of the red shoulder patches (and they are very vague in the third) that are shown by both male cyanocephala and himalayana, and that are present in five of Rothschild’s six adult intermedia, but that are always lacking in male krameri. However, of all the features in which Sane’s three birds differ from AMNH intermedia, none is more telling than the broken orange-chestnut neck ring of the former (Table 2), which (assuming that the birds are hybrids) can hardly have come from any source other than krameri or the much larger Alexandrine Parakeet P. eupatria. Also, in both of the individuals with the central rectrices present, the feathers have very small pale tips, consistent only with the latter two species. In 1990, one of two ‘intermedia’ then alive in Sane’s possession was photographed in Bombay by R. Wirth. The published photo (Wirth 1990) shows a bird very similar to BNHS 26758 and Sane’s uncatalogued specimen, and from the date it may be either the bird still living as of 1997 or the uncatalogued specimen; it possesses the same suite of features consistent with its being a hybrid krameri x cyanocephala (or possibly krameri x roseata). A description of the second live bird was not provided, but Sane considered both to be intermedia, and Wirth (in litt. 1997) noticed no differences between 11.5%, 1.0; R4 vs. rectrix widths Factor 2 Factor 1 70.8%, 6.4; size (R4 uncorrelated) Fig. 1 Identity of AMNH 621545 with immature Psittacula himalayana: graphs of individual scores (circles), group means (triangles), and 95% confidence intervals (open ovals) of factor scores from principal components analysis (PCA) on measurements of adult male P. cyanocephala (C, grey-filled circles), P. himalayana (H, white), AMNH P. intermedia (black), immature P. himalayana of both sexes (diagonal hatching); P. roseata (R, diagonal cross-hatching), both populations of P. finschii (F, horizontal cross-hatching), one of Sane’s specimens (checkered), and P. krameri (K, horizontal bars). A polygon outlines the scores for AMNH P. intermedia specimens due to small sample size. Summary statistics presented in the axis labels are percent variance explained and eigenvalues, respectively, followed by important measures for each axis. Component loadings for PCA are given in Table 3A. 36 the two. PCR examined all three of Sane’s birds in the space of two days and concluded that they lacked salient differences, all being apparent krameri x cyanocephala hybrids. IMMATURE SPECIMEN AMNH 621545, considered to be a female intermedia by Rothschild (1907), but thought by others to be an immature himalayana (Biswas 1959, 1990, Forshaw 1973), shows no relevant plumage differences from the series of immature himalayana at AMNH with which we directly compared it, nor from others at USNM and other museums with which photos of it were compared. In a PCA of several measurements, AMNH 621545 falls within the 95% confidence limits of immature himalayana of both sexes (Figure 1, component loadings in Table 3, summary statistics in Appendix). In body plumage 621545 resembles himalayana in being generally cooler green and less yellowish than immature finschii, although some juveniles of the two species overlap in this. AMNH 621545 is unlike juvenile hybrid cyanocephala x himalayana (see below) in its bigger, duskier maxilla, brighter green nape, lack of yellowish collar, bluer-green overall body colour, and especially in its bright green upper tail surface with a bright yellow tip. Thus, on the basis of both plumage and measurements, all evidence supports the hypothesis that AMNH 621545 is an immature himalayana, and we therefore exclude this specimen from further analyses. BMNH SPECIMEN Comparison of photos and measurements of BMNH 1980.3.1 with those of the adult male AMNH intermedia showed that the former has several differences from all other intermedia, although it is part of the Rothschild series, and despite the seemingly inexplicable fact that the BMNH specimen was the one upon which Husain (1959) based his conclusion that intermedia was a hybrid himalayana x cyanocephala. This bird was therefore lent to AMNH for our comparisons, where we confirmed (Table 1, Appendix) that it has a smaller maxilla with only a slight reddish tinge basally (less than in allAMNH birds except 621542, the specimen said by Biswas [1959] to be completing post-juvenile moult); it has a slightly duller head with paler reddish-purple on the face (washed yellowish in front of the eye) and paler greyish-blue on the crown and nape; and it totally lacks maroon shoulder patches. Its P3 is narrower and less squared at the tip than in all adult AMNH specimens except 621544. The tail is greener at the base, more turquoise for most of its length, and has the pale tip whiter and shorter. This specimen is the only one of the Rothschild series that has a measurable, fully grown tail, so its rectrix length cannot be directly compared with the otherintermedia. The salient differences between BMNH 1980.3.1 and typicalroseata are: the former lacks reddish shoulder patches; it has an entirely pale lower mandible (though this is nearly all-pale in a few roseata; Table 4); its hindneck has a turquoise tinge; it has a slightly broader tail tip; the front of its face is slightly redder; and its P3 tip is broader (Table 1). In most statistical analyses, BMNH 1980.3.1 falls within the roseata and cyanocephala groups (Figures 2-5). REMAINING JNTERMEDIA SPECIMENS The other five AMNH specimens attributed to intermedia (including the holotype), and also Bhargava’s three specimens, are quite similar to one another. However, although most previous authors (Rothschild 1907, Hartert 1924, Husain 1959) have treated the first five under one description as if they were identical, they are in fact variable in most of the characters that separate them from any of the putative parental species (Table 1). Only Biswas (1959) mentioned variation among these five, but even he called them ‘exceedingly similar’. All have fairly large bills with varying amounts of orange at the base of the maxilla. All have nearly or entirely pale lower mandibles, P.C. RASMUSSEN AND N.J. COLLAR although AMNH 621544 has a broad black stripe down one side of the lower mandible. Each has the front of the face bright purplish- to deep pink, grading into the duller grey-blue crown, nape, and lower portions of the face. All have a pale blue-green collar, but this is highly variable in breadth and prominence, even allowing for differ- ences in preparation. In addition, all have a bluish wash of variable strength on the wing coverts and/or rump. Only three of the speci- mens now have the central rectrices present, and in none of these (contra Biswas 1959) are they fully grown (this cannot now be determined in AMNH 621540, the fourth intermedia that once had central rectrices, but the fact that they are now missing from this specimen suggests they were loosely attached and thus moulting). Thus, original tail lengths presented in previous treatments — Biswas (1959): 185, 202, 221mm; Husain (1959): about 220 mm; Forshaw (1973): 167-195 mm (mean = 180.7, n = 3), 206 (n = 1) — would be expected to be too short. However, we measured the central rectrices of the three specimens in which they are now mostly grown as 157, 200 and 170 mm, and the now-missing rectrices of AMNH 621540 were estimated at ca. 200 mm. The central rectrices of all four of these birds show (or showed) long, at least moderately broad, pale to pure yellow tips and dark or royal blue upper tail surfaces. The breadth and length of the yellow tail tip are quite variable, however, and the length of the yellow R1 tip of AMNH 621540 is estimated to have been 48 mm, compared with a mean tip length of 41.6 mm for the others (Appendix). From the photograph of Bhargava’s two specimens for which the central rectrices are present (both photo- graphed next to a cm rule), these rectrices appear to be ca. 234 and 229 mm (although it cannot be determined from the photos whether these rectrices are full-grown), while the yellowish tips are ca. 51 and 54 mm, respectively. We found no external qualitative characters in AMNH intermedia or Bhargava’s specimens that differ from those exhibited by at least one member of one of the two species groups (roseata/cyanocephala and finschii/himalayana), or that are not intermediate between them (Table 4). Among the potential parental species, roseata exhibits the most plumage features incompatible with the AMNH intermedia phenotype, while finschii also has a few characters inconsistent with intermedia, mostly in tail shape and colour. Neither cyanocephala norhimalayana shows any plumage features incompatible with their being parental species of AMNH intermedia, and a combination of the former two readily explains all plumage features of the latter. STATISTICAL RESULTS Summary statistics for measurements of the putative intermedia specimens (with the BMNH specimen treated separately), Sedgemore’s hybrids, Sane’s specimens, and comparative samples of the five putative parental species are given in the Appendix. For almost all measures, theAMNHintermedia are intermediate between cyanocephala and adulthimalayana, and in many cases also between others of the putative parental species. Bivariate scatter plots of selected measurements overwhelmingly demonstrate this pattern, e.g. Figure 2A showing culmen length from cere vs. culmen width, in which allAMNH intermedia and Sedgemore’s hybrids fall between the cyanocephala/roseata pair and the himalayana/finschii pair. Furthermore, krameri is larger than, and Sane’s two specimens are as large as, the himalayana/finschii pair. A slightly different pattern is shown in Figure 2B (wing length vs. culmen length): here again, AMNH intermedia and Sedgemore’s hybrids fall between cyanocephala/roseata and himalayana but, because of the shorter wing of finschii compared with himalayana, there is slight overlap between finschii and intermedia. Psittacula krameri is similar in wing length to himalayana but is bigger-billed, and Sane’s birds fall between cyanocephala/roseata and krameri, being considerably TAXONOMIC STATUS OF PSITTACULA INTERMEDIA 37 Culmen width Culmen from cere Culmen from cere Wing length Pale tip length of R2 R2 distal width Fig. 2 Bivariate scatter plots of measurements of putative parental species and hybrids: (A) culmen length from cere vs. culmen width; (B) wing length vs. culmen length from cere; (C) R2 pale tip length vs. R2 tip width. Symbols are as for Figure 1, with the addition of the BMNH intermedia specimen (white square) and Sedgemore’s hybrids (black squares). [a4 <x —] =| (e) O ze Z Z Z jaa) a a =) = iz < [24 YU oO. 38 “QUO UI posuN-MoOT[AA ApYSI]S,, suduToeds WOM ¢€ UT YSIN]| gs “JoyJOuR UI YstuseIs poysem ‘suouroads uJOM ¢€ UI [[NC, Seyur] Wg wos uswIoads ou ur [[Np A19A ‘adeu Jo a8pa JOMOT Iv JsosuONs sXemye ‘a]quueA, ‘Aoid aed sa10] sey YZ/Tp ‘eyed s[qrpueul Jo a10UW JO ¢ aAvy | [/E, ‘a[ed afqrpurU Jo a10UI JO Ff DALY ()Z/Sq :9UO UT aseq Jv Yyoyed YsTyor|_, ‘sotoadssadns 10yj0 ay} JO JAqUIOUI JOYIIE JO sIa}OeILYO YIM UOHeUIquIOS ul adMousYd syT]-pIpaustajUI Ue 0} 9INGLNUOS 0} ATOYI] JOU S1O}ORIeYS sa\oueP (,,) YSU}se UY $90} Ue) JaAars ‘Iayep Aais ayed UI}. MOT[EA posuny -a5uvIO ‘snooneys,. ysorym 0} uMoIg aed, ajdind-ysnyq [np ugeis-ysinyq ojed [eordiyja ‘ayesopour uoorew yep Ud9I8-YsIMo]]OA UddIS-YsIMOTIAA JYSLIG (apuewi<<) usaI3 -MOTJOA uouTUTOId,. poulyep-][om ‘Moreu on] q-ystuseis WYysLIq $90} Ue) J9AOIS ‘oy.ep Agis oyed peoiq MO][OA oind 7Yy311qG UMOIG yep onyq-ystjdind [np ugeis-ysinj{q ajed arenbs ‘aye1apow uooreul yep A19A gU99013 [009 ua913 [009 (a]JUeUUI=) dAT[O [OOS prog 0) ae1opoul ugais-ysinjq ojed 04 anjq-ystuseis ayed $90] se dues Aois-ASOJ WINIpoUul ayejmeds Apysiys ‘ayeropour MOTO ayed oe anyjq [eAo1 ugeis-ysinj{q oyed ysuenbs ‘aye1apour uooreul posun-onjq u9d13-YsIn]q (epueUt=) aAtfo-moj[a4 [INP oye. 1Opoul udo1s-anyq ayed $20} se OURSZ, Aois-ASO1 WnTpoul peoig 0} MOeU JOyIeI MoOT]9A dnd 0} ayed yor lq anyq yep 0} [eAor uad13-ysin{q ajed (%) orenbs ‘(%) aye19poul UOOJPUU-pru (%) pesun-onyq u99.15-YsIn]q (apuru=) sArfo-Moy[ad [Np peoig AIA 0} MOLIeU uso13-an{q ojed Woy Yepe Adis yrepy ur}. MO]IOA oped AroA yor[q anjq-astonbin} daap.. udd13 MOTIOA,. jeondiyja ‘morreu,, woo yep ugois oind,, udeis yYysLq sind, (afURUI=) dATTO Quou,., uddI3 SUIT]. $20} sv ures Aogis-ysryuid yep aye1opoul yoy yor lq anyq 4ep You anjq-astonb.in) ayed arenbs ‘prog A19A uooJeW 1Y31I1q udo13-ysintq A[qeiiea udeis-ysinjq A[qeiiea (epueUIZ) aArTo 1YSTIq Modeu AT[ensn ude18-YsIn]q IYs11q INo[Oo MET INO[O9 90], Ou Jo dn jo peg Ta Jo dn yo mojo5 O[PPIUU Iv yeYs TY JO nojoD Ty JO wonsesprur Jo mojoD TIV.L SJIOAOO SUIMIJOpUN JO INO[OD €d Jo adeys yoyed 3UIM pal Jo INO[OD $}19A09 IOSsa] JO INO[OD ONIM. duini jo mojo Japsog apueUl/Ie[[oD Jo INojoD Ie]JOO UO anyq Jo Waxy Ieyjoo adeu jo mojop xAdOg Aais Ayeys Wem Aoais Aye]s ploo onjq-Aois anjq-Aois PII x ,SANPU SUTUTYS adeu/uMolO Ied1 JO MO;OD Kors Ayeys Agis posun-useis yurd-ysijdind yurd-ysijdind ajed ¢ Yysiq ¢ yuld-uowyes,. per IYys11q 0} (g) yuId Yys1Iq S910] JO INO[OD MOUeU MOLIeU a)8.19pour a1eIapouls, OpIM ATOANLIAI apm AToAyeIo1 ULYS [PIIGIO o1eq JO PIM poprArp Aped ‘opi poprAip Apued ‘opin UOISIAIP JYSI[S “OprM Ayey UOISIAIP JYSI[S ‘apr Aparey UOISIAIP JYSI]s 10 OU ‘apiM AyIIey UOISTAIP JYSI[S 10 OU ‘apm ATATey 2199 Jo adeys ujoy-asue10 aed asurio ajed uwioy Wem Woy WAeM 0} [eNOU yspyor[ qs. pAois-UunI paul 9199 JO INOJOD UIOY-YsTMoT[oA WIOY-YsTMoT[oh WIOY-YsIMo][oA UIOY-YsTMo][oA > ed (¢] orig 2[qIpuew Jo mMojoD Aavay ayes Aavoy ayeJOpoul a}e.1opoul [jews [yews PI[IXBU JO 2ZIS pol-osue10 o8URIO-pol asurl0 osuvio UIOY-YsIMoy[oh (.0Z/6L) UIOY-YstMoy[oA PI[IXeBU JO aseq JO INO[OD avaH nyosuyf q puvkvjouny J spuigdy s,o10ulaspas pIpawiaqul HNINV DIDASOL Y pypydaz0uv4o gq soroadsiadns ,popeoy-Ajr]s, Dipawuiajul J sotoadsiadns ,poperoy-wossojg, ee “T Q[QRL, 998 YOIyM JO} ‘Vipaiwsajui ynq [ek 10} payeoipurl UONeLIeA, ‘(L] =u) nyosurf gq pure ‘(97 = Uv) DuDkyjouny gq (+ =U) Splaghy s,a10Wespag ‘vipauiaqul q HNIWY ‘(11 = 4) vivasos J “(17 =U) vjpydasoundo vjnovyisg JO sisouseIp pliqdy UI pasn siajoweY) fp FUL. TAXONOMIC STATUS OF PSITTACULA INTERMEDIA K—S— 10 H-=_ =— 42 juv. H—=—j—— 7 Eastern F—==- 30 Western F == 20 Sane 2 Sedgemore = 5 In@5 1 pet 40 R—=Z=— 35 el At aro Factor 1 92.1%, 15.8; bill size Fig. 3 Graph of means (squares), standard deviations (heavy bars), and ranges (narrow bars) of individual scores for putative parental species and hybrids on Factor 1, the only significant axis in a PCA of culmen length (component loading 0.95), maxilla height (0.96), and maxilla length (0.97). The number to the right of the range bar is n. Symbols are as for Figure 1 and 2; eastern and western finschii are included as separate groups. larger than intermedia. In the plot of the distal width of rectrix 2 (R2) vs. the pale tip length of that feather (Figure 2C), intermedia fall betweencyanocephala/roseata andhimalayana, but not mostfinschii. Psittacula krameriis unique in its combination of a broad, very short pale tail tip, and in this Sane’s specimen is intermediate between krameri and cyanocephala/roseata, and certainly not himalayana. On a PCA of three bill measures (Figure 3, Table 3) selected to allow inclusion of as many specimens and live hybrids as possible, the only significant axis was Factor 1, a very strong size axis. On this axis, cyanocephala and roseata had the smallest mean factor scores, with the BMNH intermedia slightly larger. The other putative intermedia and known hybrids fell between these and the succes- sively larger finschii and himalayana groups. Psittacula krameri was much the largest, and Sane’s specimens were the largest of the putative hybrid groups, again showing the influence of krameri. In a PCA for which variables were selected to allow inclusion of one of Sane’s specimens (Figure 4A, Table 3), the AMNH intermedia and Sedgemore’s birds group near each other, and between the widely spaced roseata/cyanocephala and himalayana groups, but overlap considerably with finschii. Sane’s bird, however, falls between the roseata/cyanocephala and krameri groups. Another PCA for which the variables selected allowed inclusion of Sane’s other specimen (Figure 4B, Table 3) showed AMNH intermedia grouping out halfway between cyanocephala and himalayana, with the mean of roseata falling out more distantly. In this case, the second Sane specimen falls out much closer to cyanocephala than to krameri. WING AND TAIL FORMULAE In mean shortfalls of each primary tip from the wingpoint (Figure 5A), the AMNH intermedia are completely intermediate between himalayana and cyanocephala. However, BMNH 1980.3.1 is very like the mean of cyanocephala in pattern of primary shortfalls from the wingpoint. AMNH intermedia are closer in mean primary short- falls to roseata than to cyanocephala (and thus less intermediate between roseata and himalayana; Figure 5B), but neither roseata norfinschii could be ruled out as parental species on this basis alone. Sane’s single specimen on which these characters are measurable is nearly intermediate in primary shortfall pattern between cyanocephala andkrameri (Figure SC), although again these data do not rule out some other parental combinations. On mean widths of primaries, AMNH intermedia were again intermediate between himalayana andcyanocephala except in width of P2, a measurement that is highly dependent on shortfall of P3 (Figure 6A). In spacing between tips of rectrix pairs 3-6 (Figure 6B), 39 17.7%, 1.1; R2 tip length vs. width Factor 2 a. S 0 l 2 3 Factor 1 71.5%, 4.3; size (R2 w, tip | uncorrelated) // Sane2 9 Sra / s ¢ @ @ F ————— 3.0%, 15.9; bill and Plw, notch vs. P3s, P2w So Factor 2 2 -1 “0 Factor 1 12.1%, 63.7; size (P3s, Plw uncorrelated) Sse | Q) Fig. 4 Identity of AMNH intermedia with Sedgemore’s hybrids, distinctness from Sane’s specimens, and intermediacy of all the above between putative parental species: graphs of individual scores on Factors | and 2 from PCAs on measurements of adult males of putative parental species and hybrids. Symbols are as for preceding figures. Summary statistics of PCAs are given in Table 3B and C. (A) Variables chosen to allow inclusion of Sedgemore’s hybrids and Sane’s first specimen; (B) variables chosen for inclusion of Sane’s second specimen. 40 P.C. RASMUSSEN AND N.J. COLLAR 80; == himalayana j ; 15 — 70| ™ = AMNH intermedia Aj oe cyanocephala — BMNH intermedia = 50 | = = : m= 10 — 30 = 20 10 A %, | : | | | | | | 1 (max) 1 (notch) 2 3 4 5 Primary No. 80 os mum roseata = "70 = = = AMNH mtermedia & 60) = finschii Be =~ 80 = 50 = : E = 40 Z : 2 60— e380 = : 3 = S 20 2 2 ae a 10 E B ry 2 8 20— = A B TITTLE 0 | | | l * == 2 3 4 5 krameri Rectrix No. 70 == Sane specimen e \\ 60| sini cyanocephala cS 50 ” a en ; E I = 7.5 |— “ton, 3 --" aes eosin rl 4} “Ny, , OS hd 2 yo! C 5.0 ) ) | ) | : WAL ead jeep eub Ness Nee) 12A/ | Ney Nee) eal) 1 (tip) 1 (mid) : 3 4 5 Primary No. Rectrix No. Fig. 5 Graphs of mean shortfalls of primaries from wingpoint for: (A) Fig. 6 Remex and rectrix width and spacing for Psittacula cyanocephala, Psittacula cyanocephala, P. himalayana, and P. intermedia (BMNH P. himalayana, and P. intermedia: (A) mean widths of primaries, specimen separate); (B) P. cyanocephala, P. roseata, and P. intermedia maximum and at notch for P1 (outermost), and width of P2—PS at next (BMNH specimen separate); (C) P. krameri, P. cyanocephala, and one innermost remex; (B) mean spacing between rectrices; (C) mean rectrix of Sane’s specimens. Full data are presented in Appendix. widths at next more lateral rectrix. Full data are presented in Appendix. TAXONOMIC STATUS OF PSITTACULA INTERMEDIA the AMNH intermedia are again totally intermediate, but BMNH 1980.3.1 differs from intermedia and is very like roseata (Appen- dix). In rectrix widths (Figure 6C), AMNH intermedia are more or less intermediate, although the central rectrices are somewhat closer in width to those of cyanocephala, while the outer rectrices are closer to those of himalayana; BMNH 1980.3.1 differs here as well (Appendix). STYLES OF SPECIMEN PREPARATION Examination of preparation styles of the Rothschild intermedia (Figure 7) other than the type (which predated the others) but including the juvenile himalayana (AMNH 621545), showed that one (AMNH 621541) has a longish neck and understuffed throat (vs. short necks and breast nearly touching the bill on the others); another (AMNH 621542) has the bill slightly open, while in three (AMNH 621544—-5, BMNH 1980.3.1) the maxilla is extended far beyond the mandible (vs. naturally positioned in AMNH 621540-1). The wings are positioned far forward and low on the body on two (AMNH 621544—5), and low but to the sides on two others (AMNH 621540— 1) vs. well-positioned to the sides on the remaining two (BMNH 1980.3.1, AMNH 621542). The body is compressed dorsoventrally in all but one (BMNH 1980.3.1). The tail is twisted in relation to the body in all but two (BMNH 1980.3.1, AMNH 621541); one with a twisted tail (AMNH 621542) has one of its central rectrices rotated 180° and its rectrices are spread, while they are folded tightly in the others. One specimen (AMNH 621542) is filled with dirty cotton, while another (AMNH 621544) has the body made of a tightly but roughly wound ball of coarse brown fibres each about 0.5 mm in width, and the rest of the Rothschild specimens are stuffed with rough bundles of straw. Support sticks in two (AMNH 621542, 621544) are thin (ca. 3 mm diameter), whittled, and orange-brown; thicker (3.4 mm), rougher, and grey-brown in another (AMNH 621540); and very thick (ca. 7 mm), coarse, crudely broken, and dark brown in yet another (AMNH 621545), while sticks are not visible externally in the other specimens. All the Rothschild intermedia share the following external prepa- ration features: the eyes are not stuffed and are dried shut; the breast is crudely stuffed so that the feathers of the upper breast are pushed outwards and upwards; the abdominal incision is rough; the tibiotarsi are broken medially and the feet were not secured, now being entirely missing in three (presumably having been lost after prepara- tion). Strangely, the only published photograph of any of the Rothschild intermedia specimens (AMNH 621540, in Arndt 1996) was digitally enhanced to add in a lifelike eye and periorbital skin, even though it lacks a wing on the side photographed. Radiographs (Figure 8) elucidate additional pertinent preparation features of the Rothschild intermedia specimens: AMNH 621541 and 621545 both had similar loose-woven cloth wound around the top of the support sticks and pushed into the open back of the skulls, while the others have little or no stuffing in the skulls. The X-rays confirm the similarity between the straw used in stuffing of five specimens (AMNH 621539-621541, 621545, BMNH 1980.3.1), and show that straw is lacking in two others (AMNH 621542, 621544). The body of AMNH 621540 is fusiform, while in AMNH 621544 the rear body is nearly empty. BMNH 1980.3.1 lacks a support stick altogether. In some of the specimens (AMNH 621540- 1, 621545) the support stick is jammed into the braincase, while in AMNH 621544 the tip lies between the orbits, and in AMNH 621542 it projects into the base of the maxilla. In all specimens, the wings are positioned carelessly and variably, and those of the holotype are positioned differently to the rest. Similarities among the specimens visible in the x-rays include: most or all of the radii and the entire humerus have been removed; much of the back of the 41 skull was removed but in an inconstant manner; the bones were often haphazardly broken and bone chips are embedded inside five speci- mens; and sacral vertebrae were left in five specimens. Other dataless specimens examined by us that had been prepared in this characteristic native skin style (the ‘Bombay preparation’: Rothschild 1895) include the following abnormally plumaged specimens: a partial lutinocyanocephala(AMNH 454031); a yellow- tinged (flavistic?) cyanocephala(AMNH 621491); and a near-lutino krameri (AMNH 454030). Bombay preparation Psittacula skins with normal plumage at AMNH include: three cyanocephala (621490, 621492, 621537); five himalayana (621551-621554, 621556); two finschii (621550, 621557); and one krameri (621337). Further Bombay preparation skins are now in other collections (e.g. MCZ 383245). No skins of the Bombay preparation were found among the AMNH series of other Indian Psittacula species, al- though other presumed native skin styles are represented among them. EVIDENCE FOR CAPTIVE ORIGIN We found the following features among the Rothschild intermedia that are consistent with their having been held in captivity: (1) breakage or damage of primaries in three (AMNH 621542, 621544— 5); (2) loss of central rectrices on AMNH 621539 and ongoing replacement of central rectrices on at least three others (AMNH 621541—2, 621544 and presumably 621540, showing that a high proportion of the sample is in moult); (3) irregular dark worn areas on feathers of the carpal area and heavily frayed wing coverts in AMNH 621544; (4) a featherless patch on the left side of the upper breast, and damaged feathers on the forehead of AMNH 621544 and shoulder of AMNH 621542; and (5) dirt on the feathers of the breast and/or belly of four (AMNH 621539, 621544—-5, BMNH 1980.3.1), dirt on the right wing of BMNH 1980.3.1, and apparent whitewash on the upper tail surface of BMNH 1980.3.1. PHOTOGRAPHS AND OTHER REPORTS OF CAPTIVE INTERMEDIA A photograph of a parakeet published in Herrmann (1994) as a male cyanocephala is instead much like the five AMNH adult male intermedia specimens. This individual (which has now been sold) and its mate (which has died and for which no details are available) were said to be wild-caught from an unknown locality and were held in captivity in Austria, where the photograph was taken by F. Pfeffer in 1985 (T. Arndt, in litt. 1997). Based on a copy of Pfeffer’s photograph, the male differs from cyanocephala (and agrees with AMNH intermedia) in having two-thirds of the upper mandible red- orange and fairly large; a pale yellowish lower mandible; a large head with more extensive, slatier-blue areas on the rear face and head; a weaker blue wash on nape, wing coverts, and rump; and the tail tip longer and yellowish. From the photograph of the male, it appears to differ from the AMNH intermedia in having a paler, less pure yellow tail tip; narrower central rectrices; a yellow-green area between the mantle and the hindcollar; and a dark maroon shoulder patch. A second male (also of unknown provenance but from around 1985) was recently located in captivity at Turnersee Bird Park in Austria by R. Low (in litt. 1997) and F. Pfeffer (T. Arndt in litt. 1997), and was almost immediately published (with colour photos) as a true intermedia (Fuchs 1997). Photos of this bird (which is missing parts of its toes) from all three above sources show it to be very similar to the previously mentioned captive bird in Austria, except that its bill looks smaller and less orange-red. Some of the photos clearly show very narrow central rectrices with very long pale yellow tips and whitish shafts encroaching into the blue portion at least as far proximally as the level of the R2 tips. R. Low (in litt. 1997) stated it was the same size as the female cyanocephala with 42 P.C. RASMUSSEN AND N.J. COLLAR Fig. 7 Photo of all Rothschild Psittacula intermedia skins to show preparation styles: (A) ventral view of (left to right) AMNH 621539 (holotype), 621540-2, 621544-5, BMNH 1980.3.1; (B) lateral view. TAXONOMIC STATUS OF PSITTACULA INTERMEDIA Fig.8 X-rays of all Rothschild Psittacula intermedia skins to show preparation styles and materials. Ventral view of (left to right, top row) AMNH 621544, BMNH 1980.3.1, AMNH 621539 (holotype), 621540, (bottom row) 621545, 621542, 621541. 44 which it was kept, and that the owners informed her that it had been bred from cyanocephala at Vogelpark Turnersee. However, the article about this individual (Fuchs 1997), which claimed that it was a great rarity, does not mention its parentage nor even the possibility that intermedia is a hybrid. Sane (1977) recalled that in 1972 or 1973 he had seen a bird similar to his first ‘intermedia’ , but that it had been purchased by H. H. Jamsaheb, Nawanagar, India. He had also been told in about 1976 that someone in Britain had two or three intermedia, and that 8-10 had been imported into Holland in 1976. An individual was offered for sale in 1976 as ‘probably the only known specimen in captivity in the world’ (Inskipp and Inskipp 1995). However, we know of no documentation for any of these claimed intermedia. In a poster session at the 1996 BirdLife Asia Conference in Coimbatore, India, R. Bhargava exhibited a photo of a captive bird that closely matched the AMNH specimens already examined by PCR. Bhargava (pers. comm. 1996) informed us that such birds were not rare in the pet trade within India, but that the illegality of this trade makes documentation difficult. He had obtained as many as five individuals from traders at one time, although one of these has since been ringed and released at a news conference (Anon. 1998) and three have died and been preserved as specimens. Bhargava’s three specimens appear from photographs to be indistinguishable from AMNH intermedia, but definite, verifiable data on their prov- enance appear to be lacking. KNOWN HYBRID CYANOCEPHALA X HIMALAYANA After the publication of Walters (1985), Sedgemore informed M. P. Walters (pers. comm., 1995) that he had crossed himalayana and cyanocephala in captivity and obtained intermedia-like hybrids. Beginning in the late 1970s Sedgemore tried pairing captive himalayana andcyanocephala to determine whether Husain’s (1959) hypothesis was correct (M. Sedgemore, in litt. 1997), but his adult male cyanocephala (age unknown) and five-year-old female himalayana refused to bond. In the mid-1980s he tried pairing different individuals of the same species, but again without result (M. Sedgemore, in litt. 1997). Then in 1991 he housed an immature female himalayana and immature male cyanocephala together, and these showed pair behaviour that year but did not breed. In 1992, of three eggs laid, two were clear and one was fertile but did not hatch, and in 1993 three more eggs were produced, only one of which hatched but the chick died at two weeks of age. Finally, in 1994, all three eggs laid hatched and the chicks fledged, as did chicks from two of three eggs laid in 1995 (M. Sedgemore, in litt. 1997). All the hybrids — four males and one female — were still alive and in adult plumage when we saw them in September 1997. Photos of Sedgemore’s hybrids with their parents show that the mother was a typical himalayana and the male parent a typical cyanocephala, the latter being additionally confirmed by the speci- men. All these photos and our direct examination, photographs, and measurements confirm the identity of the hybrids with AMNH intermedia. The heads of all four adult male hybrids were coloured as in the five AMNH adult specimens, with only slight variability among them. Their ceres were pale fleshy horn; their maxillae orange-red on the basal two-thirds and with yellowish tips; their lower mandibles were pale; their eyering skin was pale greyish; and their feet and claws were pale greyish-pink. All had a broad pale greenish-blue hindcollar, lesser wing coverts, ramp, and underwing coverts, the undersurfaces of the rectrices yellow with the outer webs bluish proximally, and the tail tips slightly broadened and pale yellowish from above. The single female hybrid was less distinctive but still possessed characteristics which should enable recognition of specimens or P.C. RASMUSSEN AND N.J. COLLAR captives. It would be immediately distinguishable from adult himalayana or finschii by its much paler, duller grey head and lack of a narrow black collar, and from males of the above two species by its lack of a maroon wing patch. Its head was drabber grey than adult female cyanocephala, with a pale area in front of the eye much as in juvenile cyanocephala; its upper mandible was heavier and was strongly tinged orange at the base; its lower mandible was pale; it had a slight yellowish collar on the sides of the neck (paler and duller than in adult female cyanocephala and less ochraceous than in female roseata); and it had long, slightly broadened, pale yellow tail tips, which extend much farther proximally on the feather than in cyanocephala. Sedgemore (1995) briefly described the juveniles of the first brood. His photos of four of these same hybrids as fresh-plumaged juveniles show that they would be difficult but not impossible to distinguish from those of either parental species. All four hybrids had a pale area on the front of the face (on the forehead, lores, and area around the bill base); greenish-grey auriculars; a dull green crown and nape; and a pale yellowish-green collar contrasting with the head and mantle. From below, the tails of the hybrids were narrowly yellow-tipped. The upper tail surfaces are visible only on two individuals, in which they were greenish-blue with a pale yellow tip. Note that the latter character disagrees with Tavistock’s (1933) statement (see below) that the tails of his hybrids were brighter blue than in young cyanocephala and were white-tipped. A photo of one of Sedgemore’s 1995 hybrids in direct comparison with its 1994- hatched brother (an adult by then) shows that the juvenile has the upper tail surface more turquoise than the adult male. The young hybrids differed from juvenile himalayana by their smaller bills, their lack of blackish blotches at the bases of the maxillae, their duskier ceres, yellower collar, bluer upper tail surface, and narrower rectrices, and from juvenile cyanocephala by their duller, less yellow collar, and yellowish tail tips. It is uncertain whether juvenile hybrids between cyanocephala and himalayana could reliably be discriminated from young finschii, which would be more similar in size and proportions. Successful hybridization of male cyanocephala x female himalayana was also achieved earlier by Critchley, who placed an adult male cyanocephala whose mate had just died in a mixed aviary with a placid adult female himalayana that had lived in a pet shop for several years (L. Critchley, in litt. 1998). The birds paired up during the first spring (1989) that they were kept together; from all five eggs laid, chicks hatched and were reared successfully. The fates of four of the young which were sold to a pet shop are unknown, but the remaining male moulted into adult plumage in August 1990 and was in Critchley’s aviaries as of May 1998. Several photos, including one of the adult hybrid with its parents, show that Critchley’s hybrid matches in every plumage detail the AMNH intermedia series and Sedgemore’s male hybrids, and confirm the specific identity of the parents. Finally, a male cyanocephala and female himalayana also hybridized successfully in the aviaries of Mr E. Beale (now de- ceased, M. Sedgemore, pers. comm. 1997), and the pair reared two chicks in 1980. These had yellow tail tips as juveniles (Beale 1981), but this brief description does not enable evaluation of whether they match Rothschild’s intermedia in other respects. DISCUSSION SANE’S CAPTIVES Recognizing that the bird in Sane’s collection photographed by Wirth (1990) was different from the AMNH material, Arndt (1996) stated that ‘specimens caught on the plain of the Indian state of TAXONOMIC STATUS OF PSITTACULA INTERMEDIA Uttar Pradesh and described in literature as Psittacula intermedia do not probably belong to this species’. It has been noted else- where (Inskipp and Inskipp 1995) that the insufficient descriptions and the lack of photos in Sane et al. (1987) prevent independent evaluation both of their putative intermedia and their statement that ‘each year between 1979 and 1984, one or two live specimens of this species were available in the Indian bird market most of which could not however be acquired by us. . . . The following facts, however, show their true identity: (a) all Sane’s birds share characters of krameri and cyanocephala, but not himalayana; (b) Sane (1975, 1977) himself described his immature bird as looking like a hybrid krameri x cyanocephala, with ‘some reddish colour resembling somewhat that on the nape of the Rose-ring Parra- keet’; (c) trappers informed Sane that his birds were caught with, and were natural hybrids between, krameri and cyanocephala (Sane 1975, 1977); and (d) Sane et al. (1987) were confused about soft part coloration, post-mortem changes, and sexual di- morphism in their birds (see below, also Inskipp and Inskipp 1995). Sane (1977) also stated that ‘the call of my bird is more like a Ringneck than a Blossom-head’ [here meaning cyanocephala). This evidence, coupled with our new data on Sane’s three birds, demonstrates that all the intermedia reported by Sane (1975, 1977) and Sane et al. (1987) are krameri x cyanocephala rather than intermedia. Sane (1977) identified his birds as intermedia from the description in Biswas (1959), and related that when Biswas saw Sane’s first bird, he indicated that it appeared to be intermedia but that measurements were needed for confirmation. Sane later took measurements which he said ‘con- firmed [sic] with the original’ (Sane ef al. 1987). The lack of red shoulder patches in two of Sane’s birds virtually rules out the possibility of eupatria (in which both sexes have shoulder patches) as a parental species, but it is consistent with krameri. In addition, the great size difference between eupatria and cyanocephala makes their hybridization unlikely, and the bill shapes of the Bombay specimens and live bird do not resemble that of eupatria. The only other possible combination that could result in such a phenotype would be krameri x roseata. However, all Sane’s birds have a slight bluish tinge to the nape and wing coverts, so roseata could hardly be a parent, since this colour is lacking on both krameri and roseata. The assumption by Sane et al. (1987) of reversed sexual dimor- phism in intermedia based on the presumed sex of two of their living captives has already been called into question by Biswas (1990), and Inskipp and Inskipp (1995) also queried this as well as their state- ments on post-mortem changes in bill coloration. Now that we know that these birds are krameri x cyanocephala (explaining why two of them lack shoulder patches), the only remaining published corrobo- ration for Sane ef al.’s (1987) hypothesis of reversed sexual dimorphism is, by their own account, ‘another male in the collec- tion, which . . . had mated with a female Roseringed parakeet. However, the eggs laid were infertile. Their ‘subadult female’ intermedia must presumably have been sexed by inference on the basis of its shoulder patches after Sane er al. (1987) had concluded that those lacking this feature were males. Its measurements were then compared with those published for AMNH intermedia, which Sane et al. (1987) termed ‘female’, but which surely must be males. However, even though male krameri lack shoulder patches, male cyanocephala possess them, so some male krameri x cyanocephala could have patches as well, and thus Sane et al.’s (1987) sexing of their subadult bird as a female on this basis is not upheld. In addition, the unaccessioned stuffed specimen (presumably the ‘female’) showed only a hint of a shoulder patch. Thus, even if Sane’s birds really were intermedia and even if intermedia was a valid species, it 45 is untenable to presume that it would exhibit reversed sexual dimor- phism with respect to other Psittacula species. Clearly, the specimens (number not stated) whose blood was used in the electrophoretic analyses reported in Sane et al. (1987) must have been krameri x cyanocephala, rendering the results inapplica- ble to the question of the taxonomic status ofintermedia. In addition, the four loci examined do not form an acceptable sample, the methods of analysis and interpretation of results are problematic (R. Fleischer, pers. comm. 1997), and the major differences claimed between intermedia and other Psittacula species are improbable, especially given the apparent parentage of the individuals sampled. BMNH SPECIMEN One of the original Rothschild Collection intermedia specimens, BMNH 1980.3.1, was the one on which Husain’s (1959) analysis of the hybrid origin of intermedia was primarily drawn. It is, moreover, more like the illustration of male intermedia in Inskipp and Inskipp (1995) than are any of the AMNH intermedia. However — and despite Hartert’s (1924) assertion that specimens in the Rothschild Collection are alike - BMNH 1980.3.1 differs in several respects from the remaining five adults (Table 1). It is closer in overall appearance to adult maleroseata than are the others, and is mensurally similar to both roseata and cyanocephala; except for the pale lower mandible and lack of shoulder patches it could be a hybrid cyanocephala x roseata. \t could also be an F2 hybrid, or if bred in captivity, a trigen. Its tail tips are broader than in either roseata or finschii, while the turquoise upper tail surface and tail tip coloration, shape, and length are as in roseata. The slightly brighter red on the front of its face than in roseata cannot be explained as roseata x himalayana ot finschii. However, its complete lack of reddish shoul- der patches is unique among the intermedia series, is not due to feather loss, moult, or immaturity, and defies ready explanation. On present evidence we cannot resolve its parentage, but it does not appear to be an Floffspring of a cyanocephala x himalayana cross. EVIDENCE FROM THE AMNH SPECIMENS As far as we can determine, all characters of adultAMNH intermedia are either (a) shared with the himalayana/finschii species pair or the roseata/cyanocephala pair, or (b) intermediate between one or both members of these two species groups. If for the moment we accept intermedia as a hybrid (to be further substantiated below), then we must assume (as did Husain 1959) that one member of each of the above species pairs was the parental species. Psittacula roseata cannot have been involved, as its facial coloration is not bright or deep enough to result in an intermedia phenotype, and its P3 is much too narrow (Table 4). In additionroseata lacks bluish on its hindneck, wing coverts, and rump, while most AMNH intermedia have the blue tint in these areas stronger than on the himalayana/finschii pair, and the upper tail surface of roseata is a paler, greener blue than in the other species and in AMNH intermedia. However, hybridization of cyanocephala with either himalayana or finschii would involve none of the problematic characters of roseata. Nevertheless, finschii has narrow central rectrices that make it unlikely to be a parental species, whether mated withcyanocephala or roseata, since the mean distal width of the central rectrices is greater in intermedia than for any of those three species (Appendix; see also Husain 1959). In addition, a finschii x cyanocephala or roseata cross could hardly result in the bright yellow tail tips of AMNH intermedia (as already noted by Husain 1959). Finally, unlike intermedia, finschii has a bright yellow-green band above the mantle and pale shafts on the upper tail surface. Incidentally, the tail tips of the individual illustrated as himalayana in Inskipp and Inskipp (1995) are actually those of finschii. Mensural and statistical analyses show the intermediacy of AMNH 46 intermedia between one or both members of the himalayana/finschii and roseata/cyanocephala species pairs in every character set exam- ined. Because roseata and cyanocephala are similar in size and proportions, roseata is not mensurally ruled out as a parental spe- cies, but it is ruled out on plumage (see above). However, finschii is smaller than himalayana, particularly in wing characters, and its tail proportions are different from any other species and AMNH intermedia, so it is unlikely to have been parent to the latter. We know of no cases in which a good species, which is intermediate between two patently different congeners in numerous phenotypic characters, totally lacks distinctive features of its own. Also, it is scarcely conceivable that a wild species would duplicate exactly the character states found in known hybrids between two quite distinct taxa. Thus both plumage and mensural analyses very strongly support the hypothesis that AMNH intermedia are of hybrid origin, and this is further validated by their identity with known hybrids between himalayana and cyanocephala. ARGUMENTS PREVIOUSLY USED IN FAVOUR OF SPECIFIC STATUS Hartert (1924) stated that if intermedia were a hybrid, “so many specimens would not very likely have come at the same time,* and one would expect them to vary, but they are all alike’, with foot- note 2 disclosing that ‘Our six males were selected by Mr. Dunstall, a dealer in feathers, from a greater number of these birds, he told us’. These statements have often been repeated as evidence of specific status (Biswas 1959, Walters 1985, Inskipp and Inskipp 1995), but both are flawed. First, the holotype did not originate with the other intermedia. Second, one of the six re- maining specimens is a typical immature himalayana (see above), leaving only five intermedia supposedly of similar origins and identity. Hartert’s assertion that these came from a greater number (though not “a much greater number’, contra Walters 1985) in the possession of and selected by Dunstall implies that Hartert him- self did not see additional intermedia but had taken the London plumassier’s word for it. There is nothing to indicate that Dunstall would have recognized the difference between intermedia and cyanocephala, and (our third objection) there seems no compel- ling evidence that he actually had more specimens of intermedia: the ‘greater number of these birds’ may have referred to the rest of a shipment of other Psittacula parakeets, a possibility supported by one of the six ‘intermedia’ being a juvenile himalayana. There was a considerable millinery trade in Psittacula skins around this time (Hartley 1907), and only a very small percentage would have ended up in reference collections. Fourth, preparation styles and materials used in the five Dunstall intermedia plus the immature himalayana differ strikingly among the skins. These differences strongly suggest that, although all are native skins, they were not all prepared at the same time and place. They may have come to Rothschild’s museum at the same time, but not necessarily so to Dunstall or his supplier. Additional support for staggered acquisition of the material lies in the fact that native skins of cyanocephala and himalayana strongly resemble, in style and materials, not only those of intermedia but also those of finschii, which (given their distribution) must have originated farther east. Conversely, one adult intermedia (AMNH 621541) and the imma- ture himalayana are so similar in preparation materials as to make it highly probable that they were prepared together. Many of the ‘Bombay preparation’ parakeet skins very likely came from bird markets to which captive birds had been brought from afar. Indeed, the incidence of at least three partial lutino specimens of this same preparation suggests selective breeding for this trait, which has long been highly desired by Indian aviculturists (Greene 1884). After restating Hartert’s (1924) contentions that intermedia is a P.C. RASMUSSEN AND N.J. COLLAR valid species, Biswas (1959) indicated he had concluded the same independently. However, his only further evidence was as follows: ‘Besides, if they were man-made hybrids, they would necessarily have been cage birds. But the character of their toes does not indicate this. Psittacula intermedia may, therefore, be regarded as a genuine wild species’. Besides the obvious fact that bird hybrids are not necessarily ‘man-made’, Biswas (1959) gave no indication of which features of the toes were found inconsistent with captive origin. Our examination showed that the intermedia specimens exhibit to varying degrees several conditions consistent with their having been in captivity under suboptimal conditions (Harrison and Harrison 1986). It thus seems likely that most or all of the intermedia specimens in existence were captives for some period immediately prior to their death. In India, Psittacula parakeets have long been extremely popular cagebirds (Finn 1906, Ali 1927, Dharma- kumarsinjhi 1954, Sinha 1959), and although most are taken from nests (Hume 1890), others are captive-bred commercially (H. S. A. Yahya, pers. comm. 1997), and they have long been bred for the Indian aristocracy (Greene 1884). Mutations in particular are bred in captivity in India (S. R. Sane, pers. comm. 1997), garnering up to Rs. 20,000 (Ahmed 1997). From the prices of cagebirds considered to be intermedia (Rs. 2,000 vs. less than Rs. 25 for ordinary cyanocephala: A. Rahmani in Inskipp and Inskipp 1995), it is self-evident that captive-breeding of such hybrids would be well worth the trouble. Walters (1985) drew attention to previously overlooked descrip- tions of captive-reared cyanocephala x himalayana (Tavistock 1932-1938) which do not match Forshaw’s (1973) description of intermedia. Based on this, as well as Hartert’s (1924) arguments, Walters (1985) concluded that intermedia could not be a hybrid between those species and must therefore be a valid species, and in this he has been followed by most recent authors. Presumably also on the basis of this captive-breeding event, Arndt (1996) stated ‘it has been discovered that hybrids between [himalayana and cyanocephala| differ considerably from Psittacula intermedia. . . . However, a reevaluation of the aviculturist Tavistock’s writings shows some relevant discrepancies. First, although Tavistock (1932—1938) did repeatedly pair a male himalayana with a female cyanocephala, rearing at least seven young over a period of five years, he published only a very brief description of just two of those young, which were nestmates (Tavistock 1933), and did not describe their adult plumage. Thus there is no assurance that the other hybrids resembled these two, nor is there information enabling comparison between his hybrids and AMNH intermedia. Second, while it is true that Tavistock’s (1933) description of the two young cyanocephala x himalayana as having white tail tips does not match the specimens of intermedia (the discrepancy noted by Walters), other statements Tavistock made throw doubt upon his entire account. His remark that ‘they resemble young Plumheads, but their central tail feathers are brighter blue with white tips and their heads have a dusky tinge’ is nonsensical, as young cyanocephala do have white tips to their tails, and their heads may have a dusky tinge much like juvenile himalayana. A\so, it is counterintuitive that hybrids would have brighter blue central rectrices than those of young cyanocephala, since juvenile cyanocephala have these feath- ers considerably bluer than do juvenile himalayana, which are green-tailed. Sedgemore stated that his juvenile cyanocephala x himalayana had ‘blue green’ upper tail surfaces (Sedgemore 1995), and this is confirmed by photos of them as juveniles. The central rectrices of the adult plumage of both parental species are bluer than in the juvenile plumage, while those of adult cyanocephala are bluer and less purple than for adult himalayana. Since the two young hybrids described by Tavistock (1933) had only hatched that year, TAXONOMIC STATUS OF PSITTACULA INTERMEDIA presumably in the spring or summer of 1933, they could scarcely have moulted into diagnostic adult rectrices before Tavistock’s article went to press. These inconsistencies indicate that little weight should be given to Tavistock’s rather off-hand description. Although not in connection with intermedia, Low (1992: 118) mentioned hybrids bred by the Duke of Bedford (then the Marquess of Tavistock), stating that the Duke had paired a male finschii with a female Blossom-headed Parakeet. However, Tavistock (1932) spe- cifically stated he used a male ‘Hodgson’s Slaty-headed Parakeet’ and female Plumheads (Tavistock 1932—1938).Whilecyanocephala has often gone by the common name of Blossom-headed Parakeet, to our knowledge roseata (earlier known as rosa) has not been called Plum-headed Parakeet, so he probably used cyanocephala. “‘Hodgson’s Slaty-headed’ can refer only to nominate himalayana, not finschii. However, whether or not Tavistock had used true himalayana, some progeny of a cross between yellow-tipped and white-tipped parents might well show white tail tips, and in any case the tail tips of finschii are glaucous yellow, not white. GEOGRAPHIC PROVENANCE Biswas (1959) indicated that the Rothschild Museum label of the type specimen of intermedia states ‘India Nat. Skim’ and considered it uncertain whether ‘native skin’ or ‘Native Sikkim’ (= then-autono- mous Sikkim) was meant. This was then taken by Ripley (1961) and Ali and Ripley (1969) as indicating that the type probably originated in Sikkim. However, the type’s original label (the only one borne by the specimen, in Hartert’s handwriting) clearly reads ‘Nat. Skin’, by which was meant the Bombay preparation of these trade skins, and thus there is no evidence pointing to Sikkim as the region of origin. On the basis of Rothschild (1895) and Hartert (1924), it has also been assumed that intermedia is from the Western Himalayas (Forshaw 1973, Sibley and Monroe 1990), an idea reinforced by Sane et al.’s (1987) birds that we now know are krameri x cyanocephala, although the latter were reputedly from the plains just to the south (Sane 1977, Sane et al. 1987, Knox and Walters 1994). However, there is no basis for this assumption regarding either lot of Rothschild Collection skins that contained inter- media specimens. In the description, Rothschild (1895) stated that the type came to him with two skins of Palaeornis schisticeps and, because it was shipped from Bombay, it most likely came from the “Western Provinces’. However, later Rothschild (1907) stated that speculation was useless, as the same shipment con- tained birds from various parts of the Himalayas; still later, Hartert (1924) stated that the birds evidently came from some part of the Himalayas. Subsequently it has been assumed without comment (Biswas 1959, Ali and Ripley 1969, Walters 1985) that by schisticeps Rothschild meant Psittacula himalayana of the west- ern and central Himalayas. However, Rothschild did not differentiate between himalayana and the eastern form, finschii, both of which were then known as schisticeps, in his description of intermedia. Since at least two native skins of finschii of the ‘Bombay preparation’ (AMNH 621550, 621557) are present in the Rothschild Collection, but were not identified as such until later in a different hand (the former specimen is listed in the AMNH register as himalayana ssp., the latter erroneously as Ph. himalayana), it is by no means certain which form was meant by Rothschild, and the lack of a register for his collection prior to its accession at AMNH makes it impossible to determine this now. CAPTIVE INTERMEDIA OF UNKNOWN PROVENANCE The male ‘intermedia’ located in Austria are probably both hybrids between cyanocephala and finschii, as indicated by the uniformly narrow central rectrices with very long pale yellow tips and pale 47 shafts midway up the feathers, the yellow-olive band between the bluish nape collar and olive-green mantle, and the bright yellowish- green underparts. None of these features is consistent with himalayana as a parental species. Also, R. Low (in litt. 1997) thought the Turnersee bird was the same size as the female cyanocephala with which it was kept, which further supports finschii rather than himalayana as a parental species, as does this individu- al’s small bill. Both of the Austrian ‘intermedia’ have the front of the face bright rose-red, a feature incompatible with roseata being one of the parental species. KNOWN CYANOCEPHALA X HIMALAYANA Sedgmore’s captive hybrids of known parentage are virtually iden- tical in both plumage and measurements to AMNH intermedia. The slight mensural differences shown in Figure 3 are almost certainly due to measurement error, as the live birds had to be measured with great care to avoid injuring them, and thus they are probably slightly too large. Also, slight shrinkage of museum specimens is well- known. The identity of these hybrids with the type and only known series of intermedia cannot be ascribed to coincidence. CONCLUSIONS There is no evidence that intermedia is a valid species, and there is abundant circumstantial and unambiguous direct evidence that the AMNH series is comprised of hybrid himalayana x cyanocephala specimens. The discovery in the 1990s of more birds matching the phenotype of AMNH intermedia does not negate the above, particu- larly as they may well originate in captivity. In addition, Rothschild’s original series not only contained a juvenile himalayana, but also another hybrid of uncertain parentage; Sane’s ‘intermedia’ are from a third hybrid combination (krameri x cyanocephala), and the two cage birds in Austria are probably from a fourth (finschii x cyanocephala). Thus the literature refers entirely to birds putatively of four different hybrid combinations, and the supposed species Psittacula intermedia has no taxonomic standing. ACKNOWLEDGEMENTS. Special thanks go to Mr M. Sedgemore of Codsall, near Wolverhampton, U.K., for sending photos of his captive hybrids, allowing us personally to examine the birds, and sharing his notes with us. In addition we owe thanks to T. Arndt, R. Wirth, F. Pfeffer, R. Low; to R. P. Pr¥ys-Jones, M. P. Walters, and M. Adams, The Natural History Museum, Tring, U.K. (BMNH); C. Blake, M. LeCroy, P. Sweet, and M. N. Feinberg, American Museum of Natural History (AMNH); L. Bevier, Acad- emy of Natural Sciences of Philadelphia (ANSP); D. E. Willard, Field Museum of Natural History (FMNH); R. A. Paynter, Jr., Museum of Com- parative Zoology, Harvard University (MCZ); R. B. Payne and J. Hinshaw, University of Michigan Museum of Zoology (UMMZ); G. R. Graves and S. L. Olson, National Museum of Natural History (USNM); A. Rahmani, S. Unnithan, S. R. Sane, and A Aktar, Bombay Natural History Society (BNHS); R. Bhargava and H. S. A. Yahya, Aligarh Muslim University; T. P. and C. Inskipp, K. Kazmierczak, L. Critchley, Cumbria, U.K.; and P. J. K. McGowan. Funding for PCR’s travel to Bombay on two occasions was provided by the Research Opportunities Fund, National Museum of Natural History, and by British Airways, through M. Sitnik and T. Shille, Office of Biodiversity Programs/NOAHS, Smithsonian Institution. The manuscript was substantially improved by comments from G. R. Graves, R. P. Prys- Jones and M.D. Gottfried. 48 REFERENCES Ahmed, A. 1997. Live bird trade in northern India. A TRAFFIC India Trade Study. New Delhi: World Wide Fund for Nature — India. Ali, S. 1927. The Moghul emperors of India as naturalists and sportsmen. Part 2. Journal of the Bombay Natural History Society 32: 34-47. Ali, S. & Ripley, S.D. 1969. Handbook of the birds of India and Pakistan. Vol. 3. 1st ed. Bombay: Oxford University. Anon. 1997. Three Indian birds go extinct, says [sic] experts. Free Press J., 31 May 1997. Anon. 1998a. Birdwatcher finds world’s rarest parrot in UP. The Asian Age, 25 Feb. 1998. Anon. 1998b. Lost and found. New Scientist No. 2153 (26 September 1998): 21. Arndt, T. 1996. Lexicon of parrots. Bretten, Germany: Arndt Verlag. Beale, E. 1981. Rothschilds [sic] Parrakeet (hybrid) Psittacula intermedia. Parrot Society 15: 33. Bhargava, R. 1998. The rediscovery of the Intermediate Parakeet or the Rothschild’s Parakeet Psittacula intermedia from northern India. Abstract 106. Ostrich 69 (3 & 4): 396-397. Biswas, B. 1959. On the parakeet Psittacula intermedia (Rothschild) [| Aves: Psittacidae]. Journal of the Bombay Natural History Society 56: 558-562. Biswas, B. 1990. Taxonomic status of Psittacula intermedia (Rothschild). Journal of the Bombay Natural History Society 86: 448. Collar, N.J. 1997. Family Psittacidae (parrots). Jn J. del Hoyo, A. Elliott, and J. Sargatal, eds., Handbook of the birds of the world. Vol. 4. Sandgrouse to Cuckoos, pp. 280-477. Barcelona: Lynx Edicions. Collar, N.J., Crosby, M.J. & Stattersfield, A.J. 1994. Birds to watch 2. The world list of threatened birds. BirdLife Conservation Series No. 4. Cambridge, U.K.: BirdLife International. Dharmakumarsinjhi, R.S. 1955. Birds of Saurashtra, India. Dil Bahar, Bhavnagar, Saurashtra, India: Privately published. Finn, F. 1906. Garden and aviary birds of India. Calcutta: Thacker, Spink and Co. Forshaw, J.M. 1973. Parrots of the world. \st ed. New York: Doubleday and Co., Garden City. Fuchs, W. 1997. Der Rothschilds Edelsittich. Aktuelles aus der Vogelwelt 7/97: 224— 225. Graves, G.R. 1990. Systematics of the ‘“Green-throated Sunangels’ (Aves: Trochilidae): valid taxa or hybrids? Proceedings of the Biological Society of Washington 103(1): 6-25. Greene, W.T. 1884. Parrots in captivity. London: George Bell and Sons. Harrison, G.J. & Harrison, L.R. 1986. Clinical avian medicine and surgery, includ- ing aviculture. Philadelphia: W.B. Saunders Co. Hartley, E.A.H. 1907. Ruthless importation of foreign birds. Avicultural Magazine (2)5: 221-222. Hartert, E. 1924. Types of birds in the Tring Museum. B: Types in the General Collection, IV. Novitates Zoologicae 31: 112-134. Herrmann, K. 1994. Der Pflaumenkopfsittich Psittacula cyanocephala. Gefiederte Welt 118: 170-171. Howard, R. & Moore, A. 1991. A complete checklist of the birds of the world. 2nd ed. London: Academic Press. Hume, A.O. 1890. The nests and eggs of Indian birds. 2nd ed. London: R. H. Porter. Husain, K.Z. 1959. Is Psittacula intermedia (Rothschild) a valid species? Bulletin of the British Ornithologists’ Club 79: 89-92. Inskipp, C. &. Inskipp, T.P. 1995. Little-known Oriental bird: Intermediate Parakeet Psittacula intermedia. Oriental Bird Club Bulletin No. 22: 28-31. Inskipp, T.P., Lindsey, N. & Duckworth, W. 1996. An annotated checklist of the birds P.C. RASMUSSEN AND N.J. COLLAR of the Oriental region. Sandy, U.K.: Oriental Bird Club. Juniper, T. & Parr, M. 1998. Parrots: a guide to parrots of the world. New Haven: Yale University. Knox, A.G. & Walters, M.P. 1994. Extinct and endangered birds in the collection of The Natural History Museum. Tring, U.K: British Ornithologists’ Club Occasional Publications No. 1. Low, R. 1992. Parrots: their care and breeding. 3rd rev. ed. London: Blandford. Monroe Jr., B.L. & Sibley, C.G. 1993. A world checklist of birds. New Haven and London: Yale University. Mookerjee, A. 1997. Second life for parakeet. Indian Express, 30.11.97. Peters, J.L. 1937. Check-list of birds of the world, vol. 3. Cambridge, Massachusetts: Harvard University. Rasmussen, P.C. & Collar, N.J. 1998. Reanalysis of specimens and captive birds proves hybrid origin of Psittacula intermedia. Abstract 145. Ostrich 69 (3 & 4): 412. Rasmussen, P.C. & Collar, N.J. in press. Little-known Oriental non-bird. Oriental Bird Club Bulletin. Ripley, S.D. 1953. Consideration of the origin of the Indian avifauna. National Institute of Sciences of India 7: 269-275. Ripley, S.D. 1961. A synopsis of the birds of India and Pakistan together with those of Nepal, Sikkim, Bhutan and Ceylon. \st ed. Bombay: Bombay Natural History Society. Rothschild, W. 1895. On a new parrot. Novitates Zoologicae 2: 492. Rothschild, W. 1907. [Exhibition of specimens of Palaeornis intermedia]. Bulletin of the British Ornithologists’ Club 19: 49-50. Salvadori, T. 1907. Notes on the parrots. VII. /bis (9)1: 122-151. Sane, S.R. 1975. Rothschild’s Parrakeet. Avicultural Magazine 81: 238-239. Sane, S.R. 1977. The Rothschilds [sic] Parrakeet (Psittacula intermedia): an appeal. Parrot Society 11: 136-137. Sane, S.R., Kannan, P., Rajendran, C.G., Ingle, S.T. & Bhagwat, A.M. 1987. On the taxonomic status of Psittacula intermedia (Rothschild). Journal of the Bombay Natural History Society 83 (Centenary Suppl.): 127-154. Sedgemore, M. 1995. Slatey-headed Parakeet (Psittacula himalayana himalayana) (Lesson). Parrot Society 29: 187-188. Sibley, C.G. & Monroe, Jr, B.L. 1990. Distribution and taxonomy of birds of the world. New Haven: Yale University. Sinha, R.P.N. 1959. Our birds. Delhi: Ministry of Information and Broadcasting, Publications Division. Tavistock, the Marquess of 1932. 1932; the things that didn’t come off and the new arrivals. Avicultural Magazine (4)10: 248-256. Tavistock, the Marquess of 1933. Breeding notes for 1933. Avicultural Magazine (4)11: 379-387. Tavistock, the Marquess of 1934. 1934: the things that did come off. Avicultural Magazine (4)12: 255-261. Tavistock, the Marquess of 1935. Notes for 1935. Avicultural Magazine (4)13: 227— 229. Tavistock, the Marquis of 1936. Breeding notes for 1936. Avicultural Magazine (5)1: 258-262. Tavistock, Lord 1937. Breeding results for 1937. Avicultural Magazine (5)2: 334-350. Tavistock, the Marquess of 1938. Facts and figures. Avicultural Magazine (5)3: 83— 88. Them, P. 1998. Rothschilds Edelsittich erstmals im Freiland entdeckt? Papageien 1/ 98: 6. Walters, M.P. 1985. On the status of Psittacula intermedia (Rothschild). Journal of the Bombay Natural History Society 82: 197-199. Wirth, R. 1990. Systematik: Der Rothschild-Edelsittich Psittacula intermedia. Papageien 4/90: 126-127. Wolters, H.E. 1975. Die Vogelarten der Erde. Hamburg: Paul Parey. 49 S iss} = Ss g : Ss) E ) Q ee fe) e s Nn SI = fe) Z fe) 2 eB 8 611-78 Cc 1*9'6 8 °89-6'F L'0*6'S 8 081-96 Oierecel LY 8I-S0I CEFE VL L €9C-0°CC CIF VC L ‘6 18-L'69 QEtCOL 6 O'EL-E'EL 6 CFL'89 6 ‘1S9-6'€S 9°09 6 F'9S-0'LP 6 CFE CS 6 9 SP-S'LE SCHL 6 6 VE-S 87 CCF le 6 812-991 S1+S 61 6 L'e-0'1 6 0FEC 6 00-00 0'0+0°0 (sui 9'0FIT OL “€8I-S9l LISELI Ol 697-6 €7 O'1F9'ST OL €TI-OTI VOFL TI Ol VSI-VtI vos v1 6 ‘O'0I-S'9 Gast Ol VrI-V7Zl LOFV El Ol VvC-6 77 V'OF8°EC LOUDLY 6 CS TI-S'6/9 8 11-01 9'0FS'01/S 06 01 8 OL-L t/t °6:9-9'S O'1+€°9/6'0FS'°9 OL 86-7 L/ZI “L'0I-8'L L'0¥1'8/8'0¥9'°8 SI VSI-SEI/II C9I-S TL 90FS TI/0' IFC FI €€ OIZ-SHI/LI LCC-L OL DIF 8/S [+061 LI 9'6L-1'L9/11 “6'8L-0'0L VEF8 CLIO EFC PL 81 VIL-6 19/11 “8'7L4'€9 0'€*8°99/8'C+P'L9 81 LY9-S PS/ZI “6 €9-S'SS 9°C+9'6S/8'T+0'09 61 O'09-L'EP/El “P9S-L’ Lr € €*6 19/9 CFT CS 61 9'0S-H'8E/EI “8 LP-S'6E 8 (C¥6 Eb/P C+0 PP 61 LIt-lOe/el TVLE-L le CCFO VE/D IFT VE 61 “S'87-0'0C/EI “€ 9C-C'I7 OCF VC/9' IFL'E7 VE SVS tC Sa) CIF E/0 TFL € SI 0'0-00/EI “17-00 0°0*0°0/9'0*7'0 Cl Tr-O'0/EI ‘9° L-0'0 OIFL T/T CFS 61 ‘SSI-IPl/0e “6SI-LrI CCTV 8PL/E €+9'CS I 81 S2-€°C7/1T ‘O'L7-L 7 L'0¥0'PC/0' 1#C' VT rl “S TI-V01 1S'0*6 01 02 ‘VP I-CEL/IE 9 FI-O'EI €OF8'€1/7 0F8'El 07 “L'S-6'F/DE “8°S-h'b © OFT S/EOFT'S 0¢ 6CI-WII/IE VEI-L II VOFE TI/E OFS TI Ot VEC-E 6I/CE “PH HC-C 07 6 0FP 17/0 TFL 17 A/M Myosuif (L “8°01-0'6) 91 “S'ZI-LO1 (9:01 01) 9'°0F¥' TI (8 ‘0'6-7'9) LI *L'8-9'9 (6:0F7' L) L'0F8'L (L ‘9°6-¥'L) SZ ‘TOI-1'8 (L'0F¥'8) 9'0¥1'6 (6 ‘€9I-L7I) SC ‘LLI-8'71 (V1F9'P1) T1FE'ST (L ‘8°6I-1'ST) IE “€'¥2-WLI (VIFP LD STFLIZ (6 9°SL-8'L9) ST “E'88-6'EL (1h IL) S€F0'18 (6 °6'99-L’79) LZ “8° 18-0°89 (9: TFE'P9) O'EFS'EL (6 ‘S'6S-I'PS) 87 ‘TIL-€°6S (L'1#'9S) 9°T#L'S9 (6 “6'6F-8 Eh) LZ “T'79-1'TS (6: 1#8 9p) €°T#O'9S (6 ‘60-7 SE) LZ ‘1S’ Er (8° 1#0'8€) €TF89F (L “p 1€-0'97) 97 “8 EF-E'EE (O'CFL'87) Vth LE (8 °8°CZ-6 LI) LZ ‘O'0E-8' EZ (8'1*6'61) SIFI97 (6 ‘C 8-4’) €% TL-E'E (ph 1#8'S) OTFE'S (6 ‘C'I-0'0) St ‘L'I-0'0 (F070) S070 (8 ‘I I-0'0) #7 ‘0°S-0'0 (F'0FT'0) S1FTT (6 ‘O91-OS1) 8E “L8I-VS1 (CEFT'SST) O'9FT LOI (¢ ‘0'87-7'ET) OE ‘8°8Z-6'EZ (6° TF1°S7) O'1#h'STZ CVG GOD 86 (LOFTTD S'0*601 (Gy SIEVE)! Gy G@SiSsicl (9 0F7'FL) SOFT FTI (8 ‘€'9-0'S) 8E ‘T9-8'P (F'OFL'S) p'OFS'S (6 ‘S'EI-8' 11) Gr ‘8'EI-E 11 (L:0F8°T1) 9°0F8°T1 (L ‘TV €7-8°61) Ih ‘O'P7-L'61 (TIFT 17) 6 0FL 17 pupavjpuny q (ainjeurut) I[Npe [CC HS 00 EA ai! c0*9'T LSI-8¢1 POSTS O'ET hetss 0. VOF8' II Ligt=G sl L0*9'¢e1 OCI-8 TIT L0*6 TI LCC-8 IT 90FT IC suoutsads ours € 8vI-vrl [cre 91 ST bl-€7I LOFC EL ¢ TEI-9TI € 0*8 Cl CS 0C-L'61 rm Oa OY splgsy $,d10WISpPag CS TI-6'6 O1+0TI ¢ € 01-89 V1+08 SW RSi (6'8) 7 1+8°8 y €9I-E v1 (6+1) 6 0F6'F1 C 66I-1'81 (p81) LOFT 6! CS 6 LLVTL (1'L9) © CFL GL ¢ “L'69-S'99 (Z'19) T 1#8°L9 ¢ “1'09-0°9¢ (L'pS) L1F8°8S ¢ TCS-S' Lh (8°Sp) LIFEOS CC Ere Ib (€°LE) O LF rtp ¢ 9rE-S TE (6°97) TIFIEE C ‘lr7-L 91 (E81) O'€FS'IZ c V'9-0'r (O'€) O1FH'S b 00 ($0) 0'0*0'0 61-00 (0'0) 0'1+6'0 ¢ ‘9SI-CSI (O'CH1) 9 IFS'EST ¢ 0'97-0'E7 (LT) VIFE GT v TTI-8'6 (O'0L) 01601 ¢ LEI-O'El (VTL) TOFE EL Cc 6rer (8°€) TOF9'D C SCIFI (E11) OFS TT ¢ ‘90C-S 61 (0°81) F'0*0'07 (1'€'0861 HNN) vipauatul J HNWV Ol T67L JOFES Il P9-€¢ vOF8'S Ol 867L LOFI'8 CL SHI-6'I1 LOFC El €€ OCS HI IFT SI 8€ “LTL-S'09 9°€+0'99 BE “8S9-CHS CeF9'6S LE €09-C'8r CeEFO'ES 8E O1S-C IP LCFS Sp LE O'PP-L EE OCFOLE LENGE SZOIEG O'C¥8°8C 9E €S7-0'ST crv 6l LE 98-0'€ S179 LeeLiGa0i0) 6 0FL0 1€ L700 6 0F6°0 Or “S StI-67I SrFe ET SE “LC V6 6 0FS IT €7 ‘9OII-S'8 8'0*9'6 8€ “LT7I-6 01 C0F8 TT 6€ SS-L'E VOFS TY 8€ T1I-S6 yOre OL ce “TOC-E SI Wie /E\8 DIDASOA qd GE WIS O 0' 196 YIPIM Ed 97 ‘SL-EP 8':0F7'9 WIPIM 7d GZ ‘86-7 L yojou qe 9'0FP'8 YPM Td pt ‘€SI-6'8 YIpIM pIFlel wmurnxeur [q I€ PIZ-S's1 aoursip dy C1081 -yojou [dq 97 ‘S'OL-6'LS [reyoys 9'TF9'P9 Old 87 ‘7 S9-S PS [epoys b TFC 6S 6d 6c ‘ELS-L 8b [repoys UTFS 7S 8d 87 ‘6 6b-1 th [reptoys 6 IFESP Ld O€ ‘S'Ih-9'rE [reyioys 8 1F9'LE 9d O€ “STE-6'ST [repoys C 1¥6 87 Sd I€ ‘77191 Treproys CIFL 61 td 6¢ ‘T9-1'7Z [reyztoys 6 OFE'F €d 87 ‘0'I-0'0 [repsoys 7000 Zd 6c ‘I'+-0'0 [reproys € 19] Id br OSI-971 yySug] O'S¥S' SET BUIA\ SHOINAY GNV ONIN Or ‘L'b7-S 07 YIpIa L1¥9'% 11S €Z ‘L'11-9'8 pia 6 0F8'6 a[qipuey OF ‘8 ZI-TII yIpIM €'0F0'T1 PI[IXeI Eb ‘8'S-9'E YpIM vOFL'P [eLeN Ob ‘b11-8'6 qysIoy vOFL OI BIpEXe Or ‘661-91 yysue] 8'0F8'LI uauynD davdH pipydaz0uvio gq dINSedI[ “UOTIOAS SpOUIAY] Ul USATS s]UIUTOINSPOU jo suondussaq ‘sususads nipawuiaiul HNWY JO sonsneis Areuruins SUIMOT[OJ Sasayjuared UI UDAIS [°€'O861 HNING 10} sjuowiainseayy] “(uv ‘odue1) UONRIASp plepur}s F USUI Sv UDAIS Ore [TY “awnsy d pure ‘Ajaansadsau ‘suonejndod usajsva pur wiajsom wo suoumtoads myosuyf q ‘sasayjuesed ut sapruaant way) ‘pupdpjpuiy g ynpe ‘(7 = UV) UONIET[OD S,auRg Woy SudUTIDads ‘(a[eUIay 9UO SuIpNyoUt) UON}DA][O9 S,a1OWASpas url spugdy sAq] ‘VIpauiajul J HNINY ‘Vivasos q ‘vjpydasouvd9 vjnovyIsg aeu ynpe JO (wu ‘YSIIY = Y ‘IPPIM = M “U)Sug] = |) sjUoWoINsvoul Jo sonsNeis AeuUINg xipueddy P.C. RASMUSSEN AND N.J. COLLAR 50 QE B-SL 6 E8-8'S/L ‘*S'8-1'S €0F6L 8: 0FP'9/0 1+9°9 8 ‘O'11-0'8 SI ‘Pp 6-9'L/ZI “b'6-6'L 0'1*6'6 S OFF’ 8/P 0FS'8 8 0'S=6'E OL ‘Tt-1'€/S ‘0'b-6'7 SOFT € 0F9 €/P OF PE OL ‘L'81-8'91 91 6 SI-E TST ‘O'LI-CTI 9°0F9'LI 6 0FL'b/6 0Ft'SI 8 ‘P'b-0'7 bl ‘OS-PI/IZ ‘O'EL-E'LI L'0F8'7 6 6F8'87/L EFT Th 8 ‘1'b-0'0 EI ‘O€I-06/ET “0'S9T-0'S8 (6 4 OCT FLLOL/9'6IFS STI L‘6O€I-V Il 8 ‘Y'8-T'L/S ‘L'6-9'9 6 0FETI v0F8'L/T 1F78 6 ‘871-66 ROU aEOe BSa kL O1F8'TT SOFT L/L'OFL'L 6 ‘€TI-L'8 6 ‘6 LV 9b “P8-TL TIFT TI SOFT L/S 0F9'L 6 ‘T'€1-9'6 SI ‘O'6-€'S/8I ‘8'8-L'b VIFL01 6 0FS'9/T 1#8°S e SI ‘19-8 P/6I ‘9°9-€'€ = p OFS S/LOFI'S 9 ‘P8-9°9 El ‘p'S-6E/ST “€'9-O'P SOFL'L v0F9 p/LOFTS 6 ‘7C-LI OL ‘SEI-€ 8/61 ‘O'SI-8L 8 1761 SIFSOVETFUTI 6 ‘OC-SI OL ‘€OI-1'9/6l ‘€'€I-V'L 6 1F9°L1 O'1Fb'8L/9' 196 6nge-L1 LI ‘T91-9'6/91 ‘181-901 EVFE TT 8 1F9°TI/p TF6 El 6‘Le-I€ —_ LT ‘0'69-L PI /PE ‘O'SL-L'FI O'PF6 LE P S18 11/8 €1Fh'0S 8 ‘OL-€T SI ‘OpI-€6/p ‘IZI-O1I € OCF16S 8 PIF8 91 V/L PFT PII 8 “SS7-S9I 91 ‘O9LI-ELT/61 “P81-86 OOF POI O'VIFL CEL/8°7#OLET 8 ‘pST-C7Z OL ‘ILZ-1IZ/TI ‘TOE-TET vOlFr LEZ L'61F9 0S7/6'8IFL P97 Serial Ol ‘O'II-Z 01/8 ‘8'ZI-V01 SOFTEI 9'0FT 01/8 OFE II 8S TZI-€01 6 ‘8 7I-V' 01/9 ‘TZI-T TI 80FS II 8 OFF IT/€0F9 TI MauUD1y A/M Hyosuyf (8 ‘7 L-€'9) E% “V'L-€'9 (€0FL'9) €0FO'L (8 ‘S'6-1'L) 67 ‘S'‘0I-7'8 (L'0¥9'8) S 076 (v ‘T'¥-0'€) IZ “€'F-1'€ (v'0FS'€) €0FV'E (9 ‘L'SI-8'r1) TE “P'8I-0'E1 (V'OFP'SL) VIFE'ST 6£ O'SS-0'ST GOFO'8E 8€ O'SOI-0'8S LOI+¥ 18 (8 ‘0'0I-0'L) OI ‘8'0I-8'8 (8'0*S'8) L 0FP'6 (L‘9'8-S°9) FI ‘9°6-8°L (L'0¥9'L) 9'0¥9'8 (p T6-L'L) OTS O1-O'L (9'0¥S°8) 6'0¥6'8 (C ‘L'6-1°8) Lt “6 6-T'L (8'0FL'8) L0FV'8 be O'OI-O'L L'OFS'8 O€ ‘811-08 60*L6 (8 ‘v'€I-7'6) TE “9'LI-0'01 (VIFEL1) O'CFS'EL (9 ‘1'OI-1'9) €€ ‘T'9I-O'L (S 1¥9°8) 6 1FEOI (9 ‘T'ZI-S‘OD) LE “T'It-0°6 (QOFT II) SSFEST (9 ‘707-6 V1) 8E O'E9-O'TE (6: 1*6'91) 6 9F1 br OT ‘TII-IL C 6F9' 8 Or O8I-SII CCIFI 9FI Lt 19C-LOT O'SIFL VEC (S *‘L7I-O'1D) 91 LvI-L' 11 (9'0*6'I1) 8°0*8 CI (p ‘TTI-L'8) 91 “TrI-C OL (C1+¥ 01) O'1FS 71 DUDADIDULIY (QInyeUUT) J[Npe OL (OWE Lc¥0'S1 6CI-6 Cl 0'0*6 Cl OLI-S'SI Pease) (a [tet] 9LI OCI VIT suawtoads ours b O0StOLT Tel+l Ce b 0'79-00€ S CIF 8P b 0L-09 S0FCT9 € 01-08 LT 1+t6 7 O'0S-0'cE OCIFS IV b 99I-StI 8°8+0'SST [ry ‘911-96] v €L7-9VT 9 T1*0°C9C spuighy $,d10WISpoIg 6 SESH (0'9) L:0*0'8 € 88-€8 (SL) 70F9'8 € 1-67 L0F0'€ € 6vI-L v1 (9°S1) 1:08 'r1 SW GaaGC (S11) PTFE STZ t ‘Tpb-16€ (6'€£) SEFO' Ih ¢‘Oll-8 O IFES CP 6-9'L L0F78 C‘L6-VL 1 1#78 ¢ ‘0'6-0'L (9°9) L'0F8'L € ¢lL79 (S'S) S0FO'L eo FREL (19) TI1FE8 ¢ TVEI-C ll (701) 8'0F6 TI C‘SOI-LL (€°6) O1*6'8 ¢ ‘7 EI-T 01 (68) TIFL II ¢ ‘O'rh-8'81 (1°07) O';01F7'SE ¢ ‘8eI-901 (ZOI) V€1F0O°L7Z1 (+781) +007 “+OLT +ZS1 C*LTI-€ ll C OFT CI ¢ 6 7I-001 (5 av, |1)! (1°€'0861 HNN) Dipausajul q@ HNNV Cc ‘OL-19 v0+9°9 €l C6-OL 9'0*6'L Il 1t-L7 VOFCE cE 6 SI-O'TI 6 0F9'E1 9€ ‘O'EI-0'S OCFLL le “LS7-€L OVFS OT Ol ‘76-99 SO0FEL OL €8-8'S L:0*0'L OL 18-7 8079 87S L-E'S COFCO TE 89-1 P 9'0FTS 0¢ L'9-6'€ DOFES LE ‘Vel-c9 eC IFe Ol 9€ ‘1'SI-0'9 S1FS8 9€ VEI-L9 CIFL6 CES Sh-8 SI SSFE VC LI VILL €v BLFB CS LE “I€1-68 18+9'€OT €€ 6CC-I¥1 VSI+I 991 OL ‘O'11-06 9°0*C 01 Ol v0I-S'8 C OFE 6 DIDISOL I €1 OL-G'S COFI9 €C O6-1TL C 0F6'L £7 SP VT VOFEE €€ 19l-V7l 60FC V1 Iv ‘CET-E'S SvFl El 9€ “S$ 8E-0 TI LOFE6C vo S6-EV C1+9'L St £6-9'S 60F0'L Sc LOI-wS O'1+8°9 €€ L8-1S L'0+0'9 0€ “6 L-6'F L'0*0°9 Lt TOI-6' 4 CIFLL ce PrI-69 L1+9'6 CE CIE Tcre Ll 0€ LOI-C9 c 1+9°8 tr L'6r-0'01 V6FS VE €t 60I-LY 6 TI+9L'88 LE 61-16 eLIFl SII Ce “TET99I 8 SIFT C07 91 071-66 9°00 TT v7 811-88 LO0*V OL pjpydas0uvio q y3ug] Me[SPuTH y3ue] MPTSPTIAL UIPIM sns.iey WINUITUTy wsue] snsiey, aI TGNIH yysue] dn 7y wsugy dn 1y IpIM Ca WIpIM ra WPIM ca WPIM Ta (pI jewrxoid ,.y pi Teasip Ta aoudIAyFIp 9a-Sa oUdIOTFIP cara QOUudIOTIIP pa-ca 2oudIOTFIpP eu-ca JoudIATJIpP ca-la ysug] ca ysugy 1a SSOMLOTA (PIM Cd WIPIM td amnseayy panuyuo) xipusddy Bull. nat. Hist. Mus. Lond. (Zool.) 65(1): 51-72 Issued 24 June 1999 A review of the genus Bargmannia Totton, 1954 (Siphonophorae, Physonecta, Pyrostephidae) PR.PUGH xX (SIEC)\. SouthaMpton Oceanography Centre, Empress Dock, Southampton, Hants, SO14 3ZH, UK CONTENTS ITEICYCLC TOM reeks se cesorsoses=cexaceccnrstdectscsscaueacrscesageesscsancvercesuscsssanasoe> Family Pyrostephidae Moser, 1925 ..............:.::.scesccssecserseeseee Genus Bargmannia Totton, 1954 .....eeceeesessetseeeeseeneeeeeeee Bargmannia elongata totton, 1954 .......:cccccsseesesseseereees BOREMANMIG GMOENG SP. TWOV: :...00:s-tn-+vasensedoxcegavedostvss-n>s Bargmannia lata Mapstone, 1998 ...........ccsesescenseneeees (BAF 2IIANIIGIEISAS | SP MONS «ose cae ctaca-Sedececaedscnnrsda-eadecersnode TAS GIS SIO faced ee execs ceed o waaawns damnsen te dase oavadenns itches snensuevesastet eonee PN GRAIO WE SELCIUS ceceasd. catenins saat ernesntaveecsnevevns qaracsatnasebonsacadepc PR GECHOMGE Sa sase cre arn zest ssa cc suendnaae ta evavanserih cise sone sesatquses veaetssseecvar> Sor oS nt oes ate hee estan ea a aac euc Recalls ua caca ths dus lyeeten 00 NoPE rowan E este nas eee vance dated ec cet str stresccesstssaostee 00 Be ae ee cea reece nn acta octa cn ds cadaver nisi atone -veaicen dns 00 Waeannanild Wane es ex we uaa uae Tuas savas eens Ws pccnsdaceasceccatncstevdsceset suesaanasacver 00 Lecce Cece REREAD OA ae CACC CR ETE DP REE ee ae 00 So esacanapie vances Coren iatan ees sartusccvansustvagsvesiipecundunesiuncvesrsearaestecreereonrrcae 00 des geetavtnastevepoetacone rover strcrorcen-trevttensadecnsiccrsaereseg nanny sasteeenntnasttterets 00 SYNOPSIS. Two new species of the physonect siphonophore genus Bargmannia are described, and B. elongata Totton (1954) and B. lata (Mapstone, 1998) are redescribed. The status of the genus and its retention within the family Pyrostephidae are discussed. INTRODUCTION Totton (1954) established the genus Bargmannia, named after his colleague Dr Helene Bargmann, to include the single species, B. elongata; nectophores of which he had found in thirteen Discovery samples, plus one from the Michael Sars Expedition (Leloup, 1955). Because the structure of the nectophores differed so markedly from those of all other known physonect siphonophores, Totton did not give a detailed description of them; remarking only that the lateral radial canals on the nectosac had straight courses. Totton (1965) later noted that, although B. elongata was one of the most easily recognised siphonophore species, nothing more had been published on it since its original description. In fact, by the time of publication of Totton’s monograph, only Alvarifio (1963, 1964) had mentioned it; and then only in lists of siphonophore species collected in the western Pacific. Totton included a brief description of a further specimen collected at Discovery St 4246 (37°50'N, 13°22'W), re- marking on the orange coloration of the stem. Since that time several authors have reportedly identified this species from various collections. However, examination of both Totton’s material and that from more recent Discovery collections (Mackie, Pugh & Purcell, 1987) appeared to indicate that Totton’s (1954, 1965) illustrations of Bargmannia elongata could be referred to two species, and that his material also included a third species. However, it was not until submersibles collected speci- mens of this genus that this contention could be proved beyond doubt. Study of this submersible material, together with that from the Discovery collections, shows that there are at least four spe- cies that may be referred to the genus Bargmannia. The second species that Totton illustrated under the name B. elongata has recently been described under the name B. lata. More detailed descriptions of both these species, together with descriptions of © The Natural History Museum, 1999 two previously undescribed species, are given herein. Totton (1954) did not refer the genus Bargmannia to any of the physonect families, although his description appears at the end of a section dealing with various species of the family Agalmatidae. Later, Totton (1965) placed the genus in the family Pyrostephidae, which previously had been monotypic for the species Pyrostephos vanhoeffeni Moser, 1925. However, his diagnosis of that family applied only to the genus Pyrostephos, and included such features as marked bends in the dorsal and lateral radial canals on the nectosac of the nectophores. This character alone would exclude the genus Bargmannia. Since then, Stepanjants (1967) placed the genus in the catch-all family Agalmatidae, whereas Daniel (1974) retained it within the family Pyrostephidae. Now that intact specimens have been collected by submersibles it is possible to review the systematic position of the genus Bargmannia. It is concluded that, for the presentat least, it should be retained within the family Pyrostephidae, the diagnosis of which is adjusted accordingly. Family PYROSTEPHIDAE Moser, 1925 DIAGNOSIS. Long-stemmed physonect siphonophores. Necto- phores with large triangular thrust block; with lateral wedge-shaped processes reduced or absent. With apico-, infra- and vertical (meso-) lateral ridges; apico-laterals divide above ostial level. Adaxial wall of nectosac lacking musculature; deeply hollowed. Long pallial canal; short pedicular canal, giving rise, on nectosac, to only dorsal and ventral radial canals; lateral radial canals arise separately from dorsal. Dorsal and lateral radial canals either looped or straight. Tentillum with straight (or twisted, but not tightly coiled) cnidoband; lacking an involucrum; with terminal filament. Dactylozooids either absent or modified to form peculiar palpacle-less oleocysts. Indi- vidual specimens of single sex (dioecious), with gonophores budded 52 one from another to form a small gonodendron; female gonophores contain two or more eggs. REMARKS. In Pyrostephos vanhoeffeni, the triangular thrust block is best seen on smaller nectophores. On larger, preserved ones it is bent up dorsally (see also Discussion section). Genus BARGMANNIA Totton, 1954 DIAGNOSIS. Pyrostephids with distinctive elongate nectophores. Mature nectophores with large, triangular thrust block; without apical wedge-shaped processes; with extensive ventro-lateral wings. Basic ridge pattern may be augmented by additional ridges branch- ing from apico-laterals. Nectosac basically cylindrical; dorsal and ventral radial canals straight; lateral radial canals arise separately, but in close proximity, from the dorsal canal. Pheumatophore with- out apical pore. Siphosome diffuse; devoid of fully formed dactylozooids. Bracts specifically variable in shape. Each cormidium; with simple tenta- cle-like structure attached to stem midway between successive gastrozooids; with single gonodendron; with four bud-like struc- tures (?vestigial dactylozooids) with sexually dimorphic arrangement. Second tentacle and fifth bud occasionally present proximal to a gastrozooid. REMARKS. The meso-lateral ridges on the nectophores, as referred to in the above diagnosis, are homologous with the vertical lateral P.R. PUGH ridges, as defined by Pugh and Youngbluth (1988), found on the nectophores of certain agalmatid species. In these latter species these ridges run vertically, or slightly obliquely, between the apico- and infra-lateral ridges, although they may not reach the latter. However, in Bargmannia spp. their arrangement is strikingly differ- ent in that they have a very oblique course; and it is the infra-lateral ridges that may or may not join them basally. For these reasons the term meso-lateral ridges will be used herein. In contrast, the outer of the two branches of the apico-lateral ridges should not be compared with the lateral ridges of agalmatid species, as defined by Pugh and Youngbluth (1988). They more closely resemble the near-ostial branching of the apico-laterals in agalmatid species such as Lychnagalma utricularia (Claus, 1879) (see Pugh & Harbison, 1986) and Halistemma transliratum Pugh & Youngbluth, 1988, which also possess normal lateral ridges. The long, median canal that runs up the thrust block (see Figure 2), just below its ventral surface, has been variously referred to as a pallial (e.g. Daniel, 1974) or a pedicular canal (e.g. Daniel, 1985). In accord with the definitions given by Totton (1965) here the canal will be referred to as the pallial canal; and the short canal, passing through the mesogloea from the stem to the nectosac, the pedicular canal. Recently, it has been brought to my attention (Dr S. Haddock, personal communication) that the generic name Bargmannia was used by Herre (1955) in a description of a genus of an extinct sala- mander. Bargmannia Totton, 1954 clearly has priority of publication. Fig. 1 Bargmannia elongata. A. Photograph (reproduced by kind permission of Larry Madin, WHOI) of live specimen collected during Alvin Dive 961. B. Photograph (reproduced by kind permission of Steve Haddock, UCSB) of live specimen collected during JSL I Dive 2673. Nectosomal length c. 9 cm. BARGMANNIA REVISION Bargmannia elongata Totton, 1954 (Figures 1—S) Bargmannia elongata Totton 1954 (Text-Figure 28 A-—D only); Totton 1965 (Figure 45, A—D only); Kirkpatrick & Pugh, 1984: Figure 11. HOLOTYPE. BMNH 1952.11.19.7, designated by Totton (1954): one nectophore from Discovery II St. 699; 14° 27.3'N, 30° 02.3'W; 14-v-1931; 0-370m. The specimen was figured by Totton (1954, text-Figure 28 C, D; 1965, Figure 45 C, D). PARATYPES. As designated by Totton (1954): eighteen nectophores from the same sample as the holotype. BMNH 1952.11.19.8—25. MATERIAL EXAMINED. The holotype and paratype material have been re-examined in order to establish to which of the presently recognised Bargmannia spp. the name elongata should be applied. Totton’s (1954) other material also has been re-examined and, although all the material is in poor condition, it appears that only the nectophores from two other Discovery stations belong to this spe- cies. These are St. 681 (21°13'S, 29°55.25'W; 1—v—1931) where a TYFYV net was fished over a depth range 1500—1000m; and St. 107 (43°03'S, 17°03'E; 4—xi-1926) where the net used was a N450 and the depth range was 850—950m. The nectophore from the former of these stations was figured by Totton (1954, text-Figure 28 A, B; 1965, Figure 45 A, B). The other nectophore, from Discovery St. 1769, also illustrated in the same figures (E, F) does not belong to B. elongata, but to B. lata. Several nectophores of this species have been found in more recent Discovery collections, as is discussed below. However, the major part of the redescription will be based on two specimens collected by DSRV Alvin off San Diego, California, U.S.A. in 1979, during Dives 961 (32°14'N 117°22'W; 5-ix-1979; water depth 833m) and 966 (33°04'N 118°16'W; 8-ix-1979; water depth 747m). The Alvin Dive 961 specimen, preserved in Steedman’s solution, has been deposited in The Natural History Museum London (BMNH 1998.2163). The exact depths of collection for both Alvin specimens were not recorded. DIAGNOsIS. Nectophores with central thrust block broadly rounded or obliquely truncate apically. Pair of short ridges, directed toward mid-line, branch from apico-laterals where latter bend out sharply at a right angle. Outer branches of apico-laterals end, basally, on, or just apical to, enlarged processes lateral to ostium. In preserved specimens ostium opens dorso-basally and nectosac, with appar- ently dense musculature, has distinct dorso-ventral undulations. The ratio of the overall length of the nectophore to the length of the nectosac averaged 1.31. Delicate, foliaceous bracts; typically with patches of ectodermal cells on distal half of dorsal surface. DESCRIPTION. A photograph of the living specimen collected during Alvin dive 961 is shown in Figure 1A. By the time it was taken, in a tank on board the mother ship, several nectophores had become detached and the siphosomal stem had contracted. A second living specimen, collected during Johnson-Sea-Link (JSL) I Dive 2673 (27°02.7'N, 85°01.5'W, depth 780m), is shown in Figure 1B. PNEUMATOPHORE. The pneumatophore measured c. 2.2 mm in length and 1 mm in width, but was distorted and ruptured. No pigmentation was apparent. In theAl/vin dive 961 specimen, the main gas cavity, the pneumatosaccus (height 1.8 mm), was separated from the small gas secreting region, the pneumadenia, by a narrow collar. Below the pneumatophore was a long stalk, up to 7.6 mm in length. Immediately above the nectosome, this stalk narrowed and was flattened to form a hinge-like structure, which could facilitate the 513) use of the pneumatophore as a means of orientating the animal. NECTOPHORE (Figures 2—3). The nectophores had a biserial, stag- gered arrangement down the nectosome (Figure 1). Forty two nectophores were found with the Alvin dive 961 specimen, though many were small or immature; and 26, mostly mature ones, were found with the A/vin dive 966 specimen. The mean dimensions, for the fully developed nectophores of each specimen, were:- length: 21.29 + 0.93 mm and 16.49 + 0.75 mm; width: 9.58 + 0.56 mm and 7.40 + 0.32 mm; and the ratios of total length of the nectophore to the length of the nectosac were 1.29 + 0.02 and 1.34 + 0.04, respectively. For net collected nectophores, damage and distortion by preserva- tion, particularly to their basal halves, made it difficult to assess this ratio accurately. The nectophores of the dive 966 specimen were noticeably smaller than those from dive 961 but, as will be seen in the description of the following species, the size range of the nectophores can vary greatly between individual specimens. In general, the thrust block was roundly, and often slightly asymmetrically, truncate (Figure 2A, tb; 2B), although for a few of the nectophores of the smaller specimen it was distinctly tapered. The latter was also apparent on several net collected nectophores where the apex of the thrust block was drawn out to form a small digitiform process that could be folded over ventrally. Fig. 2 Bargmannia elongata. A. Upper, B. lower, and C. lateral views of mature nectophore. Scale bar = 5 mm. bi, bo: inner and outer branches of apico-lateral ridge; mp: mouth-plate; n: nectosac; o: ostium; pc: pallial canal; pedc: pedicular canal; ral, ril, rml: apico-, infra- and meso- lateral ridges; sb: side branch; tb: thrust block. 54 The basic Bargmannia ridge pattern is supplemented by a pair of short ridges (Figures 2A, sb; 3 A) that branch from the apico-laterals (Figure 2A, ral) at the point where the latter bend sharply, through 90°, away from the mid-line. This sharp bend typically can be seen in less well preserved specimens and is characteristic for this species. The side branches are directed, for a short distance, toward the deep median furrow. In many specimens, particularly net col- lected material, they were difficult to discern but often can be seen after staining. Basally, the inner branch of each apico-lateral ridge curves inwards and then down to reach the ostium (Figure 2 A, bi), except for immature nectophores (Figure 3A) where it ends slightly above that level. Each outer branch (Figures 2A, bo; 3C) typically terminates on or just above one of the small, but prominent, lateral processes on either side of the ostium. Basal extensions of the meso-lateral ridges form the baso-lateral margins of the bilobed mouth-plate (Figure 2B, mp; 2C), each lobe being thickened ventrally, particularly toward the mid-line. Basally, the two lobes typically overlap and unite, in the mid-line, at about half the height of the mouth-plate (Figure 3C). The lower nerve tract (see Mackie, 1964), which can be traced down the nectophore, beneath its ventral surface in the mid-line, recurves at this point and P.R. PUGH continues obliquely to the baso-ventral margin of the ostium (Figure 3C). In immature nectophores the mouth-plate is not thickened and has a U-shaped emargination in the mid-line (Figure 3A, B) whichis deepest in the youngest nectophores. Above the mouth-plate, the basal extensions of the meso-lateral ridges curve round toward the mid-line, on the ventral surface of the nectophore (Figure 2 B), before looping back outwards as the meso- laterals proper (Figure 2C, rmil). The infra-laterals are weakly defined in the region where they divide from the meso-laterals, and in younger nectophores clearly terminate before reaching the latter (Figure 3 B). The meso-laterals curve up, obliquely, across the lateral surface to reach the junction with the other main ridges at a level slightly below the apex of the nectosac (Figure 2C). The connection with the other ridges is weak, and often the meso-laterals appear to end slightly below the junction, as was found for younger nectophores (Figure 3 A). The infra-laterals (Figure 2B, ril) demarcate the ventral margins of the thickened walls of the more basal part of the ventro-lateral wings. In lateral view these wings are slightly emarginate in outline. Apical to where the infra-laterals curve up to join the other ridges, the wings remain well developed and are thickened with mesogloea. Fig. 3. Bargmannia elongata. A. Upper and B. lower views of young nectophore; C. detail of ostial region of mature nectophore. Scale bar = 1 mm. BARGMANNIA REVISION 55 Fig. 4 Bracts of Bargmannia elongata. Scale bar = 1 mm. This thickening diminishes in the region of the thrust block, but there is still a shallow median gutter that enfolds the nectosomal stem in the region of attachment of the nectophore (Figure 2 B). In the preserved nectophores, the nectosac is a dorso-ventrally undulating tube (Figure 2B, n; 2C), with prominent dorso-lateral extensions in the mid region, and ventro-lateral ones both apically and basally. However, this arrangement is not apparent in the nectophores of the living animal (Figure 1). The nectosac is broadest at about two-thirds its length, narrowing slightly towards its apex. It has a distinct apical emargination; U-shaped in the younger nectophores (Figure 3 A). Typically, the ventral, adaxial region towards the apex of the nectosac is distinctly undercut and, from a level just basal to the point of insertion of the pedicular canal, its wall is devoid of musculature (Figure 2 B). The musculature of the remainder of the nectosac appears well developed and gives it a distinctly opaque appearance. The ostium, in the preserved material, opens onto the dorso-basal (abaxial) surface (Figure 2C, 0) and is roughly rhomboidal in shape. However, this probably is distortion due to preservation (see Figure 1). In the Alvin specimens it has a large velum, with a relatively small central opening, but in net collected material often the velum is destroyed. The lateral walls of the ostium extend out to form lateral processes (Figure 3C) that, typically, are covered by patches of ectodermal cells of varied size. Further such patches are present on the ventral margin of the velum, but not on the dorsal margin, except for the youngest nectophores. Some, if not all, of these cells probably produce bioluminescent material since this has been found to be the case in another Bargmannia spp. (Dr S. Haddock, personal communication). The long pallial canal (Figure 2B, pc) extends up into the median thrust block, where it ends with a short dorsal inflection into the 56 P.R. PUGH Fig.5 Bargmannia elongata. A. Young tentilla with stenoteles (st) at proximal end of cnidoband (magn. 50x); B. Part of siphosome showing three siphosomal tentacles (¢) and several buds (b) (magn. 16x); C. Male gonophores (magn. 30x). mesogloea. At its base the lower nerve tract can be seen to leave its proximity and to continue down beneath the ventral surface of the nectophore to reach the ostium. The short pedicular canal (Figure 2B, pedc) extends through the mesogloea, from the base of the pallial canal, to the nectosac. There it gives rise to only the dorsal and ventral radial canals. The lateral canals arise separately, but in close proximity to each other, from the dorsal canal, and initially are directed toward the apico-lateral margins of the nectosac. They then continue down the lateral margins of the nectosac and, although their courses show undulations (Figures 1C, 2C), they are merely follow- ing the dorso-ventral undulations in the nectosac itself; the latter being a preservation artefact. The youngest nectophores (Figure 3A, B) typically show the absence of a median thrust block, and the apico-lateral margins are demarcated by the apico- and infra-lateral ridges. The basal portions of the apico-lateral ridges are particularly well marked, and the inner branches are distinctly broadened, often appearing almost bifurcate at their basal ends, which lie just above the ostium (Figure 3A). There are two short tracts of cells extending out from the lateral processes of the ostium just ventral to the outer branch of the apico- lateral ridges. These could not be discerned in the mature nectophores. BRACT (Figure 4). The bracts are extremely delicate, foliaceous structures, the largest of which measures 9 mm in length. The dorsal surface is slightly convex, the ventral one slightly concave. For many the proximal region is bent up dorsally, or one side is folded over the other resulting in a distinct asymmetry. The bracteal canal extends, approximately in the mid-line and in close proximity to the ventral wall, to about four-fifths the length of the bract. The distal end of the bract is slightly truncate and bears two lateral processes, which vary in shape from merely rounded corners to distinct teeth. The region between them usually is roundly pointed. Additional processes may be present on the lateral margins of the bract. Again these can form distinct teeth, but quite often are indiscernible. The maximum number of lateral processes found was two on one side, and one on the other. The distal half of the dorsal side of the bract is dotted with distinctive patches of small round ectodermal cells. These patches are densely packed on the smallest bracts; but more spread out on the larger ones, where some patches have been lost by abrasion. These cells probably are sites of bioluminescence. GASTROZOOID AND TENTACLE. The larger gastrozooids in the Alvin material measured up to 10 mm in length. They are brown in colour, in their preserved state, and are comprised of a short, narrow basigaster, to the base of which the tentacle is attached; a large, expanded stomach, the inside of which is covered with thickened patches of endodermal cells; and a long proboscis, with longitudinal endodermal hepatic stripes. Several younger, smaller gastrozooids BARGMANNIA REVISION also are present, which are largely colourless and transparent, with only small patches of endodermal cells in the stomach region. No mature tentilla remained with the specimens. The immature tentilla (Figure SA) conformed to the basic Bargmanniadesign, with the cnidoband ranging from being straight to curved or slightly twisted. There was a maximum of only six large nematocysts, probably stenoteles, thatmeasurec. 120 by 80um, irregularly arranged onthe proximal region of the cnidoband. Abouthalf the circumference ofthe cnidoband is covered withrows of two othertypes of nematocysts; one is ovoid, measuring c. 16 x 11 um; and the other is spherical, measuring c. 8.5um in diameter. Similar nematocysts are also present on the terminal filament. It has not been confirmed that these nematocysts are the acrophores and desmonemes that are typically found on the terminal filaments of many physonect species. It is, however, unusual to find such small nematocysts on the cnidoband. The terminal filament obviously can extend to a considerable extent, but in the preserved specimens it is generally tightly coiled. SIPHOSOMAL TENTACLES AND BUDS (Figure 5B). Midway between each gastrozooid a peculiar tentacular structure is attached directly to the siphosomal stem. In the preserved specimens they are usually tightly coiled, but some relaxed ones can reach lengths of 8 mm. Along one side there is a biserial arrangement of spherical nematocysts, c. 13 um in diameter, similar to those on the cnidoband of the tentacle. The gastrovascular canal is lined by an irregular honeycomb of large endodermal cells. In addition to the siphosomal tentacle, small bud-like structures were noted protruding from the stem. Because the siphosome in both 57 specimens was tightly coiled it was not possible to assess the precise disposition of these buds. However, their arrangement may be similar to that which will be described for the following species. GONOPHORE (Figure SC). Both the Alvin specimens are male and bear numerous gonodendra at various stages of development. The gonophores measure up to 4 mm in length, including the pedicel. They appear to bud one from another to form a small gonodendron. If the gonophores becomes detached, their thin-walled pedicels remain, giving the false impression of the presence of gonopalpons. Again, since the stem is highly contracted, it is difficult to ascertain their exact disposition. REMARKS CONCERNING THE IDENTIFICATION OF BARGMANNIA ELONGATA. Complete and well-preserved specimens of B. elongata easily can be distinguished from the other Bargmannia spp., particu- larly as the form of the bracts is very distinctive. For the nectophores, the arrangement of the apico-lateral ridges, with their distinct right- angled bend and the presence of the short extra ridges branching from them, also are characteristic features. However, in the case of net collected material, which is usually damaged or distorted, the nectophores of this species may be difficult to distinguish from those of the following species, as is discussed further after its description. Bargmannia amoena sp. nov. HOLOTYPE. BMNH 1998.2164, preserved in Steedman’s solution, collected during JSL Il Dive 1458 off Dry Tortugas, Florida; 24°00.6'N 82°17.4'W; 3.ix.1987; 841m. ie S&S pe YY gut } iy ), \ \ @. Fig.6 Bargmannia amoena sp. nov.. Photographs (taken by Ron Gilmer) of live specimen collected during JSL II Dives 968 (A) and 1687 (B). Nectosomal lengths: A. c. 5 cm, B. c. 7 cm. 58 P.R. PUGH Fig. 7 Bargmannia amoena sp. nov. A. Upper, B. lower, and C. lateral views of mature nectophore from type specimen collected during JSL II Dive 1458. Scale bar = 5 mm. PARATYPE. BMNH 1998.2165, preserved in Steedman’s solution, collected during JSL I Dive 2636 off The Bahamas; 25°53.2'N 77°48.3'W; 5—xi-1989; 890m. MATERIAL EXAMINED. 67 specimens have been collected during 40 dives by the submersibles JSL I and II. Of these, 52 have been re- examined for this description. Some of the material originally ascribed to B. elongata by Totton (1954) probably belongs to this species. In addition, some poorly preserved nectophores have been found in recent Discovery collections. DIAGNOsIS. Apico-lateral ridges of nectophores smoothly curved, without pronounced bends; their outer branches terminating well above the ostium, before reaching the relatively small lateral ostial processes. No additional ridges. in smaller specimens central thrust block pointed with small digitiform process apically; in larger ones, latter folded over ventrally so that, in upper view, thrust block appears roundly truncate. In preserved specimens, ostium opens basally. Nectosac more translucent than that of B. elongata. Ratio of overall length of nectophore to that of nectosac averages 1.42, varying slightly according to the size of specimen. Bracts of two types; both delicate and foliaceous, with two pairs of lateral teeth; without patches of large ectodermal cells. DESCRIPTION. Photographs of living specimens collected during JSL Il dives 968 and 1687 are shown in Figure 6. The specimens from the JSL collections fall within three size classes, based on the length of the mature nectophores, but also reflected by the degree of sexual maturity. All the smaller specimens were colourless, while the largest ones had bright orange-red basigasters; the basal part of the gastrozooid. PNEUMATOPHORE. The pneumatophore measured approximately 3 mm in height and 1.5 mm in width, but was highly distorted and BARGMANNIA REVISION 59 Fig. 8 Bargmannia amoena sp. nov. A. Upper, B. lower, and C. lateral views of mature nectophore from small specimen collected during JSL II Dive 976. Scale bar = 1 mm. ruptured by the expansion of the gas within it. No pigmentation is apparent. The pneumatophore is inserted onto the apical end of a long stalk that, depending on the degree of contraction, can be 5—6 mm in length. As in B. elongata, this stalk is flattened at its base, where it joins the nectosome, to form a hinge-like structure. NECTOPHORE (Figures 7-9). The nectophores had a biserial, stag- gered arrangement down the nectosome (Figure 6). The number of nectophores found with each specimen varied from 5 to 32. Depend- ing on the mean length of their nectophores, these specimens can be divided into three size categories. Seven specimens, all collected during the same cruise in 1984, bore c. 10 relatively small nectophores whose lengths were less than 8 mm. The mean length, for the mature nectophores, was 7.41 + 0.43 mm; the mean width 3.10 + 0.22 mm; and the ratio of the overall length to that of the nectosac averaged 1.41 + 0.06. None of these specimens was sexually mature. The bulk of the specimens was included in second size category, where the length of the mature nectophores ranged from 9 to 19 mm. These specimens bore distinct, but immature, gonophores. Each specimen averaged about 20 nectophores, whose mean length was 13.70 + 60 1.86 mm; mean width 6.71 + 0.88 mm; and the ratio of the overall length to that of the nectosac averaged 1.42 + 0.06. Finally six specimens had even larger nectophores and were sexually mature. They averaged 13.5 nectophores, whose mean length was 20.94 + 2.36 mm: mean width 10.58 + 2.18 mm; and the ratio of the overall length to that of the nectosac averaged 1.44 + 0.05. As was the case for B. elongata, the apex of the thrust block of the smaller specimens was drawn out to form a small digitiform process (Figure 8A). In the larger specimens, this process usually became folded over onto the ventral side of the nectophore (Figure 7C), so that, in upper view, the thrust block appeared roundly truncate (Figure 7A). The apico-lateral ridges are, in their preserved state, smoothly curved and have no pronounced bend or side branches (Figures 7 & 8), as was found for B. elongata. After these ridges divide, the inner branches extend obliquely down to reach the ostium; while the outer branches curve down the sides of the nectophore, but peter out well above ostial level. The latter is particularly marked on the smaller nectophores (Figure 8C). Basal extensions of the meso-lateral ridges form the baso-lateral margins of the mouth-plate (Figures 7 & 8). The structure of the mouth-plate varies with the size of the mature nectophore. In the smallest specimens, the mouth-plate is only slightly truncate basally (Figure 8A). In the middle size range of specimens, the mouth-plate becomes more and more emarginate and, in the largest ones, it has a narrow U-shaped median indentation stretching up to the ostium (Figure 7A). The mouth-plates of the immature nectophores of all sizes of specimens show the same features as the corresponding mature ones (Figure 9). Above the mouth-plate, in the small and medium sized speci- mens, the basal extensions of the meso-lateral ridges curve slightly in toward the mid-line (Figure 8B), before curving out again to form the meso-laterals proper. In addition the infra-laterals do not unite with the latter. On the largest specimens, there is no inward curve of the meso-laterals (Figures 7B, 9B), but the infra-laterals have a very weak connection with them (Figure 7B); However, the apical junc- tion of the meso-laterals with the other ridges is always clearly defined. The arrangement of the infra-lateral ridges, in the small (Figure 8C) and medium sized specimens, is very similar to that described for B. elongata. However, in the largest specimens, the ventro-lateral wings are more extensive in the region where the infra-laterals curve up to join the other ridges. The ventral margins of these wings are distinctly emarginate. The nectosac, in its preserved state, appears as a dorso-ventrally undulating tube; but this is probably a preservation artefact. The dorso-lateral extensions, in the mid region of the nectosac, are slightly more extensive than in B. elongata. At its apex the nectosac has a shallow U-shaped indentation, and the adaxial wall is distinctly undercut and devoid of musculature. On the remainder of the nectosac the musculature appears much less dense that of B. elongata, and the nectosac is considerably more translucent. The arrangement of the pallial and pedicular canals, and the radial canals on the nectosac is similar to that of B. elongata. In the preserved specimens, the ostium opens almost basally and has a large velum. Its lateral walls are only slightly extended to form small lateral processes. The pattern of the patches of ectodermal cells is similar to that of B. elongata, but the cells are more uniform in size, and the patches more diffuse laterally. In addition, there are two ventro-lateral patches of deeply granulated cells that are rela- tively large and almost spherical. The youngest nectophores (Figure 9) typically show the absence of a median thrust block. The inner branches of the apico-lateral ridges reach the ostium. The degree of emargination of the apex of P.R. PUGH Fig. 9 Bargmannia amoena sp. nov. A. Upper, B. lower, and C. lateral views of young nectophore from specimen collected during JSL II Dive 1449. Scale bar = 2 mm. the nectosac is variable, according to the developmental stage. It ranges from a narrow, median U-shaped indentation to a marked emargination across most of the width of the nectosac. As noted above the shape of the mouth-plate varies according to the size of the specimen. On either side of the ostium there is a tract of small ectodermal cells extending up toward the end of the outer branch of the apico-lateral ridges. These tracts are longer than those seen on the young nectophores of B. elongata and, again, are difficult to discern on the adult nectophores. BRACT (Figure 10). There are three pairs of bracts per cormidium. Each is thin and leaf-like, with a slight thickening in the central region of the proximal half. The dorsal surface is slightly convex, and the ventral one slightly concave. In general their size is in proportion with that of the nectophores, with those of the largest specimens measuring up to 18 mm in length. No patches of ectoder- mal cells were observed. However, in each cormidium, each successive pair of bracts tends to be slightly larger than the pair proximal to it. The proximal part of each bract is slightly asymmetri- cal to allow for insertion onto the stem. The bracteal canal extends to about two-thirds to four-fifths the length of the bract. It remains in close contact with the ventral wall of the bract at all times. There is much variation in the shape and form of the bracts, but two basic types can be distinguished; both having two pairs of lateral teeth. In one type, which make up the first two pairs of bracts in each cormidium, the bracts are relatively symmetrical. The more distal pair of lateral teeth are very variable in shape, ranging from being virtually absent to being quite marked (Figure 10A, B, D). In the second type (Figure 10C, E), which are the distal pair, the bracts are asymmetrical, and the bracteal canal can have a distinct proximal curve. The distal pair of lateral teeth are well developed and closer BARGMANNIA REVISION A Fig. 10 Bracts of Bargmannia amoena sp. nov. Scale bars: A, B, C = | mm, D, E=2 mm. together than on the first type, so that the distal end of the bract is relatively narrow. One of the proximal pair of teeth is usually more developed than the other, and on that side the lateral wall of the proximal part of the bract often extends out as a rounded notch. GASTROZOOID AND TENTACLE (Figure 11A, C). The largest gastrozooids measure up to 10 mm in length. In the preserved state they are suffused with a brown coloration with, in the largest specimens, the basigaster having bright orange-red pigmentation. The latter (Figure 11C, bg), typically, is cup-shaped, enclosing the base of the stomach region, and is covered in large rounded ecto- dermal cells. The stomach region (Figure 11C, s) appears relatively thin and the endodermal hepatic villi can be seen within. The proboscis region can be extended to some distance. The tentacle can be several centimetres in length. It is a simple, narrow, unsegmented tube, bearing a haphazard and irregular arrange- ment of the two sorts of small nematocysts that are also found on the tentilla. In the present specimens, only a few tentilla, up to 10, remain attached close to its base. In their preserved state, the tentilla (Figure 11A) typically are highly contracted and are comprised of a short pedicel; an irregularly twisted cnidoband; and, for the most 61 part, a regularly coiled terminal filament. The cnidoband is a simple tube that, in life, is generally straight or slightly curved, and can extend to a length greater than 0.5 cm. One side of the cnidoband appears to consist of a primitive elastic strand. It is not tightly folded, as is the case in some other physonect species, but a few pleats are present. The other side of the tentillum is comprised of numer- ous rows of small nematocysts of two types, as was the case in B. elongata. These are ovoid, measuring 20 x 14 um and 12 x 11 um, and occur in roughly equal proportions and possibly in alternat- ing rows, although this could not be determined with certainty. Similar nematocysts are found along the length of the terminal filament. Again, it has not been determined whether these nemato- cysts are the acrophores and desmonenes found in other physo- nect siphonophores. At the proximal end of the cnidoband there is a paired series of up to 26 stenoteles that measure 135 x 105 um. SIPHOSOMAL TENTACLES AND BUDS (Figure 11B,C). As several of the siphosomal stems of the specimens examined remained relaxed, it was possible to study the disposition of the siphosomal tentacles and buds in detail. The primary siphosomal tentacle (Figure 11B, #) is inserted midway between successive gastrozooids and can be tightly coiled or extend to several millimetres in length. As in B. elongata, its surface is covered in large ectodermal cells and there is a paired series of nematocysts along one side. On each cormidium there are, at least, four solid bud-like struc- tures, whose arrangement is sexually dimorphic. In the female specimens (Figure 11B), the first bud (b/) lies a short distance distal to the gastrozooid (gz/), while the second (b2) is inserted about one quarter the length of the cormidium. The gonodendron is then inserted between the latter and the central siphosomal tentacle (7). The third bud (b3) lies a short distance distal to this tentacle, and the last (b4) is inserted immediately proximal to the next gastrozooid (gz2). In the male specimens (Figure 11C), the gonodendron is situated immedietly distal to the gastrozooid. The first bud (b/) then lies distal to the gonodendron at about one quarter the length of the cormidium; that is approximately in the same position as the second bud on the female specimens. The second bud lies immediately proximal to the central siphosomal tentacle; and the third midway between that tentacle and the next gastrozooid. The fourth, as in the female specimens, is inserted immediately proximal to the next gastrozooid. These arrangements pertain in the great majority of the specimens examined, but in the largest ones another tentacle, and possibly another bud, are found in close proximity to the fourth bud. Usually, this tentacle is much smaller than the central tentacle, but otherwise appears to be identical; including the double row of nematocysts. GONOPHORE. (Figure 11B, C). As noted above, the degree of sexual maturity of the specimens appears to be directly related to their size, as assessed by the length of the nectophores. Thus in the smallest specimens, at most, only gonophore buds can be seen. The major group of medium sized specimens have better developed gonophores, while the largest are obviously sexually mature. All seven of the largest specimens are male. There is only a single gonodendron per cormidium. In male specimens the gonodendron lies immediately distal to a gastrozooid and proximal to the first siphosomal bud. The mature male gonophores (Figure 11C, mg) measure up to 5.5 mm in length and 1.1 mm in diameter The female gonophores (Figure 11B, fg) are attached to the stem by a short stalk that is inserted approximately midway between the second siphosomal bud and the central siphosomal tentacle. Between one and six gonophores, in various stages of development, are attached to it by short pedicels. Each 62 P.R. PUGH Fig. 11 Bargmannia amoena sp. noy. A. Mature tentillum, with stenoteles (st) at base of cnidoband (magn. 25x); B. Cormidium of siphosome, with gastrozooids detached (gz' and gz’ mark their attachment points) showing the siphosomal tentacle (t), four buds (b'*) and female gonophores (fg) (magn. 25x); C. Male gonophores (mg) attached just distal to gastrozooid, with its basigaster (bg) and stomach (s), and proximal to the first bud (b’) (magn. 12.5x). BARGMANNIA REVISION gonophore contains two eggs. This is a highly unusual situation as, according to Carré and Carré (1995) all other physonect siphono- phores have only one egg in each gonophore. However, Totton (1965) states that the gonophores of Pyrostephos vanhoeffeni con- tain from three to five eggs. DISTRIBUTION. Much of Totton’s (1954, 1965) Bargmannia material, from early Discovery collections, is so poorly preserved that it is difficult to be certain to which species it belongs. However, as noted earlier, the nectophores from Discovery Sts. 699 and 681, from the South and North Atlantic Ocean respectively, belong to B. elongata; as does that from Discovery St. 107 from south of South Africa. Several other of his nectophores probably belong to B. amoena, but this has not been established with certainty. The great majority of the 8500+ nectophores of Bargmannia spp. that have been identified from over 300 recent Discovery samples, mainly from the North east Atlantic Ocean, belong to either B. elongata or B. amoena. However, these identifications were made before it was realised that two similar species were present. A re- examination of some of the material, however, typically showed that the material was too poorly preserved for specific identification. However, it was clear that B. amoena, not B. elongata, was the predominant species of the genus at c. 44°N, 13°W, where an extensive series of collections was made (Pugh, 1984). Collectively, both species have a widespread distribution in the North-east Atlan- tic Ocean; from the equator to 60°N, with possibly B. elongata being more common at lower latitudes and B. amoena at higher ones. Nectophores have been collected at all depths from the surface to 4520 m, but the vast majority were found in samples from between 200 and 600 m. Most of the 67 specimens of B. amoena collected by the sub- mersibles JSL I and Il came from a relatively small area in the region of The Bahamas, from 25°03' to 26°36'N and 77°23’ to 78°44'W. Five others were collected near the Dry Tortugas, between Florida and Cuba, at c. 24°30'N, 83°45'W. All were collected over a wide depth range, from 435 to 910 m, with a mean depth of 625 + 130 m. This mean depth is slightly deeper than the depth range for both B. elongata and B. amoena found in Discovery net collections. How- ever, both figures probably are biased because, in the case of the submersible, most observations and collections were made within the 600-900 m depth range, while at 44°N, 13°W for instance, most of the net sampling was concentrated in the top 600 m of the water column. REMARKS CONCERNING THE IDENTIFICATION OF BARGMANNIA AMOENA. Complete and well-preserved specimens of B. amoena easily can be distinguished from the other Bargmannia spp. as they have very distinctive bracts. However, if only poorly preserved nectophores are present, then there may be some difficulty in separating this species from B. elongata. They cannot be confused with B. gigas, because of the relatively enormous size of the latter’s nectophores; and should not be confused with B. lata. The much narrower nectosac of the latter species, together with the greater depth of the furrow between its deep lateral wings, should easily distinguish it. As noted above, the best feature distinguishing B. elongata and B. amoena is the arrangement of the apico-lateral ridges. InB. elongata these have a pair of side branches, running down toward the mid- line, while in B. amoena such side branches are absent. In addition, in B. elongata, at the point where these side branches arise, the apico-laterals bend sharply outwards, through 90°, while in B. amoena the apico-laterals only curve gently away from the mid-line. Unfortunately, it is this region of the nectophore that most often becomes distorted in poorly preserved specimens and these distin- 63 guishing features can be masked. This can result in the distinct, right-angle bend in the apico-lateral ridges of B. elongata coming to resemble the much smoother curve in B. amoena or, conversely, those of the latter species becoming distorted to form distinct bends resembling those of the former. The side branches to the apico- laterals in B. elongata often are difficult to discern, but usually show up after staining. Another obvious difference, despite its subjectiveness, is the density of the musculature on the nectosac. Nectophores with almost opaque nectosacs appear to belong to B. elongata, while those with translucent nectosacs belong to B. amoena. In addition, the ratio of the total length of the nectophore to that of the nectosac may be of use. In B. elongata this ratio, in well-preserved specimens, averages 1.31, while in B. amoena it averages 1.42. However, with poorly preserved material, particularly when the basal half of nectophore is damaged, both measurements could be underestimated, which would lead to a corresponding increase in the ratio. Other features of the nectophore that might be considered include the fact that in B. amoena the outer branch of the apico-lateral ridges peters out higher above the ostium than in B. elongata. Also, the lateral processes to the ostium are much larger in B. elongata. In addition, the angle at which the ostium opens is very characteristic in well-preserved material. In B. elongata the ostium is directed dorso-basally while in B. amoena it opens basally. However, all these features might be difficult to discern in poorly preserved nectophores The structure of the mouth-plate may be important but, as has been shown for B. amoena, this may vary according to the size of the specimens. More well preserved specimens of B. elongata are needed in order to assess this. Similarly, this applies to the arrange- ment of the meso-lateral ridges and their basal extensions; and to the profile of the ventral margins of the ventro-lateral wings. ETYMOLOGY. Amoena is Latin for ‘pleasing, delightful’. Bargmannia lata Mapstone 1998 Bargmannia elongata Totton 1954 (partim) (text-Figure 28, E-F only), 1965 (partim) (Figure 45 E-F only). Bargmannia lata Mapstone 1998: 141-146, figs 1-3. HOLoTyYPE. In the collections of the British Columbia Provincial Museum, BCPM 996-203-1; one nectophore and one bract collected at St. LC10 (48°22.4'N 126°20.2'W; 24-iv-1987; 700-Om) off British Columbia, Canada; PARATYPES. As designated by Mapstone; 1. 7 nectophores and 7 bracts (BCPM 996-204-1#1), and 2. 6 nectophores and 6 bracts (BCPM 996-205-1#2) from the same sample as the holotype; 3. 11 nectophores (BCPM 996-206-1#3), and 6. 1 bract (BMNH 1996.1239-1240#6) from St.A4 (48°15'N 126°40'W; 21.iii.87; 500 m); 4. 8 nectophores (0O-700m; BCPM 996-207-1#4), and 5. 14 nectophores and 2 bracts (BMNH 1996.1234-1238#5) from St. LB17 (47°56.5'N 126°26.1'W; 21.11.87; 700m). MATERIAL EXAMINED. One specimen collected during Alvin Dive 966 off San Diego, California, U.S.A.; 33°04'N 118°16'W; 8-ix— 1979; water depth 747m. The depth of collection of the specimen was not recorded. Two nectophores collected at Discovery St. 1769, and figured by Totton (1954, Text-Figure 28, E, F; 1965, Figure 45, E, F) as B. elongata; 33°43.3'S 8°38.5'E; 20—v—1936; 1000-750 m; NHML 1957.5.15.110. In addition, the specimens that Totton included under the name B. elongata have been re-examined. Although not all are in good condition, the following appear to belong to B. lata:- 64 Table 1 Geographical distribution of Bargmannia lata from recent Discovery collections. Station Date Position 8565# 1 1—viti-74 3°03.1'N 23°14.3'W 6662#37 21-11-68 10°34.9'N 19°43.7'W 6662#32 20-11-68 10°45.3'N 19°51.7'W 6662#30 19-11-68 10°47.4'N 19°52.7'W 6662#15 16-11-68 10°57.0'N 19°56.6'W 6662#20 17-11-68 10°57.5'N 19°49.0'W 6662#22 17-11-68 10°57.6'N 19°57.3'W 6662#16 16-11-68 10°59.4'N 19°52.1'W 7824#39 10-11-72 11°01.1'N 19°55.8'W 6662#10 15-11-68 11°03.1'N 19°59.2'W 6662# 7 14-11-68 11°04.6'N 19°48.2'W 6662# 8 15-11-68 11°08.2'N 19°47.8'W 7831# 1 16-11-72 13°18.4'N 25°33.2'W 7803# 2 19-11-72 18°06.4'N 25° 8.1'W 11261#16 28—vi-85 31°13.1'N 25°18.3"W 8281#29 17-i1-73 31°42.5'N 63°43.6'W 11794#36 26-vi-88 47°14.2'N 19°31.2"'W 11794#83 2-vii-88 47°17.9'N 19°21.4"°W 11794#31 26—vi-88 47°27.1'N 19°18.0'W 12096# 2 3-vi-90 47°57.8'N 16°49.6'W 4 nectophores from John Murray Expedition St. 34; 13°05.6'N, 46°24.7'E; 16—x—33; 0-1022 m. BMNH 1949.11.10.378; and 4 nectophores from Discovery St. 206; 16°36'S, 6°25.1'W; 1—v— 37; 1900-1500 m. BMNH 1957.5.15.111. Several nectophores and bracts also have been identified from more recent Discovery material from the N.E. Atlantic (Table 1). DIAGNOSIS. Nectophores with relatively long median thrust block; with extensive ventro-lateral wings, emarginate on ventral margins. Nectosac a relatively short, narrow tube without any pronounced dorso-ventral undulations; squarely truncate apically. Ratio of over- all length of nectophore to that of nectosac, on average, exceeds 1.59. Bracts large, robust, distally truncate; with semicircular dorsal ridge connecting tips of baso-lateral processes and delimiting a dorso-basal facet; with prominent tooth on outer lateral margin. DESCRIPTION. A photograph of the live specimen collected by Alvin and taken on board the mother ship is shown in Figure 12. Unfortunately, prior to photography, several nectophores had be- come detached and the siphosomal stem had contracted. No pigmentation is apparent in the preserved specimen, but the original colour photograph indicates that the whole of the endodermal lining of the stem was suffused with an orange-red colour. PNEUMATOPHORE. The highly distorted pneumatophore of the Alvin specimen measured 2 mm in height. It is borne on a very short, but probably highly contracted, stalk. NECTOPHORE (Figures 13-15). A total of 16 nectophores was found with the Alvin specimen. The mean dimensions for the fully developed nectophores of this specimen were:- length 19.44 + 2.73 mm (range 14.79—23.86 mm); width 9.72 + 1.07 mm, and the ratio of total length to that of the nectosac was 1.59 + 0.09. Nectophores of the earlier Discovery material, in the NHM, are somewhat larger with an average length of 24.94 + 6.02 mm (range 17.74-31.94 mm). The ratio of total length to that of the nectosac also was slightly greater ratio (1.67 + 0.10; n = 10); the same as that for Mapstone’s (1998) material. Similarly, the nectophores of more recent Discov- ery material also are larger: length 22.89 + 4.04 mm (range 15.8—29.6 mm; n = 48). These nectophores have the greatest length:nectosac- length ratio of 1.73 + 0.11. However, this increase in ratio probably P.R. PUGH Depth (m) Nects Bracts 700— 800 2) 1060-1300 24 16 1210-1450 3 4 730-795 17 600— 695 DH 810-900 8 610— 680 25 3 810— 890 18 DB 895-1000 3) 910-985 4 715— 800 33 910-985 Ti 10-1000 12 1015-1250 13 1000-1100 16 1259-1500 26 1200-1300 10 1300-1395 7 2500-2750 17 2 1100-1200 16 20 Fig. 12 Bargmannia lata. Photograph (reproduced by kind permission of Larry Madin, WHOJ) of live specimen collected during Alvin Dive 966. Nectosomal length c. 5 cm. BARGMANNIA REVISION 65 Fig. 13 Bargmannia lata. A. Upper, B. lower, and C. lateral views of mature nectophore collected during DSRV Alvin Dive 966. D. Ventral view of the apex of another nectophore. Scale bar = 2 mm. is due to the fact that the base of the nectophore frequently is damaged, resulting in an underestimate of both measurements, and a consequent increase in their ratio. The central thrust block forms an extensive triangular process whose apex is often roundly pointed Figure 13A). However, on several nectophores one side is drawn out to form a small digitiform process that may be folded over laterally or ventrally (Figure 13D). The ridge pattern conforms to the basic Bargmannia design, with no extra ridges being present. From their junction with the meso- and infra-lateral ridges on the ‘shoulder’ of the nectophore, the apico-laterals are directed ob- liquely toward the mid-line. They closely approach each other, and continue for some distance in a basal direction; leaving a narrow median furrow between them. At about one quarter the length of the nectophore, in the Alvin material, they rapidly curve out laterally, before giving rise to the typical inner and outer branches (Figures 13-15). The inner branch curves obliquely toward the mid-line and joins the ostium on its dorsal surface. The outer branch curves down and then round and ends on the lateral margin of ostium, although it can be difficult to discern basally. The angle between the apico- lateral ridge and its inner branch is acute (Figure 13A). However, in less well preserved nectophores, this pronounced angle is not always apparent (Figure 14B) and the inner branch can appear as a simple continuation of the main ridge. The mouth-plate is small and made up of two rounded lobes that unite in the mid-line, slightly basal to the ostium. The ventro-lateral margins of these lobes are, as usual, formed by basal extensions of the meso-lateral ridges. Above the ostium, on the ventral surface of the nectophore, these basal extensions curve round toward the mid- line, before looping back out as the meso-laterals proper and continuing apically. After a relatively long distance, in comparison with B. elongata and B. amoena, the infra-laterals branch from them (Figures 13C, 15C). The meso-laterals then continue obliquely up and across the lateral margins of the nectophore to join the apico- and infra-laterals on the ‘shoulder’ of the nectophore. The junctions with the other ridges, both apically and basally, are obvious, unlike in B. elongata. In the basal two-thirds of the nectophore, the infra-lateral ridges form the ventral margins to the ventro-lateral wings (Figure 13C, 15C). These wings are relatively large in comparison with those of B. elongata and B. amoena, occupying more than half the depth of the nectophore. They are distinctly emarginate in the mid region of 66 P.R. PUGH -- 7 ‘ - = oo --- -- =< it ‘ -— L--- Noes -2 ere Fig. 14 Bargmannia lata. A. Lower and B. upper views of nectophore from Discovery St. 1769. Scale bar = 5 mm. Fig. 15 Bargmannia lata. A. Upper, B. lower and C. lateral views of young nectophore collected during DSRV Alvin Dive 966. Scale bar = 1 mm. BARGMANNIA REVISION the nectophore. In the region where the infra-laterals leave the ventral margins of the ventro-lateral wings, the latter begins to thicken toward the mid-line. These thickened, rounded, unridged lateral walls continue up to the apex of the nectophore and, on the thrust block, delimit a narrow gutter that enfolds the nectosomal stem (Figures 13B, 14A). The nectosac is a relatively short and narrow tube without any marked dorso-ventral undulations (Figures 13, 14) in the preserved specimens. Its apex lies approximately on a level with the ‘shoulder’ of the nectophore. The nectosac occupies only c. 38% of the width of the nectophore, as measured across its ‘shoulder’. This results from the fact that the extensive ventro-lateral wings not only increase the depth of the nectophore, but also increase its relative width. The apex of the nectosac is squarely truncate, without any marked indentation. Its adaxial surface is distinctly undercut and, typically, is devoid of musculature. The remaining musculature on the nectosac appears much less dense in comparison with that of B. elongata. The canal system follows the basic Bargmannia plan. The long pallial canal ends, apically, with a short dorsad inflection into the mesogloea. On the nectosac the pedicular canal gives rise to only the dorsal and ventral radial canals. In contrast to B. elongata and B. amoena, in the preserved material the lateral radial canals have straight courses down the lateral margins of the nectosac. However, afi wire er 67 in life, their courses appeared to be slightly undulating (Figure 12). In the original colour photograph there are indications that the pallial, dorsal, ventral, ostial ring, and proximal parts of the lateral canals were suffused with a light orange-red colour. The ostial opening, in the preserved nectophores, typically is displaced slightly dorsad and has a well-developed velum, but no pronounced lateral processes. There are no marked patches of ectodermal cells, although some nectophores show a single row around the basal half of the ostium and/or a short, narrow band of small cells that lies just dorsal to the outer branch of the apico-lateral ridge. These cells, again, are presumed to be sites of biolumines- cence. BRACT (Figure 16). Only seven bracts were retained with theAlvin material. However, because of their very characteristic shape, sev- eral more have been identified from recent Discovery material. The bracts measured from 13.5 to 27 mm in length and were remarkably robust. They had a convex dorsal and a concave ventral surface. In the Alvin specimen, there were two types of bract, with one type being represented by only a single small bract. The key feature that distinguishes them is the presence of only a single lateral tooth on the outer margin (Figure 16A, B) of the larger ones; while the smaller one has lateral processes on both sides (Figure 16C). The Fig. 16 Bracts of Bargmannia lata collected during DSRV Alvin Dive 966. Scale bar = 2 mm. 68 larger ones also were distinctly asymmetric proximally; and the bracteal canal made a right-angled bend. These differences may be just the result of growth, or may related to their point of attachment on the cormidium, as was noted for B. amoena. Both types of bract are distally truncate, and possess a semicircular dorsal ridge that delimits a rounded distal facet. The ridge connects the tips of two distal processes. The shape of the distal margin of the larger bracts varied consid- erably. On some, the inner margin of one of the distal process extended up the ventral side of the bract forming a flap-like struc- ture; while on others this flap was absent. Another small ventral flap can be present, approximately in the mid line, in the proximal half of the bract. The bracteal canal lies just above the ventral surface of the bract and ends close to its distal margin. The original colour photo- graph of the Alvin dive 966 specimen indicated that, in life, some of the bracteal canals had an orange-red pigmentation like that of the remainder of the stem. GASTROZOOID AND TENTACLE. Only a small portion of the proxi- mal end of the siphosome was preserved from the Alvin specimen, so that only a few young gastrozooids were present. These measured up to 7 mm in length, and showed no distinguishing features. No pigmentation is apparent in the preserved material, but in life they had a deep red pigmentation. The tentacles attached to the gastrozooids mainly bore young tentilla; with a c. 1 mm pedicel; a 2.5 mm cnidoband, apparently devoid of large nematocysts; and a c. 4 mm uncoiled terminal filament, apparently without a terminal process. However, a few more mature tentilla had cnidobands extending to more than 8 mm, with 4-6, possibly more, large nematocysts (stenoteles), arranged biserially and arranged alternately, at their proximal ends. Small nematocysts, possibly of two types, are present throughout the cnidoband and terminal filament. SIPHOSOMAL TENTACLES AND BUDS. The peculiar tentacular proc- esses, previously noted in specimens of B. elongata and B. amoena, are present on the small part of the stphosomal stem remaining but, because of the contracted state of the latter, it was not possible to ascertain their exact disposition. They are narrow, delicate struc- tures, measuring up to c. 4 mm in length, and are covered in large, rounded ectodermal cells. Small nematocysts are present but, with- out destroying the tentacle, it was no possible to assess whether they had a biserial arrangement, as noted for the previously described species. Siphosomal buds also appear to be present, but their arrange- ment could not be discerned. GONOPHORE. A few loose male gonophores are present with the Alvin material. They are identical to those previously described for B. elongata and B. amoena. DISTRIBUTION. A total of 288 nectophores and 27 bracts of B. lata have been found in recent Discovery collections (Table 1). The data indicated that, in the North-east Atlantic Ocean, B. lata was more commonly collected at lower latitudes and at deeper mesopelagic depths; with a mean depth of c. 1000 m. Totton’s material came from two sites in the SouthAtlantic Ocean and one in the Gulf of Aden; the Alvin material came from off San Diego, California, USA; and Mapstone’s (1998) from off British Columbia, Canada; thus indicat- ing a widespread geographical distribution for this species. REMARKS CONCERNING THE IDENTIFICATION OF BARGMANNIA LATA. B. lata can now be easily distinguished from other Bargmannia spp. The nectophores are relatively broad, with extensive ventro-lateral wings, that are markedly emarginate along their ventral margins. The median thrust block is relatively long so that the ratio of the total P.R. PUGH length of the nectophore to that of the nectosac is high, c. 1.6, or more for net collected material, as compared with c.1.31 in B. elongata and c. 1.41—1.44 in B. amoena. The nectosac appears as a relatively short, narrow, straight-sided tube, without any pronounced dorso-ventral undulations, and squarely truncate apically. It occu- pies only c. 38% of the width of the nectophore, as compared with 45% in B. elongata. The large, robust bract, with a semicircular dorsal ridge connecting the tips of the baso-lateral processes, also is distinctive. Despite these differences it is clear that Totton (1954, 1965), had little reason to suspect that he was dealing with at least two Bargmannia spp., particularly as he had so few, poorly preserved nectophores. However, with the collection of complete specimens of Bargmannia spp. by submersibles the specific differences between the various types of nectophore that Totton illustrated can now be established. Bargmannia gigas sp. nov. HOLOTYPE. BMNH 1998.2166 one nectophore, preserved in Steedman’s solution, collected at Discovery St. 8560#2 (0°03.1'N, 22°44.2'W; 27-vii-1974; 1510-2000 m; RMT8 net). PARATYPES. Three nectophores, preserved in Steedman’s solu- tion, from the same Discovery sample. BMNH 1998.2167-69. MATERIAL EXAMINED. The type and paratypes, and a further ten nectophores from the same station, which are retained in the Discov- ery collections at the Southampton Oceanography Centre. All the nectophores are presumed to have originated from a single speci- men. DIAGNOSIS. Nectophores relatively enormous, up to 52 mm in length; with large ventro-lateral wings; with small mouth-plate deeply divided. Basic ridge pattern supplemented by three pairs of ridges, all dividing from apico-laterals; two pairs short, directed toward mid-line; third pair directed laterally. Nectosac without dorso-ventral undulations; apex only slightly emarginate; ostial opening large. Ratio of overall length of nectophore to that of nectosac averages 1.63. DESCRIPTION. NECTOPHORES (Figures 17 and 18). The relatively enormous nectophores varied in size from 14.5 x 8 mm (length x width) for the smallest, immature one, to 52 x 20 mm, respectively, for the largest. The mean dimensions for the fully developed nectophores were length: 41.0 + 6.96 mm and width: 19.1 + 2.52 mm, and the ratio of total length to that of the nectosac was 1.63 + 0.10. The nectophores, in their present state of preservation, are devoid of pigmentation and, in most cases, the muscular lining of the nectosac has become detached and/or lost. The large, thickened, central thrust block is roundly truncate. The basic pattern of the prominent ridges conforms with that of the genus, and both the inner and outer branches of the apico-laterals appear to reach the dorso-lateral margins of the ostium. In addition there are three pairs of ridges that branch from the apico-laterals (Figures 17A, 18). Two of these pairs of ridges are very short and run down into the shallow median gutter, towards the mid-line. One pair arises at a level of about two-fifths the length of the nectosac, while the other originates from the inner branches of the apico-laterals, close to the ostium. The other pair arises from the outer branches of the apico-laterals and extends up the sides of the nectophore between the apico- and meso-laterals. These ridges peter out approximately at the mid-length of the mature nectophore. Below them the lateral walls of the nectophore often show prominent thickenings (Figure iG): BARGMANNIA REVISION ye ~ x 4 4 * ee, ee ee eee - 4 > 7 -<-- Fig. 18 Bargmannia gigas sp. nov. Upper views of A. smallest, and B. slightly larger nectophores. Scale bar = 0.5 cm. 69 70 The broad, but relatively short, mouth-plate consists of two rounded processes whose basal margins are marked by basal exten- sions of the meso-lateral ridges. These ridges peter out, without apparently uniting, on the lower surface of the nectophore a short distance above ostial level. The infra-lateral ridges branch from the meso-laterals approximately on a level with the mid-length of the nectosac. The meso-laterals then run obliquely up the sides of the nectophore to join the other ridges, approximately on a level with the top of the nectosac. The infra-laterals form the ventral margins of the ventro-lateral wings up to a level just above the top of the nectosac. They then bend through 90° and run up to join the apico- and meso- laterals (Figure 17C). The thickened ventro-lateral wings are well developed and enclose a deep groove, which at its deepest occupies half the height of the nectophore (Figure 17C). They are roundly truncate apically at about four-fifths the length of the nectophore. The nectosac is a long tube, with only a slight apical emargination, that occupies most of the main body of the nectophore, and has no obvious dorso-ventral undulations. It is distinctly undercut adaxially and is presumed to have a muscle-free zone in that region, although this could not be established with certainty. The ostial opening is very large and is only slightly directed towards the upper surface. The pedicular canal (Figure 17B) typically only gives rise to the dorsal and ventral radial canals. The course of all the canals is straight. Typically, the youngest nectophores show the absence of a central thrust block (Figure 18A), but with a clearly defined ridge pattern. A slightly larger nectophore shows the gradual development of the thrust block and the ventro-lateral wings (Figure 18B). REMARKS CONCERNING THE IDENTIFICATION OF BARGMANNIA GIGAS. B. gigas is known only from the nectophores of what is presumed to be a single specimen, collected in the equatorial Atlantic at a depth of 1510—2000m. The nectophores easily can be distinguished from other Bargmannia sp. by their incredible size and the distinctive pattern of ridges. ETYMOLOGY. The specific name gigas refers to the giant size of the nectophores. DISCUSSION As was noted in the Introduction, the content of the genus Barg- mannia is debated. Although Totton (1965) included it in the family Pyrostephidae, some of the characters that he listed in his diagnosis of that family apply exclusively to the scope accorded to the genus Pyrostephos, which is monotypic for P. vanhoeffeni. In particular, these are the looping of the lateral radial canal on the nectosac of the nectophore, and the three to four marked bends of the dorsal canal. In Bargmannia all the canals are held to be straight, or only slightly sinuous. Other characters, such as the number of tentilla on the tentacle, and the structure of the bracts and gastrozooids probably are more specific than familial. However, in both genera the nectosome is long but again this is not a good familial character. At first glance, the nectophores of Pyrostephos vanhoeffeni (see Totton, 1965, Figure 41) and Bargmannia species look strikingly different. However, there are several similarities. Specimens of P vanhoeffeni have been collected recently by SCUBA divers (G.R. Harbison, personal donation) and by net (Pages, Pugh & Gili, 1994). It is apparent from these that the mature nectophores can vary greatly in size; ranging from 8 x 5 mm (length x width) in the SCUBA collected material (Figure 19B) to 15 x 18mm, respectively, in the net collected specimens (Figure 19C). Such large size ranges of the mature nectophores of physonect species havenotoften been observed, P.R. PUGH although such is so in Nanomia bijuga (delle Chiaje, 1841) (Pugh, pers. obs.). Itis also known to be so in at least two Bargmannia spp. In B. amoena (Figure 19A), the size variation of mature nectophores is even greater than that of P. vanhoeffeni, ranging from c. 6 x 3 mm (length x width) to 25.5 x 12.5 mmrespectively. Although the general shape of mature Bargmannia nectophores does not change with size, it appears that that of P. vanhoeffeni does. In the smaller specimens of the latter (Figure 19B) there is a large triangular thrust block, reminiscent of that on mature Bargmannia nectophores. However, in the larger, preserved specimens (Figure 19C) the thrust block is folded upwards producing a deep transverse furrow on the dorsal surface, just basal to it. Neither P. vanhoeffeni nor Bargmannia spp. have large apico-lateral processes. The nectophores of both genera have the same basic ridge pattern; comprising apico-, infra, and vertical (meso-) laterals, but no lateral ridges. In addition, in both, the apico-laterals divide into two branches close to the ostium. The inner branch (‘frontal ridge’) of the larger nectophores of Pyrostephos vanhoeffeni (see Totton, 1965, Figure AQ) is relatively short, in comparison with Bargmannia spp., and directed only toward the mid-line. However, the present material, particularly that of the smaller specimens, shows that these ridges can curve round basally and continue for a short distance towards the ostium before petering out. Nonetheless, the species of these two genera are not the only physonects to show this basic pattern of ridges. It is also found on the nectophores of two others namely, Frillagalma vityazi Daniel, 1966 (see Pugh, 1998) and Erenna richardi Bedot, 1904 (P.R.Pugh, personal observation). In addition, an even simpler arrangement, in which the vertical lateral ridges are absent, is found in two Marrus species, namely M. antarcticus Totton, 1954 and M. orthocanna Kramp (1942). For these, the branching of the apico-lateral ridges is weak and difficult to discern. A third species, namely M. orthocannoides, that Totton (1954) include in the latter genus probably does not belong there as its nectophores do not have an adaxial muscle-free zone on the nectosac. Species referred to both Bargmannia and Pyrostephos have an adaxial zone on the nectosac of the nectophore that is muscle-free and deeply embayed. In addition, the lateral radial canals arise separately from the dorsal canal. These appear to be important characteristics. Of the other species previously mentioned Marrus antarcticus and M. orthocanna show all of these characters. However, in Frillagalma vityazi, there is no deeply embayed, muscle-free adaxial zone; al- though the lateral radial canals do arise separately from the dorsal one, albeit very close to the point of insertion of the pedicular canal (Pugh, 1998). Further, this species has many marked differences from the others under consideration and need not be considered further in this discussion. Erenna richardi does have a muscle-free zone, but it lies at the apex of the nectosac, whichis not deeply embayed adaxially. Thus, from the basic arrangement of the ridges and nectosac, the nectophores of Bargmannia, Pyrostephos and Marrus species are very similar. Another common feature is that they all have relatively short pedicular and relatively long, ascending pallial canals. But how do their siphosomal elements compare? Most siphonophores are believed to be hermaphrodite (mono- ecious), bearing both male and female gonophores. However, specimens of Physalia, the Portuguese Man O'War, and probably all other cystonect siphonophores, are single sexed (dioecious). It should be noted that Mackie, Pugh & Purcell (1987, p.100) used the terms monoecious and dioecius erroneously. In physonect siphonophores, species of the benthic family Rhodaliidae appear to be dioecious (Pugh, 1983), as are Marrus antarcticus, Pyrostephos vanhoeffeni (Totton, 1965), and from the present study Bargmannia spp. According to Andersen (1981) M. orthocanna is monoecious, but the male gonophores he illustrated were only minute, bud-like BARGMANNIA REVISION Tt Fig. 19 Nectophores of A. Bargmannia amoena sp. nov. (magn. 10); B, C. Pyrostephos vanhoeffeni collected by SCUBA (B, magn. 11x) and by net (C, magn. 7.5x). structures. Since only female gonophores could be identified on submersible collected specimens, this point could not be checked (P.R.Pugh, personal observation). Whether Erenna richardi is monoecious or dioecious remains unknown. Nonetheless, it is of interest to note that, of the physonect species whose female gonophores are known, only those of P. vanhoeffeni and B. amoena contain more than one egg; 3—5 in the former species (Totton, 1965) and two in the latter. The structure of the tentillum is another feature in which close similarities between Bargmannia species and Pyrostephos vanhoeff- eni appear. In both the cnidoband is straight, or slightly twisted, but not tightly coiled, and is without a basal involucrum. In addition, they both have long terminal filaments. Even more striking is the presence of large nematocysts, probably stenoteles, only in the Te proximal region of the cnidoband of both species. However, those of Bargmannia spp. are considerably larger than those of P. vanhoeffeni, which measure c. 40 x 28 mm. Further, the other nematocysts present on the cnidoband and terminal filament are very similar. Two types of small nematocysts were found in Bargmannia spp. and similar ones, measuring 13-17 x 9.5—10.5 ym and 6.5 x 6.5 um, were found in P. vanhoeffeni. Although the tentillum of Erenna has a straight cnidoband, and that of Marrus is loosely coiled or straight, the types and distribution of the nematocysts are quite different. The cnidoband of Evenna is massively armed with two types of elongate nematocysts, measuring c. 160 x 37 um and c. 35 x 18 um, while the terminal process appears to be devoid of any nematocyst. The cnidoband of Marrus contains heteronemes and haplonemes, meas- uring c. 55 x 20 um and c. 35 x 7 um of the type often seen in other agalmatid species. The terminal filament of the latter species con- tains only small nematocysts, probably desmonemes, acrophores or anacrophores, measuring c. 16 x 9.5 um and c. 10 x 10 um. These differences in the nematocyst types alone seem sufficient to indicate that Bargmannia and Pyrostephos are more closely related to each other than either is to Marrus or Erenna. Despite all these similarities between Bargmannia spp. and Pyro- stephos vanhoeffeni, there are at least two major differences: Bargmannia spp. are the only physonect siphonophores known to have siphosomal tentacles, though apolemiid species have nectosomal ones; they also lack dactylozooids, although the bud-like structures may be vestigial ones. In addition, P vanhoeffeni is the only species known to have highly modified dactylozooids, the oleocysts, without palpacles. The only other species in which dactylozooids are thought to be absent is Marrus orthocanna (Andersen, 1981). However, Totton (1965) reported that palpons are present on the gonodendra of M. antarcticus. Further work needs to be carried out on well-preserved specimens of theseMarrus species in order to investigate this apparent difference, and whether each is monoecious or dioecious. Although there are major differences between Bargmannia spp. and Pyrostephos vanhoeffeni, there appear to be sufficient similari- ties to warrant the retention of the genus Bargmannia in the family Pyrostephidae. The alternative would be to propose a new family for it, since the genus certainly does not fit neatly into the family Agalmatidae. This might also apply to the genera Marrus and Erenna but their species are too little known. ACKNOWLEDGEMENTS. [am extremely grateful to Drs Richard Harbison and Edie Widder for inviting me to participate in several cruises involving the use of submersibles, and for donating the siphonophore material collected to me. I thank Dr Paul Cornelius for his helpful comments on the manuscript. P.R. PUGH REFERENCES Alvarino, A. 1963. Chaetognatha, Siphonophora, and Medusae in the Gulf of Siam and the South China Sea. Report on the results of the NAGA Expedition — South East Asia Research Project. Scripps Institution of Oceanography: 104-108. 1964. Report on the Chaetognatha, Siphonophorae, and Medusae of the MONSOON Expedition to the Indian Ocean. Report. Scripps Institute of Oceanog- raphy. SIO Ref. Ser. 64/19: 103—108,209-212. Andersen, O.G.N. 1981. Redescription ofMarrus orthocanna(Kramp, 1942) (Cnidaria, Siphonophora). Steenstrupia 7: 293-307. Carré, C. & Carré, D. 1995. Ordre des Siphonophores. pp. 523-596 in Traité de Zoologie. Anatomie, Systématique, Biologie. Tome III. Fascicule 2. Cnidaires. Cténaires. ed. D. Doumenc. Masson, Paris. Daniel, R. 1974. Siphonophora from the Indian Ocean. Memoirs of the Zoological Survey of India 15(4): 1-242. —— 1985. Coelenterata: Hydrozoa Siphonophora. The fauna of India and adjacent countries, Zoological Survey of India, 440 pp. Herre, W. 1955. Die Fauna der miozanen Spaltenfiillung von Neudorf a.d.Match (CSR.), Amphibia (Urodela). Sitzungsberichte der Osterreichische Akademie der Wissenschaften, Mathematisch-naturwissenschaftliche Klasse. Abteilung I, 164: 783-803. Kirkpatrick, P.A. & Pugh, P.R. 1984. Siphonophores and Velellids. Synopses of the British Fauna (New Series) 29: 1-154. Leloup, E. 1955. Siphonophores. Report on the Scientific Results of the ‘Michael Sars’ North Atlantic Deep-Sea Expedition 1910 5(11): 1-24. Mackie, G.O. 1964. Analysis of locomotion of a siphonophore colony. Proceedings of the Royal Society of London, B 159: 366-391. —— Pugh, P.R. & Purcell, J.E. 1987. Siphonophore biology. Advances in Marine Biology 24: 97-262. Mapstone, G.M. 1998. Bargmannia lata, an undescribed species of physonect siphonophore (Cnidaria, Hydrozoa) from Canadian Pacific waters Zoologische Verhandelingen 323: 141-147. Pages, F, Pugh, P.R. & Gili J-M. 1994. Macro- and megaplanktonic cnidarians collected in the eastern part of the Weddell Gyre during summer 1979. Journal of the Marine Biological Association of the United Kingdom 74: 873-894. Pugh, P.R. 1983 Benthic Siphonophores. A review of the Family Rhodaliidae (Siphonophore, Physonectae). Philosophical Transactions of the Royal Society of London B 301: 165-300. —— 1984. The diel migrations and distributions within a mesopelagic community in the North east Atlantic. 7. Siphonophores. Progress in Oceanography 13: 461-489. — 1998. A re-description of Frillagalma vityazi Daniel 1966 (Siphonophorae, Agalmatidae). Scientia Marina 62: 233-245. & Harbison, G.R. 1986. New observations on a rare physonect siphonophore, Lychnagalma utricularia (Claus, 1879). Journal of the Marine Biological Associa- tion of the United Kingdom 66: 695-710. & Youngbluth, M.J. 1988. A new species of Halistemma (Siphonophora, Physonectae, Agalmidae) collected by submersible. Journal of the Marine Biologi- cal Association of the United Kingdom 68: \1—14. Stepanjants, S.D. 1967. Siphonophores of the seas of the USSR and the north western part of the Pacific Ocean. Opredeliteli po Faune SSSR 96: 1-216. Totton, A.K. 1954. Siphonophora of the Indian Ocean together with systematic and biological notes on related specimens from other oceans. Discovery Reports 27: 1— 162. — 1965. A Synopsis of the Siphonophora. London: British Museum (Natural History). . GULDE FOR AUTHORS Aims and scope. The Bulletin of the British Museum (Natural History) Zoology, was established specifically to accommodate manuscripts relevant to the Collections in the Department of Zool- ogy. It provides an outlet for the publication of taxonomic papers which, because of their length, prove difficult to publish elsewhere. Preference is given to original contributions in English whose contents are based on the Collections, or the description of speci- mens which are being donated to enhance them. Acceptance of manuscripts is at the discretion of the Editor, on the understanding that they have not been submitted or published elsewhere and become the copyright of the Trustees of the Natural History Mu- seum. All submissions will be reviewed by at least two referees. Submission of manuscripts. Initially three clear, complete cop- ies should be submitted in the style and format of the Bulletin. The text must be typed double-spaced throughout, including references, tables and legends to figures, on one side of A4 paper with 2.5 cm margins. All pages should be numbered consecutively, beginning with the title page as p. 1. SI units should be used where appropriate. Whenever possible a copy of the text, once the paper has been accepted, should also be provided on floppy disc (see below). Discs should only be sent after final acceptance, as papers generally need revision after refereeing. If it is impossible to provide an appropriate disc please ensure that the final typescript is clearly printed. Authors are requested to ensure that their manuscripts are in final format, because corrections at proof stage may be charged to the author. Additions at proof stage will not normally be allowed. Page proofs only will be sent. Word-processor discs. _ Please follow these instructions. 1. Ensure that the disc you send contains only the final version of the paper and is identical to the typescript. 2. Label the disc with the author’s name, title of the paper and the word-processor programme used. Indicate whether IBM or Apple Mac (IBM preferred). 3. Supply the file in the word-processor format; if there is a facility to save in ASCII please submit the file in ASCII as well. 4. Specify any unusual non-keyboard characters on the front page of the hard copy. 5. Do not right-justify the text. 6. Do not set a left-hand margin. 7. Make sure you distinguish numerals from letters, e.g. zero (0) from O; one (1) from | (el) and I. 8. Distinguish hyphen, en rule (longer than a hyphen, used without a space at each end to signify ‘and’ or ‘to’, e.g. the Harrison— Nelson technique, 91-95%, and increasingly used with a space at each end parenthetically), and em rule (longer than an en rule, used with a space at each end parenthetically) by: hyphen, two hyphens and three hyphens, respectively. Be consistent with rule used paren- thetically. 9. Use two carriage returns to indicate beginnings of paragraphs. 10. Be consistent with the presentation of each grade of heading (see Text below). Title. The title page should be arranged with the full title; name(s) of author(s) without academic titles; institutional address(es); sug- gested running title; address for correspondence. Synopsis. Each paper should have an abstract not exceeding 200 words. This should summarise the main results and conclusions of the study, together with such other information to make it suitable for publication in abstracting journals without change. References must not be included in the abstract. Text. All papers should have an Introduction, Acknowledge- ments (where applicable) and References; Materials and Methods should be included unless inappropriate. Other major headings are left to the author’s discretion and the requirements of the paper, subject to the Editors’ approval. Three levels of text headings and sub-headings should be followed. All should be ranged left and be in upper and lower case. Supra-generic systematic headings only should be in capitals; generic and specific names are to be in italics, underlined. Authorities for species names should be cited only in the first instance. Footnotes should be avoided if at all possible. References. References should be listed alphabetically. Authori- ties for species names should not be included under References, unless clarification is relevant. The author’s name, in bold and lower case except for the initial letter, should immediately be followed by the date after a single space. Where an author is listed more than once, the second and subsequent entries should be denoted by a long dash. These entries should be in date order. Joint authorship papers follow the entries for the first author and an *&’ should be used instead of ‘and’ to connect joint authors. Journal titles should be entered in full. Examples: (i) Journals: England, K.W. 1987. Certain Actinaria (Cnidaria, Anthozoa) from the Red Sea and tropical Indo- Pacific Ocean. Bulletin of the British Museum (Natural History), Zoology 53: 206-292. (ii) Books: Jeon, K.W. 1973. The Biology of Amoeba. 628 p. Academic Press, New York & London. (iii) Articles from books: Hartman, W.D. 1981. Form and distribution of silica in sponges. pp. 453-493. In: Simpson, T.L. & Volcani, B.E. (eds) Silicon and Siliceous Structures in Biological Systems. Springer- Verlag, New York. Tables. Each table should be typed on a separate sheet designed to extend across a single or double column width of a Journal page. It should have a brief specific title, be self-explanatory and be supple- mentary to the text. Limited space in the Journal means that only modest listing of primary data may be accepted. Lengthy material, such as non-essential locality lists, tables of measurements or details of mathematical derivations should be deposited in the Biological Data Collection of the Department of Library Services, The Natural History Museum, and reference should be made to them in the text. Illustrations DRAWINGS — Figures should be designed to go across single (84 mm wide) or double (174 mm wide) column width of the Journal page, type area 235 x 174 mm. Drawings should be in black on white stiff card with a line weight and lettering suitable for the same reduction throughout, ideally not more than 40%. After reduction the smallest lettering should be not less than 10 pt (3 mm). Tracing paper should ideally be avoided because of the possibility of shadows when scanned. All artwork must have bulletin, author and figure number included, outside of the image area, and must be free of pencil, glue or tape marks. PHOTOGRAPHS — All photographs should be prepared to the final size of reproduction, mounted upon stiff card and labelled with press-on lettering (eg Letraset). They can be mounted on white or black background; a black background must be evenly black all over; any background must be free of all pencil and glue marks within the image area. All figures should be numbered consecutively as a single series. Legends, brief and precise, must indicate scale and explain symbols and letters. Photos, when components of figure-plates should be abutted, trimmed as regular rectangles or close trimmed up to edge of specimen. Joins etc. can be removed at the scanning stage but at extra cost. Cropping instructions, if any, should be indicated on an overlay or marked on a photocopy of the figure. SIZE — Maximum size of artwork for use of flatbed scanners is A3. Larger artwork has to be reduced photographically prior to scanning, therefore adding to expense. Symbols in text. Male and female symbols within the text should be flagged somehow within curly brackets to enable setter to do a swift global search. Reprints. 25 reprints will be provided free of charge per paper. Orders for additional reprints can be submitted to the publisher on the form provided with the proofs. Later orders cannot be accepted. CONTENTS 1 Phylogenetic relationships of Toad-headed lizards (Phrynocephalus, Agamidae) based on morphology E.N. Arnold 15 Rita sacerdotum, a valid species of catfish from Myanmar (Pisces, Bagridae). C.J. Ferraris, Jr. 23 Indian Ocean echinoderms collected during the Sinbad Voyage (1980-81): 4. Crinoidea J.I.M. Crossland and A.R.G. Price 31. Onthe hybrid status of Rothschild’s Parakeet Psittacula intermedia (Aves, Psittacidae) PC. Rasmussen and N.J. Collar 51 A review of the genus Bargamannia Totton, 1954 (Siphonophorae, Physonecta, Pyrostephidae) PR. Pugh Bulletin of The Natural History Museum | ZOOLOGY SERIES Vol. 65, No. 1, June 1999