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