X

C/J

£ n
z

m ^ m ^ X^oqq^X fn

ITHSONIAN~INST!TUTiON£/? N0linillSNl“NVIN0SHimS S3 I HVH8 iVu B RAR ! ES SMITHS0NIAN_INS1
2 -v w z S’ 3 xy- £2 z

< 4Xn S Xs3lS$x 1 . . 5 <

<V4Wt,

•1'^^ I I | | 2 | l ^

IN0SHI!IAISWS3 I HVH 9 LI B RAR I ES^SMSTHSONIAN2 INSTITUTION NOSinillSNI NVINOSHlIWS^Sa I

m — ___ m 5 ___ « ___ 5 M

‘ O 'Xg _ __ ... _ ^ »\ j

IITHSOmAN^INSTITUTION^NOlinillSNI^NVINOSHimS S3 I HVH 8 II LI B RAR I ES SMITHSONIAN INS!
r- z r~ , z r° Z . "

2 XS$X 03 . 00 O

OT

— V

m -" w 7 vjvas>-^ rn x^ps^jz ™ ni

_ to ■*' — C/> £ w

HN0SH1IWS S3 1 HVH 811 LIBRARIES SMITHSONIAN INSTITUTION NOIlfUllSNi NVIN0SH1IINS S3 1
in z , w _ z » w 5 £?

| 4^^ 1 | 11^1 1 |

'l^P/ 1“ 2 >’* ~ z

IITHS0NIANZ|NSTITUTI0NC/,N0linillSNI_NVIN0SHllWS<°S3iaVMan LIBRARIES SMITHSONIAN INS1
— t n _ z X ^ 5 w z

&, I i £ JJgk s

c GOSSIP 2 '$&,>&■/ s life,,#*/ 2 t*? /vi- h

5 o 2 Xjvasv^X 5 X^f®^ “ ■’ O

Z J z ~™ J Z

IN0SH1IWS ^SBIBVHSSl LIBRARIES SMITHSONIAN INSTITUTION NOIifUllSNI NVIN0SH1IWS S3 I

S»^ I „.^aA !0^ 1 itiHfek 5 Jfc °

j= xb

ay

cn N^uiiT^X m xx dxx ^ '^S'r m x^iiqssx rn

/I ITHSOI^TfAN- INSTITUTION ^ NOIinillSNl”NVINOSHllWS rS3 I H VH 8 1 1~L I B R AR I ES SMITHSONIAN INS

XV ^ < x^sovTx S , < \Vv . > s ^osvaTx. <

i/INOSHHWS saiavuan LIBRARIES SMITHS0NIAN_INSTITUTI0N NOIinillSNI_NVINOSHllWS S3

tn ~ ay m — . -» c.v .■ ^

LxJ

NX ^

03 oq

O \»« oty ~ W' o nj< dx^X „ xfijus^X q X =

M ITHSOnIan”1 INSTITUTION^ NO I iniliSN I "^NVINOSHHIAIS S3 I HVH 9 11 LI B RAR I ES SMITHSONIAN^ I NS

w m >W' ™ m ^ rn

Z CO ' 3 CO _ w

IVIN0SHlllNSU,S3IHVHan LIBRARIES SMITHSONIAN INSTITUTION NOIifUllSNI NVINOSHlIWS^Sa

^ < Xj. S y<SsT»T(i^X. <C A S si&Mfr , S /oO A?5

rl

X HE, o >

;°*5c

m m m ^ 71 Njams^x m

05 2 05 £ 05 \ 5 05

N NOliniiiSNI NVIN0SH1IWS S3IHVH8n LIBRARIES SMITHSONIAN INSTITUTION NOliniiiSNI
'« 5" to z tft Z >v.. CO <

_ > -V” S 2 >

LI B RAR I ESt°SMITHSONIAN2|NSTITUTION</>NOIiniliSNI NVIN0SHI!lNs“s3 S HVB S LS B RAR I Es‘

“ 05 5 , 05 . OT 5 _ <

OJ 77 W 2 .4SS\ .5. OJ xtSwJTx ^ *

2

05

O NO* pC^X “ ^‘^ar o NO* DCJ^ “ Nl^-inS^X Q ^ " N^Uast^X O

'i NOliniiiSNI ~Winqshiiins2S3 iavaanJUB rar s eszsmithsoni’an”1institution 3NouniiiSNf

2 r° *, z r* 3? r* z t

i

l“Li B R AR I ESW SMITHSONIAN^INSTITUTION^NOIinillSNI^NVINOSHlIWS^Si I B VH 8 IT ~LI B RAR I ES

Z •_ 05 3= M» OS z •, £

I NOIinilISNI_NVINOSHl!WS S3IBVBan LIBRARIES SMITHSONIAN INSTITUTION NOliniUSNf

05 — » N 05 “ tfjp)i —= t fi Z

X- S • D ^?rirN Z

05

UJ

05

£0 .

05

UJ

i LIBRARIES SMITHSONIAN INSTITUTION NOliniiiSNI NVIN0SH1I1AIS S3IBVB8n LIBRARIES
«■=> z r~ z r- 5» m ;

05

>1 NOliniiiSNI NVIN0SH1IIAIS S3IHVH8n LIBRARIES SMITHSONIAN INSTITUTION NOliniiiSNI

^ W Z 05 Z c/5

z
o

05 ' 05 ^

1 LI BRAR i ES^SMSTHSONIAN^ INSTITUTlON^NOlinillSNI^NVINOSHilWS^SS I B V d 3 II ^ L I B R A R I Es‘

— m — CO ~ (/) — <

l

^ NOliniiiSNI ■JNVIN0SH1UNS^S3 I dVd 8 11 ^ LI B RAR I E SMITHSONIAN^ INSTITUTION 2 NOliniiiSNI
Z c , z r- z f~ z

r

~ t* — 05 « 05 2 c

1 LIBRARIES SMITHSONIAN INSTITUTION NOliniiiSNI NVIN0SH1IINS S3lUVH8n LIBRARIES

A

H Palaeontology

VOLUME 31 • PART 3 AUGUST 1988

Published by

The Palaeontological Association ■ London
Price £23 00

THE PALAEONTOLOGICAL ASSOCIATION

The Association was founded in 1957 to promote research in palaeontology and its allied sciences.

COUNCIL 1988-1989

President: Dr J. D. Hudson, Department of Geology, University of Leicester, Leicester LEI 7RH
Vice-Presidents: Dr L. B. Halstead, Department of Geology, University of Reading, Reading RG6 2AB
Dr P. W. Skelton, Department of Earth Sciences, The Open University, Milton Keynes MK7 6AA
Treasurer: Dr M. E. Collinson, Department of Biology, King’s College, London W8 7AH
Membership Treasurer: Dr H. A. Armstrong, Department of Geology, University of Newcastle,

Newcastle upon Tyne NE1 7RU

Institutional Membership Treasurer: Dr A. W. Owen, Department of Geology, University of Dundee, Dundee DD1 4HN
Secretary: Dr P. Wallace, The Croft Barn, Church Street, East Hendred, Oxon 0X12 8LA
Circular Reporter: Dr D. Palmer, Department of Geology, Trinity College, Dublin 2
Marketing Manager: Dr V. P. Wright, Department of Geology, University of Bristol, Bristol BS8 1RJ
Public Relations Officer: Dr M. J. Benton, Department of Geology, The Queen’s University of Belfast, Belfast BT5 6FB

Editors

Dr M. J. Benton, Department of Geology, The Queen’s University of Belfast, Belfast BT5 6FB
Dr J. E. Dalingwater, Department of Environmental Biology, University of Manchester,, Manchester M13 9PL
Dr D. Edwards, Department of Plant Sciences, University College, Cardiff CF1 1XL
Dr C. R. C. Paul, Department of Earth Sciences, University of Liverpool, Liverpool L69 3BX
Dr P. A. Selden, Department of Extra-Mural Studies, University of Manchester, Manchester M13 9PL
Dr P. D. Taylor, Department of Palaeontology, British Museum (Natural History), London SW7 5BD

Other Members

Dr J. A. Crame, Cambridge Dr C. Hill, London

Dr G. B. Curry, Glasgow Dr E. A. Jarzembowski, Brighton

Dr R. A. Spicer, London

Overseas Representatives

Australia: Professor B. D. Webby, Department of Geology, The University, Sydney, N.S.W., 2006
Canada: Dr B. S. Norford, Institute of Sedimentary and Petroleum Geology, 3303-33rd Street NW., Calgary, Alberta
Japan: Dr I. Hayami, University Museum, University of Tokyo, Hongo 7-3-1, Bunkyo-Ku, Tokyo
New Zealand: Dr G. R. Stevens, New Zealand Geological Survey, P.O. Box 30368, Lower Hutt
U.S. A.: Dr R. J. Cuffey, Department of Geology, Pennsylvania State University, Pennsylvania 16802
Professor A. J. Rowell, Department of Geology, University of Kansas, Lawrence, Kansas 66045
Professor N. M. Savage, Department of Geology, University of Oregon, Eugene, Oregon 97403
South America: Dr O. A. Reig, Departamento de Ecologia, Universidad Simon Bolivar, Caracas 108, Venezuela

MEMBERSHIP

Membership is open to individuals and institutions on payment of the appropriate annual subscription. Rates for 1988 are:

Institutional membership
Ordinary membership .
Student membership
Retired membership

£50-00 (U.S. $79)
£21-00 (U.S. $38)
£11-50 (U.S. $20)
£10-50 (U.S. $19)

There is no admission fee. Correspondence concerned with Institutional Membership should be addressed to Dr A. W. Owen,
Department of Geology, The University, Dundee DD1 4HN. Student members are persons receiving full-time instruction at
educational institutions recognized by the Council. On first applying for membership, an application form should be obtained
from the Membership Treasurer: Dr H. A. Armstrong, Department of Geology, University of Newcastle, Newcastle upon Tyne
NE1 7RU. Subscriptions cover one calendar year and are due each January; they should be sent to the Membership Treasurer.
All members who join for 1988 will receive Palaeontology , Volume 31, Parts 1-4. Back numbers still in print may be ordered
from Basil Blackwell, Journals Department, 108 Cowley Road, Oxford OX4 1JF, England.

Cover: The brachiopod Meristina obtusa (J. de C. Sowerby, 1823), a life position assemblage from the Much Wenlock
Limestone Formation, Abberley Hills, Hereford (Specimen no. BB52671, x 1). Photograph by Harry Taylor of the British

Museum (Natural History) Photographic Studio.

RARE TETRAPOD REMAINS FROM THE
LATE TRIASSIC FISSURE INFILLINGS OF
CROMHALL QUARRY, AVON

By N. C. FRASER

Abstract. Disassociated assemblages from the Mesozoic of South-west Britain display considerable variation
both in the numbers of species present and in their distribution. Triassic fissure deposits at Cromhall Quarry,
Avon have yielded abundant reptilian remains which for the most part are readily identified to generic level.
These sediments have also revealed some very rare and quite unusual skeletal elements, including jaw bones
and a procoelous vertebra. These could be prolacertiform, thalattosaurian, or pterosaurian remains, but the
nature of the material makes taxonomic diagnoses difficult.

Vertebrate-bearing Mesozoic fissure deposits are widespread throughout the Avon and South
Glamorgan areas, and most probably range in age from the Norian to Sinemurian (Fraser 1985).
Research has largely centred upon the abundant mammalian remains since they constitute some
of the earliest known members of the group (D. M. Kermack et al. 1956, 1968; K. A. Kermack
et al. 1973, 1981 ). However, the sediments are also notable for the wealth of small reptilian remains
which have only recently been studied in some detail (Evans 1980, 1981; Fraser 1982; Fraser and
Walkden 1983, 1984; Crush 1984; D. Kermack 1984; Whiteside 1986).

Generally, the fossils occur as highly concentrated assemblages of completely disassociated
bones, which are frequently quite fragmentary, although some exquisite articulated and associated
skeletons are known (e.g. D. Kermack 1984; Fraser, in press). In terms of the quantity of material
and total numbers of different genera at a single locality, Cromhall Quarry (ST 704 916) is perhaps
the most prolific of the English localities. Here, the occurrence of ten or more different species
within a single stratum is commonplace and the individual fragments of each species must be
separated. To a large extent, the most abundant species can be restored with some confidence. In
the first instance, the relative abundance of individual elements forms a useful guideline for the
recognition of each species; and then the nature and orientation of articulation facets can be
analysed to test the suspected associations (see e.g. Fraser 1982). But with the rarest species,
represented by the occasional isolated element, it may prove impossible to deduce precise
relationships, but they should be properly documented to complete the record of the assemblages.
The purpose of this paper is to describe some of these rare elements from the Cromhall assemblages.

THE CROMHALL ASSEMBLAGES

The series of fissures at Cromhall Quarry and their associated Mesozoic reptile faunas are well
documented (Robinson 1957; Fraser and Walkden 1983; Fraser 1985). The most abundant genera
are two sphenodontids Planocephalosaurus (Fraser 1982; Fraser and Walkden 1984) and Clevosawus
(Robinson 1973; Fraser, in press). Two rarer sphenodontid genera are sufficiently abundant to
allow partial descriptions and the definition of some diagnostic characters (Fraser 1986). A fifth
sphenodontid, Diphydontosaurus, described by Whiteside (1986) from abundant remains at the
neighbouring locality of Tytherington Quarry, is relatively common. There are also isolated
fragments of Kuehneosaurus, a gliding diapsid reptile described by Robinson (1962) from similar
fissure localities in Somerset. Included within the material awaiting full description there are well-
preserved specimens of a procolophonid and abundant archosaurian remains. The latter include

| Palaeontology, Vol. 31, Part 3, 1988, pp. 567-576.|

© The Palaeontological Association

568

PALAEONTOLOGY, VOLUME 31

C

text-fig. 1. The fused premaxillae AUP 1 1305 in a, lateral, b, dorsal, and c,
ventral aspects. The scale bar represents 0-5 mm.

a terrestrial crocodile and two thecodontians. On the basis of various diagnostic criteria, twelve
distinct reptilian taxa have been recognized, and their taxonomic relationships can be at least
partially assessed. By contrast, a few quite characteristic elements have been recovered that are
extremely rare indeed. From 1-5 tonnes of rock processed at Aberdeen University Geology
Department, which have yielded in the region of 10 000 identifiable bone fragments, two different
types of premaxillae, two maxillae, and a procoelous vertebra are exceptionally rare— only six
specimens of the vertebra have been found, and there are even fewer examples of the four jaw
bones. By contrast, the same quantity of sediment produced 150 Planocephalosaurus maxillae and
120 premaxillae. The rare forms are quite distinct from the more ubiquitous genera in the deposits,
and they are consequently very difficult to treat taxonomically. It is undesirable to erect new genera
or species on such isolated material, yet they merit description as additional taxa.

JAW BONES

Premaxilla I

Five specimens of a long, slender, bilaterally symmetrical bone represent fused premaxillae (text-fig. 1). Four
originate from levels M, K, and L of site 4, and one from Fevel A of site 5 (for details of the fissure
stratigraphy and nomenclature, see Fraser 1985). The largest specimen is 6 0 mm long and the smallest
4-5 mm. Two tooth rows are exposed in ventral aspect. They meet at the sharply angled anterior end, but
diverge somewhat posteriorly to leave a narrow channel between the two dental rami (text-fig. lc). In the
few instances where the teeth are preserved, they are acutely conical and set in very shallow alveoli which
have a slightly higher lateral than medial wall. When restored, it is estimated that there were between ten
and twelve tooth positions in each row. Each tooth alveolus is produced into a slight lateral bulge so that
in dorsal view the margins of the bone are faintly scalloped (text-fig. 1b). In lateral aspect, the bone exhibits
a low profile, and both sides are deeply cmarginated posteriorly by separate openings, presumably representing
the external nares (text-fig. I a). The posterior boundaries of the bones are incomplete in all five specimens;
as a result the full extent of the bone above and below each narial opening is unknown. Nevertheless, in one
specimen (AUP 11305), the posterior process passing beneath the left naris appears to be almost complete
(text-fig. 1a). On the dorsolateral surface of this process there is a small notched facet which presumably
formed the articulation with the maxilla, and indicates a limited contact between the two elements (text-fig.
1b). Each specimen bears a variable number of small nutrient foramina, usually three or four on each side,
which lie in a line just above the level of the tooth rami.

The general outline of this element is most reminiscent of a pterosaur. However in pterosaurs,
including the known Norian rhamphorhynchoid forms (Wild 1978), the ventral border of the
external naris is almost entirely formed by the maxilla and there are characteristically only three

FRASER: TRIASSIC FISSURE REPTILES

569

text-fig. 2. Rhamphorhynchoid pterosaur skulls in lateral
aspect. A, Eudimorphodon and b, Dorygnathus (after Wild
1978).

or four premaxillary teeth (text-fig. 2). The tooth implantation of pterosaurs is generally considered
to be thecodont or possibly subthecodont (Edmund 1969; Wild 1978). In the element under
discussion there is insufficient depth of bone to support a ‘deep-rooted’ thecodont dentition.
Bearing in mind that the lateral wall of the dental groove appears to be slightly higher than the
medial side, there is reason to speculate that the tooth implantation may be a modified subthecodont
type correlated with the low lateral profile and miniaturization of the jaw.

The tooth morphology and implantation is similar to Kuehneosaurus , but the overall shape of
the element is quite different. The elongated form is not dissimilar to a miniature crocodile or
thalattosaur (text-fig. 3). However in crocodiles, the nares are generally terminal and face dorsally.

A

B

text-fig. 3. Thalattosaur skulls in lateral view, a, Thalat-
tosaurus and b, Askeptosaurus. (a, after Merriam 1905; b,
after Kuhn (-Schynder) 1952.)

570

PALAEONTOLOGY, VOLUME 31

text-fig. 4. Maxilla I. AUP 1 1293 in a, lateral aspect and b, medial view. Scale bar represents 0-5 mm.

pal. sf

pm. sf

pal

Tooth implantation in thalattosaurs apparently varies from thecodont in Askeptosaurus and
Thalattosaurus (Kuhn (-Schnyder 1952), to acrodont in Hescheleria (Peyer 1936/d, and either
pleurodont or acrodont in Clarazia (Peyer 1936a; Rieppel 1987). In addition, the known
thalattosaurs are much larger than the material under discussion, the premaxillae are apparently
separate, and the premaxillary dentition is restricted to the anterior part of the element.

Maxilla I

A maxilla of a size and form consistent with the fused premaxillae is represented by four specimens, all from
Site 4 (Levels M, K, and J). It is a relatively short but slender element (text-fig. 4) not exceeding 5 mm long,
and when restored probably possessed a maximum of twelve teeth. In all specimens, the rather short ascending
process is incomplete. It bears a facet on its medial aspect where it presumably overlapped the nasal or
prefrontal (text-fig. 4b). There is an additional notched facet, positioned towards the posterior margin on the
lateral face of the ascending process (text-fig. 4a). It is quite conceivable that this facet received the lachrymal
or prefrontal, and this in turn suggests that an antorbital fenestra was unlikely. Judging by the gentle
posterodorsal slope and slight concavity of the anterior margin of the bone, the external nares were elongate.
In medial view there is a prominent faceted flange set obliquely to the anterior edge of the dental groove
(text-fig. 4b). This presumably formed the articulation with the premaxilla (or possibly the vomer). Posteriorly,
the element broadens into a medial shelf which is poorly preserved in all four specimens, although it
presumably formed an articulation with the palatine. Immediately above this shelf there is a fairly prominent
foramen, the suborbital foramen, which transmitted the palatine nerves and blood vessels. Where preserved,
the teeth are acutely conical and only slightly recurved. They are circular in cross-section and appear hollow
and thin-walled. The implantation is of the same type as the fused premaxillae described above.

In terms of overall structure, tooth morphology, and size, it is tempting to suggest that these
maxillae belong to the same species as the fused premaxillae. Their relative abundance and
distribution within the deposits is also consistent with this view. However, because the material is
so scarce the link between the two elements remains tenuous.

Premaxilla II

The two remaining jaw bones to be described are a single premaxilla and an isolated maxilla, both from
Level M of Site 4, and both having similar tooth implantation to the forms described above.

The premaxilla is from the left side, and the entire tooth ramus would appear to be present, consisting of
nine alveoli (text-fig. 5b). Four teeth are preserved, three complete, and one missing the distal end; they are
ankylosed at every other tooth position. Within the constraints of current inadequate definitions, the tooth
implantation is best described as a shallow subthecodont type — each tooth set in a very shallow depression
and with a slightly higher lateral than medial wall. The teeth themselves are subcircular in cross-section, and
they are only very slightly recurved. The smooth surfaces of the teeth are relieved by fine longitudinal
striations covering the distal third of each complete tooth. In lateral profile, the anterior margin of the bone
is straight and extends posterodorsally at an angle of approximately 45° to the dental ramus (text-fig. 5a).
The medial surface forms an elongate, almost vertical, symphysis (text-fig. 5b) that presumably articulated
with its counterpart, and together they would have formed an acutely pointed snout. The bone is emarginated

FRASER: TRIASSIC FISSURE REPTILES

571

text-fig. 5. Premaxilla II. AUP 1 1294 in a, lateral and b, medial aspect. Scale bar represents 0-5 mm.

posteriorly by the external naris. The full extent of the process above the naris is unknown. Ventral to the
naris the element is developed into a short medially directed ledge. A shallow depression on the dorsolateral
surface of this ledge is satisfactorily interpreted as the maxillary facet. Situated immediately anterior to the
narial opening, a posteriorly facing foramen probably transmitted branches of the maxillary artery and nerve.

In general terms, the outline of the premaxilla is perhaps most like a prolacertiform. However,
in macrocnemid prolacertiforms at least, the external nares are placed further up on the dorsal
surface of the snout and the premaxillae meet the maxillae in extended sutures. Tanystropheus is
similar to the macrocnemids in this respect (text-fig. 6a). Although the arrangement in Prolacerta
is perhaps closest to the new form (text-fig. 6b), the premaxillary tooth count of Prolacerta , like
Tanystropheus , rarely exceeds five. The tooth implantation of the new form is comparable
with the kuehneosaurids, a pattern which Robinson (1962) and Colbert (1970) referred to as
subpleurodont. Wild (1973, 1980) also classifies the teeth of Macrocnemus and Tanystropheus as
subpleurodont (or pleurothecodont), yet the tooth implantation of these two genera is rather
different from the kuehneosaurids. Definitions of tooth implantation need to be much stricter if
comparisons between the dentitions of such genera are to be meaningful.

Maxilla II

The last jaw element to be described here is interpreted as a left maxilla (text-fig. 7). The bone is preserved
as two fragments, but only the extreme anterior and posterior limits of the bone are missing. There are eight
partially preserved teeth and a total of ten tooth positions. The teeth are acutely conical, slightly recurved,
and display an overall similarity to those of kuehneosaurids and the dentitions already described. The most
notable characteristic of the teeth is their exceptional size relative to the depth of the bone, yet they are only
ankylosed in shallow alveoli by a minimum of spongy bone of attachment. Longitudinal striae are most
pronounced towards the distal extremities of the teeth, and the lateral wall of the dental groove is marginally
higher than the lingual wall. An exceptionally narrow ascending process bears no obvious prefrontal
or lachrymal facets, and this may indicate the presence of an antorbital fenestra. The short section of
the dental ramus extending anterior to the ascending process exhibits a marked medial flexure. This hints at
a snout that was somewhat shorter and blunter than those species represented by the two premaxillae
described above. On the medial surface, approximately a third of the length from the anterior end of the
specimen, there is a prominent foramen which presumably transmitted the palatal vessels. Immediately below

text-fig. 6. Prolacertiform skulls in lateral view, a, Tanystropheus and B, Prolacerta. (a, after Wild 1978; B,

after Kuhn (-Schynder) 1952.)

572

PALAEONTOLOGY, VOLUME 3 1

A

B

text-fig. 7. Maxilla II. AUP 1 1303, in a, lateral and b, medial view. Scale bar represents 0-5 mm.

the foramen, the bone is developed into a faceted medial shelf which is considered to have contributed to
the palatine articulation. Further posteriorly the element bears an elongate slot facet on the external surface.
The jugal might be expected to articulate with the maxilla in this region, and there is apparently no other
potential jugal facet. Nevertheless some doubt exists since the articulation between these two elements in
other reptiles is more usually located on the medial surface of the maxilla. If this particular species possessed
an antorbital fenestra, it is possible that the facet could have received the lachrymal and that the jugal facet
is not preserved in this specimen. In any event, the evidence suggests that this new maxilla represents a form
with a lightly built, highly fenestrated skull such as that exhibited by the pterosaurs or the 'thecodontian'
Megalancosaurus (Calzavara et al. 1980).

I have already mentioned that current definitions of reptilian tooth implantation are somewhat
nebulous. Consequently, in the case of the new jaw material a consideration of tooth implantation
as a diagnostic characteristic is not thought to be appropriate. Nevertheless, recurved teeth have
been considered as one of the characters of the archosaur/prolacertiform group of diapsid reptiles
(Benton 1985) (cf. the peg-like teeth of the outgroups Rhynchosauria and Lepidosauromorpha),
and certainly the dentitions described herein are generally somewhat recurved and acutely conical.
It may seem somewhat anomalous to imply archosauromorph relationships for the new jaw bones
when they were also shown to be comparable with kuehneosaurid dentitions (a group normally
supposed to have squamate affinities) (Robinson 1962, 1967; Carroll 1977; Estes 1983). However,
Evans (1984) pointed out that kuehneosaurids lack the basic lepidosauromorph characters of
single-headed ribs on all dorsal vertebrae, accessory facets on the neural arch, and postfrontals
entering into the borders of the upper temporal fenestrae. Benton (1985) also expressed some
doubts concerning the assignment of the Kuehneosauridae to the Lepidosauromorpha, and there
is good reason to suppose that they may have closer affinities to the Archosauromorpha. These
include reduction of the postfrontal, the laterally placed carotid foramina, and the contribution of
the basisphenoid to the lateral walls of the braincase. Unfortunately, the ankle joint, which is
crucial to the question, is unknown in all kuehneosaurs. The rarity and very fragmentary nature
of the new material does not permit a detailed taxonomic study. Generally these jaw bones exhibit
a mosaic of characteristics which cannot be readily reconciled with any one particular taxon. It is
also likely that the overall features are associated with adaptations towards miniaturization and
insectivory and they are therefore not necessarily indicative of taxonomic affinities.

The Procoelous vertebra

Different jaw bone types are readily identifiable within the assemblages, and variation in dental
morphology is at least a good indicator of the number of genera, if perhaps not necessarily
diagnostic. By contrast, it is by no means apparent with which other elements in a disassociated
assemblage isolated postcranial bones should be grouped. This can be particularly true of the axial
skeleton where some taxa are known to exhibit marked variation in basic structure within the
length of the vertebral column (e.g. the Chelonia, where the cervicals may be a mixture of

FRASER: TRIASSIC FISSURE REPTILES

573

table I . The distribution of the procoelous vertebrae and small jaw bones within the Cromhall fissure
deposits. (For details of fissures and horizons see Fraser 1985.)

Total

Site 4
Level J

Level K

Level L

Level M

Site 5
Level A

Premaxilla I

5

2

1

1

1

Maxilla I

4

1

1

2

Premaxilla II

1

1

Maxilla II

1

1

Procoelous vertebra

6

1

1

2

2

procoelous, amphicoelous, and opisthocoelous). Therefore, the occurrence of a most unusual and
rare procoelous vertebra within the Cromhall assemblages poses its own special problems.

The great majority of vertebrae in the assemblages are of the amphicoelous or notochordal
amphicoelous type, but the new specimens are quite distinctive and it is not clear whether they are
representative of a species partially described previously on the basis of other material, or indicate
the occurrence of a new form. The six specimens are of uniform size, attaining a length of 6 mm,
a height of 4 mm, and a width of 4 mm. These dimensions are likely to be consistent with the
species represented by the fused premaxillae, and the occurrences of the two elements follow similar
distribution patterns (Table 1). Although it is tempting to suggest that they may represent the
same species, there is no other evidence to support this view. All six specimens have an identical
structure, and they are therefore assumed to originate from exactly the same region of the vertebral
column. In addition, the lack of any further remains of procoelous vertebrae strongly suggests
that the remainder of the vertebral column may have been more typical, and perhaps fragments
of indeterminate amphicoelous vertebrae are representative of the major portion of the axial
skeleton. Other workers have noted that there is a tendency for small braincases to exhibit a certain
degree of structural convergence towards vertebrae (A. R. I. Cruickshank and O. Rieppel, pers.
comm.), and the possibility that these specimens might represent a rather unusual braincase has
been investigated. Whilst certain features can be reconciled with such an identification (e.g. a
possible parasphenoid rostrum), there are no apparent paroccipital processes, and the specimens
are unreservedly considered to be vertebrae by virtue of the definite anterior and posterior
articulation facets.

The new vertebra (text-fig. 8) is rather elongate, a condition accentuated by the extension of the centrum
posteriorly beyond the level of the zygapophyseal articulation. The diameter of the neural arch is some two
to three times that of the centrum, the latter taking the form of a slender conical frustum. A narrow keeled
hypopophysis, produced below the centrum, is incomplete in all specimens, but it appears to have extended
beyond the intercentral articulation so that it passed under the anterior end of the succeeding vertebra. The
procoelous intercentral articulation is unusual in that the anterior concavity, the cotyle, is approximately
kidney-shaped, and it is inclined ventrally. The opposing convex posterior facet, the condyle, is saddle-like
and faces posterodorsally. The overall intercentral articulation is therefore rather like the heterocoelous
condition in birds, but lacking the bilateral expansions of the cotyle and condyle. The zygapophyses are quite
unusual in that they are inclined towards the vertical plane. This would have tended to restrict lateral
movement, but at the same time facilitated flexure of the vertebral column in the vertical plane. The level of
the zygapophysial articulation is set forward from the intercentral articulation. There are no accessory
intervertebral articulations comparable to those of lepidosauromorphs. There appear to be separate
diapophyses and parapophyses. The diapopysis, although incomplete in all specimens, apparently formed a
short pedicel with a small circular distal rib facet. A short bony ridge connects this pedicel to a V-shaped
articular surface which is presumed to be the parapophysis. The apex of the putative parapophysis is directed
anteriorly and is situated immediately above and lateral to the cotyle on the centrum. This particular

574

PALAEONTOLOGY, VOLUME 31

A

B

pre

pop

pre.

C □ —

text-fig. 8. The procoelous vertebra. AUP 11362 in a, lateral, b, dorsal, and
c, ventral aspects. Scale bar represents 05 mm.

arrangement is also consistent with the view that these V-shaped articular surfaces represent pre-exapophyses,
but the apparent lack of complementary postexapophyses does not lend any further support to this
identification.

The affinities of these specimens are not immediately apparent. The procoelous condition
approaches the heterocoelous articulation of birds, but they are not identical since laterally the
cotyle and condyle flare considerably in birds. On the one hand, separate parapophyses and
diapophyses are more generally associated with archosauromorphs than lepidosauromorphs, and
the lack of accessory intervertebral articulations on the mid-line of the neural arch provides further
support for an assignment to the archosauromorphs. On the other hand, affinities with non-diapsid
groups cannot be discounted.

It is interesting to note certain similarities between the new vertebra and the cervical vertebrae
of Pterodactyloidea, as described by Howse (1986). In particular, they share a shallow centrum
extending posteriorly well beyond the limits of the postzygapophyses. Howse noted that Creta-
ceous pterodactyloids were normally characterized by the presence of exapophyses associated with
the cotyle and condyle, and a hypopophysis situated towards the anterior ventral surface of the
centrum. Whilst there is a remote possibility that exapophyses are present in the new vertebra, the
hypopophysis is positioned on the posterior ventral surface of the centrum, and although the new
vertebra may possess certain characters indicative of pterodactyloid affinities, age considerations
are not consistent with this view. The known Triassic pterosaurs belong to the Rhamphorhyncho-
idea, and peterodactyloids do not appear in the geological record until the Upper Jurassic.
Rhamphorhynchoid cervical vertebrae are immediately distinguishable from those of pterodactyl-
oids (Howse 1986). Apart from the procoelous nature of the pleurocentral articulation, the only
character that the new vertebra might conceivably share with rhamphorhynchoids is the possible
occurrence of pneumatic foramina. Immediately below the pedicel of the neural arch, each of the
new specimens exhibits either one or two small foramina which may lead into larger internal
excavations.

SUMMARY

Isolated elements from a disassociated vertebrate assemblage are difficult to treat taxonomically.
Often the rarest components of such assemblages are only recognizable from jaw bone fragments,

FRASER: TRIASSIC FISSURE REPTILES

575

yet their structure alone is generally insufficient to enable us to make substantial claims with regard
to their relationships. Although jaw elements may exhibit certain diagnostic characteristics, they
also reflect dietary habits, and it has been shown here that the use of reptilian tooth implantation
as a fundamental taxonomic criterion is open to criticism. Accordingly, only very broad taxonomic
statements have been made with respect to the rarest faunal elements, but the possible occurrence
of prolacertiform, thalattosaurian, or pterosaurian remains within the Cromhall assemblages
should not be overlooked.

Acknowledgements. I should like to thank Drs M. J. Benton, A. R. I. Cruickshank, P. J. Currie, S. E. Evans,
R. E. Molnar, O. Rieppel, H.-D. Sues, and R. Wild for their helpful comments on the identification of the
bones. The management of Amey Roadstone Corporation Ltd. kindly provided access to Cromhall Quarry.
I thank Girton College, Cambridge for the financial support of a Research Fellowship.

REFERENCES

benton, m. j. 1985. Classification and phytogeny of the diapsid reptiles. Zool. J. Linn. Soc. 84, 97-1 64.
calzavara, m., muscio, G. and wild, r. 1980. Megalancosaurus preonensis n.g., n.sp., a new reptile from the
Norian of Friuli, Italy. Gortania. Atti Museo Friul. Storia not. 2, 49 64.
carroll, r. l. 1977. The origin of lizards. In Andrews, s. m„ miles, r. s. and walker, a. d. (eds. ). Problems
in vertebrate evolution. Linn. Soc. Symp. Ser. 4, 359-396.
colbert, E. H. 1970. The Triassic gliding reptile Icarosaurus. Bull. Am. Mus. nat. Hist. 143, 89 142.
crush, p. j. 1984. A late Triassic sphenosuchid crocodilian from Wales. Palaeontology , 27, 131 157.
edmund, a. g. 1969. Dentition. In gans, c., bellairs, a. d’a., and parsons, t. s. (eds.). Biology of the Reptilia
1, 117-200. Academic Press, London.

estes, R. 1983. Encyclopedia of Paleoherpetology. 10a, Sauria terrestria , Amphisbaenia. Gustav Fischer Verlag,
Stuttgart.

evans, s. e. 1980. The skull of a new eosuchian reptile from the Lower Jurassic of South Wales. Zool. J.
Linn. Soc. 70, 203-264.

— 1981. The postcranial skeleton of the Lower Jurassic eosuchian Gephyrosurus bridensis. Ibid. 73, 81-
116.

1984. The classification of the Lepidosauria. Ibid. 82, 87-100.

fraser, n. c. 1982. A new rhynchocephalian from the British Upper Trias. Palaeontology , 25, 709-725.

1985. Vertebrate faunas from Mesozoic fissure deposits of South-west Britain. Mod. Geol. 9, 273-300.

1986. New Triassic sphenodontids from South-west England and a review of their classification.
Palaeontology, 29, 165-186.

— In press. The osteology and relationships of Clevosawus (Reptilia: Sphenodontida). Phil. Trans. R.
Soc. B.

— and walkden, g. m. 1983. The ecology of a late Triassic reptile assemblage from Gloucestershire,
England. Palaeogeogr., Palaeoclimatol., Palaeoecol. 42, 341 365.

— 1984. The postcranial skeleton of Planocephalosaurus robinsonae. Palaeontology , 27, 575-595.
howse, s. c. b. 1986. On the cervical vertebrae of the Pterodactyloidea (Reptilia: Archosauria). Zool. J. Linn.
Soc. 88, 307-328.

kermack, d. 1984. New prosauropod material from South Wales. Ibid. 82, 101-117.
kermack, d. m., kermack, k. a. and mussett, f. 1956. New Mesozoic mammals from South Wales. Proc.
geol. Soc. Lond., 1533, 31.

— 1968. The Welsh pantothere Kuehneotherium praecursoris. Zool. J. Linn. Soc. 47, 407-423.
kermack, k. a., mussett, f. and rigney, H. w. 1973. The lower jaw of Morganucodon. Ibid. 53, 87 175.
1981. The skull of Morganucodon. Ibid. 71, 1-158.

kuhn (-schynder), e. 1952. Askeptosaurus italicus Nopcsa. In peyer, b. (ed.). Die Triasfauna der Tessiner
Kalkalpen XVII. Schweiz, palaont. Abh. 69, 1 73.

merriam, j. c. 1905. The Thalattosauria, a group of marine reptiles from the Triassic of California. Mem.
Calif. Acad. Sci. 5, I -52.

peyer, b. 1936n. Die Triasfauna der Tessiner Kalkalpen, X. Clarazia schinzi nov. gen. nov. sp. Schweiz,
palaont. Abh. 57, 1 61.

1936 b. Die Triasfauna der Tessiner Kalkalpen, XI. Hescheleria rubeli nov. gen. nov. sp. Ibid. 58, 1 48.

576

PALAEONTOLOGY, VOLUME 31

rieppel, o. 1987. Clarazia and Hescheleria: a re-investigation of two problematical reptiles from the middle
Triassic of Monte San Giorgio (Switzerland). Palaeontographica A , 1987, 101-129.
robinson, p. L. 1957. The Mesozoic fissures of the Bristol Channel area and their vertebrate faunas. Zool. J.
Linn. Soc. 43, 260-282.

- 1962. Gliding lizards from the upper Keuper of Great Britain. Proc. geol. Soc. Loud. 1061, 137-146.
1967. Triassic vertebrates from lowland and upland. Sci. Cult. 33, 169-173.

- 1973. A problematic reptile from the British Upper Trias. J. geol. Soc. Lond. 129, 457 479.
whiteside, d. i. 1986. The head skeleton of the Rhaetian sphenodontid Diphydontosaurus avonis gen. et sp.

nov. Phil. Trans. R. Soc. Lond. B , 312, 379-430.

wild, r. 1973. Die Triasfauna der Tessiner Kalkalpen XXIII. Tanystropheus longobardicus (Bassani) (Neue
ergebnisse). Schweiz, palaont. Abh. 95, 1-162.

1978. Die Flugsaurier (Reptilia, Pterosauria) aus der Oberen Trias von Cene bei Bergamo, Italien. Boll.
Soc. paleont. ital. 17, 176-256.

Typescript received 29 May 1987
Revised typescript received 3 July 1987

N. C. FRASER

Department of Zoology
Downing Street
Cambridge CB2 3EJ

ABBREVIATIONS

cond

condyle

n./pfr. f

nasal and/or prefrontal facet

cot

cotyle

n.s

neural spine

d. pop

diapophysis

pal. sf

palatine shelf

ex. n

external naris

pm. sf

premaxillary shelf

fo. mx

foramen for maxillary artery and nerve

post, zy

postzygapophysis

h. pop

hypopophysis

P pop

parapophysis

1. f

possible lachrymal facet

pre. zy

prezygapophysis

m.c

median channel

s. orb. fo

suborbital foramen

mx. f

maxillary facet

symp

symphysis

n. fo

nutrient foramen

to. al

tooth alveolus

HYPOSTOMES AND VENTRAL CEPHALIC
SUTURES IN CAMBRIAN TRILOBITES

by H. B. WHITTINGTON

Abstract. Restorations of the cephala of species of each of eighteen genera show the hypostome and cephalic
sutures; new photographs are given of these features in Holmia, Bathynotus , Paradoxides , Fieldaspis ,
Ptychoparia , Conocoryphe , and Agraulos. It is considered that probably in all trilobites the tip of the upwardly
directed anterior wing of the hypostome was situated close beneath the ridge formed on the internal surface
of the cephalon by the axial furrow, in a position immediately in front of where the eye ridge or eye lobe
met this furrow. This position of the hypostome may be observed in species in which the hypostome was
attached either by a suture to the cephalic doublure, or fused to the rostral plate. In species in which the
hypostome was detached from the cephalic doublure it is assumed that it was situated in a morphologically
similar position. In forms in which it was attached, the hypostome was thus braced against the dorsal
exoskeleton of the cephalon so that movement was not possible; such movement was probably restricted in
detached forms. During development the close connection between anterior wing and a particular site in the
axial furrow was maintained, hence the hypostome may have been detached in the early stages but attached
in the holaspis, or vice versa. Fusion of hypostome to rostral plate in holaspids is known only in Cambrian
trilobites. Progressive reduction in transverse width of the rostral plate, culminating in a median suture, is
not known in an evolutionary series. Until more is known of the hypostome, cephalic doublure, and ventral
sutures in Cambrian trilobites, these features will have only limited value in discriminating familial and higher
taxa, compared with their importance in such characterization of post-Cambrian forms; this particularly
applies to species having the hypostome detached.

A review by Rasetti (1952), and the Treatise (Harrington in Moore 1959, figs. 42.1-13, 44, 486-
d, g, i) give only ventral (external) views of a limited range of hypostomes of Cambrian trilobites.
Hence the convexity of the hypostome, the size and inclination of the anterior wing, and how the
hypostome was related to the rest of the cephalic exoskeleton, are not revealed and their significance
cannot be appreciated. A limited amount of silicified Cambrian material (that retains the original
convexity) has been isolated from the matrix and described, a few entire exoskeletons with the
hypostome in place illustrated, and isolated hypostomes recorded and figured. In many of the
latter the anterior wing is either hidden in shadow, or has not been exposed from the matrix. The
appearance of the dorsal (internal) side of the hypostome is virtually unknown, and there have
been only limited attempts to excavate the doublure of the free cheek to show if and where it may
have ended at a median or a connective suture. In addition, lateral or anterior views which show
the convexity of the cephalon of uncompressed specimens are lacking in many publications. The
present review embraces taxa selected to represent major groups (orders and superfamilies of the
Treatise) of Cambrian trilobites for which information is adequate to provide a reconstruction of
the cephalic exoskeleton, and covers a wide range of morphology. Sources are mainly post- 1959,
supplemented by new investigations (text-fig. 2; Pis. 52-55). The results are summarized in text-
figs. 1, 3, 5-20, drawings that show not only dorsal and ventral aspects of the cephalon, but a
right lateral view and a sagittal section combined with a right lateral view of the hypostome. The
latter also includes a heavy dashed line in the position of the crest of the ridge formed on the inner
surface of the cephalon by the axial furrow. This line helps to show how the attitude and position
of the hypostome were related to the size and form of the anterior wing, and the position of the
axial furrow. It may also be seen that knowledge of the convexity of the exoskeleton is essential
to making the reconstruction, and that an estimation may be made of the probability that the

| Palaeontology, Vol. 31, Part 3, pp. 577-609, pis. 52-55.|

© The Palaeontological Association

578

PALAEONTOLOGY, VOLUME 31

hypostome was attached by a suture to the cephalic doublure, or was detached from it. All the
requisite information for such reconstructions may not be known for a single species (preferably
the type) of a particular genus. In such cases I have combined information from two or more
species into a drawing of an indeterminate species, for which a generalized stratigraphical range
is given. The descriptive section deals with the basis for each figure and elements of uncertainty
that may obtain. The investigation has shown that in forms in which rostral plate and hypostome
were fused, this fused sclerite was firmly braced by the anterior wing of the hypostome against the
rest of the cephalic exoskeleton. In species that had the hypostome attached at a hypostomal
suture, the anterior wing appears to have fulfilled a similar role. In perhaps the majority of species
of Cambrian trilobites, the hypostome was detached from the rest of the exoskeleton and inserted
into the un-mineralized ventral integument; muscles linking the anterior wing to the dorsal
exoskeleton served to hold it in place. The reasoning lying behind these findings, and their
implications, are discussed, together with the relationships between them and those on the
hypostomes of post-Cambrian trilobites (Whittington, in press).

FIGURES AND TERMINOLOGY

Text-figs. 1, 3, 5-20, give a, a partial dorsal, B, a partial ventral, and c, a right lateral view of the cephalic
exoskeleton; d is a sagittal section of this exoskeleton combined with a right lateral view of the hypostome.
Each sagittal section incorporates a gap where it is crossed by a suture, to clarify the position of the suture
(no gap is shown where the section follows a median suture). A heavy dashed line indicates the position of
the crest of the ridge formed by the axial furrow on the inner surface of the exoskeleton. Scale bars are in
millimetres. Certain of these figures are of indeterminate species, being based on more than one species
assigned to the genus; the stratigraphical range given is that of species of the genus. Such ranges follow the
definitions of Lower, Middle, and Upper Cambrian by Palmer (1977).

Abbreviations used in the plates and text-figures are listed below, and are for terms used earlier (Whittington
and Evitt 1954, p. 13; Harrington el al. in Moore 1959). An attached hypostome was one that was attached
to the cephalic doublure and/or the rostral plate at the hypostomal suture, or was fused to the rostral plate;
a detached hypostome was not so attached, but inserted into the un-mineralized ventral integument of the
cephalon. Harrington (in Moore 1959, p. 058) used the term rostral-hypostomal plate for the fused rostral
plate and hypostome (Henningsmoen 1959, p. 157, proposed ‘rostri-hypostomal plate’); in text-figs. 8 and
10, the two portions of this plate are labelled separately.

Abbreviations used, aw, anterior wing, subtriangular or rounded extension of anterolateral border of
hypostome; cs, connective suture, one of pair of sutures that extend from junction of rostral and facial suture
to inner margin of cephalic doublure, and hence isolates the rostral plate; d, doublure of cephalon; gr, genal
ridge, the low ridge that runs from the inner, anterior corner of the gena out toward the genal angle; /;,
hypostome, mineralized plate on anterior, ventral surface of cephalon; hs, hypostomal suture separating
hypostome from anterior cephalic doublure; imd, inner margin of cephalic doublure; me, macula, an ovate
area situated adjacent to the outer, anterior margin of the posterior lobe of the middle body of the hypostome;
ms, median suture was directed sagittally and connected dorsal facial and hypostomal sutures; pa, panderian
opening, a small perforation or notch in posterolateral doublure of cephalon; pi, pit in anterior border furrow
of cranidium, corresponding pit in doublure; pr, perrostral suture in olenelloids traverses ventral cephalic
doublure between genal angles and bounds rostral plate; pw, posterior wing of hypostome, an extension of
the doublure; rp , rostral plate, the plate isolated between the rostral and connective sutures, may be bounded
along all or part of the posterior edge by the hypostomal suture (in olenelloids isolated from the cephalic
doublure by the perrostral suture); rs, rostral suture joins distal ends of anterior branches of facial suture
and bounds rostral plate anteriorly; s, suture, here used for the dorsal facial suture and its extension on to
the doublure, or the marginal suture; wp, wing process, the rounded or thorn-like process on the inner surface
of the anterior wing of the hypostome, that corresponds with a pit in the external surface of the wing.

In the text, the altitude of the hypostome refers to the angle at which the external surface was held relative
to the horizontal, the posterior margin of the occipital ring having been orientated vertically in the figures.
Thus a downward attitude refers to a downward and backward slope, an upward attitude to an upward and
backward slope. In the descriptive section, reference of a genus to a particular family follows the Treatise
(Moore 1959) unless otherwise noted. In discussing supra-generic relationships the termination ‘-oid’ is given

WHITTINGTON: CAMBRIAN TRILOBITES

579

to a particular generic name to imply a relationship with other genera above the family level, e.g.
corynexochoid, ptychoparioid.

SYSTEMATIC DESCRIPTIONS OF HYPOSTOMES AND VENTRAL SUTURES

Family eodiscidae Raymond, 1913/?

Genus pagetia Walcott, 1916

Pagetia ocellata Jell, 1970
Text-fig. 1

Jell (1970; 1975, pp. 50 51) has described the silicified material on which the present drawing is based.
Contrary to the views expressed by Jell (1975, p. 22) the hypostome is like that of many other Cambrian
trilobites in having a narrow band along the anterior edge bent to incline forward and ventrally, long anterior
and shorter posterior wings. As the sagittal section shows, if the tip of the anterior wing was held close
beneath the axial furrow immediately in front of the eye ridge, there would have been ample room between
hypostome and glabella for the soft parts of the animal. If a flat, crescentic rostral plate were present that
extended inward to a position beneath the border furrow, like that described in one agnostid by Hunt (1966),
I agree with Jell that the hypostome could not have been joined to it by a hypostomal suture. There would
have been a considerable gap between the anterior margin of the hypostome and the inner edge of such a
rostral plate, and the downwardly flexed anterior edge of the hypostome makes it unlikely that there
was any such junction. Pagetia is placed in the family Eodiscidae, accepting the arguments of Jell (1975,
pp. 14, 30).

Family holmiidae Hupe, 1953c/

Genus holmia Matthew, 1890

Holmia kjerulfi (Linnarsson, 1871)

Text-figs. 2 and 3

Holm (1887) and Kiaer (1916) described the morphology and ontogeny of this species, the type, from the
type locality, and Bergstrom ( 1973) and Nikolaisen ( 1986) have described additional specimens. Two specimens
figured here show the cephalon in dorsal aspect, one with part of the left anterior wing of the hypostome
exposed (text-fig. 2b); in a second (text-fig. 2a), an external mould of the slightly displaced hypostome has
been partially exposed. In the latter specimen a left lateral portion of the external mould of the rostral plate
is preserved, and it appears that the anterior wing curved upward free of the rostral plate, so that the
extremity lay close beneath the axial furrow in front of the eye ridge (text-fig. 3). Two of the specimens
illustrated by Holm (1887, pi. 15, figs. 13 and 14) confirm this shape of the anterior wing, but it is concealed
in Holm’s (1887, pi. 14, fig. 2) restoration. The form and position of the anterior wing of the hypostome of
the related genera Schmidtiellus and Wannerial, as drawn by Bergstrom (1973, figs. 12 and 18), were similar.

text-fig. 1. Pagetia ocellata Jell, 1970. Beetle Creek Formation, T5-2 miles north of Mount Murray, at
approximately 21° 50' south latitude, 139° 58' east longitude, north-western Queensland; early Middle
Cambrian. After Jell (1970, 1975). Scale bar in millimetres. See p. 578.

580

PALAEONTOLOGY, VOLUME 31

text-fig. 2. Holmia kjerulfi (Linnarsson, 1871). Holmia Shale, Tomten, Ringsaker, Norway; Lower Cambrian.
a, BMNH H20673, incomplete internal mould of cephalon broken to show portion of rostral plate (rp);
glabella excavated to show right side and posterior half of external mould of hypostome, x 2. b, BMNH
1150, internal mould of cephalon; glabella broken anteriorly to show anterior wing (aw) of hypostome, x 3.

After Holm (1887), Kiaer (1916), and text-fig. 2. Scale bar in millimetres. See p. 578.

Holm (1887, pi. 15, figs. 13 and 14) and Kiaer (1916, pi. 7, fig. 4) figured the hypostome attached to the
rostral plate (the 'hypostomal attachments’ of Kiaer), and Kiaer considered that no hypostomal suture was
present (cf. Resser in Stubblefield 1936, fig. 7, p. 422). However, Kiaer noted the 'fine, raised line' between
the two sclerites, and figured an isolated hypostome (Kiaer, 1916, pi. 7, fig. 5). Two incomplete hypostomes
were figured by Nikolaisen (1986, fig. \d, e ), one of which shows the impressed line dividing the hypostome
from the narrow (sag. and exs.) median portion of the rostral plate. The occurrence of isolated hypostomes
suggests that at particular times during the holaspid period the hypostomal suture may have been functional,
at others not; hence the presence of the suture is questioned in text-fig. 3.

WHITTINGTON: CAMBRIAN TRILOBITES

581

text-fig. 4. Olenellus gilberti Meek, 1874. Combined Metals
bed, Pioche Shale, Pioche Mining district, Lincoln County,
Nevada; Lower Cambrian, a, developmental stage III, ventral
view, after Palmer (1957, text-fig. 6, III, a; pi- 19, figs. 2 and
3). b, developmental stage V, ventral view, after Palmer (1957,
text-fig. 7, V, e; pi. 19. figs. 16 and 18). Scale bars in millimetres.

The classification of olenclloid trilobites continues to be a matter of debate (e.g. Ahlberg et al. 1986); here
I have followed Bergstrom (1973).

Family olenellidae Vogdes, 1893
Genus olenellus Billings, 1861

Olenellus gilberti Meek, 1874

Text-fig. 4

The development of the cephalon, including the hypostome of Olenellus has been revealed by silicified material
(Palmer 1957). Palmer suggested that the hypostome may have been attached to the rostral plate along an
extremely short (tr.) hypostomal suture. Two of his developmental stages are drawn here (text-fig. 4) with
the hypostome placed so that the anterior wing is situated below the anterior margin of the large eye lobe.
It is then apparent that a gap separates the bent-down anterior edge of the hypostome from the inner margin
of the rostral plate. A small median projection is present on the anterior margin of the hypostome, and the
inner margin of the rostral plate has a slight median backward projection; it appears unlikely that these
projections were in contact if the hypostome was situated as shown. As growth proceeded the anterior wing
of the hypostome became broad and merged with the large, inflated anterior body, so that in isolated
specimens the anterior margin of wing and hypostome formed a continuous curve (Walcott 1910, pi. 35, fig.
7; Palmer 1957, pi. 19, fig. 9). In some species a narrow border along this margin may have been down-
curved. Attachment in large holaspids of Olenellus can only have been at an extremely short (tr.) suture. In
Paedumias (= Olenellus , see Fritz 1972, p. 11), Walcott illustrated (1910, pi. 34, figs. 5-7; cf. Resser and
Howell 1938, pi. 9, figs. 6 and 7) a narrow (tr.), presumably mineralized, median strip connecting rostral
plate and hypostome. The original of Walcott’s fig. 6 is similar in size to that of text-fig. 4b. Such a median
strip was evidently present in some species of Olenellus , at least in the developmental stages.

Family bathynotidae Hupe, 1953 a
Genus bathynotus Flail, 1860

Bathynotus holopygus (Hall, 1859)

Plate 52; text-fig. 5

A single species of this genus is known from only one locality in the Lower Cambrian of Vermont (Shaw
1955, p. 778). Hall’s (1859, pp. 61-62, fig. 3) type specimen was recorded as missing by Resser and Howell
(1938, p. 230), but twelve topotype specimens in the US National Museum include those on which Walcott
(1886, pp. 191 193, pi. 31, figs. 1 and la; 1890, p. 646, pi. 95, figs. I and la) based his description, and the
original of Rasetti’s (1952, pi. 1, fig. 5) drawing of the cephalic doublure and hypostome. These and additional
specimens are re-figured here as the basis for a reconstruction; the convexity shown in this reconstruction is

582

PALAEONTOLOGY, VOLUME 31

text-fig. 5. Bathynotus holopygus (Hall, 1859). Parker Slate, Parker Quarry, Georgia, north-western Vermont;
Lower Cambrian. After Plate 52. The eye surface is unknown, the possible form being shown by a dashed
line; see text for discussion of queries. Scale bar in millimetres. See p. 578.

conjectural, since the specimens are partially flattened and distorted, preserved in a dark, iron-stained
micaceous shale.

The best-preserved cranidium (PI. 52, fig. 4) shows the long (exs.), curved, gently convex palpebral lobe,
and the anterior branch of the suture directed inward and forward, curving, with the two branches confluent
along the anterior margin. The low tubercle on the outer portion of glabellar L2 is unique to this specimen;
the low median occipital tubercle and granulation on the glabella are better preserved in other specimens. In
a cephalon exposed from the dorsal side (PI. 52, fig. 6) the genal regions are crushed, so that the eye surface
is not preserved, hence the dashed outline of its possible form in text-fig. 5. Outside the eye lobe the external
surface appears to be preserved, showing a narrow (tr.) librigenal area and a gently convex border, laterally
granulose and traversed by terrace lines subparallel to the margin. These lines continued inside the anterior
sutural margin of the cranidium and on the long genal spine. Specimens showing an internal mould of the
broad cephalic doublure and hypostome (PI. 52, figs. 2, 3, 5, 7) show the outline and convexity of the
doublure, and the convex (ventrally), inner, marginal band that extends from the hypostome to the broadest
(tr.) portion of the doublure at the genal angle. Medially this doublure is crossed by the two sections of the
hypostomal suture, each directed straight outward and backward from the anterior margin, the angle between
the two sections being slightly oblique.

The hypostome was subpenlagonal in outline, its length (sag.) about equal to the width (tr.) at the
midlength; the example (PI. 52, fig. 3) used by Rasetti appears to have been elongated by distortion. The
triangular, anterior portion of the hypostome lying between the sutures (PI. 52, fig. 8) was traversed by
terrace lines continuous with those on the doublure, directed subparallel to the margin. The middle body of
the hypostome was gently convex, externally smooth, and subdivided by a faint, middle furrow. The narrow,
convex lateral border, separated by a broad, shallow depression from the middle body, was continued by a
less convex posterolateral and posterior border. Terrace lines traverse these borders, and a broader
posterolateral area. In this area, inside the convex border, all the specimens show a subcircular depression
of varying depth. At the anterolateral angle of the hypostome, adjacent to the convex inner border of the
doublure and the anterior end of the lateral border, the hypostomal exoskeleton was bent dorsally and

EXPLANATION OF PLATE 52

Fig. 1-8. Bathynotus holopygus (Hall, 1859). Parker Slate, Parker Quarry, Georgia, north-western Vermont;
Lower Cambrian. 1, USNM 15409 (255(7), internal mould of dorsal exoskeleton, external mould of cephalic
doublure and genal spines, dorsal view, x 2; original of Walcott (1886, pi. 31, fig. 1) and of Resser and
Howell (1938, pi. 12, fig. 6). 2 and 8, USNM 15409 (255p), ventral views of internal mould of cephalic
doublure and hypostome, x2 and x6 respectively; original of Walcott (1886, pi. 31, fig. la). 3, USNM
419926, internal mould of cephalic doublure and hypostome, ventral view, x 2; original of Rasetti (1952,
pi. I, fig. 5). 4, USNM 15408, internal mould of exoskeleton lacking free cheeks, dorsal view of anterior
portion, x 2-5. 5, USNM 419927, internal mould of cephalic doublure and hypostome, ventral view, x 3.
6, USNM 15409 (255o), anterior portion of internal mould of dorsal exoskeleton, dorsal view, x 2. 7,
USNM 419928, internal mould of cephalic doublure and hypostome, ventral view, x 2.

PLATE 52

rs} 3*?$' ^

*gS§|

wil

WHITTINGTON. Bathynolus

584

PALAEONTOLOGY, VOLUME 31

extended as the anterior wing. This wing is poorly preserved in all the specimens, but presumably extended
upward and outward so that the tip lay close beneath the axial furrow at the anterolateral angle of the
glabella. The acute angle between each section of the hypostomal suture and the anterior margin of the
doublure shows that the two sections of this suture met at the margin in the median line.

The complete cephala, though flattened, show that the preglabellar area was short (sag. and exs.). Walcott’s
original (PI. 52, fig. 1) combines an internal mould of glabella and fixed cheeks with an external mould of
the doublure, and indicates that the inner margin of the doublure lay beneath the anterior slope of the
glabella. The preglabellar area of the cranidium (PI. 52, fig. 4) appears to be of a length (sag. and exs.) such
that the confluent anterior sutural branches were situated on the anterior margin of the cephalon. In the
reconstruction a triple junction is therefore shown between the confluent dorsal sutures and the two sections
of the hypostomal suture. Rasetti (1952, p. 890) suggested that a short (sag.) median suture might intervene
between hypostomal and dorsal sutures; but if so it would have been extremely short. He went on to state
that there was no rostral plate, nor was the rostral plate fused to the hypostome. However, Harrington (in
Moore 1959, p. 067), in defining a bathynotid type of sutural pattern, thought that the rostral plate was
probably fused with the hypostome, and that the inverted V-shaped suture described here as hypostomal
represented a pair of connective sutures diverging backward from the anterior margin of the doublure. The
triangular area enclosed by these sutures is traversed by raised terrace lines continuous with those on the
adjacent cephalic doublure (PI. 52, fig. 8). This likeness lends credence to the view that this triangular area
represents the rostral plate, which was fused to a hypostome of subrectangular outline, wider (tr.) than long.
Text-fig. 5 is labelled with queries because a choice between these conflicting views is hindered by lack of
evidence. There is no indication, such as the change in slope in Paradoxides (PI. 53, figs. 1, 3, 4) or Fieldaspis
(PI. 54, figs. 1 and 3) between rostral plate and hypostome, of a boundary between the supposed fused
sclerites. The gently convex middle body of the hypostome projects into the triangular area enclosed by the
suture, and the terrace lines appear to cross this anterior edge of the middle body. For convenience I have
referred to the hypostome in the sense of Rasetti (1952), whether or not this sclerite included the rostral
plate.

In the originals of Plate 52, figs. 2, 3, 8 and USNM 419925, the free cheeks and hypostome are slightly dis-
placed from one another, but in normal relation to the thorax and pygidium, with the cranidium missing. In the
original of Plate 52, fig. 5, free cheeks and hypostome are only slightly displaced from one another, but lie across
the pygidium of an articulated thorax and pygidium. In USNM 419928 and 419929, free cheeks and hypostome
are slightly displaced from one another, but the entire unit is inverted relative to the rest of the exoskeleton
(419929), or to the thorax and pygidium (419928). These specimens are presumably all moults, and suggest that
the free cheeks and hypostome were released as a unit, the dorsal facial suture and the articulation between
cranidium and thorax being the most important places of opening in ecdysis. Nevertheless, the slight displace-
ments at the two sections of the hypostomal suture show that this suture was functional.

Family redlichiidae Poulsen, 1927
Genus redlichia Cossman, 1902

Redlichia sp. indet.

Text-fig. 6

The transversely wide rostral plate has a row of pits in the external surface that correspond in position with
pits in the border furrow and, as suggested by a photograph in Zhang et al. (1980, pi. 20, fig. 9), an extension
that narrows posteriorly to the junction with the hypostome. This junction was along the mid-anterior margin
of the hypostome; outside it the margin was continuous with that of the anterior wing; the latter curved
upward and outward, the tip being close below the axial furrow in front of the eye ridge. This reconstruction
is similar to that of Kobayashi and Kato (1951, pi. 5, fig. 6) in showing the anterior wing as a structure
distinct from the rostral plate. Schindewolfs (1955, fig. 2) reconstruction was based on an incomplete
specimen (Schindewolf and Seilacher 1955, pi. 6, fig. 8) that led him to consider that anterior wing and rostral
plate were fused together. I take Opik’s photographs (1958, pi. 5, fig. 1; pi. 6, figs. 4 and 5), together with
those of Zhang et al. (1980, pi. 14, fig. 3; pi. 20, fig. 9), as evidence of the presence of connective and rostral
sutures and of the form of the anterior wing. Opik (1958, p. 28) showed how the pits in the external surface
of the rostral plate and border furrow formed interlocking cones; he considered that the junction between
rostral plate and hypostome was fused. Published photographs show both rostral plate and hypostome linked
together, and the two plates isolated. The latter may result from breakage, or indicate that a hypostomal

WHITTINGTON: CAMBRIAN TRILOBITES

585

text-fig. 6. Redlichia sp. indet. Lower to Middle Cambrian. After Zhang el at. (1980) and
Opik (1958). Scale bar in millimetres. See p. 578.

suture was present; hence the question in text-fig. 6. In the related Sardoredlichia Rasetti, 1972, there appears
to have been a hypostomal suture, but the interlocking pits of doublure and border are absent. The Chinese
material of Redlichia appears to have been flattened, so that Opik’s (1958, pi. 3) figures were used to indicate
the convexity.

The ventral structures of the emuellids (Pocock 1970) are similar to those of redlichiids, to which they are
considered to be related, but the rostral plate is transversely narrower.

Family dolerolenidae Kobayashi in Kobayashi and Kato, 1951
Genus dolerolenus Leanza, 1949

Dolerolenus sp. indet.

Text-fig. 7

In his description of the type species of the genus, Rasetti (1972, pp. 57 58 ) figured the isolated rostral plate
and the hypostome; Sdzuy (1961, pp. 542-544) described specimens preserved in relief, including the
hypostome, of a different species. If the tip of the anterior wing of the hypostome was situated as shown in
text-fig. 7, then the hypostome could not have been attached to the wide rostral plate by a suture, but was
inserted into the ventral integument to leave a wide gap between them.

A transversely wide rostral plate, the inner edge of which underlies the border furrow, is a character
shared by Dolerolenus and Ellipsocephalus (Snajdr 1958, fig. 14; pi. 7, fig. I), but the hypostome of the
latter is not known.

text-fig. 7. Dolerolenus sp. indet. Lower Cambrian. After Rasetti (1972)
and Sdzuy (1961). Scale bar in millimetres. See p. 578.

586

PALAEONTOLOGY, VOLUME 31

rs

text-fig. 8. Paradoxides davidis Salter, 1863. Manuels River Formation, Manuels River,
Newfoundland; Middle Cambrian. After Bergstrom and Levi-Setti (1978, figs. 5 and la\
pi. 3, fig. 4; pi. 5, figs. 3, 6-8) and Plate 53, figs. 1, 3, 8. Scale bar in millimetres. See p. 578.

Family paradoxididae Hawle and Corda, 1847
Genus paradoxides Brongniart, 1822

Paradoxides davidis Salter, 1863

Plate 53, figs. 1-3, 8; text-fig. 8

The dorsal exoskeleton has become well known through the work of Bergstrom and Levi-Setti (1978), but
their illustrations of the fused rostral plate and hypostome show only the proximal portion of the anterior
wing. An example from Newfoundland (PI. 53, figs. 1-3, 8) shows the form and size of the steeply inclined
anterior wing, and the rounded margin of the tip. The anterior margin of the wing is separated from the
inner side of the gutter-shaped distal portion of the rostral plate. The boundary between rostral plate and
hypostome is a shallow furrow that follows the change in slope between the two fused sclerites. The terrace
lines (cf. Bergstrom and Levi-Setti 1978, pi. 3, fig. 4; pi. 5, fig. 7) are continuous across the boundary. Text-
fig. 8 shows that the tip of the anterior wing, which lay close to the vertical, inner portion of the rostral
plate, must also have been situated close beneath the ventrally projecting axial furrow (cf. Bergstrom and
Levi-Setti 1978, pi. 9, fig. 3, where the same relationship is seen). The convexity of the hypostome and rostral
plate is shown here (PL 53, figs. 3 and 8) and examples of other species (Sdzuy 1967, pi. 2, fig. 9; Westergard
1936, pi. 3, fig. 1; pi. 6, fig. 4; pi. 9, fig. 3) have been used to suggest the convexity of the cephalon dorsally.

Snajdr (1958, pp. 102-103; PI. 53, fig. 4 shows a Bohemian example) regarded the fusion of rostral plate
and hypostome as diagnostic of Paradoxides , but in other species attributed to this genus by Westergard
(1936, p. 33, footnote) a hypostomal suture is developed. Such species are placed in the new genera (or
subgenera) proposed by Snajdr (1958), but how these names are to be used is disputed (cf. Sdzuy 1967,

explanation of plate 53

Figs. 1-3, 8. Paradoxides davidis Salter, 1863. Manuels River Formation, L5 miles east of Elliot Cove and
north of Foster’s Point, Random Island, Newfoundland; Middle Cambrian. SM A. 105203, rostral-
hypostomal plate. I and 8, ventral and left lateral views, x 3. 2, enlargement of right macula, x 7. 3,
oblique view, x 5.

Fig. 4. Paradoxides gracilis (Boeck, 1827). Jince Formation, Jince, Czechoslovakia; Middle Cambrian. SM
A. 49780, internal mould of incomplete rostral-hypostomal plate, ventral view, x 3.

Figs. 5-7. Sao hirsuta Barrande, 1846. Skryje Formation, Skryje, Czechoslovakia; Middle Cambrian. SM
X. 11469, internal mould of cranidium and anterior thoracic segments; anterior, left lateral, and dorsal
views respectively, x 3.

PLATE 53

WHITTINGTON, Paradoxides, Sao

588

PALAEONTOLOGY, VOLUME 31

text-fig. 9. Xystridura sp. indet. Middle Cambrian. After Opik (1975) and Palmer and Gatehouse (1972).

Scale bar in millimetres. See p. 578.

p. 94; Bergstrom and Levi-Setti 1978, p. 15; Snajdr 1985, p. 169). Here I follow Snajdr’s diagnosis of
Paradoxides.

Family xystriduridae Whitehouse, 1939
Genus xystridura Whitehouse, 1936

Xystridura sp. indet.

Text-fig. 9

Opik (1975) illustrated in detail many species of Xystridura (and its subgenera), including the rostral plate
and the isolated hypostome (his figs. 8 and 10; pi. 2, figs. 2 and 4; pi. 9, fig. 3; pi. 13, fig. 1; pi. 22, fig. 4; pi.
32, fig. 2), and examples of the two plates in position relative to each other (his pi. 15, fig. 3; pi. 30, fig. 2).
Opik (1975, pp. 35 36) asserted that in X. (Inosacotes) browni no hypostomal suture was present in the
holaspis, the hypostome being fused to the rostral plate; a hypostomal suture in Xystridura is thus shown
with question in text-fig. 9. He described the tip of the large anterior wing as braced against the axial furrow,
and the much smaller projection of the presumed posterior wing. The anterior margin of the hypostome was
bent to slope downward and forward; a small macula was developed. All Opik’s material was flattened, so
that the convexity shown in text-fig. 9 is somewhat conjectural, but supported by specimens from Antarctica
described by Palmer and Gatehouse (1972, pi. 2, figs. 18, 20, 23, 25). Opik (1975, pp. 25-26) regarded
Xystridura as related to Paradoxides , remarking on the unusual extension of the rostral plate on to the dorsal
surface, and the olenelloid-like extension of the rostral plate rearward. The form of the hypostome supports
this view of their relationships.

Family zacanthoididae Swinnerton, 1915
Genus fieldaspis Rasetti, 1951

Fieldaspis celer (Walcott, 1917)

Plate 54, figs. 1-3; text-fig. 10

Rasetti (1957, pp. 957 958, pi. 1 18, figs. I 8; text-fig. 4) showed the long, steeply upwardly directed anterior
wing of the hypostome. He considered that the anterior wing was part of the rostral plate and hence not

EXPLANATION OF PLATE 54

Figs. 1-3. Fieldaspis celer (Walcott, 1917). Mounl Whyte Formation, Mount Field, British Columbia, locality
W28fg of Rasetti (1957); Middle Cambrian. BMNH U4570, rostral-hypostomal plate 1, ventral view, x 8.
2 and 3, left lateral and oblique views, x 6.

Figs. 4 8. Ptychoparia striata (Emmrich, 1839). Jince Formation, Vinice, near Jince, Czechoslovakia; Middle
Cambrian. 4, 5, 8, SM A.51043o, internal mould of cranidium, dorsal, anterior, and right lateral views
respectively, x 3. 6 and 7, SM A. 1574, internal mould, cephalon broken to show displaced external mould
of doublure (d), connective suture (cs), and rostral plate (rp), anterior and dorsal views respectively, x 3.

PLATE 54

mmm

WHITTINGTON, Fieldaspis , Ptychoparia

590

PALAEONTOLOGY, VOLUME 31

text-fig. 10. Fieldaspis celer (Walcott, 1917). Mount Whyte Formation, Mount Field,
British Columbia; Middle Cambrian. After Rasetti (1957) and Plate 54, figs. 1-3. Scale

bar in millimetres. See p. 578.

homologous with that of other trilobites. A topotype specimen of the rostral-hypostomal plate collected by
Rasetti (PI. 54, figs. 1-3) shows the gutter-shaped lateral portion of the rostral plate to be continuous with
a convex, narrower (sag. and exs.) median portion. A shallow furrow at a change in slope indicates the fused
boundary between rostral plate and hypostomc; a shallow median indentation is present at the anterior
margin of the hypostome. The doublure of the free cheek is unknown, but I assume that it was convex and
anteriorly similar in form to the lateral portion of the rostral plate. The anterior wing of the hypostome is
triangular in outline and the anterior margin straight, with the wing meeting the inner, vertical portion of
the rostral plate at almost a right angle. At about half the length of the posterior margin of the wing it is
crossed by a short, sharp flexure; the tip of the wing is rounded and there is no pit in the external surface
that would indicate the presence of a wing process. Rasetti’s text-fig. 4 suggests the presence of such a pit,
rather than the flexure. The middle furrow, strongly convex macula, and gently inflated posterior lobe of the
middle body are shown by the present specimen. It is slightly smaller, and the middle body is considerably
less convex, than Rasetti’s example. The present specimen is partially exfoliated (completely so in the mid-
region of the anterior lobe of the middle body), but shows the prominent terrace lines on the lateral portion
of the rostral plate, and fainter ones on the hypostome. The change in intensity of the terrace lines takes
place at the anterior edge of the furrow between rostral plate and hypostome. Between the terrace lines there
are minute pits. The structure of the rostral plate in Fieldaspis is thus like that in Paradoxides (PI. 53, figs.

I -4, 8; text-fig. 8), and the anterior wing of the hypostome homologous with that of other trilobites. The tip
of the wing (text-fig. 10) would have extended up to a point just beneath the axial furrow, immediately in
front of where it was met by the eye ridge.

Rasetti (1951, 1957) placed Fieldaspis in the Dolichometopidae, but in 1959 (in Moore, p. 0227) in
Zacanthoididae. In the Treatise (Moore 1959, p. 0217) all trilobites placed in the Order Corynexochida are
said to have the rostral plate and hypostome fused. It is by no means certain that this is so (Opik 1982,
pp. 6 7), and in his conception of the corynexochoid family Dolichometopidae Opik includes the subfamily
Horonastinae, which is characterized by having the hypostome and rostral plate as separate sclerites. Here
I retain the family assignment of Fieldaspis used in the Treatise.

Family ptychopariidae Matthew, 1888
Genus ptychoparia Hawle and Corda, 1847

Ptychoparia striata (Emmrich, 1839)

Plate 54, figs. 4-8; text-fig. 1 1

Snajdr (1958, pp. 185-190) figured specimens which showed the doublure of the cephalon, isolated hypostomes,
and one hypostome exposed almost in place. Additional specimens (PI. 54, figs. 4-8) show the doublure of
the displaced free cheek, the rostral plate, and the connective suture, and include a cranidium which shows
the convexity and that the rostral suture ran along the doublure just inside the anterior margin. The doublure
was gently convex, widest (sag. and exs.) anteriorly, with the inner edge underlying the border furrow. It was
crossed by the connective suture at a position that lay well outside the projected line of the axial furrow,
giving a relatively wide (tr.) rostral plate. The hypostome (Barrande 1852, pi. 14, fig. 3; Snajdr 1958, pi. 39,

WHITTINGTON: CAMBRIAN TRILOBITES

591

text-fig. 11. Ptychoparia striata (Emmrich, 1839). Juice Formation, Vinice, near Jince,
Czechoslovakia; Middle Cambrian. Modified from Snadjr (1958) and after Plate 54,
figs. 4 8. Scale bar in millimetres. See p. 578.

fig. 4; pi. 40, fig. 5) appears to have been moderately convex, and the lateral border well defined. The anterior
wing is poorly known, but assuming it was situated beneath the axial furrow immediately in front of the eye
ridge (text-fig. 1 1), there was a considerable gap between the edges of the rostral plate and the hypostome.
This accords with Barrande’s restoration (1952, pi. 2b, fig. 26) and some specimens of Snajdr (1958, pi. 39,
fig. 7; pi. 40, fig. 3) but not his fig. 40 (in which the hypostome is portrayed as joined to the rostral plate at
a hypostomal suture). This type of ventral structure of the cephalon— a wide (tr.) rostral plate, with the
hypostome not being attached by a hypostomal suture but inserted into the ventral cuticle some distance
behind the rostral plate— has been regarded (Rasetti 1951, p. 140; Opik 1963, p. 77) as a basic character of
ptychoparioids.

Family conocoryphidae Angelin, 1854
Genus conocoryphe Hawle and Corda, 1847

Conocoryphe sulzeri (Schlotheim, 1823)

Plate 55, figs. 1, 3, 6, 7; text-fig 12

Snajdr (1958, fig. 32) showed the cephalon of this type species and its doublure in dorsal view. The present
cranidium (PI. 55, figs. 1, 3, 6) shows the course of the suture posterolaterally and, though partially flattened,
gives some idea of the convexity; other species illustrated by Sdzuy (1961, pi. 31, fig. I; 1967, pi. 9, fig. 9)
are better in this latter respect. The cephalon (PI. 55, fig. 7) shows how the anterior doublure, striated with
concentric terrace lines, extended some distance horizontally inward before being cut by the rostral suture.
The displaced right free cheek shows the steeply upturned, smooth inner portion of the doublure that extended
upward under the border furrow to the margin of the gena; it terminates adaxially at the convexly curved
connective suture. The rostral plate was similarly shaped, and terminated under the anterior margin of the
convex preglabellar field (Snajdr 1958, pi. 34, fig. 3). The anterior margin and anterior wing of the hypostome
(Snajdr 1958, pi. 34, figs. 4, 8, 9) are not well known, but if it was situated so that the anterior wing lay
below the axial furrow immediately in front of the genal ridge, it appears (text-fig. 12) that it was not joined
by a hypostomal suture to the rostral plate, but inserted into the ventral integument above and behind the
rostral plate. Snajdr (1958, fig. 32) did not show the hypostome, but this arrangement is similar to that shown
(Snajdr 1958, fig. 35) for the related Ctenocephalus , which has a considerably longer (sag.) preglabellar field.
Barrande (1852, pi. 2b, fig. 24) interpreted the hypostome of Conocoryphe as joined by a hypostomal suture
to the rostral plate, but my restoration (that takes account of the convexity of the cephalon) makes this
unlikely and accords with the view of Poulsen (in Moore 1959, p. 0242).

Family solenopleuridae Angelin, 1854
Genus sao Barrande, 1846

Sao hirsuta Barrande, 1846

Plate 53, figs. 5 7; text-fig. 13

Snajdr’s (1958, pis. 43-45) many illustrations of this type species include two of specimens showing an
external mould of the hypostome exposed beneath the broken glabella, and internal moulds of the isolated

592

PALAEONTOLOGY, VOLUME 31

text-fig. 12. Conocoryphe sulzeri (Schlotheim, 1823). Jince Formation, Jince, Czechoslovakia;
Middle Cambrian. After Snajdr (1958) and Plate 55, figs. 1, 3, 6, 7. Scale bar in millimetres.

See p. 578.

text-fig. 13. Sao hirsute i Barrande, 1846. Skryje Formation, Skryje, Czechoslovakia;
Middle Cambrian. After Snajdr (1958) and Plate 53, figs. 5-7. Small spines and
tubercles on external surface omitted. Scale bar in millimetres. See p. 578.

hypostome. The largest isolated hypostome (Snajdr 1958, pi. 45, fig. 13), a ventral view, gives a suggestion
of the convexity and shows the anterior wing and the downward and forward sloping anterior edge of the
wing and middle body. The only lateral view of the cephalon is Barrande’s (1852, pi. 7, fig. 29), that shows
the convexity and the upwardly arched outline of the border that is confirmed by the original of Plates 53,
figs. 5-7. An anterolateral view (Whittington 1957a, pi. 1 15, fig. 22) of a somewhat crushed specimen revealed
the external mould of the cephalic doublure and rostral plate. The lateral and anterior borders and doublure
together formed a tubular border to the cephalon, with the inner edge of the doublure curving up beneath
the outer edge of the border furrow; the curving connective suture crossed the doublure in the projected line
of the axial furrow. Snajdr (1958, fig. 45) did not include the hypostome in his reconstruction, but it seems
evident (text-fig. 13) that, because of the convexity and anterior arching of the cephalon, the hypostome was
inserted into the ventral integument some distance behind and below the inner edge of the rostral plate.

EXPLANATION OF PLATE 55

Figs. 1, 3, 6, 7. Conocoryphe sulzeri (Schlotheim, 1823). Jince Formation, Jince, Czechoslovakia; Middle
Cambrian. 1, 3, 6, SM A. 51042, internal mould of cranidium, anterior, dorsal, and right lateral views
respectively, x 3. 7, SM X.l 1477, internal mould, cephalon broken to show doublure (d) anteriorly and
laterally, of displaced free cheek (cs, connective suture), dorsal view, x 4.

Figs. 2, 4, 5, 8, 9. Agraulos ceticephalus (Barrande, 1846). Skryje Formation, Skryje, Czechoslovakia; Middle
Cambrian. 2, 4, 5, SM X.l 1475, internal mould of cranidium, anterior, left lateral, and dorsal views
respectively, x4-5. 8 and 9, SM X.l 1474, internal mould of cephalon broken to show external mould of
doublure (d) and rostral plate (rp) separated by the connective suture; glabella excavated to reveal external
mould of incomplete hypostome (h); oblique and dorsal views respectively, x 7.

PLATE 55

WHITTINGTON, Conocoryphe , Agraulos

594

PALAEONTOLOGY, VOLUME 31

text-fig. 14. Agraulos ceticephalus (Barrande, 1846). Skryje Formation, Skryje, Czechoslovakia; Middle
Cambrian. After Plate 55, figs. 2, 4, 5, 8, 9. Scale bar in millimetres. See p. 578.

Family agraulidae Raymond, 1913a
Genus agraulos Hawle and Corda, 1847

Agraulos ceticephalus (Barrande, 1846)

Plate 55, figs. 2, 4, 5, 8, 9; text-fig. 14

An internal mould of the cepfialon and part of the thorax (PI. 55, figs. 8, 9) has been prepared to show the
external mould of an incomplete hypostome below the glabella. The anterior wing has a convex border along
the posterior margin, continuous with the lateral border of the middle body; posteriorly this body is convex,
sloping almost vertically to the narrow posterior border. Barrande (1852, pi. 10, figs. 12 and 13) illustrated
a much smaller example as having a broad, flat posterolateral border; despite the incompleteness, there seems
no doubt that in the present example the inflated posterior portion of the middle body had a steep posterior
slope and narrow border. Also preserved is an external mould of the cephalic doublure that is narrow and
flat adjacent to the outer margin and traversed by concentric terrace lines; a much wider inner portion
extended dorsally close beneath the outer slope of the genal region and preglabellar area. The connective
suture is curved convexly adaxially, the rostral plate having a narrow (sag.), flat outer portion and a wider
inner portion that bulged forward beneath the preglabellar area. An internal mould of a cranidium (PI. 55,
figs. 2, 4, 5) shows that the rostral suture was situated close to the anterior margin of the doublure, and
other features, including the faint impressions of glabellar furrows and the eye ridge. Text-fig. 14 shows that
the hypostome cannot have been attached by a suture to the rostral plate; Snajdr (1958, pi. 37, figs. 8 and
13) figured the rostral plate and a poorly preserved hypostome, but did not show the latter in his reconstruction
(fig. 37).

I have followed Henningsmoen (in Moore 1959, p. 0278) in using the family Agraulidae; Snajdr (1958)
used a broader classification, while Opik (1961, p. 142) stressed the similarities between Agraulos and
Ellipsocephala (cf. Ahlberg and Bergstrom 1978, pp. 9-11).

Family menomoniidae Walcott, 1916
Genus bolaspidella Resser, 1937

Bolaspidella sp. indet.

Text-fig. 15

The reconstruction is based primarily on silicified material of the type species described by Robison (1964,
pp. 552-554, pi. 88, figs. 16-21); other species were described by him (1964, pp. 554-555, pi. 88, figs. 7 15;
pi. 89, figs. 1 1 1, 14-17), and one by Rasetti (1967, pp. 94-96, pi. 13, figs. 17-30). The hypostome is notable
in that the pit in the external surface of the anterior wing formed a wing process. The posterior margin of
the rostral plate curved convexly forward more strongly than the anterior margin of the hypostome, which
bounded the ventrally inclined anterior edge of the hypostome and anterior wing. Robison gave no lateral
views of cranidia, but that of Rasetti (1967, pi. 13, fig. 18) suggests how steeply the preglabellar area descended
to the border furrow. When a hypostome of appropriate size is placed with the wing process close

WHITTINGTON: C A M B RI AN TR I LO B ITES

595

text-fig. 15. Bolaspidella sp. indet. Middle to Upper Cambrian (Robison 1976, text-fig. 4). After Robison
(1964) and Rasetti (1967). Scale bar in millimetres. See p. 578.

below the axial furrow, immediately in front of the eye ridge, it appears unlikely that the hypostomc was
attached by a suture to the rostral plate; and the curvature of the opposing edges does not suggest such an
attachment.

Family ceratopygidae Linnarsson, 1869
Genus proceratopyge Wallerius, 1895

Proceratopyge sp. indet.

Text-fig. 16

The hypostome of the specimen figured by Rushton (1983, pi. 19, fig. 12; text-fig. 6) is poorly preserved and
displaced, but the wide (sag. and exs.) doublure of the cephalon in front of it, and the median suture, are
preserved. Text-fig. 16 is based on this specimen, supported by Westergard’s (1947, pi. 2, figs. 1-5, 9) figures,
and the assumption that the hypostome was like undetermined specimens thought to be of a ceratopygid
and figured by Shergold (1980, p. 89, pi. 20, fig. 10; 1982, p. 53, pi. 7, fig. 8). It is also assumed that the
hypostome was attached at a hypostomal suture to the backward-projecting doublure. Illustrations by Jago
(1987, pi. 26, figs. 3, 5, 6; pi. 27, figs. 7 and 8) support this reconstruction. Species of Diceratopyge are like
those of Proceratopyge , and an example figured recently (Lu and Qian 1983, pi. 10, fig. 12) has a partial
external mould of the hypostome exposed beneath the glabella, attached at the suture to the doublure
anteriorly, which is crossed by the median suture.

It has been argued (Fortey and Owens in Owens et al. 1982, pp. 14 15, pi. 2, figs, d /, h /; cf. Shergold in
Shergold and Sdzuy 1984, pp. 94-95) that Macropyge is a ceratopygid, and the form of the hypostome, its
attachment to the doublure, and the median suture support this view. An exceptional example of Macropyge ,
showing the cephalic doublure and hypostome, has been figured by Lu and Qian (1983, pi. 13, fig. 3).

Poulsen (in Moore 1959, p. 0363) stated that in the ceratopygid Hysterolenus the hypostome was probably
fused with the rostral plate. This statement may be based on Moberg and Segerberg’s (1906, pi. 4, fig. 36)
illustration of the hypostome of the type species H. toernquisti. I suggest that Moberg and Segerberg were

text-fig. 16. Proceratopyge sp. indet. Late Middle to Upper Cambrian, and Tremadoc.
After Rushton (1983) and Westergard (1947). Scale bar in millimetres. See p. 578.

596

PALAEONTOLOGY, VOLUME 31

rp

rs

rs

text-fig. 17. Aphelaspis sp. indet. Upper Cambrian (Dresbachian). After Rasetti (1965) and Palmer (1962a,

1965). Scale bar in millimetres. See p. 578.

outlining the anterior border and anterior wings of an incomplete specimen of the hypostome, and not a
‘rostral plate’, so that Poulsen’s supposition is unlikely to be correct.

Family pterocephalidae Kobayashi, 1935
Genus aphelaspis Resser, 1935
Aphelaspis sp. indet.

Text-fig. 17

The hypostome, free cheek, and probable rostral plate were illustrated by Palmer (1962a, pi. 6, figs. 15-19;
1965, pi. 8, figs. 16, 17, 21, 24); Rasetti's (1965) many illustrations include the type species (pi. 18, figs. 10-
20), lateral views of cranidia, and external views of hypostomes. My reconstruction (text-fig. 17) is based on
these figures, and when the hypostome is positioned as shown it cannot have been attached by a suture to
the rostral plate.

The rostral plate and hypostome of Cedaria (Palmer 1962a, pi. 6, figs. 13 and 14), type genus of the
Cedariidae, were like those of Aphelaspis , with a similar wide gap between the two plates. Middle Cambrian
trilobites that had a similar rostral plate, but in which the hypostome is unknown, are Ekathia (Robison
1964, pi. 85) and Modocia (Robison 1964, pi. 87, figs. 5-19); the rostral plate of the latter has a pointed
median projection on the posterior edge.

Family lonchocephalidae Hupe, 1953 h
Genus welleraspis Kobayashi, 1935

Welleraspis swartzi (Tasch, 1951)

Text-fig. 18

Rasetti (1954, p. 601, pi. 62, figs. 11 14; text-fig. 16, c) diagnosed the genus and described the material on
which the reconstruction is based. His ventral view of the free cheek shows the course of the connective
suture and the genal spine; the hypostome is described as having an angulate anterior outline, and the anterior
edge is bent to slope downward and forward. When the tip of the anterior wing is placed as in text-fig. 18
it appears probable that the median, anterior margin of the hypostome may have been joined by a hypostomal
suture to the posterior edge of the narrow (tr.) rostral plate.

Family olenidae Burmeister, 1843
Genus parabolinella Brogger, 1882

Parabolinella sp. indet.

Text-fig. 19

Henningsmoen (1957, pp. 135 137, pi. 12, figs. 1-5) described the cranidium, free cheek, and hypostome of
the type species, and Ludvigsen (1982, pp. 58-65, figs. 48, 49, 50 a-o, q , r) silicified cranidia, free cheeks

WHITTINGTON: CAMBRIAN TRILOBITES

597

text-fig. 18. Welleraspis swartzi (Tasch, 1951). Warrior Limestone, road cut 2-5 miles east of Bedford,
Pennsylvania; Upper Cambrian (Dresbachian). After Rasetti (1954). Scale bar in millimetres. See p. 578.

joined by the median, channel-form section of the doublure, and the hypostome of other species. Text-fig. 19
is based on specimens in Ludvigsen’s fig. 48, and if the hypostome is placed as shown, it cannot have been
connected by a suture to the inner, anterior edge of the doublure. Further, this inner edge of the doublure
(Ludvigsen 1982, p. 60) was finely toothed, each tooth connecting with a pit in the anterior border furrow.
The anterior lobe of the middle body of the hypostome was convex, with the triangular anterior wing sloping
steeply upward; whether or not this wing bore a wing process is not shown by any of the illustrations.

In their diagnosis of Olenidae, Nikolaisen and Henningsmoen (1985, p. 2) stated that the hypostome was
not attached by a suture to the cephalic doublure; they considered that in most species connective sutures
were absent (cf. Henningsmoen 1957, pp. 90-92). An exception is that Rushton (1983, p. 124, pi. 17, figs. 2
and 3) has recognized the rostral plate in a species of Oleniis.

Family eurekiidae Hupe, 19536
Genus eurekia Walcott, 1916

Eurekia ulrichi (Rasetti, 1945)

Text-fig. 20

Taylor (1978) redescribed the type and other species of Eurekia , and reconstructed one species showing
(without comment) a median suture extending downward from where the two anterior branches met on the
outer edge of the border at an oblique angle. The existence of such a suture was not well documented by the
specimens, but a silicified specimen illustrated by Ludvigsen (1982, fig. 61 g) shows the truncated sutural edge
of the doublure of the free cheek. Two other specimens in exterior view (Ludvigsen, 1982, fig. 616, i) suggest
that this suture may be median, and not bounding a narrow (tr.) rostral plate. Text-fig. 20 is based on these
and other specimens in Ludvigsen’s fig. 61, which includes examples of the hypostome. The anterior edge of

text-fig. 19. Parabolinella sp. indet. Upper Cambrian (Trempealeau and Tremadoc). After Henningsmoen
(1957) and Ludvigsen (1982). Scale bar in millimetres. See p. 578.

598

PALAEONTOLOGY, VOLUME 3 1

hs

text-fig. 20. Ewekia ulrichi (Rasetti, 1945). Rabbitkettle Formation, Broken Skull River, western District
of Mackenzie. Upper Cambrian (Trempealeau). After Ludvigsen (1982). Scale bar in millimetres. See p. 578.

the hypostome is sharply bent to slope ventrally and forward; the anterior wing is small and displays no
wing process. In the reconstruction I tentatively show the hypostome as attached for a short distance medially
by a suture to the inner edge of the doublure. A steeply downward attitude of the hypostome would have
resulted from this attachment and the juxtaposition of the anterior wing and the ridge formed by the axial
furrow. The convexity of the cephalon, combined with the strong anterior arch, were sufficient to conceal
the hypostome in lateral view (text-fig. 20c) despite the steep attitude.

Taylor (1978, p. 1062) noted a notch in the inner margin of the doublure of the free cheek, close to the
genal angle. This notch is clearly visible in Ludvigsen’s fig. 61g. Taylor termed it a vincular notch, but it may
well be associated with the panderian opening.

DISCUSSION

Relation of hypostome to dorsal exoskeleton

When found in a specimen in what is generally considered to be the original position, the hypostome
of Cambrian trilobites is either fused to the rostral plate (PI. 53, figs. 1, 3, 4; PI. 54, figs. 1-3) or
attached at a hypostomal suture to the anterior cephalic doublure (PI. 52, figs. 2, 3, 5, 7, 8; text-
figs. 3, 6, 9, 16, 18, 20). As shown here, in these and other examples, an anterior wing extended
upward and outward from the anterolateral corner of the hypostome. It is considered that the tip
of this wing lay immediately beneath the axial furrow at the anterolateral edge of the glabella. The
level of the crest of the ridge, formed by the axial furrow on the inner surface of the cephalic
exoskeleton, is shown by a heavy dashed line in text-figs. Id, 3d, 5d-20d. The position and attitude
of the hypostome beneath the cephalon shown in these text-figures is determined by the juxtaposition
of the tip of the wing and this ridge, and in attached forms by this juxtaposition in combination
with the hypostomal suture. The eye ridge, or the anterior end of the palpebral lobe, abuts against
the axial furrow immediately behind the tip of the wing, and if an anterior pit is developed in the
axial furrow it depresses the axial furrow inward (as a boss) toward the tip of the wing. In species
in which neither anterior pit nor eye ridge are evident, the tip of the anterior wing lay in a
homologous position, in a transverse line passing in front of the eye lobe. In post-Cambrian
trilobites the position of the hypostome is related in the same way to the dorsal exoskeleton, and
in particular groups a wing process is developed, the tip of which rested against the boss formed
by the anterior pit (Whittington, in press). Similar devices, less prominently developed, are known
in a few well-preserved Cambrian specimens. Thus the attached hypostome in trilobites appears
to have had a constant relation, via the anterior wing and a particular site in the axial furrow, to
the dorsal exoskeleton. Hence the hypostome lay beneath the anterior portion of the glabella.

In a large number of Cambrian, and fewer post-Cambrian, trilobites the hypostome was detached
from the inner edge of the anterior cephalic doublure. The morphology of such hypostomes was
similar, and a prominent anterior wing developed (PI. 55, figs. 8 and 9). In this example and others

WHITTINGTON: CAMBRIAN TRILOBITES

599

(Robison 1972, figs. 2 d-f and 3a, b; Palmer 1962a, pi. 6, figs. 13 and 15; Snajdr 1958, pi. 35, fig.
11; pi. 37, fig. 8; pi. 39, fig. 7; pi. 40, fig. 3) the hypostome lies below the anterior portion of the
glabella, with the tip of the anterior wing beneath the axial furrow. Such specimens are exoskeletons
of whole animals, not moults, and the hypostome presumably was held approximately in its
original position by muscles and the ventral integument into a late stage of decay. This assumption,
tacitly made by earlier authors, is made here, and implies that in holaspid trilobites in which the
hypostome was detached, the anterior wing lay close beneath the axial furrow at a site immediately
in front of where the eye ridge abutted against the furrow. This relationship of hypostome to
dorsal exoskeleton appears likely to have been universal among trilobites, and text-figs. 1, 7, Il-
ls, 17, 19 have been drawn accordingly.

My restorations of post-Cambrian trilobites with attached hypostomes (Whittington, in press,
figs. 2-10, 12, 17?, 18-26) include two drawings that may appear to cast doubt on the universality
of this relationship. That of Triarthrus (Whittington, in press, fig. 2) is based on inadequate
information; in the pyritized specimens described by Whittington and Almond (1987) the anterior
portion of the hypostome is concealed, and the anterior wing therefore unknown. It may well have
been larger (compare the olenid Parabolinella , text-fig. 19) and have extended beneath the axial
furrow anteriorly. In the case of Remopleurides (Whittington, in press, fig. 6), the tip of the long,
slim wing process reached close not only to the tip of the doublure process, but also to the anterior
boss. This boss (Whittington 1959, pi. 2, fig. 25) had a pit at its crest, a feature known in other
trilobites that suggests a close connection with the lip of the wing process. The doublure process
in Remopleurides is an additional, and apparently associated, internal process, while its wing
process and anterior boss were like those of other trilobites.

General morphology of Cambrian hypostomes

Many isolated hypostomes from Cambrian rocks have been illustrated, too many to refer to in
detail here. In the genera having a detached hypostome (text-figs. 1, 7, 11-15, 17, 19) the outline
is subrectangular and elongate sagittally; the anterior wing is a relatively large, dorsally and
outwardly directed projection, triangular or elongated in outline. The narrow (sag. and exs.)
anterior border of the hypostome is flat, and bent to slope ventrally and forward. The lateral and
posterior borders are narrow and convex, with the lateral border projecting moderately or slightly
at about two-thirds the length (exs.), so that there is a broad, shallow lateral notch (in ventral
view) between anterior wing and projection. Whether or not a posterior wing extended upward
from the doublure of this projection (as revealed in Crassifimbria by Palmer 1958, pi. 25, figs. 12
and 14) is not known because of the rarity of specimens showing the internal aspect of the
hypostome. The middle body is convex, with a depression dividing a larger anterior lobe from the
smaller (and in some cases strongly convex) posterior lobe. In attached forms in which the glabella
expanded forward (text-figs. 3, 8-10) the anterior lobe is wide and merges into the large triangular
anterior wing. The lateral border is short (exs.), with a slight posterior projection; in oblique view
(PI. 53, fig. 3; PI. 54, fig. 3) there is a rounded lateral notch between anterior wing and projection.
Wide lateral and posterior borders are developed in such Upper Cambrian genera as Dikelocephalus
(Ulrich and Resser 1930, pi. 10, figs. 1 and 2), Palaeodotes (Opik 1967, fig. 129; pi. 50, fig. 3), and
Polycyrtaspsis (Opik 1967, p. 384; pi. 9, fig. 4), and in the upper Middle Cambrian Iranoleesia and
Chelidonocephalus (Wittke 1984, pi. 1, fig. 5; pi. 3, fig. 4). In Polycyrtaspis , Opik noted a deep pit
in the anterior wing distally— presumably the external expression of a wing process.

The anterior wing of Cambrian hypostomes has been so poorly illustrated that any generalization
about presence or absence of a wing process is questionable. There is no evidence of the wing
process in the form of a pit in the exterior surface at the tip of the anterior wing in Pagetia (Jell
1975, pi. 28, figs. 1 and 2), Holmia (text-fig. 2), Olenellus (Palmer 1957, pi. 19, fig. 9), Redlichia
(text-fig. 6), Dolerolenus (text-fig. 7), Paradoxides (PI. 53, fig. 3), Xystridura (Opik 1975, pi. 2, fig.
2; pi. 13, fig. 1), or Eokaolishania (Wittke 1984, pi. 6, figs. 5 and 10). In Fieldaspsis (PI. 54, figs. 2
and 3) a fold in the anterior wing formed a ridge on the inner surface that may have acted as a
wing process; the exact form of the anterior wing of the hypostome of other corynexochoids is

600

PALAEONTOLOGY, VOLUME 31

unknown. Silicified material has revealed a low wing process in Crassifimbria (Palmer 1958, p. 162,
pi. 25, figs. 12-14), but in other genera which had the hypostome detached in the holaspid stage
(text-figs. 11-14, 17) no wing process is known; however, silicified material of Bolaspidella (text-
fig. 15) reveals such a process. In Parabolinella (text-fig. 19) the presence of a wing process is
uncertain, and it was absent in Eurekia (Ludvigsen 1982, fig. 61 o-q).

The macula is visible in many Cambrian hypostomes and takes the form of an elongate raised
area adjacent to the posterior edge of the middle furrow, on the crescentic posterior lobe of the
middle body. Lindstrom (1901, p. 64, pi. 5, figs. 33 and 34) noted that the macula of a species of
Paradoxides was elongated, with a broken area along the crest, which he suggested was originally
covered by a Thinner membrane’. In P. davidis the macula is similar, and though the right one is
partly exfoliated in the figured specimen (PI. 53, figs. 1 and 2), it appears that the crest bore closely
spaced, circular depressions that had a central mound. In Fieldaspsis (PI. 54, fig. 1) the prominent
right macula appears smooth externally, but may bear minute tubercles along the crest.

Relation of hypostome to cephalic doublure , connective and median sutures

Text-figs. 1, 7, 11-15, 17, 19 show that in more than half the holaspid species of genera chosen
for illustration the hypostome was detached. This condition contrasts with that prevailing in post-
Cambrian trilobites in which fewer genera (Whittington, in press, figs. 1, 11, 14) have the hypostome
detached. In Pagetia (text-fig. 1) there may have been a crescentic ventral plate, like that known
in one species of agnostid (Hunt 1966); in the latter there were no connective sutures. In
Parabolinella (text-fig. 19), as in other olenid genera, the doublure was not crossed by connective
sutures, but in one olenid species such sutures have been described. In the other genera (text-figs.
7, 11-15, 17) connective sutures isolate a rostral plate; in Dolerolenus (text-fig. 7) this plate is wide
(tr.), but is less so in the other examples. Ptychoparia (text-fig. 11) has a moderately wide rostral
plate; this character, and the detached hypostome, were regarded by Opik (1963, p. 77) as important
features of a superfamily or other taxa of higher rank centred on Ptychoparia. The 'ptychopariid
type’ of ventral sutures as defined by Harrington (in Moore 1959, p. 067, fig. 48c) is highly
misleading. Whether the presence of a rostral plate, and a detached hypostome, should be regarded
as cardinal features of any ptychoparioid group, is a matter for consideration.

In Paradoxides (PI. 53, figs. 1, 3, 8) and Fieldaspis (PI. 54, figs. 1-3) the hypostome and rostral
plate are fused into a single sclerite, the rostral-hypostomal plate. The evidence for fusion is that
specimens of the two plates are not found separately, but invariably fused, with a change in slope
marking the boundary between them, across which terrace lines may run continuously. The same
evidence is seen in species of genera of corynexochoids related to Fieldaspis (e.g. Rasetti 1948;
Palmer and Halley 1979), and Moore (in Moore 1959, p. 0217) regarded this fusion as an ordinal
character. Included in this order were the oryctocephalids; more recent work (Chernysheva 1962,
pi. 6, fig. 1; Shergold 1969, pi. 1, fig. 4; Lu and Qian 1983, pi. 3, fig. 7) has provided excellent
examples that support Rasetti's drawing (1952, pi. 1, fig. 1) of the cephalic doublure, including the
rostral plate with the hypostome fused to it. The two components of the plate are separated only
by a change in slope, ill-defined medially, and have not been found to occur separately. Shergold
(1969, pi. 2, fig. 4) also showed the anterior wing of the hypostome, which must have extended up
close beneath the axial furrow immediately in front of the eye ridge.

A hypostomal suture was present in trilobites of widely differing morphology. The evidence for
such a suture having been functional includes the occurrence of rostral plate and hypostome in
isolation, and possession by the hypostome of a well-defined anterior margin appropriately shaped
to fit against the margin of the cephalic doublure. In the cases of Holmia (text-fig. 3), Redlichia
(text-fig. 6), and Xystridura (text-fig. 9) I have questioned the existence of a functional hypostomal
suture because specimens are known that show rostral plate and hypostome linked together, with
an impressed line at the junction. Such specimens have been considered (Stubblefield 1936, p. 413;
Harrington in Moore 1959, p. 066, fig. 44 a, b ) to indicate that the suture was in a state of
symphysis, and was not functional in ecdysis. However, in species of each of these genera isolated
hypostomes have been figured which appear to cast doubt on this interpretation. It may be that

WHITTINGTON: CAMBRIAN TRILOBITES

601

an isolated hypostome was entombed as a part of a completely dissociated, moulted exoskeleton,
whereas when found in place in a complete exoskeleton, the specimen is of a dead, whole animal
entombed between moults. That is, the hypostomal suture was functional only at ecdysis, the
hypostome having been attached to the rostral plate by the closed suture between moults. Such
an argument may be applicable to the originals of text-fig. 2, for example, but apparently not to
the Redlichia figured by Opik (1958, pi. 4, fig. 2). He argued that this specimen was a moult that
had the cranidium displaced, but rostral plate and hypostome linked to each other and to the free
cheeks. A parallel case is that of the specimens of Bathynotus described above that were also
presumably moults. In these the cranidium may be displaced, but the hypostome and free cheeks
were separated as a unit from the rest of the exoskeleton, and may be the right way up or inverted
in relation to it. Yet there is clearly a slight displacement between free cheeks and hypostome,
indicating a functional hypostomal suture that was not apparently a primary line of separation at
ecdysis. It appears likely that slightly greater disturbance of the moult would have resulted in
greater separation of parts and the occurrence of an isolated hypostome. The taphonomy of
trilobite exoskeletons needs further study, and claims that particular sutures were in a state of
symphysis need re-examination.

Isolated hypostomes of olenelloids other than Holmia have been figured (Palmer 1957; Poulsen
1958, pi. 7, figs. 8 and 9; Fritz 1972, pi. 3, figs. 11 and 12; pi. 14, fig. 14), which also show the
large anterior wing and the evenly curved anterior margin; whether the holaspid hypostome was
detached, or a short (tr.) hypostomal suture was functional, is not known. In such redlichioids as
Sardoredlichia (Rasetti 1972), and in emuellids (Pocock 1970), the evidence for a functional
hypostomal suture is that cited above, as it is for the originals of text-figs. 16, 18, and 20, and
more post-Cambrian trilobites. The queries in text-fig. 5 of Bathynotus relate to interpretation, not
function. If the hypostome is regarded as inserted into the doublure at an inverted V-shaped
hypostomal suture, it is an arrangement without parallel. If the hypostome was fused with the
rostral plate, and the inverted V-shaped sutures are connective, a broadly triangular rostral plate
having the apex anteriorly directed is otherwise unknown.

A median suture crossing the doublure, to which the hypostome was joined at the hypostomal
suture, is present in various late Middle and Upper Cambrian trilobites. Specimens of Dikelocephalus
(Ulrich and Resser 1930, pi. 10, fig. 2; pi. 14, figs. 3 and 4; pi. 15, figs. 2 and 9) offer ambivalent
evidence on the presence or absence of a median suture, but the presence of such a suture has
been cited (Ludvigsen and Westrop 1983, p. 28) as characteristic of Dikelocephalacea, a superfamily
that is regarded as including saukiids and ptychaspidids. The evidence for a median suture in
saukiids is not compelling (even in silicified specimens described by Ludvigsen 1982, fig. 58 a-j),
but is more satisfactory in ptychaspidids (Ludvigsen 1982, figs. 58A-p, 59, 60 a-k; Ludvigsen and
Westrop 1986, fig. 4f). The hypostome of Dikelocephalus (e.g. Ulrich and Resser 1930, pi. 10, figs.
2 and 3; pi. 11, fig. 4) is transversely rectangular in outline, with broad lateral and posterior
borders; that of saukiids is poorly known, and the 'hypostome' attributed to the ptychaspidid
Kathleenella (Ludvigsen 1982, p. 92, figs. 31, 59 o-r, 60 ci e) is not a hypostome (R. Ludvigsen,
pers. comm. 25 March 1987). In Eurekia (text-fig. 20) the evidence for the ventral structure of the
cephalon is adequate; relationships of the eurekiids are considered (Ludvigsen and Westrop 1983,
p. 28) uncertain. Another group having a median suture is represented by Proceratopyge (text-fig.
16). A median suture is known in the Upper Cambrian genera Theodenisia (Rasetti 1954, fig. 3 a),
Leiocoryphe , Plethometopus , and Stenopilus (Rasetti 1959, p. 385), as well as Housia (Rasetti 1952,
p. 892); in none of these is the hypostome known.

Altitude of hypostome , relation to mouth , and possible movement

In the holaspid of species of genera in which the hypostome was not attached to the cephalic
doublure (text-figs. 1, 7, 11-15, 17, 19) each restoration assumes that the tip of the anterior wing
lay close beneath the axial furrow, and that an approximately horizontal attitude of the hypostome
was reasonable. In Pagetia (text-fig. 1) the hypostome projected below the plane in which the
lateral margins of the cephalon lay, but in others the convexity of the cephalon (text-figs. 15 and

602

PALAEONTOLOGY, VOLUME 31

17) and the upward anterior arch (text-figs. 13 and 19) kept the hypostome above this level.
Attachment at the hypostomal suture in Holmia (text-fig. 3) and Xystridura (text-fig. 9), and the
tip of the anterior wing lying close to the ridge formed by the axial furrow, implies that the
hypostome would have been held firmly in a horizontal attitude. Symphysis of the hypostomal
suture would have contributed to this firmness. The convexity of the middle body of the hypostome
of both these genera was such that it was partially visible in the lateral view of the cephalon. In
Redlichia (text-fig. 6) the rostral plate was held firmly in place by the device of the interlocking
pits, and the hypostome attached to it in a horizontal attitude, with the tip of the anterior wing
close against the axial furrow. Symphysis at the hypostomal suture would have contributed to
holding the hypostome in place. Fusion of rostral plate and hypostome (text-figs. 8 and 10), and
a close fit at the rostral and connective sutures, meant that the hypostome was held horizontally
with the tip of the anterior wing close against the ridge formed by the axial furrow. This arrangement
seems devised to hold the hypostome rigid, the close-fitting rostral and connective sutures being
at right angles to one another. Whether the sutures in Bathynotus (text-fig. 5) are regarded as
connective or hypostomal, a close fit along them, combined with the position of the anterior wing,
would have braced the hypostome rigidly. The hypostomes of Kootenia and Olenoides (Whittington
1975, pp. 121-122, 135) were fused with the rostral plate, and in Olenoides there was a wing process
that probably lay close to the boss formed by the anterior pit (Whittington 1975, fig. 25). These
devices, if closely linked, helped to hold the hypostome firmly in position. The furrow in the long
anterior wing of Fieldaspsis (PI. 54, figs. 2 and 3) may have formed a ridge on the inner surface,
that lay against the ridge formed by the axial furrow (no anterior pit is developed) and functioned
in the same way. In Welleraspis (text-fig. 18) and Eurekia (text-fig. 20) anterior wings and presumed
attachment held the hypostome firmly, the attitude being steeply downward in the latter. In
Proceratopyge (text-fig. 16) the ventral aspect recalls that of asaphids (Whittington, in press, figs.
3 and 4), though in contrast the attachment was along a relatively short (tr.) hypostomal suture
that lay almost in one plane, and there was a relatively large anterior wing that braced the
hypostome. The different ventral structures found in this wide range of taxa having an attached
hypostome all appear to have braced the hypostome rigidly against the rest of the cephalic
exoskeleton.

An anterior pit in the axial furrow has rarely been described in holaspid Cambrian trilobites.
Such a pit is regarded as characteristic of dorypygids (Poulsen in Moore 1959, p. 0217), and Opik
(1982) observed them in some dolichometopids. Apparently an anterior boss, which lay close to a
wing process and aided in bracing the hypostome firmly against the rest of the exoskeleton, was
not as widespread and important a device in Cambrian as in post-Cambrian trilobites (Whittington,
in press). I consider, however, that in Cambrian trilobites the function of the anterior wing was
to brace the hypostome, though it may have lacked this particular device. The presence in some
species of a wing process, but not apparently an anterior boss, suggests that the latter was developed
subsequently as a complementary structure.

I have reviewed (Whittington, in press) the evidence for believing that the backward-facing
mouth of the trilobite lay above and just behind the posterior margin of the hypostome. Restorations
of Olenoides (Whittington 1975) and Triarthrus (Whittington and Almond 1987) suggest that firm
bracing of the hypostome may have aligned the mouth axially with the coxal gnathobases that
brought food forward along the ventral mid-line. In species of genera in which the hypostome was
detached, it was less firmly so aligned, but the position and manner of any link between the anterior
wing and dorsal exoskeleton would have been important in positioning the mouth. Possible
movements of the hypostome— vibratory or swinging up and down — have long been discussed
(Stubblefield 1936, p. 410; Whittington, in press). In species in which the hypostome was fused to
the rostral plate, as in Paradoxides and Fieldaspsis (text-figs. 8 and 10), such movements would
have been impossible if there was a close fit at the rostral and connective sutures; the anterior
wing braced the hypostome firmly. In species that appear to have had a hypostomal suture (text-
figs. 3, 6, 16, 18, 20), movement up and down about this suture would have required extension
and contraction of muscles at the tips of the wings. A rocking motion of the hypostome about the

WHITTINGTON: CAMBRIAN TRILOBITES

603

tips of the wings would have required a membrane along the hypostomal suture capable of
extension and contraction. Silicified material representing many families of post-Cambrian trilobites
has shown that there was a close fit along the hypostomal suture at flat faces that cut across the
thickness of the exoskeleton (Whittington, in press), and movement of the types mentioned above
at such sutures appears unlikely. If the fit at the hypostomal suture was similarly close in Cambrian
species, any movement of the hypostome appears to have been equally unlikely, and the anterior
wing would have served to brace the hypostome against the rest of the exoskeleton. More exact
knowledge of the nature of the attachment at the hypostomal suture in Cambrian trilobites may
clarify this question of possible movement. In genera in which the hypostome was detached, the
nature and amount of any movement would have depended on arrangement of muscles and
flexibility of the integument.

Evolution of hypostome and ventral cephalic sutures in Cambrian and post-Cambrian trilobites

This review and that on post-Cambrian hypostomes (Whittington, in press) has shown the basic,
conservative similarity of all trilobite hypostomes— the convex, subdivided middle body with the
macula; the presence of anterior, and probably posterior wings; the lateral notch; and convex
lateral and posterior borders. The wing process is present in the Lower Cambrian Crassifimbria
(Palmer 1958), but not linked in the holaspis to an anterior boss. The wing process is most widely
developed in post-Cambrian forms; in cheirurids, encrinurids, pliomerids, and calymenids it is
associated with a prominent anterior boss to form a device that assisted in bracing the hypostome
firmly. This device is not confined to these groups but appears, for example, in some proetids and
some lichids. The lateral notch in Cambrian hypostomes was wide (exs.) and shallow, extending
between anterior wing and projection of the lateral border, in contrast to the deep, narrow notch
and conspicuous shoulder in the post-Cambrian cheirurids and their allies. It appears probable
that the antenna passed through this notch as it curved downward and forward (Whittington, in
press). Wide lateral and posterior borders have been described in the hypostome of a small number
of late Middle and Upper Cambrian genera, the posterior border in Palaeodotes (Opik 1967, pi.
50, fig. 3) having a median notch. Similar features occur more commonly in various post-Cambrian
trilobites such as asaphids, remopleuridids, and lichids. The rhynchos, a raised median area on
the middle body of the hypostome, related to enrolment, is only known in post-Cambrian trilobites.
It appears from the present drawings that the attitude of the hypostome in most Cambrian trilobites
was approximately horizontal; a steep downward attitude is suggested for the late Cambrian
Eurekia ; an upward attitude is not known. After the Cambrian, groups in which the hypostome
was detached are fewer, and were much reduced in the Upper Palaeozoic; fusion of rostral plate
and hypostome is not known. Broadly, evolution of the hypostome is in the direction of attachment,
and of the development of more diverse special structures and attitudes in post-Cambrian forms
that are characteristic of particular families, and may reflect particular habits and adaptations.
These special structures include elongation of the anterior wing and elaboration of the wing process
and distal tip of the wing, structures that aided in bracing the hypostome in a particular attitude
against the rest of the cephalic exoskeleton.

Studies on the ontogeny of trilobites (Whittington 19576; Palmer 1957, 1958, 19626; Chatterton
1980) have revealed the relatively large, spinose hypostome as typical of protaspides. In Sao
(Whittington 19576, fig. 6g), Crassifimbria (Palmer 1958, pi. 26, figs. 5 and 6), and Bathvuriscus
(Robison 1967) the hypostome was attached, and fused to the rostral plate in Bathyuriscus; a
shallow anterior pit was developed. Attached hypostomes (in one asaphid fused with the rostral
plate in its earliest stages: Tripp and Evitt 1986) are characteristic of protaspides of some post-
Cambrian trilobites; in such examples the glabella extended far forward, close to the anterior
margin. During development, the hypostome in Sao (text-fig. 13), Crassifimbria (Palmer 1958,
p. 162), and Aphelaspis (text-fig. 17; Palmer 19626, fig. 2a) became detached, as the convexity of
the cephalon and the length (sag.) of the preglabellar field increased and the hypostome retained its
relation to the glabella and position of the anterior pit (the pit disappears). In Bathyuriscus and
many post-Cambrian trilobites such detachment did not occur, the preglabellar field being short

604

PALAEONTOLOGY, VOLUME 31

(sag.) or absent. It is possible that heterochrony (Robison 1967; McNamara 1986) was one of
many factors in lines of evolution that led, for example, to retention of the fused rostral plate and
hypostome in the holaspid stages. The relationship between length of preglabellar field, cephalic
convexity, and detachment of the hypostome is illustrated in the ontogeny of Olenellus (text-fig.
4). The hypostome appears not to have been attached during known developmental stages, when
a preglabellar field is present. If the developmental stages of Holmia (text-fig. 3) were similar,
attachment of the hypostome may have taken place in step with anterior expansion of the glabella
to bring it in contact with the doublure of the anterior border.

It has been suggested (Stubblefield 1936, p. 410, fig. 2; Hupe 1954, p. 15, fig. 9; Robison 1964,
p. 520; Opik 1967, p. 214) that an evolutionary trend may have led to the reduction in width (tr.)
of the rostral plate, resulting in the median suture. Diagrams illustrating this view are a
morphological, not a phylogenetic, series (as Stubblefield pointed out) and Palmer’s diagram (1960,
fig. 8), referred to by Robison and Opik, was not presented as illustrating an evolutionary trend.
It showed a median suture crossing the doublure in Housia and a relatively narrow (tr.) rostral
plate in Prehousia. Neither Palmer’s illustrations of species of these genera (1960, pi. 7, figs. 1-19;
1965, pi. 12, figs. 1-11, 16-26; pi. 13, figs. 1-18) nor earlier work offer clear evidence for his
diagram; Rasetti (1952, p. 892) referred to specimens giving evidence of a median suture in Housia.
The hypostome of neither genus is known. I conclude that Palmer’s diagram (1960, fig. 8) needs
substantiation, but that in any event it was not intended to show an evolutionary lineage; no such
lineage appears to have been demonstrated. Thus the origin of the median suture, and whether it
took place more than once, appears to be unknown. Various late Middle and Upper Cambrian
trilobites having such a suture have been mentioned above, and two in which the hypostome is
known are illustrated (text-figs. 16 and 20). These various trilobites may not be closely related,
but until we know more of ventral sutures and hypostomes relationships will remain uncertain. It
might be expected, for example, that Richardsonella and its allies, if they are remopleuridids, would
have a median suture. However, in a specimen from Alaska (Palmer 1968, pi. 14, fig. 8) no suture
crosses the doublure.

The hypostome and supra-generic relationships

Text-figs. 1, 3, 5-20 reflect the limitations of knowledge, and suggest that, in contrast to hypostomes
of post-Cambrian trilobites, those of the Cambrian are not so readily distinctive of family groups.
The olenelloid hypostome (text-figs. 2 and 3) may prove to be distinctive, as may that of Paradoxides
(PI. 53, figs. 1-4) and its allies, although it was fused with the rostral plate in species of Paradoxides
(in the restricted sense of Snajdr 1958). The Fieldaspis (PI. 54, figs. 1-3) type of rostral-hypostomal
plate is readily recognizable, and may prove of value in determining family relationships within
the corynexochoids. Opik (1982, p. 7) included within dolichometopids (which he regarded as
corynexochoids) a subfamily in which rostral plate and hypostome were not fused. Supposed
symphysis of the hypostomal suture or the rostral-hypostomal plate do not appear to be characters
of high taxonomic rank.

The detached hypostomes of various genera (text-figs. 1,7, 11-15, 17, 1 9) do not exhibit distinctive
features, but the morphology of other parts of the exoskeleton shows that some clearly belong to
different family groups (e.g. Pagetia , Dolerolenus , Paraholinella). The lack of distinctive features
in the hypostome is reflected in many publications on Cambrian trilobites in which unassigned, or
doubtfully assigned, isolated hypostomes are described. Detachment in the holaspis results from
a variety of factors, including width (sag. and exs.) and form of the doublure, length of preglabellar
field, and convexity of the cephalon, and this single character cannot be taken as characteristic of
any particular group. Thus to suggest (Rasetti 1952, p. 894) that trilobites with a detached
hypostome be referred to as ‘the ptychopariid type’ may be misleading, and the definition of this
‘type’ given by Harrington (in Moore 1959, p. 067) is too general to be useful. The difficulties in
defining the ptychoparioid type of trilobite are notorious, and are well illustrated by Palmer’s
(1958) description of Crassifimbria. This genus is regarded as a ptychoparioid (in Moore 1959) in
agreement with Palmer, who noted (1958, p. 159) the similarity to Agraulos (text-fig. 14). As Opik

WHITTINGTON: CAMBRIAN TRILOBITES

605

(1961, p. 143) pointed out, the cephalon of Crassifimbria was far more like that of Agraulos than
it was like Ptychoparia (text-fig. 11), and it might be regarded as an ellipsocephalid, and hence
transferred to the redlichioid group.

Acknowledgements. Helpful comments on earlier drafts by Sir James Stubblefield. FRS, Dr R. A. Fortey,
and Dr C. P. Hughes are gratefully acknowledged, as are those by anonymous reviewers. Dr R. A. Fortey
loaned to me specimens from the British Museum (Natural History) (BMNH), Dr R. B. Rickards material
in the Sedgwick Museum, University of Cambridge (SM); Dr J. E. Almond kindly examined and photographed
the type specimens of Bathynotus in the US National Museum of Natural History (USNM). and Mr F. J.
Collier loaned additional specimens. 1 am indebted to Mrs Sandra Last for preparing the typescript, to Miss
Sheila Ripper for drawing the figures from my sketches, and to the Leverhulme Trust for their support. This
is Cambridge Earth Sciences Publication no. 1020.

REFERENCES

ahlberg, p. and Bergstrom, j. 1978. Lower Cambrian ptychopariid trilobites from Scandinavia. Sver. geol.
Unders. Ca 49, 1 40.

— and johansson, j. 1986. Lower Cambrian olenellid trilobites from the Baltic faunal province. Geol.
For. Stockh. Fork. 108, 39-56.

ANGELIN, N. p. 1854. Pcilaeontologia Scandinovica , 92 pp. Lund.

barrande, J. 1846. Notice preliminaire sur le Systeme Silurien et les trilobites de Boheme , vi + 97 pp. Leipsic.

- 1852. Systeme Silurien du centre de la Boheme. Iere partie. Recherches paleontologiques, Vol. I. Crustaces,
Trilobites , xxx + 935 pp., 51 pis. Prague and Paris.

bergstrom, j. 1973. Classification of olenellid trilobites and some Balto-Scandian Species. Norsk geol. Tidsskr.
53, 283-314.

- and levi-setti, r. 1978. Phenotypic variation in the Middle Cambrian trilobite Paradoxides davidis
Salter at Manuels, southeast Newfoundland. Geol. Palaeont. 12, 1 40.

billings, E. 1861. On some new or little-known species of Lower Silurian fossils from the Potsdam Group
(Primordial zone). Geol. Surv. Canada , Palaeozoic Fossils , 1, 1-24.
boeck, c. 1827. Notitser til laeren om trilobiterna. Magazin Naturv. Christ. 8, 1, 11 44.
brogger, w. c. 1882. Die Silurischen etagen 2 und 3 im Kristianiagebeit und auf Eker , 376 pp. Kristiania
[Oslo],

brongniart, a. In brongniart, a. and desmarest, a. g. 1822. Histoire naturelle des Crustaces fossiles, 154 pp.
Paris.

burmeister, h. 1843. Die organisation der trilobiten , xii+ 148 pp. Berlin.

chatterton, b. d. e. 1980. Ontogenetic studies of Middle Ordovician trilobites from the Esbataottine
Formation, Mackenzie Mountains, Canada. Palaeontographica , A171, 1 74.

Chernysheva, n. e. 1962. Cambrian trilobites of the family Oryctocephalidae. In Problemy neftegazonosti
Sovetskoi Arktiki. Paleontologiya i biostratigraphiya. 3. Trudy nauchno-issled. Inst. Geol. Arkt. 127,
3-64.

cossman, m. 1902. Rectifications de la nomenclature. Revue crit. Paleozool. 16, 52.
emmrich, h. f. 1839. De trilobitis, 56 pp. Berolini, Berlin.

fritz, w. h. 1972. Lower Cambrian trilobites from the Sekwi Formation type section, Mackenzie Mountains,
northwestern Canada. Bull. geol. Surv. Can. 212, 1 58.
hall, j. 1859. Contributions to the palaeontology of New York. 12th Ann. Rep. NY St. Cab. nat. Hist.,
pp. 7-64.

— 1860. 13th Ann. Rep. NY St. Cab. nat. Hist., p. 118.
hawle, i. and corda, a. j. c. 1847. Prodrom einer Monographic der bohmischen trilobiten , 176 pp. Prague.
henningsmoen, g. 1957. The trilobite family Olenidae. Skr. norske Vidensk-Akad. mat. nat. Kl. 1957, 1,
1 303.

- 1959. Rare Tremadocian trilobites from Norway. Norsk geol. Tiddskr. 39, 153 173.
holm, G. 1887. Om Olenellus kjerulfi Linrs. Geol. For. Stockh. Fork. 9, 493-522.

hunt, a. s. 1966. Submarginal suture and ventral plate of an agnoslid trilobite. J. Paleont. 40, 1238 1240.
hupe, p. 1953a. Contribution a l’etude du Cambrien inferieur et du PreCambrien III de l’Anti-Atlas Marocain.
Notes Mem. Serv. geol. Maroc, 103, I 402.

— 1953 b. Classe des trilobites. In piveteau, j. (ed.). Traite de Paleontologie, 3, 44-246. Masson, Paris.

606

PALAEONTOLOGY, VOLUME 31

hupe, p. 1954. Classification des trilobites. Annls Paleont. 39, 59 168.

jago, J. b. 1987. Idamean (Late Cambrian) trilobites from the Denison Range, south-west Tasmania.
Palaeontology, 30, 207-231.

jell, p. a. 1970. Pagetia ocellata , a new Cambrian trilobite from northwestern Queensland. Mem. Qd Mus.
15, 303-313.

1975. Australian Middle Cambrian eodiscoids, with a review of the Superfamily. Palaeontographica ,
A 150, 1-97.

kiaer, j. 1916. The Lower Cambrian Holmia fauna at Tomten in Norway. Skr. norske Vidensk-Akad. mat.
nat. Kl. 10, 1-140.

kobayashi, t. 1935. The Cambro-Ordovician formations and faunas of South Chosen. Palaeontology.
Pt. III. J. Fac. Sci. Tokyo Univ. (Sect. 2), 4, 49-344.

— and kato, F. 1951. On the ontogeny and the ventral morphology of Redlichia chinensis with description
of Alutella nakamurai new gen. and sp. Ibid. 8, 99-143.

leanza, a. f. 1949. Olenopsis Ameghino 1889 (un Reodor) versus Olenopsis Bornemann 1891 (un trilobite).
Revta Asoc. geol. argent. 4, 1 .

lindstrom, G. 1901. Researches on the visual organs of the trilobites. K. svenska VetenskAkad. Handl. 34,
1-86.

linnarsson, j. G. o. 1869. Om vestergotlands Cambriska och Siluriska aflagringar. Ibid. 8, 1-89.

— 1871. Om nagra forsteningar fran sveriges och Norges ‘primordialzon’. Ofvers. K. VetenskAkad. Fork.
Stockh. 1871, 789-796.

lu yan-hao, and qian yi-yuan. 1983. Cambro-Ordovician trilobites from eastern Guizhou. Palaeont.
cathay ana , 1, 1 105.

ludvigsen, r. 1982. Upper Cambrian and Lower Ordovician trilobite biostratigraphy of the Rabbitkettle
Formation, western District of Mackenzie. Contr. Life Sci. Div. R. Ont. Mus. 134, 1 188.

— and westrop, s. R. 1983. Franconian trilobites of New York State. Mem. NY St. Mus. 23, 1-82.

— 1986. Classification of the Late Cambrian trilobite Idiomesus Raymond. Can. J. Earth Sci. 23,
300-307.

mcnamara, k. J. 1986. The role of heterochrony in the evolution of Cambrian trilobites. Biol. Rev. 61, 121-
156.

matthew, G. f. 1888. Illustrations of the fauna of the St John Group. No. 4. Pt. I. Description of a new
species of Paradoxides (Paradoxides regina). Pt. 2. The smaller trilobites with eyes (Ptychoparidae and
Ellipsocephalidae). Proc. Trans. R. Soc. Can. 5 (sect. 4), 115-166.

— 1890. On Cambrian organisms in Acadia. Ibid. 7 (sect. 4), 135-162.
meek, F. B. 1874. Preliminary report upon invertebrate fossils. In white, c. a. (ed. ). US geol. geogr. Swv., W.
100th Meridian, Report, pp. 5-27.

moberg, j. c. and segerberg, c. o. 1906. Bidrag till kannedomen om Ceratopygeregionen. Lunds Univ. Arsskr.
nf 2, 2, 1-113.

MOORE, R. c. (ed.). 1959. Treatise on invertebrate paleontology. Part O. Arthropoda 1, xix + Ol-560. Geological
Society of America and University of Kansas Press, New York and Lawrence, Kansas.
nikolaisen, f. 1986. Olenellid trilobites from the uppermost Lower Cambrian Evjevik Limestone at Tomten
in Ringsaker, Norway. Norsk geol. Tiddskr. 66, 305-309.

— and henningsmoen, G. 1985. Upper Cambrian and lower Tremadoc olenid trilobites from the Digermul
peninsula, Finnmark, northern Norway. Norg. geol. Unders. 40(1, I 49.

opik, a. A. 1958. The Cambrian trilobite Redlichia : organization and generic concept. Bull. Bur. Miner.
Resour. Geol. Geophys. Aust. 42, 1-50.

1961. The geology and palaeontology of the headwaters of the Burke River, Queensland. Ibid. 53,
1 249.

1963. Early Upper Cambrian fossils from Queensland. Ibid. 64, 1-107.

1967. The Mindyallan fauna of north-western Queensland. Ibid. 74 (1), 1 104, (2) 1 167.

1975. Templetonian and Ordian xystridurid trilobites of Australia. Ibid. 121, 1-184.

1982. Dolichometopid trilobites of Queensland, Northern Territory, and New South Wales. Ibid. 175,

1-85.

owens, R. m., fortey, r. a., cope, j. c. w., rushton, a. w. a. and bassett, m. g. 1982. Tremadoc faunas
from the Carmarthen district. South Wales. Geol. Mag 119, 1-38.
palmer, a. r. 1957. Ontogenetic development of two olenellid trilobites. J. Paleont. 31, 105 128.

1958. Morphology and ontogeny of a Lower Cambrian ptychoparioid trilobite from Nevada. Ibid. 32,
154 170.

WHITTINGTON: CAMBRIAN TRILOBITES

607

I960. Trilobites of the Upper Cambrian Dunderberg Shale, Eureka District, Nevada. Prof. Pap. US
geol. Surv. 334-C, 53-109.

1962a. Glyptagnostus and associated trilobites in the United States. Ibid. 374-F, 1 49.

1962 b. Comparative ontogeny of some opisthoparian, gonatoparian, and proparian Upper Cambrian
trilobites. J. Paleont. 36, 87-96.

— 1965. Trilobites from the Late Cambrian pterocephaliid biomere in the Great Basin, United States.
Prof. Pap. US geol. Surv. 493, 1 105.

1968. Cambrian trilobites of east-central Alaska. Ibid. 559-B, 1115.

1977. Biostratigraphy of the Cambrian System— a progress report. Ann. Rev. Earth Planet. Sci. 5,
13-33.

and gatehouse, c. G. 1972. Early and Middle Cambrian trilobites from Antarctica. Prof. Pap. US geol.
Surv. 456-D, 1 36.

— and halley, r. B. 1979. Physical stratigraphy and trilobite biostratigraphy of the Carrara Formation
(Lower and Middle Cambrian) in the southern Great Basin. Ibid. 1047, i v, 1 — 131.

pocock, k. j. 1970. The Emuellidae, a new family of trilobites from the Lower Cambrian of South Australia.
Palaeontology, 13, 522-562.

poulsen, c. 1927. The Cambrian, Ozarkian and Canadian faunas of northwest Greenland. Meddr Granland ,
70, 237-343.

1958. Contribution to the palaeontology of the Lower Cambrian Wulff River Formation. Ibid. 162,
1-24.

rasetti, f. 1945. New Upper Cambrian trilobites from the Levis Conglomerate. J. Paleont. 19, 462-478.

— 1948. Middle Cambrian trilobites from the conglomerates of Quebec. Ibid. 22, 315-339.

— 1951. Middle Cambrian stratigraphy and faunas of the Canadian Rocky Mountains. Smithson, misc.
Colins, 116, I 270.

1952. Ventral cephalic sutures in Cambrian trilobites. Am. J. Sci. 250, 885-898.

— 1954. Phylogeny of the Cambrian trilobite family Catillicephalidae and the ontogeny of Welleraspis. J.
Paleont. 28, 599 612.

— 1957. Additional fossils from the Middle Cambrian Mt. Whyte Formation of the Canadian Rocky
Mountains. Ibid. 31, 955-972.

— 1959. Trenrpealeauian trilobites from the Conococheague, Frederick and Grove limestones of the central
Appalachians. Ibid. 33, 375-398.

— 1965. Upper Cambrian trilobite faunas of northeastern Tennessee. Smithson, misc. Colins , 148, 1 127.

— 1967. Lower and Middle Cambrian trilobite faunas from the Taconic sequence of New York. Ibid. 152,
I 111.

— 1972. Cambrian trilobite faunas of Sardinia. Atti Accad. naz. Lined Memorie , Ser. 8, Sec. Ila, 11,
I 100.

Raymond, p. e. 191 3a. A revision of the species which have been referred to the genus Bathyurus. Bull.

Victoria meml Mus. geol. Surv. Can. 1, 51-69.

19136. On the genera of the Eodiscidae. Ottawa Nat. 27, 101 106.

resser, c. e. 1935. Nomenclature of some Cambrian trilobites. Smithson, misc. Colins, 93, 1-46.

1937. Third contribution to nomenclature of Cambrian trilobites. Ibid. 95, 1 29.

— and howell, b. f. 1938. Lower Cambrian Olenellus zone of the Appalachians. Bull. geol. Soc. Am. 49,
195-248.

robison, r. a. 1964. Late Middle Cambrian faunas from western Utah. J. Paleont. 38, 510-566.

— 1967. Ontogeny of Bathyuriscus fimbriatus and its bearing on affinities of corynexochid trilobites. Ibid.
41, 213-221.

1972. Hypostoma of agnostid trilobites. Letlmia, 5, 239-248.

— 1976. Middle Cambrian trilobite biostratigraphy of the Great Basin. Geology Stud. Brigham Young
Univ. 23, 93-109.

rushton, a. w. a. 1983. Trilobites from the Upper Cambrian Olenus zone in central England. In briggs,
d. e. G. and lane, p. d. (eds.). Trilobites and other early arthropods: papers in honour of Professor H. B.
Whittington, F.R.S. Spec. Pap. Palaeont. 30, 107-139, pis. 14 19.
salter, j. w. 1863. On the discovery of Paradoxides in Britain. Q. Jl geol. Soc. Lond. 19, 274-277.
schindewolf, o. h. 1955. Uber Hypostom und gesichtsnacht bei Redlichia (Trilob.). Neues Jb. Geol. Palaont.
Mh. 1955, 130-136.

— and seilacher, a. 1955. Beitrage zur kenntnis des Kambriums in der Salt Range (Pakistan). Abh. math-
naturw. Kl. Akad. IViss. Lit. 1955 (10), 1 90.

608

PALAEONTOLOGY, VOLUME 31

schlotheim, e. f. 1823. Nachtrage zur Petrefactenkunde II, 1 14 pp. Gotha.

sdzuy, K. 1961. Das Kambrium Spaniens. Pt. II: trilobiten. Abh. math.-naturw. Kl. Akad. \ Viss. Lit. 1961
(7, 8), 501 -693.

1967. Trilobites del Cambrico Medio de Asturias. Trab. Geol. Fac. cienc. Univ. Oviedo . 1, 77 133.
shaw, a. b. 1955. Paleontology of northwestern Vermont. V. The Lower Cambrian fauna. J. Paleont. 29,
775-805.

SHERGOLD, j. H. 1969. Oryctocephalidae (Trilobita: Middle Cambrian) of Australia. Bulk Bur. Miner. Resour.
Geol. Geophys. Aust. 104, I -66.

1980. Late Cambrian trilobites from the Chatsworth Limestone, western Queensland. Ibid. 186,
1-111.

— 1982. Idamean (Late Cambrian) trilobites, Burke River Structural bell, western Queensland. Ibid. 187,
1-69.

— and sdzuy, k. 1984. Cambrian and early Tremadocian trilobites from Sultan Dag, central Turkey.
Senckenberg. leth. 65, 51-135.

snajdr, m. 1958. The trilobites of the Middle Cambrian of Bohemia. Rozpr. ustfed. Ust. geol. 24, 1 -230.
1985. Two new paradoxid trilobites from the Jince Formation (Middle Cambrian, Czechoslovakia).
Vest. ustr. Ust. geol. 61, 169-174.

Stubblefield, c. J. 1936. Cephalic sutures and their bearing on current classifications of trilobites. Biol. Rev.
11, 407-440.

swinnerton, h. h. 1915. Suggestions for a revised classification of trilobites. Geol. Mag. ns (Dec. 6), 2, 487-
496, 538-545.

tasch, p. 1951. Fauna and paleoecology of the Upper Cambrian Warrior Formation of central Pennsylvania.
J. Paleont. 25, 275-306.

taylor, m. e. 1978. Type species of the Late Cambrian trilobite Eurekia Walcott. 1916. Ibid. 52, 1054-
1064.

tripp, R. p. and evitt, w. r. 1986. Silicified trilobites of the family Asaphidae from the Middle Ordovician
of Virginia. Palaeontology. 29, 705-724.

ulrich, E. o. and resser, c. e. 1930. The Cambrian of the upper Mississippi valley. I, Trilobita: Dikelocephalinae
and Osceolinae. Bull pub!. Mus. Milwaukee , 121, 1-122.
vogdes, a. w. 1893. Bibliography of the Palaeozoic Crustacea. Occ. Pap. Calif. Acad. Sci. 4, I 412.
walcott, c. d. 1886. Second contribution to the studies on the Cambrian faunas of North America. Bull.
US geol. Surv. 30, 1 369.

1890. The fauna of the Lower Cambrian or Olenellus zone. Rep. Dir. U.S. geol. Surv. 10, 1 774.

— 1910. Olenellus and other genera of the Mesonacidae. Smithson, misc. Co/Ins , 53, 231-422.

— 1916. Cambrian geology and paleontology. III. No. 5. Cambrian trilobites. Ibid. 64, 303-456.

1917. Fauna of the Mount Whyte Formation. Ibid. 67, 61 114.

wallerius, I. D. 1895. Undersdkningar qfver zonen med Agnostus laevigatus / Vestergotland. 72 pp. Lund.
westergard, a. h. 1936. Paradoxides oelandicus beds of Oland. Sver. geol. Unders. Avh., Ser. C, 394,
1-66.

— 1947. Supplementary notes on the Upper Cambrian trilobites of Sweden. Ibid. 489, I 34.
whitehouse, f. w. 1936. The Cambrian faunas of north-eastern Australia. Parts 1, 2. Mem. Qd Mus. 11, 59-

112.

— 1939. The Cambrian faunas of north-eastern Australia. Part 3: the polymerid trilobites. Ibid. 179-
282.

Whittington, H. B. 1957 a. Ontogeny of EUiptocephala. Paradoxides , Sao. Blainia and Triarthrus (Trilobita).
J. Paleont. 31, 934 946.

\951b. The ontogeny of trilobites. Biol. Rev. 32, 421-469.

— 1959. Silicified Middle Ordovician trilobites. Remopleurididae, Trinucleidae, Raphiophoridae, Endy-
mioniidae. Bull. Mus. comp. Zool. Harv. 121, 371 496.

— 1975. Trilobites with appendages from the Middle Cambrian Burgess Shale, British Columbia. Fossils
Strata. 4,97 136.

— In press. Hypostomes of post-Canrbrian trilobites. New Mex. Bur. Mines Miner. Resour.

and almond, j. e. 1987. Appendages and habits of the Upper Ordovician trilobite Triarthrus eatoni.
Phil. Trans. R. Soc. B317, 1-46.

- and evitt, w. r. 1954. Silicified Middle Ordovician trilobites. Mem. geol. Soc. Am. 59, 1-137.
wittke, h. w. 1984. Middle and Upper Cambrian trilobites from Iran: their taxonomy, stratigraphy and
significance for provincialism. Palaeontographica, A183, 91-161.

WHITTINGTON: CAMBRIAN TRILOBITES

609

ZHANG WENTANG, LU YANHAO, ZHU ZAOLING, QIAN YIYUAN, LIN HUANLING, ZHOU ZHIYI, ZHANG SENGUI and

yuan jinliang. 1980. Cambrian trilobite faunas of southwestern China. Palaeont. sin. 159 (New Series B
16), 1 497.

H. B. WHITTINGTON

Sedgwick Museum
Department of Earth Sciences
Downing Street

Typescript received II May 1987 Cambridge CB2 3EQ

Revised typescript received 14 September 1987 England

THE ENIGMATIC ARTHROPOD DUSLIA FROM
THE ORDOVICIAN OF CZECHOSLOVAKIA

by IVO CHLUPAC

Abstract. A restudy of Duslia insignis Jahn, 1893 from the Upper Ordovician of the Barrandian area,
Bohemia, indicates that this trilobate arthropod, originally referred to polyplacophorans, cannot be assigned
to true trilobites but shows some morphological analogies with Cheloniellon , Pseudarthron , and Triopus.
Duslia inhabited a nearshore shallow marine environment and was probably a benthic animal which lived
buried in sandy substrate near the sediment-water interface.

The Upper Ordovician Letna Formation of the Barrandian area, central Bohemia, has yielded,
apart from trilobites and other fossils, some remarkable remains of unusual arthropods; these were
partly described by Barrande (1872) and later tentatively assigned to the aglaspids, xiphostirids,
and eurypterids (Chlupac 1965; discussion in Bergstrom 1968; Eldredge 1974). One member of this
group of enigmatic arthropods is Duslia insignis , originally described by Jahn (1893) as a chitonid
mollusc. It was recognized as an arthropod by Pilsbry (1900) and Fritsch (1908), although
Knorre (1925) still assigned it to the polyplacophorans. After the exclusion of Duslia from the
polyplacophorans by Pompeckj (1912) and Quenstedt (1932a, b ), it was referred with some doubt
to the burlingiid trilobites (letter of A. Liebus cited by Quenstedt 19326; Broili 1933) and later to
the cheloniellid arthropods (Chlupac 1965), being definitely rejected from polyplacophorans by
Smith and Hoare (1987). Duslia was omitted from the Treatise on Invertebrate Paleontology and
the present revision, based on new and previously unstudied material, is the first since its original
establishment in 1893.

The new reference material includes collections made at Vesela, near Beroun, at the turn of the
century and recently by Dr M. Snajdr and the author; all are deposited in the National Museum,
Prague (L) and in the Geological Survey, Prague (ICh).

SYSTEMATIC PALAEONTOLOGY
ARTHROPODA

Genus duslia Jahn, 1893

Type and only known species. Duslia insignis Jahn, 1893, from the Upper Ordovician, Barrandian area,
Czechoslovakia.

Diagnosis. Trilobed and dorsoventrally flattened, thin exoskeleton of oval outline, with conspicuous
spinose fringe. Cephalic region large, with distinctly differentiated conical glabellar area and smooth
genal areas lacking eyes. A continuous fringe composed of flat spines borders the entire cephalon,
including the posterolateral margin. Trunk composed of ten tergites with clearly defined rhachis
and flat pleurae; pleural furrows shallow. Pleurae arranged radially; first two expand anterolaterally,
third laterally, the more posterior posterolaterally. Abaxial extremities of pleurae bordered by
spines continuing on the posterior pleural margins. Trunk terminated by short telson and spinose
furcal rami of medium length. Other appendages unknown.

Occurrence. Duslia occurs sporadically on the north-west slope of Ostry hill near Beroun, and at Vasela in
the gorge and on the ridge north-west of the former Vesela farm. At the latter locality, Duslia is concentrated

| Palaeontology, Vol. 31, Part 3, 1988, pp. 611-620, pis. 56-57.|

© The Palaeontological Association

612

PALAEONTOLOGY, VOLUME 31

in greater abundance in a distinct layer. Both localities belong to the fossiliferous biohorizon within the
upper part of the Letna Formation (Chlupac 1965).

Duslia insignis Jahn, 1893

Plates 56 and 57; text-fig. 1

1893 Duslia insignis Jahn, pp. 592-599, pi. 1, figs. 1 3.

1900 Duslia insignis Jahn; Pilsbry, p. 434.

1908 Duslia ; Fritsch, p. 9.

1912 Duslia insignis Jahn; Pompeckj, p. 357.

1925 Duslia insignis Jahn; Knorre, pp. 497 499, text-fig. 1.

1 932a Duslia insignis Jahn; Quenstedt, p. 555.

1932fi Duslia insignis Jahn; Quenstedt, p. 86.

1933 Duslia insignis Jahn; Broili, pp. 30-31.

1960 Duslia insignis Jahn; Smith, p. 174.

1965 Duslia insignis Jahn; Chlupac, p. 31.

1987 Duslia insignis Jahn; Smith and Hoare, p. 34.

Type material. Holotype (by monotypy), L26148, an internal mould (original of Jahn 1893, pi. 1, fig. 1;
refigured here as PI. 56, fig. 2), from Ostry hill (north-west slope), near Beroun, central Bohemia,
Czechoslovakia. Upper part of Letna Formation (oldest fossiliferous horizon with Deanaspis goldfussi
(Barrande) distinguished by Chlupac 1965); Lower Berounian (late Llandeilian or early Caradocian).

Other material. L27106 from the type locality. All other material from Vesela: five slabs (L26149, L26150,
L26157, L26160, ICh522) each bearing two specimens (designated a and />); twelve other specimens (L26151-
26156, L26158, L26159, L26161, L27105, ICh521, ICh7003), preserved as internal moulds and counterparts
in impure sandstones or quartzites.

Description. Dorsal exoskeleton broadly oval in outline, length/width ratio 11 I -3. Cephalic region large and
flat, subsemicircular in outline. Subconical, posteromedially placed and gently convex glabellar area is
markedly delimited by broad and laterally pit-like, deepened but indistinct furrows. Transverse lobation less
distinct; apart from the incompletely differentiated occipital ring, two or three anterior lobes are slightly
indicated in some specimens (L26151, L26155, L26161) by shallow transverse furrows. Entire glabellar region
depressed, and convexity of glabella does not exceed that of lateral parts of cephalic shield. Preglabellar and
genal regions smooth, without any traces of eyes. Indistinct shallow depressions (usually two) radiate from
glabellar region anterolaterally in some specimens (holotype L26148; best in L26150), suggesting differences
of convexity in genal regions. Narrow, shallow, rather sharp furrow, running parallel to posterior margin
and fading before reaching lateral cephalic margin, delimits the posterior border. Entire outer margin of
cephalic shield bears marked fringe of closely spaced, flat spines of almost equal length, becoming only very
slightly longer towards the genal angles. Fringe sharply separated from flat surface of cephalic shield by
continuous furrow which in some specimens shows gentle increase of anterior convexity frontomedially (PI.
56, figs. I, 2, 4, 5; PI. 57, fig. 2). Spines arranged radially and, as shown by L26161 (PI. 57, fig. 2), they
continue along posterolateral angles of cephalic shield up to posterior border, shortening markedly adaxially.
Flat anterior doublure, gently broadened medially, is shown by L26160A

On the trunk, ten tergites with clearly differentiated rhachis are defined by narrow and shallow lines which
continue from rhachis without interruption onto the lateral pleural regions; these lines evidently represent
intertergite boundaries. Rhachis composed of gently convex anteriorly curved rings and depressed, sagittally

EXPLANATION OF PLATE 56

Figs. 1-5. Duslia insignis Jahn, 1893. Letna Formation, Upper Ordovician; Vesela (figs. 1, 3-5), Ostry
(fig. 2), Bohemia, Czechoslovakia. 1, L26158, internal mould, x0-8. 2, L26148, holotype, internal mould
with partly exposed doublures of left pleurae, x0-8. 3, L26160a, internal mould, less deformed, x IT.
4, L26156, incomplete cephalic shield and anterior pleurae with spinose fringe, xO-9. 5, L26151, internal
mould, slightly bent laterally, x0-8.

PLATE 56

CHLUPAC, Duslia

614

PALAEONTOLOGY, VOLUME 31

text-fig. 1. Duslici insignis Jahn, 1893. Reconstruction of
dorsal exoskeleton.

shorter articulating facets. Rather broad and deep depressions separate the rhachis from markedly transversely
broader pleural regions.

Pleurae arranged radially; the two anterior expand anterolaterally, the third laterally, and subsequent ones
posterolaterally with successively diminishing angle of divergence. Pleurae widen slightly abaxially and bear
shallow and indistinct pleural furrows that subdivide anterior pleurae into unequal bands; posterior bands
usually more convex and narrower than anterior bands (best shown by L26160«, PI. 56, fig. 3). Details of
convexity of individual pleural parts generally obscured by overlapping and flattening of tergites. Abaxial
extremities of pleurae bordered by closely spaced spines which lengthen slightly towards posterolateral angles,
along which the spines continue up to posterior pleural margins, partly overlapping subsequent pleura. The
spines represent extensions of pleural margins and, although in shape and arrangement they resemble trilobite
appendages, they evidently belong to the dorsal exoskeleton as distal extremities of pleurae. Details of spines
generally obscured by coarse-grained sediment; they are clearly evident in L26151 (PI. 56, fig. 5; PI. 57, fig.
1), L26152, L26154, L26158, ICh521, and L26160a (PI. 56, fig. 3; PI. 57, fig. 4). These specimens show
differences in length of individual spines, the longest and strongest being at posterolateral pleural tips.

In specimens with partly exfoliated outer surface, flat pleural doublures are seen (L26153; PI. 57, fig. 3;
also L261606).

Convexity of pleural regions generally low, but exceeds that of rhachis— in transverse section, rhachis lies
in broad depression (Jahn 1893, pi. 1, fig. 3). Spinose fringe gently upraised in some specimens (e.g. L26158,
L26161). Although these features might have been accentuated by compaction, the lack of deformation
suggests their primary nature.

Posterior termination of trunk usually inadequately preserved in the coarse-grained sediment. L26160«,

EXPLANATION OF PLATE 57

Figs. 1-5. Duslia insignis Jahn, 1893. Letna Formation, Upper Ordovician; Vesela, Bohemia, Czechoslovakia.
All internal moulds. 1, L26151, enlarged part showing spinose fringe continuing along the posterolateral
extremities of pleurae, x 1 -5. 2, L26161 , specimen showing spinose fringe continuing along the posterolateral
angle of cephalic shield, surface slightly weathered, x 1-3. 3, L26153, damaged trunk with partly exposed
doublures, x TO. 4, L26160a, enlarged posterior part of the trunk with telson and furcal rami, x2-3.
5, L26156, enlarged spinose cephalic fringe, x 1-5.

PLATE 57

CHLUPAC, Duslia

616

PALAEONTOLOGY, VOLUME 31

whose relief is most distinct, exhibits behind last rhachial ring a suboval and posteriorly narrowed plate—
probably a telson— to which are attached (articulated?) two prolonged lanceolate lamellae, interpreted as
furcal rami (PI. 57, fig. 4). These are somewhat narrower and more convex than the last pleurae and bear
spines analogous to those of pleurae (indicated in L26153, L26160a, L26149n— here markedly convex furcal
rami).

Holotype has (slightly extrapolated) length of 86 mm and maximum width of 75 mm (including fringe).
Other specimens are 85 1 10 mm long, 75-95 mm wide (extrapolated, including fringe). Largest incomplete
specimen (ICh521) suggests length of c.120 mm.

Discussion. All the known specimens of Duslia are articulated and, if not broken after removal
from the rock, they are preserved as complete dorsal exoskeletons. This suggests a rather tight
connection between individual elements of the exoskeleton, especially in the rhachial region. As
shown by the dorsoventrally flexed specimen ICh7003, the pleurae could have been removed up
to the rhachial furrows. Other specimens (e.g. holotype L26148, PI. 56, fig. 2; L26151, PI. 56,
fig. 5) show gently detached pleurae in abaxial parts of the posterior segments. The rather firm
connection in the rhachial region seems to be a characteristic feature of Duslia , and contrasts with
trilobites.

The carapace is markedly thinner than in the associated trilobites; the trilobite remains generally
show no marked deformations and are preserved in limonite (which replaces the calcium carbonate
of unweathered Letna Formation). In contrast, the remains of Duslia lack a thicker limonite cover
and are preserved in a similar manner to the associated conulariids and other thin-shelled fossils
with an originally phosphatic shell. Although the cuticle of Duslia has been completely dissolved,
this mode of preservation may reflect a different original composition of the carapace to that of
trilobites.

Some specimens show exoskeletal parts that are gently deflexed in a horizontal plane; the
holotype (L26148) is slightly curved to the right in the posterior part of the trunk (PI. 56, fig. 2),
L26I51 is gently bent laterally (PI. 56, fig. 5), and similar patterns are shown to a lesser degree by
some other individuals. Some flexibility should therefore be considered likely.

The thin carapace of Duslia was commonly affected by compaction which resulted, for example,
in a gently differing convexity of the rhachis— although the general flatness of the entire exoskeleton
remains characteristic. Asymmetrical irregularities are in most cases caused by coarser rock grains,
fossil partings, or ichnofossils beneath the thin carapace (e.g. right part of L26151a, L26153),
although a pathological cause cannot always be excluded (see e.g. the elevation on the fifth pleura
of L26 150(7).

AFFINITIES

Duslia cannot be ranged with the polyplacophoran molluscs because it exhibits a different
morphology of the cephalic shield (with differentiated glabellar region), a larger number (ten) of
tergites with distinct trilobation and pronounced rhachis and pleurae, and because the posterior
termination of the trunk differs markedly from all polyplacophorans. The spinose fringe along the
dorsal exoskeleton may resemble the girdle of polyplacophorans, but its nature is quite different
(spines are projections of individual dorsal segments of Duslia).

Duslia shows certain similarities to trilobites, and especially to the unusual non-trilobite
arthropods Triopus draboviensis Barrande, Cheloniellon calmani Broili, and Pseudarthron whitting-
toni Selden and White. Duslia shares with trilobites a marked longitudinal trilobation of the
exoskeleton, the trilobite-like cephalic shield with its clearly differentiated glabellar region, and the
configuration of rhachis and pleurae. The basic difference, however, is the absence of a pygidium,
and the trunk being terminated by a telson with furcal rami. Other less important features
differentiating Duslia from trilobites are the absence of facial sutures, the radial arrangement of
trunk tergites and their firm connection along the sagittal axis, the inter-tergite boundaries
maintaining their course at the dorsal furrows, and the uniform spinose fringe bordering the dorsal
exoskeleton.

CHLUPAC: ORDOVICIAN ARTHROPOD

617

T. draboviensis Barrande, 1872 is based on a single specimen from Drabov (Ded) hill near
Beroun. It may be from the same stratigraphical horizon as Duslia or be somewhat younger; it is
preserved in a yellowish quartzite with fragments of Dalmanitina socialis (most common in the
upper fossiliferous horizon of the Letna Formation, as distinguished by Chlupac 1975). The
holotype of T. draboviensis , LI 6736, newly recognized in the collections of the National Museum,
Prague, represents an incomplete trunk showing nine radially arranged tergites with a rhachis
differentiated by shallow and ill-defined dorsal furrows. Pleurae are smooth, slightly overlapping,
and widen markedly abaxially; their posterolateral extremities are sharp, each being produced into
a short spine. Pleural furrows are only faintly indicated near the anterior margin of some pleurae.
The most posterior (left) pleura preserved extends behind the end of the rhachis in a manner that
leaves no place for a pygidium, and a spine-like telson may be postulated. Pleural regions are
inclined steeply abaxially.

The one specimen of Triopus is preserved as an internal mould, compacted in a longitudinal
direction. Anteriorly it is obscured by weathering and the coarseness of the sediment to such an
extent that it is unclear whether the most anterior portion represents the posterior part of the
cephalic shield or a remnant of a trunk tergite. Longitudinal depressions on the rhachis close to
the dorsal furrows are irregular and evidently caused (or at least accentuated) by secondary
deformation. In this respect, and also in the right extremities of pleurae, Barrande’s original figure
(1872, pi. 5, fig. 41) is idealized and restored (cf. text-fig. 2 herein).

Although Triopus resembles Duslia in its trilobation and gross morphology of radially arranged
tergites, it differs markedly in convexity of the exoskeleton, the different proportions of the rhachis
and pleurae, and the absence of a spinose fringe. The supposed cephalic shield of Triopus (Zonozoe
or Drabovaspis ; Chlupac 1965; Bergstrom 1968) exhibits a completely different morphology to that
of Duslia.

text-fig. 2. Triopus draboviensis Barrande, 1872. a, original drawing by Barrande (1872, pi. 5, fig. 41), from
the holotype L16736. b, photograph of holotype. c, schematic drawing of holotype; broken and obscure lines

dashed.

618

PALAEONTOLOGY, VOLUME 31

The systematic position of Triopus is uncertain. Barrande (1872) regarded it as a trilobite, Novak
(1885) reassigned it to non-trilobite arthropods, while Jahn (1893) ranged it with the chitonids.
Chlupac (1965) stressed its similarity to aglaspids and combined it tentatively with prosomas
described as Zonozoe Barrande or Drabovaspis Chlupac. Bergstrom (1968) considered Triopus to
be the trunk of Drabovaspis and referred it to xiphosurids. Although it is clearly non-trilobite and
possibly xiphosuran (according to the less distinct trilobation, the radial arrangement of abaxially
widened pleurae, and the evident absence of pygidium), its affinities remain obscure (especially
because of its indifferent preservation).

C. calmani Broili, 1932 from the Lower Devonian (lower Emsian) Hunsriick Shale of Germany
resembles Duslia in its oval outline, distinct trilobation, radial arrangement of tergites, shallow
and indistinct axial furrows, flat pleurae, and possession of a telson and furcal rami. The cephalic
shield of Cheloniellon , however, is markedly smaller and sagittally shorter, the eyes are well
developed, the posterior pleurae widen considerably, and the furcal rami are notably prolonged
(details in Stunner and Bergstrom 1978). No spinose fringe, so characteristic of Duslia , is present
in Cheloniellon.

Cheloniellon is the only arthropod comparable with Duslia in which appendages have been
preserved, but even this feature fails to conclusively resolve its systematic position. Broili (1932,
1933) referred Cheloniellon to a separate subclass of Crustacea, Stormer (1959) assigned it to the
subclass Trilobitoidea of the Trilobitomorpha (which appears to be a heterogeneous group;
Whittington 1979; Briggs 1983), while Stunner and Bergstrdm (1978) concluded that it occupies a
position intermediate between trilobitomorphs and chelicerates (cf. also Bergstrom 1979, 1980).

The small Upper Silurian P. whittingtoni Selden and White, 1984, from the Ludlovian lagoonal
deposits of Scotland, resembles Duslia in its oval outline, distinct trilobation, and gently radiating
flat pleurae. The incompletely preserved cephalic shield, however, was evidently much smaller than
in Duslia , with the number of trunk tergites being only seven or eight, the rhachis broader, and
the pleural furrows sharper. The exoskeleton of Pseudarthron does not show the fringe which is
so typical of Duslia. The systematic position and affinities of Pseudarthron are uncertain, although
its non-trilobite nature is clear.

As the appendages of Duslia are unknown, its systematic position and affinities remain doubtful.
The morphology of the dorsal exoskeleton points to Trilobitomorpha but not to the class Trilobita
proper. The systematic position of Duslia may be analogous to that of Cheloniellon.

ENVIRONMENT AND PALAEOECOLOG Y

Duslia occurs in marine deposits characterized by an alternation of lighter grey (yellow and brown
if weathered) sandstones and subgreywackes with markedly darker sandy and clayey siltstones in
beds several centimetres to several tens of centimetres thick. Sandstone beds (locally quartzites)
show frequent sedimentary and biogenic structures on bedding planes, and the siltstones commonly
exhibit traces of bioturbation. Most fossils are concentrated in thicker sandstone layers with
organic debris and siltstone and claystone pebbles. Graded bedding is infrequent and the siltstones
are laminated in some layers. According to Kukal (1958, 1963), the lithology suggests shallow
water, nearshore sedimentation within the reach of river-borne material, and a periodically changing
climate (possibly seasonal fluctuations).

Both Duslia- bearing localities, Ostry hill and Vesela, have been well-known palaeontological
localities since Barrande's time. Sandstone slabs with Duslia contain scattered, disarticulated
remnants of trilobites Deanaspis goldfussi (Barrande) and Dalmanitina socialis (Barrande), ostracods,
sporadic Metaconularia anomala (Barrande), and rare orthoconic nautiloids. At both localities,
some sandstone layers are rich in fragmented trilobites: apart from the dominant Deanaspis
goldfussi and Dalmanitina socialis, less common forms include Selenopeltis buchi (Barrande),
Opsimasaphus ingens (Barrande), Pharostoma pulchrum mendax Vanek, Zelizskella hawlei (Bar-
rande), Eccoptochile clavigera (Beyrich), Stenopareia panderi (Barrande), and Primaspis primordialis
(Barrande). Associated with these trilobites are orthid brachiopods Drabovia redux (Barrande),

CHLUPAC: ORDOVICIAN ARTHROPOD

619

Drabovinella draboviensis (Barrande), and the less common Petrocrania obsolete i (Barrande),
gastropods, bivalves, conulariids, and nautiloids, while echinoderm debris occurs in some layers.
Ichnofossils commonly include the vertical and oblique burrows Monocraterion, Skolithos ,
Arenicolites , and Diplocraterion , rarely the fasciculate Phycodes , and very frequently the epistratal
Palaeophycus, Gordia , and (ubiquitous) Planolites.

The composition and preservation of the fauna indicates a shallow water and high energy Benthic
Assemblage 3 in Boucot’s (1975) classification, and the same is postulated from the ichnofossils
ranged within the Craziana Ichnofacies (cf. Chlupac 1965; Havlicek 1982; Chlupac and Kukal, in
press).

The occurrence of Duslia as complete and articulated exoskeletons contrasts strikingly with the
fragmentary preservation of associated fossils and deserves particular attention. Complete trilobite
exoskeletons have been recovered from sandstones at Vesela, but they are very rare.

The fourteen specimens of Duslia (some with counterparts) collected at Vesela at the turn of the
nineteenth century (originally housed at the Technical universities in Prague and Brno, but now
in the National Museum and Geological Survey, Prague) all most likely derive from the same layer
of light grey, impure sandstone, 60-90 mm thick. This is evidenced not only by lithology but also
by analogous lower and upper bedding planes of the Duslia- bearing slabs of rocks. The relative
position of the Duslia exoskeletons seems to be uniform, with the dorsal side turned towards the
flatter bedding plane, and the same position is confirmed by pairs of specimens preserved in close
proximity on the same slab of rock (L26149o, b , L26150a, b , L26157r/, b, L26160a, b , ICh522<7,
b). Due to a lack of primary documentation during collection it is not clear whether the specimens
were deposited dorsal side upwards or downwards. All known specimens of Duslia lie within the
sandstone layers and none was found directly on the bedding plane proper. Although freshly killed
soft-bodied arthropods can survive turbulent transport over substantial distances, as recently
shown by Allison (1986) under experimental conditions, the exclusive preservation of complete
exoskeletons of Duslia and their uniform position within the sandy layers suggest that the specimens
were not subjected to any significant transport and are preserved in life position.

The firm connection of exoskeletal elements in the axial part suggests that Duslia was incapable
of enrolment. The most characteristic feature of Duslia— the spinose fringe bordering the whole
exoskeleton— may have protected limbs or other soft parts in the sandy environment. The thin
carapace of Duslia was evidently not suitable for a high energy environment on the substrate itself
and, in view of its broadly oval shape and flat morphology, the absence of eyes, and other features
mentioned above, I conclude that Duslia lived buried in the unconsolidated sandy substrate close
to the sediment-water interface.

Acknowledgements. I thank Dr D. E. G. Briggs (University of Bristol) for a kind revision of the manuscript,
and Dr V. Turek (National Museum, Prague) for a critical reading of earlier drafts and for preparing
photographs. Dr R. Prokop (National Museum, Prague) arranged access to materials deposited in the
collections of the National Museum (Natural History), Prague. Mr I. Kolebaba helped in preparing the text-
figures.

REFERENCES

Allison, p. a. 1986. Soft-bodied animals in the fossil record: the role of decay in fragmentation during
transport. Geology, 14, 979 981.

barrande, j. 1872. Systeme silurien du centre de la Boheme. lere parlie. Recherches paleontologiques.

Supplement au Vol. 1. Trilobites, crustaces divers et poissons. Prague.

Bergstrom, j. 1968. Eolimulus, a Lower Cambrian xiphosurid from Sweden. Geol. For. Stockh. Fork. 90,
489 503.

— 1979. Morphology of fossil arthropods as a guide to phylogenetic relationships. In gupta, a. p. (ed.).
Arthropod phytogeny. Van Nostrand Reinhold, New York.

— 1980. Morphology and systematics of early arthropods. Abh. Verb, naturw. Ver. Hamburg, 23, 7-42.
boucot, a. j. 1975. Evolution and extinction rate controls. Elsevier, Amsterdam.

620

PALAEONTOLOGY, VOLUME 31

briggs, d. E. G. 1983. Affinities and early evolution of the Crustacea: the evidence of the Cambrian fossils,
1-22. In schram, F. r. fed.). Crustacean phytogeny. A. A. Balkema, Rotterdam.
broili, F. 1932. Ein neuer Crustacee aus dem rheinischen Unterdevon. Sber. bayer. Akad. Wiss. 1932, 27-38.

1933. Ein zweites Exemplar von Cheloniellon. Ibid. 1933, 11-32.
chlupaC, i. 1965. Xiphosuran merostomes from the Bohemian Ordovician. Sb. geol. Ved Praha , Rada P, 5,
7-38.

— and rural, z. In press. Possible global events and the stratigraphy of the Barrandian Palaeozoic
(Cambrian Devonian, Czechoslovakia). Ibid., Rada G, 43.

eldredge, n. 1974. Revision of the Suborder Synziphosurina (Chelicerata, Merostomata) with remarks on
merostome phylogeny. Am. Mus. Novit. 2543, I 41.

fritsch, a. 1908. Problematica silurica, I 28. In Systeme silurien du centre de la Boheme. Bellmann, Prague.
havlicer, v. 1982. Ordovician in Bohemia: development of the Prague Basin and its benthic communities.
Sb. geol. Ved , Praha , Rada G, 37, 103 -136.

rnorre, h. 1925. Die Schale und die Riickensinnesorgane von Trachydermon ( Chiton ) cinereus und die
ceylonischen Chitonen der Sammlung Plate (Fauna et Anatomia ceylanica, 3, Nr. 3). Jena Z. Naturw. 61,
469-632.

jahn, j. j. 1893. Duslia , eine neue Chitonidengattung aus dem bohmischen Untersilur, nebst einigen
Bemerkungen fiber die Gattung Triopus Barr. Sber. Akad. IViss. Wien, 102, 591 603.
rural, z. 1958. Petrograficky vyzkunr letenskych vrstev barrandienskeho ordoviku. [The petrographic
research of the Letna Beds of the Barrandian Ordovician; English summary.] Sb. Ustred. Ust. Geol.
24, Odd. geol. 1 , 7-111.

— 1963. Vysledky sedimentologickeho vyzkumu barrandienskeho ordoviku. [The results of the sedi-
mentological investigation of the Ordovician in the Barrandian area; English summary.] Sb. geol. Ved
Praha , Rada G, I, 103-138.

novar, o. 1885. Studien an Hypostomen bohmischer Trilobiten, III. Sber. K. bohm. Ges. Wiss. Math.-nat.
Kl. 1885, 581 587.

pilsbry, h. a. 1900. Polyplacophora Blainville. Chitons, 433-436. In zittel, R. a. and Eastman, c. r. (eds.).

Textbook of Palaeontology, Vol. 1. Macmillan, London.
pompecrj, j. f. 1912. Amphineura, 347-357. In roschelt, e. et al. (eds.). Handworterbuch der Naturwissen-
schaften , Erste Auflage, Sechster Band. Jena.

quenstedt, w. 1932c/. Loricata, 552-555. In FISCHER, E. (ed.). Handworterbuch der Naturwissenschaften , Zweite
Auflage, Sechster Band. Jena.

1932 b. Die Geschichte der Chitonen und ihre allgemeine Bedeutung (mit Zusatzen). Paldont. Z. 14,
77-96.

smith, a. G. 1960. Amphineura, 14 1 -176. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part I.
Mollusca 1. Geological Society of America and University of Kansas Press, Boulder, Colorado and
Lawrence, Kansas.

— and hoare, r. D. 1987. Paleozoic Polyplacophora: a checklist and bibliography. Occ. Pap. Calif. Acad.
Sci. 146, 1-71.

selden, p. a. and white, d. e. 1984. A new Silurian arthropod from Lesmahagow, Scotland. Spec. Pap.
Palaeont. 30, 43 49.

stormer, L. 1959. Trilobitomorpha: Trilobitoidea, 022-037. In MOORE, R. c. (ed.). Treatise on invertebrate
paleontology. Part O. Arthropoda 1. Geological Society of America and University of Kansas Press,
Boulder, Colorado and Lawrence, Kansas.

sturmer, w. and bergstrom, j. 1978. The arthropod Cheloniellon from the Devonian Hunsriick Shale.
Paldont. Z. 52, 57-81.

Whittington, h. b. 1979. Early arthropods, their appendages and relationships. In house, m. r. (ed.). The
origin of major invertebrate groups. Spec. Vol. Syst. Ass. 12, 253-268.

ivo chlupaC
Geological Survey
Malostranske namesti 19

Typescript received I April 1987 11821 Praha I

Revised typescript received 29 June 1987 Czechoslovakia

UPPER ORDOVICIAN TRILOBITES FROM THE ZAP
VALLEY, SOUTH-EAST TURKEY

by w. t. dean and zhou zhiyi

Abstract. In the extreme south-east of Turkey the §ort Tepe Formation rests disconformably on Arenig
strata of the highest Seydi§ehir Formation. Trilobites are described from the stratotype of the §ort Tepe
Formation and a supplementary section, both on the north-east side of the Zap Valley, 40 km south-south-
west of Hakkari. The twenty taxa described include three new species: Sinocybelel fluminis , Calymenesun
longinasuta , and Paraphillipsinella pilula. The assemblages, of early Ashgill, probably Pusgillian or possibly
Cautleyan, age, consist mainly of genera widespread in Europe, Scandinavia, and Asia, but Sinocybelel,
Calymenesun, and Paraphillipsinella have not previously been recorded outside China.

The south-eastern corner of Anatolia, known in antiquity as the Hakkari, is a mountainous,
traditionally isolated region, bounded to the east by Iran and to the south by Iraq (Text-fig. I ). It
is traversed by the River Zap (known sometimes as the Great Zab) which rises near the Iranian
border, flows south-west and then south in the vicinity of the towns of Hakkari and £ukurca, and
crosses into Iraq where it forms a tributary of the River Tigris.

Between Hakkari and £ukurca the Zap cuts a deep valley to expose two inliers of Cambrian
and Ordovician sediments, mostly elastics, that form part of the Arab Platform. Until recently the
only available account of the Lower Palaeozoic strata was that of Altinli (1963, pp. 60-61), who
divided the pre-Devonian rocks into two parts, of Cambrian(?) and Silurian(?) age; the latter, over
1000 m thick and termed Giri Formation, were said to contain Cruziana , possibly of Ordovician
age. The latter record led Dean (1980, p. 7) to suggest that, by comparison with the Taurus
Mountains, the term Seydi§ehir Formation should be used in place of Giri Formation. Following
initial reconnaissance work, the region was re-mapped by the Turkish Petroleum Corporation
(TPAO), whose maps formed the basis of a reassessment of the Cambrian and Ordovician
stratigraphy by Dean et al. (1981). The Lower Palaeozoic rocks were noted briefly by Janvier et
al. (1984, p. 148, fig. 1), whose account included TPAO’s map showing the two inliers to represent
the cores of east-west folds, the larger, northerly one termed the Zap anticline, and the other the
(jmkurca anticline. In Dean et al. (1981) it was demonstrated that shales and sandstones of Upper
Cambrian and lower Ordovician age represented the Seydi§ehir Formation, described from the
western Taurus Mountains but widespread in the eastern Taurus, south-eastern Turkey, and
neighbouring parts of Iraq. Disconformably overlying strata, mainly shale, mudstone, and quartzite
with very minor limestone, were named the §ort Tepe Formation and shown to be of Ashgill age.

STRATIGRAPHY AND FOSSIL LOCALITIES
a. Section at §ort Tepe.

The type section of the §ort Tepe Formation is located at the eponymous hill, 7-5 km north-west
of (jmkurca, where the disconformable junction with the underlying Seydi§ehir Formation is
exposed high on the north-east side of the Zap Valley (text-fig. 1 ). The line of contact is sharp and
planar, with no irregularity or conglomerate at the base of bed a (text-fig. 2), 30 cm of grey oolitic
limestone in which no macrofossils were found. Bed b, grey shale 1 m thick, proved sparsely
fossiliferous, yielding only a few small brachiopods ( Aegiromenal sp.) and fragments of diplograptid
graptolites, but no trilobites. No macrofossils of any kind were found in the L5 m of bioturbated,

| Palaeontology, Vol. 31, Part 3, 1988, pp. 621-649, pis. 58-62.)

© The Palaeontological Association

622

PALAEONTOLOGY, VOLUME 31

text-fig. 1. Left: sketch map of south-eastern Turkey, showing principal place-names. Right: geological map
(after TPAO) of the Zap Valley between Hakkari and (jrikurca. 1 = Zabuk and Sadan formations,
undifferentiated (Cambrian); 2 = Koruk, Seydijehir, and §ort Tepe formations, undifferentiated (Cambrian,
Ordovician); 3 = Upper Palaeozoic rocks (undifferentiated); 4 = Mesozoic and Tertiary rocks (undifferenti-
ated); Thr = thrust; Fit = fault.

grey siltstone that make up bed c. The most varied assemblages at this section came from bed d,
grey shale 5 m thick in which trilobites were found at two levels (Iocs. Z.33-3, Z.33-4), though
much less abundantly than at §ort Dere.

The 2 m of grey-green siltstone of bed e mark a transition from the shale of bed d to a succession
of resistant quartzites, mostly thickly bedded but some finely laminated, that are grey-green when
fresh but weather to form a distinctive, whitish feature in the hillside. In the higher part of the
measured section a small fault, with downthrow to the north-east, cuts the quartzites, whose
outcrop continues north-westwards to the adjacent road, where a succession 25 m thick was seen.
Silurian rocks are unknown from the area and Devonian strata overlie both Seydi§ehir and §ort
Tepe formations with low angular unconformity in the vicinity of Kopriilii, north-west of §ort
Tepe (text-fig. 1; see also Janvier et al., 1984, pp. 149-151).

Faunal lists : Z.33-1. Aegiromenal sp., fragments of diplograptid graptolites; Z.33-3. Lonchodomas
sp., Dindymenel sp., Prionocheilus cf. obtusus ; Z.33-4. Dindymenel sp., Calymenesun altinasuta,
Birmanites latus.

b. Section at §ort Dere

About 1-5 km south-east of §ort Tepe the small valley of §ort Dere intersects the east bank of the
River Zap. Approximately 200 m north-east of the intersection, the base of the §ort Tepe Formation
rests disconformably on silty shale and quartzite of the Seydi§ehir Formation, 40 m of which were
seen between this point and the Zap. The succession (text-fig. 2), which differs in detail from that
at §ort Tepe, is more accessible and better exposed but much less complete and so was used to
supplement the stratotype. Again, the interformational boundary is planar, but in this case the
basal unit, bed a', is a high energy grainstone, 75 cm thick, ferruginous and oolitic, containing
reworked quartzitic fragments derived from the Seydi§ehir Formation (Monod in Dean et al. 1981,
p. 277). No macrofossils were found in the two succeeding units, beds b/ and c', comprising,

DEAN AND ZHOU: ORDOVICIAN TRILOBITES FROM TURKEY

623

text-fig. 2. Measured sections in the §ort
Tepe Formation showing stratigraphic posi-
tions of fossil localities.

SORT TEPE SORT DERE

respectively, 11 m of grey shale and 40 cm of sandy limestone, but bed d' yielded numerous
specimens at two levels, Z.34 and Z.36. The rocks are mainly brown-weathering, green-grey shale
and silty mudstone, often exhibiting bioturbation; 8 m were seen and the highest beds are faulted
against Seydi§ehir Formation, though the contact is masked by rock falls. Most of the fossils were
found at loc. Z.34, the great majority of them trilobites, with a few poorly preserved brachiopods
and sporadic machaeridian plates and echinoderm debris. All but one (PI. 62, fig. 8) of the trilobites
were disarticulated, and most came from a layer 5 cm thick, though Z.34 is taken here to include
also material from the overlying 40 cm of sediment. The assemblage proved to be diverse, and the
asaphinid B. latus , the largest form present, was easily the most abundant.

Faunal lists: Z.34. Ampyxl sp., Lonchodomas sp., Raphiophorusl sp., Hibbertia sp., Sinocybelel
fluminis , Ovalocephalus tetrasulcatus , Duftonia sp., C. longinasuta , P. cf. obtusus , Paraphillipsinella
pilula , B. latus, Harpidella sp., Phorocephala sp., Amphitryon ? sp., Lichas atf. laciniatus, Dicranopeltis
sp., Diacanthaspis sp., Miraspis sp.; Z.36. Lonchodomas sp., Dindymenel sp., Paraphillipsinella
pilula, Amphitryon l sp., Stenopareia sp.

AGE AND RELATIONSHIPS OF THE TRILOBITES

During recent years it has become increasingly apparent that many Ordovician trilobite genera
once thought to have a limited vertical distribution within the system have, in fact, very long
ranges, and that their lateral distribution may be restricted by changes in facies. These comments
apply to at least eight trilobites of the §ort Tepe Formation, assessment of whose age depends
heavily on only a few genera and species. For present purposes the trilobites are divided into three
groups: a, genera and species restricted to the Ashgill Series, though sometimes widely distributed;

624

PALAEONTOLOGY, VOLUME 31

b , genera with a longer stratigraphic range but previously reported only from China; and c, genera
with both long stratigraphic range and wide geographic distribution.

Group a. Duftonia was described first from the Pusgillian Stage, lowest Ashgill Series, of northern
England but has been recorded from the higher Ashgill (pre-Hirnantian) in Wales and Bohemia
(Kraluv Dvur Formation), though not from the Caradoc or the Silurian.

Birmanites as now interpreted (see Zhou el al. 1984, p. 17 for synonymy) has a long vertical
range, from Tremadoc to Ashgill. B. latus is known only from the Ashgill of Vastergotland,
Sweden, where it occurs in the Red Tretaspis Mudstones, strata equated by Kielan (1960, p. 78)
with the combined zones of Eodindymene pulchra and of Staurocephalus clavifrons in the lower
and middle portions of the Polish Ashgill. In Scania, southern Sweden, a zone of Opsimasaphus
(now B.) latus and Dicellograptus complanatus was employed by Glimberg (1961, p. 83). Jaanusson
(1963, pp. 163, 164) equated the O. latus Zone with the E. pulchra Zone, and the D. complanatus
Zone with the lower half of his Jerrestad Stage (subsequently termed Jerrestadian, Jaanusson 1982,
p. 8), underlain by the Pleurograptus linearis Zone.

Ovalocephalus tetrasulcatus was described (as Hammatocnemis ) from the S. clavifrons Zone of
the Ashgill in Poland (Kielan 1960, p. 141) and has not been reported elsewhere in Europe. The
record from the §ort Tepe Formation increases considerably the geographic distribution of the
species, but the genus already had an extended history in China, where it occurs from the Arenig
to the Ashgill (Lu and Zhou 1979).

Although the Turkish specimens of Lichas are specifically undeterminable, species assigned to
the genus by Tripp (1958, p. 575) occur only in the Ashgill and the lower and middle Silurian.
Dicranopeltis is not recorded below the Ashgill, in which it is poorly represented, and also includes
several species from the Middle and Upper Silurian (Tripp 1958, p. 575).

Group b. Three genera are of particular interest as they are unrecorded from Europe but are
well represented in China where, however, their vertical range extends far below the Ashgill.
The type species of Calymenesun came from the Shihtzupu Formation (Llandeilo Series) of
Guizhou Province, but the genus is recorded also from low in the Ashgill (Zhou et al. 1984,
p. 29).

Paraphillipsinella (see review in Zhou and Dean 1986, p. 767) was founded on material from the
Caradoc of Sichuan Province but occurs also in the lower Ashgill of the Yangtze region.

The generic position of trilobites here termed Sinocybelel is uncertain but closest comparison is
with Chinese species, all of which have three pairs of pygidial pleurae in contrast to four pairs in
European and Scandinavian species of the possibly related Atractopyge.

Whether Amphitryon ? should also be included in group b is debatable. The genus is recorded
from the higher Caradoc and the Ashgill Series in Europe, but species in which the preglabellar
field of the cranidium is triangular in plan have been described only from China, where the
character occurs much earlier, in the Llanvirn, and may merit generic recognition.

Group c. Remaining genera in the §ort Tepe Formation contribute no precise evidence of age. Of
the Raphiophoridae, Ampyx and Lonchodomas extend from Arenig to Ashgill. The range of
Raphiophorus is uncertain; the type species came from the Black Tretaspis Shale, approximately
late Caradoc, of Sweden but the genus is well represented in the Ashgill and Silurian (species listed
by Thomas 1978, p. 53). A supposedly Arenig species was excluded by Thomas, and Raphiophorus
sp. from the Meadowtown Beds (upper Llandeilo or lowest Caradoc) of Shropshire (Whittard
1955, p. 23, pi. 2, figs. 13-16) comprises poorly preserved meraspids of uncertain position. Hibbertia
occurs in both Caradoc and Ashgill, and dindymeninids have a long range within the Ordovician.
Harpidella ranges from Ashgill to lower Devonian (Thomas and Owens 1978, p. 72), and species
of Phorocephala that lack a preglabellar field are reported from both Caradoc and Ashgill (Zhou
and Dean 1986, p. 751). The single cranidium of Stenopareia sp. is of little significance and
Prionocheilus cf. obtusus , though closely resembling an Ashgill species from Britain and Sweden,
represents a genus that changed relatively little during the Ordovician and whose oldest representa-
tives occur in the lower Arenig of southern France (Dean 1966, p. 300). Odontopleurids constitute

DEAN AND ZHOU: ORDOVICIAN TRILO BITES FROM TURKEY

625

only a minor element in the §ort Tepe assemblages and Diacanthaspis is known from both Caradoc
and Ashgill strata; Miraspis has a long range, from lower Ordovician to upper Silurian in Europe
(Bruton 1968, p. 42).

To summarize, evidence given in group a favours a lower Ashgill age, corresponding to the
lower half of the Jerrestadian Stage in terms of the Swedish succession and the Dicellograptus
complanatus Zone in terms of the standard British graptolite zones. Correlation of the Ashgill
stages with corresponding graptolite zones is imprecise, but according to Williams et al. ( 1972) the
D. complanatus Zone falls within approximately the upper half of the Pusgillian, though no
distinctively Pusgillian trilobites were found in the §ort Tepe assemblages. Although B. latus and
O. tetrasulcatus occur together in the §ort Tepe Formation, the latter species is found slightly
higher (, Staurocephalus clavifrons Zone) in Poland, so it is possible the Turkish strata may extend
above the D. complanatus Zone, into the Cautleyan Stage.

The relationship of the §ort Tepe Formation to successions elsewhere in south-eastern Turkey
is not yet established, but the rocks may represent a continuation of the transgressive sequence,
represented by the Bedinan Formation, that began in the middle Caradoc and persisted probably
into the Ashgill in the Derik-Mardin region, 320 km west of £ukurca (Dean et al. 1981, p. 278).

SYSTEMATIC DESCRIPTIONS

Terminology is essentially that used in the Treatise on Invertebrate Paleontology (Harrington et al. in Moore
1959, pp. 0117 0126), with the addition of eye socle (Shaw and Ormiston 1964) and baccula (Opik 1967).
Stratigraphic position of fossil localities at §ort Tepe and §ort Dere is shown in text-fig. 2. Figured and cited
specimens are deposited in the Department of Palaeontology, British Museum (Natural History), London,
and their numbers carry the prefix It.

Family raphiophoridae Angelin 1854
Genus ampyx Dalman 1827

Type species. Ampyx nasutus Dalman 1827, Asaphus Limestone (upper Arenig) of Vastana, Ostergotland,
Sweden.

Ampyx sp.

Plate 58, figs. 2?, 8?, 9

Figured specimens. It. 19494 (PI. 58, fig. 2), It. 19497 (PI. 58, fig. 8), It. 19499 (PI. 58, fig. 9).

Locality. §ort Dere, Z.34.

Description and discussion. The pygidium resembles that of A. nasutus , refigured by Whittington (1950,
pi. 74, figs. 3 9). but is slightly shorter and the anterior pleural furrows are almost straight. The cranidium
is slightly crushed but the shape of the glabella is generally similar to that of A. nasutus in having three pairs
of depressed muscle scars on the glabellar flanks, and in the form of the anterior border, defined by a shallow
border furrow that is continuous with the axial furrows. It differs from A. nasutus in having the anterior
branches of the facial suture less convergent forwards. In all these characters the Turkish species is comparable
with A. abnormalis Yi (1957, p. 557, pi. 5, fig. 3 a-e\ Lu 1975, p. 414, pi. 39, figs. 5-11; pi. 40, figs. 1 7), from
the upper Arenig to Llandeilo of western Hubei, China, but the pygidium of the latter has a narrower, more
gently tapered axis and the pleural regions have four pairs of pleural furrows. Records of Ampyx from
Ashgill strata are rare and the Turkish material may represent a new species, but is too poorly preserved for
formal description.

Genus lonchodomas Angelin 1854

Type species. Ampyx rostratus Sars 1835, Ampyx Limestone, 4a/3 (Llandeilo or lowest Caradoc) of Bygdoy,
Oslo, Norway.

626

PALAEONTOLOGY, VOLUME 31

Lonchodomas sp.

Plate 58, figs. 5, 6, 11

Figured specimens. It. 19495 (PI. 58, fig. 5), It 19496 (PI. 58, fig. 6), It. 19498 (PI. 58, fig. 1 1).

Localities. §ort Tepe, Z.33-3; §ort Dere, Z.34 and Z.36.

Description and discussion. The incomplete, slightly compressed cranidium is characterized by the wide (tr.)
triangular outline, the short (sag.) anterior projection of the glabella, and the slight curvature, abaxially
concave, of the axial furrows. These features suggest comparison with L. tecturmasi (Weber 1932, p. 6, pi.
4, fig. 43; 1948, p. 18, pi. 2, figs. 20-22, 26; Chugaeva 1958, p. 32, pi. 2, figs. 3-5) from the Anderken Horizon
(Caradoc) of the Chu-Ili Mountains, Kazakhstan and L. jiantsaokouensis Lu (1975), p. 421, pi. 41, figs. 1 1
and 12) from the Jiantsaokou Formation (low Ashgill) of northern Guizhou, China. But the Turkish specimen
is inadequate for reference to either of these species.

L. portlocki (Barrande 1846, p. 9; 1852, p. 636, pi. 30, figs. 24 -28; Olin 1906, p. 69, pi. 4, figs. 5-8; Kielan
1960, p. 169, pi. 33, fig. 8; pi. 35, fig. 4), from the Ashgill of Bohemia, Sweden, and Poland, also bears some
resemblance to the Turkish form, but is distinguished by its narrower (tr.) cranidium and what appears to
be a slightly depressed preoccipital lobe.

Genus Raphiophorus Angelin 1854

Type species. Raphiophorus setirostris Angelin 1854, Lower Tretaspis Shale (Ashgill) of Dragga bro, Dalarne,
Sweden.

Raphiophorus ? sp.

Plate 58, figs. I, 3, 4, 7

Figured specimens. It. 19490 (PI. 58, fig. 1), It. 19491 (PI. 58, fig. 3), It. 19492 (PI. 58, fig. 4), It. 19493
(PI. 58, fig. 7)

Locality. §ort Dere, Z.34.

Description and discussion. The glabella is similar to that of R. setirostris Angelin (see Whittington 1950,
p. 553, pi. 74, figs. 1 and 2), a species in which, according to Whittington, triangular bacculae are probably
present, as they are in the Turkish material. However, the wide (tr.), long (exsag.) fixigenae, the forward
curvature of the distal portions of posterior border and furrow, and the short anterior projection of the
glabella in front of the fixigenae distinguish the Turkish specimens from R. setirostris. These features suggest,
rather, a comparison with the type species of Taklamakania , T. tarimensis W. Zhang (1979, p. 1003, pi. 1,
fig. 9), from the Engou Formation (Caradoc) of Keping, Xinjiang, China. A specifically identical specimen
from the same locality and horizon was later described as a new genus and species Xinjiangia yinganensis T.
Zhang (1981, p. 199, pi. 74, fig. 1 1 ). However, Taklamakania has a larger, longer pygidium than Raphiophorus
and its thorax comprises only three segments. The pygidium of the Turkish species generally resembles that

EXPLANATION OF PLATE 58

Figs. I, 3, 4, 7. Raphiophorus ? sp. Loc. Z.34. 1, cranidium. It. 19490, x 5-5. 3, cranidium. It. 19491, x 6.

4, cranidium. It. 19492, x 5-5. 7, pygidium, It. 19493, x 8.

Figs. 2 and 8. Ampyxl sp. 2, cranidium. It. 19494, x 5, loc. Z.36. 8, pygidium. It. 19497, x 5, loc. Z.34.
Figs. 5, 6, IF Lonchodomas sp. 5, cranidium. It. 19495, x 6, loc. Z.34. 6, pygidium. It. 19496, x4, loc. Z.33-
F.3. 1 1, hypostoma. It. 19498, x 6, loc. Z.36.

Fig. 9. Ampyx sp. Loc. Z.34. Cranidium, It 19499, x 5.

Figs. 10, 12, 13, 17. Hibbertia sp. Loc. Z.34. 10, dorsal surface of left genal prolongation. It. 19500, x 3-5.
12, ventral surface of left genal prolongation. It. 19501, x4. 13, ventral surface of left genal prolongation,
It. 19502, x 4. 17, part of right genal region of cranidium, It. 19503, x 3-5.

Figs. 14 16, 18. Dindymene ? sp. 14, cranidium. It. 19504, x 5-5, loc. Z.33-4. 15, cranidium. It. 19505, x 6,
loc. Z.36. 16, cranidium. It. 19506, x 5, loc. Z.33-3. 18, cranidium. It. 19507, x 6, loc. Z.33-3.

PLATE 58

DEAN and ZHOU, Turkish Ordovician trilobites

628

PALAEONTOLOGY, VOLUME 31

of Raphiophorus but has three distinct axial rings, traces of a fourth ring, and three pairs of wide (exsag.)
pleural furrows. In R. setirostris there are two axial rings and one well-defined pair of pleural furrows.

Family harpetidae Hawle and Corda 1847
Genus hibbertia Jones and Woodward 1898

Type species. Harpes flanaganni Portlock 1843, Bardahessiagh Beds (early Caradoc) of Pomeroy, County
Tyrone, Northern Ireland.

Hibbertia sp.

Plate 58, figs. 10. 12, 13, 17; Plate 59, fig. 1

Figured specimens. It. 19500 (PI. 58, fig. 10), It. 19501 (PI. 58, fig. 12), It. 19502 (PI. 58, fig. 13), It. 19503
(PI. 58, fig. 17), It. 19508 (PI. 59, fig. 1).

Locality. §ort Dere, Z.34.

Description and discussion. The specimens exhibit a uniformly very wide brim with large genal prolongations
that narrow gently posteriorly, features suggestive of Hibbertia. The cheek roll is narrow, widens slightly
medially, and is well defined by the distinct girder, which dies out before reaching the posterior margin. The
remains of the glabella suggest that it was slightly pointed frontally. The eye ridge is narrow and the ala
depressed, defined by a distinct alar furrow as in H. sanctacrucensis Kielan (1960, p. 157, pi. 34, figs. 4 and
6; pi. 35, fig. 8), from the Ashgill Series, S. clavifrons Zone, of Brzezinski, Poland. The fringe is finely and
densely pitted with small pits of almost uniform size, and the cheek is covered with radiating, anastomosing
ridges.

Family encrinuridae Angelin 1854
Subfamily cybelinae Holliday 1942
Genus sinocybele Sheng 1974

Type species. Sinocybele baoshanensis Sheng 1974, Lower Pupiao Formation (Llandeilo to Caradoc Series),
south of Shihtien, western Yunnan, China.

Sinocybylel fiuminis sp. nov.

Plate 59, figs. 2-6, 77, 8, 9

Diagnosis. Sinocybele ? species with four pairs of large tubercles sub-equispaced along median area
of glabella. Palpebral lobes sited opposite posterior half of 2p glabellar lobes and about midway
between glabella and lateral border furrow. Posterior branches of facial suture are very slightly
curved, convex forwards, and meet the margin just in front of genal angles. Strongly developed
eye ridges end opposite 3p furrows. Pygidium with three pairs pleurae that curve strongly backwards

EXPLANATION OF PLATE 59

Fig. I. Hibbertia sp. Loc. Z.34. Ventral surface of cephalic fringe. It. 19508, x 3.

Figs. 2-6, 77, 8, 9. Sinocybele0! fiuminis sp. nov. Loc. Z.34. 2, cranidium. It. 19509, x 5. 3, pygidium. It.
19510, x 5. 4, pygidium. It. 19511, x 6. 5, pygidium. It. 19512, x 5. 6, pygidium. It. 19513, x 5. 7,
hypostoma referred questionably to species. It. 19514, x 6. 8, left librigena with associated pygidium of
Paraphillipsinella, It. 19515, x 5. 9, cranidium. It. 19516, x 5. 5 is holotype; remainder (excluding fig. 7)
are paratypes.

Figs. 10, 12 16. Ovalocephalus tetrasulcatus (Kielan 1960). Loc. Z.34. 10, cranidium, It. 19477, x 5. 12,
pygidium. It. 19517, x 5. 13, cranidium. It. 19478, x 6. 14, cranidium, It. 19518, x 6. 1 5, hypostoma,
It. 19519, x 5. 16, cranidium. It. 19520, x 5.

Figs. II, 17, 18. Duftonia sp. Loc. Z.34. 11 and 18, internal mould and latex cast of cranidium. It. 19521,
x3-5. 17, cranidium, It. 19522, x 3.

PLATE 59

i mm

' . V.. ■ •. ■

Mil

I:

16 17 18

DEAN and ZHOU, Turkish Ordovician trilobites

630 PALAEONTOLOGY, VOLUME 31

to end in long free points. Axis subtriangular with three complete and about nine partially defined
axial rings.

Holotype. It. 19512 (pi. 59, fig. 5).

Paratypes. It. 19509 (PI. 59, fig. 2), It. 19510 (PI. 59, fig. 3), It. 19511 (PI. 59, fig. 4), It. 19513 (PI. 59, fig. 6),
It. 19515 (PI. 59, fig. 8), It. 19516 (PI. 59, fig. 9).

Locality. §ort Dere, Z.34.

Description. The more complete of the two paratype cranidia has an estimated overall breadth (excluding
fixigenal spines) of 1 1 -6 mm and an estimated length of 3-7 mm. Even allowing that the specimen is dorsally
compressed, the long (tr.), acutely triangular outline of the posterior areas of the fixigenae is noteworthy.
The combined glabella and preglabellar field is about as broad as long, its outline broadening slightly so that
the basal breadth is 0-8 of the frontal. The boundary between glabella and preglabellar field is indicated by
a pair of short (tr.), shallow grooves (PI. 59, fig. 2) that run adaxially forwards from the axial furrows. Three
inequisized pairs of lateral glabellar lobes, lp the smallest, are separated by short (tr.), deep lateral furrows,
the lp pair of which turn backwards adaxially so that the lp lobes are of ‘cat’s ear’ outline, linked to the
median area by narrow necks. Most of the glabella, if not all, is covered with fine granules. In addition there
are four pairs of large tubercles, subequally spaced longitudinally and sited opposite the rear half of the 2p
lobes, the centre of the 3p lobes, and then at almost equal intervals between the 3p lobes and the anterior
border. The last named is low, well defined by a broad (sag.) anterior border furrow, and carries about six
(estimated) large tubercles dorsally; it is incompletely preserved medially and the presence of a median
projection has not yet been demonstrated.

The more complete cranidium shows the anterior branches of the facial suture to be straight, converging
forwards slightly from eyes that are sited opposite the lp glabellar furrows and approximately midway
between the axial furrows and the lateral border furrow. The eye lobes are joined to the axial furrows opposite
the 3p glabellar furrows by strongly developed, smooth eye ridges. One of the latter is better seen in the
second paratype cranidium (PI. 59, fig. 9), in which, owing to crushing, it appears to be directed more strongly
backwards.

A single left librigena (PI. 59, fig. 8) shows the eye to be small, possibly pedunculate, and the lateral border
well defined; the pitted surface carries a few large tubercles like those on the lateral border.

The pygidium bears a marked general resemblance to that of Atractopyge Hawle and Corda 1847 (for
examples, see Dean 1971 and Ingham 1974), the most obvious difference being that in S.l fluminis the pleural
regions are composed of three pairs of pleurae instead of four. The outline of the axis is subtriangular,
slightly constricted at the third ring furrow, and apparently ends in a sharp point that represents, in fact, the
post-axial ridge, set slightly below the parabolic, diminutive, true terminal piece (PI. 59, figs. 5 and 6). The
first three ring furrows are complete but subsequent furrows (nine, possibly ten are visible) are incomplete
both immediately adjacent to the axial furrows and medially, where a smooth median band occupies the
axial third of the axis. Three pairs of pleurae that end in long free points curve strongly backwards and
inwards so that the third pair almost meet behind the axis, where they are separated by the small post-axial
ridge. Each pleura is divided by a distinct pleural furrow into unequal anterior and posterior bands, the
latter about twice as wide (tr.) as the former. Anterior bands are clearly visible on the first two pairs of
pleurae but scarcely or not at all on the third pair, so that the second and third pairs of posterior bands
appear to be separated by a single furrow, in which a row of granules may be visible. Surface of the pygidium
is granulose.

The hypostoma of S. baoshanensis remains undescribed and the present small specimen (PI. 59, fig. 7) is
therefore assigned only questionably to the new species. It is of encrinurid type and its outline (excluding
anterior wings) is suboval, with maximum breadth 2-0 mm. and median length 2-5 mm. The anterior two-
thirds of the middle body are divided into three longitudinal lobes by a pair of straight, parallel furrows.
The central lobe so formed occupies just over half the breadth of the middle body and projects forwards of
it; in life position it would have underlain the centre of the anterior border. Temple (1954, p. 318) suggested
that in Encrinurus the generally comparable central lobe may have accommodated the ventral surface of the
pygidium during enrolment and the present specimen may have functioned similarly. The middle body is
circumscribed by a narrow, rim-like border that broadens to form the incompletely preserved anterior wings.

Discussion. S. baoshanensis Sheng (1974, p. 110, pi. 7, fig. 6a, b) was founded on a single cranidium,
illustrated as both internal and external moulds, of cybelinid type in which the anterior border

DEAN AND ZHOU: ORDOVICIAN TRILOBITES FROM TURKEY

631

was produced forwards to form a prominent, flat-topped, horn-like protuberance whose length,
though incomplete, was at least two-thirds that of the glabella. Three, possibly four, pairs of large
tubercles are visible on the central area of the glabella and the remainder of the surface is covered
with small tubercles.

Zhou et al. (1984, p. 27), in discussing the relationships of Cybelurus ? from the Shihtzupu
Formation in Guizhou Province, noted that Sinocybele has branching 3p glabellar furrows like
Cybelurus and that if the pygidium of Sinocybele proves to have three pairs of pleurae, their
Cybelurus ? sp. would be assignable to Sinocybele even though the anterior cranidial projection is
only small. A. sinensis Lu (1975, pp. 233, 444, pi. 45, figs. 15 and 16), from the Shihtzupu Formation
of Zunyi and based on a single cranidium, was assigned to Cybelurusl by Zhou et al. (1984,
p. 27). The holotype has long (tr.), acutely triangular, postocular fixigenae generally similar to
those of S.fluminis, but the eyes of sinensis (here also placed questionably in Sinocybele ) are set
very far back, opposite the lp glabellar lobes; both species have the fixigenal spines directed only
slightly backwards. SP. sinensis has five pairs of large tubercles on the median area of the glabella,
and there are additional, paired tubercles on the abaxial parts of the combined frontal glabellar
lobe and preglabellar field.

Subfamily dindymeninae Henningsmoen in Moore 1959
Genus dindymene Hawle and Corda 1847

Type species. Dindymene fridericiaugusti Hawle and Corda 1847, from the Kraluv Dvur Formation (Ashgill
Series), Kraluv Dvur, Czechoslovakia.

Dindymene! sp.

Plate 58, figs. 14 16, 18

Figured specimens. It. 19504 (PI. 58, fig. 14), It. 19505 (PI. 58, fig. 15), It. 19506 (PI. 58, fig. 16), It. 19507
(PI. 58, fig. 18).

Localities. §ort Tepe, Z.33-3 and Z.33-4; §ort Dere, Z.36.

Description and discussion. The Turkish specimens, inadequate for specific determination, comprise mostly
fragments of cranidia, though one piece of shale (It. 16063, not illustrated) shows vestiges of a partly
disarticulated thorax and pygidium of dindymeninid type, the pleural tips extended to form long, slim spines.
The posterior border and border furrow are narrow, transversely straight, and the genal angles are produced
to form fixigenal spines, seen in It. 19507. The same specimen shows a broken spine, c. 2 mm long, apparently
extending from the glabella, and It. 19505 (PI. 58, fig. 15) may retain the spine base. An analogous though
shorter spine is seen on some cranidia of D. hughesiae Reynolds 1894 (Ingham 1974, pi. 18, figs. 1, 4, 8 10),
from the Ashgill (Rawtheyan) of northern England and of D. cordai Nicholson and Etheridge 1878 (Ingham
1974, pi. 18, fig. 18), from the Rawtheyan of Scotland; a spine base is visible on the holotype of D.
fridericiaugusti (original of Hawle and Corda 1847, pi. 1, fig. 3; see also Horny and Bastl 1970, pi. 15, fig.
1). The remainder of the glabellar surface is smooth except for a few, widely spaced tubercles, some possibly
paired. By contrast the fixigenal surface is corsely pitted, with more numerous large tubercles than on the
glabella. Broadly similar ornamentation is seen on cranidia of D. ornata Linnarsson 1869 from Sweden and
Poland illustrated by Kielan (1960, pi. 26, fig. 6; pi. 27, fig. 4), though the glabella is more granulose and
has more tubercles than the Turkish species.

Family hammatocnemidae Kielan 1960
Genus ovalocephalus Koroleva 1959

Type species. Ovalocephalus kelleri Koroleva 1959, from the late Caradoc of northern Kazakhstan. The close
resemblance of Ovalocephalus and Hammatocnemis was noted by Zhou and Dean ( 1986, p. 776) and the two
genera are considered here as synonyms.

632

PALAEONTOLOGY, VOLUME 31

Ovalocephalus tetrasulcatus (Kielan 1960)

Plate 59, figs. 10, 12-16

Hammatocnemis tetrasulcatus Kielan 1960, p. 141, pi. 25, fig. 3; pi. 26, figs. 2-4; pi. 27, figs. 6 8; text-figs. 38
and 39.

Figured specimens. It. 19477 (PI. 59, fig. 10), It. 19478 (PI. 59, fig. 13), It. 19517 (PI. 59, fig. 12), It. 19518

(PI. 59, fig. 14), It. 19519 (PI. 59, fig. 15), It. 19520 (PI. 59, fig. 16),

Locality. §ort Dere, Z.34.

Description and discussion. All the Turkish specimens are small, but the cranidia match closely those illustrated
from Poland. According to Kielan’s (1960, p. 142) original account the frontal breadth of the glabella is
equal to three times its breadth in front of the preoccipital segment, but her illustrations of undistorted
specimens (Kielan 1960, pi. 26, fig. 4; pi. 27, fig. 6; text-fig. 38) show that it is only 2-3 times as broad. In
the largest, slightly compressed, Turkish cranidia the corresponding figures are 2-2 and 2-4, and the specimens
closely resemble the holotype and one paratype (Kielan 1960, pi. 26, figs. 2 and 4).

The hypostoma of Ovalocephalus is not well known and that of O. tetrasulcatus has not been described,
but the present Turkish example (PI. 59, fig. 15) is attributed to the species on account of its resemblance to

the hypostoma of O. decorus (Lu in Lu and Chang 1974) figured by Lu and Zhou (1979, pi. 3, fig. 6). The

specimen is almost as long (2-8 mm) as broad (3-0 mm), of low convexity, pentagonal in outline with the
transverse anterior margin slightly convex. The subparallel lateral margins occupy 0-57 of the overall length,
and the straight posterolateral margins converge to meet at an angle of 100°. A low, narrow rim runs around
the lateral and posterior margins and widens (sag.) slightly to form a small point at the posterior extremity
of the hypostoma.

The only available Turkish pygidium (PI. 59, fig. 12) is very small, with an estimated breadth and median
length of 4 0 mm and 1-4 mm respectively, and closely resembles the Polish examples. The axis has three
distinct axial rings and a fourth is less well defined. In the Polish type material the pleural regions comprise
four pairs of pleurae, the first three distinct, and four pairs of free points were said to be present, though
these are not seen in all the illustrations. In the Turkish example there are three distinct pleurae plus faint
traces of a fourth. First and second pleurae are bounded by broad (exsag.), deep, interpleural furrows and
end in short free points; third pleurae show no free points and only the adaxial half of the third interpleural
furrow is clearly defined. All three pleurae have a node developed immediately outside the axial furrow; a
similar structure was described by Kielan (1960, p. 143) and evidently corresponds to nodes on the thorax
(see also Lu and Zhou’s illustrations 1979, pi. 4, figs. 3 and 4 of the thorax of O. decorus (Lu in Lu and
Chang 1974)).

Present evidence suggests that O. tetrasulcatus has been found as yet only in Poland and south-eastern
Turkey. O. tetrasulcatus as recorded by Lu and Zhou (1979, pi. 2, figs. 10 and 1 1) from the Qilang Formation
(Caradoc) of Keping, Xingjiang Province, China, has since been described as O. kanlingensis (T. Zhang 1981,
p. 209, pi. 77, figs. 5-7).

Family dalmanitidae Vogdes 1890
Genus duftonia Dean 1959

Type species. Duftonia lacunosa Dean 1959, Dufton Shales (Ashgill; Pusgillian Stage) of northern England.

Duftonia sp.

Plate 59, figs. 11, 17, 18

Figured specimens. It. 19521 (PI. 59, figs. 1 I and 18), It. 19522 (PI. 59, fig. 17).

Locality. §ort Dere, Z.34.

Description and discussion. The Turkish cranidia differ from D. lacunosa Dean (1959, p. 144, pi. 19, figs. 2,
5, 6, 8) in having: frontal glabellar lobe, though slightly compressed, proportionately longer, greater than
half the glabellar length, compared with about half; rear ends of the eye set proportionately further from the
axial furrows and opposite the mid-points of the 2p glabellar lobes, compared with opposite the lp glabellar
furrows. In the two species both the palpebral lobes and the well-defined, strongly sigmoidal palpebral

DEAN AND ZHOU: ORDOVICIAN TRILOBITES FROM TURKEY

633

furrows extend to the axial furrows, in D. lacunosa at points opposite the 3p glabellar furrows and anterolateral
tips of the 3p lobes, and in D. sp. well in front of the 3p furrows. The Turkish material shows traces of a
very low, thin rim that represents the anterior border, a structure scarcely discernible in D. lacunosa.

Dalmanites morrisiana Barrande (1852, p. 559, pi. 27, figs. 8 and 9) was assigned to Duftonia by Dean
(1967a, p. 38) and the original of Barrande’s fig. 8 was selected as neotype by Marek (in Horny and Bastl
1970, p. 210). Dalmanites morrisiana , from the Kraluv Dvur Formation (Ashgill), has slightly smaller,
narrower palpebral lobes than Duftonia sp. and these structures extend from just behind the 3p furrows to
end at points relatively further from the posterior border furrow than in either D. sp. or D. lacunosa , and a
greater distance from the glabella. Evidence for a median occipital tubercle in D. lacunosa is equivocal, but
one is visible both in D. sp. and in Barrande’s illustrations of Dalmanites morrisiana.

Family calymenidae Burmeister 1843
Subfamily reedocalymeninae Hupe 1955

Lu (1975, pp. 445 458) included Reedocalymene , Ca/ymenesun , and Neseuretus (a senior subjective synonym
of Synhomalonotus) in the Reedocalymeninae, with the tacit implication that Synhomalonotinae Kobayashi
1960 is a junior synonym. We follow this classification provisionally here as Calymenesun has several
characters in common with Neseuretus , and we add Vietnamia Kobayashi 1960 and Neseuretus ( Neseuretinus )
Dean 1967b. The position of Reedocalymene Kobayashi 1951, with anterior projection of the frontal area
still longer than that of Calymenesun. is less clear and the genus is in need of revision.

Genus calymenesun Kobayashi 1951

Type species. Calymene tingi Sun 1931, Shihtzupu Formation (Llandcilo) of Feilaishi, Zunyi, Guizhou, China.

Calymenesun longinasuta sp. nov.

Plate 60, figs. 1-3, 5, 6, 8 10, 12, 13

Diagnosis. Calymenesun species with glabellar outline straight-sided laterally and frontally. Anterior
border steeply inclined forwards, well defined by anterior border furrow that is deep abaxially but
broad (sag.) and shallow medially. Median third of anterior border of cranidium is produced to
form stout spine. Lateral border wide, well defined.

Holotype. It. 19527 (PI. 60, figs. 8 and 9).

Par a types. It. 19523 (PI. 60, figs. 1 and 2), It. 19524 (PI. 60, fig. 3), It. 19525 (PI. 60, fig. 5), It. 19526 (PI. 60,
figs. 6, 10, 12, 13).

Locality. §ort Dere, Z.34.

Description. The length ol the glabella is almost equal to, or slightly less than, its basal breadth; there are
three inequisized pairs of lateral lobes, and glabellar outline tapers evenly to a transversely straight anterior
margin. Anterior border is produced in same plane to form a frontal spine at least as long as the preglabellar
field. Pedunculate palpebral lobes stand higher than glabella, and are sited opposite the 2p furrows and 3p
lobes. Weakly developed eye ridges extend to the axial furrows opposite, or slightly in front of, the 3p
furrows. Anterior branches of facial suture are straight and convergent. Axial furrows widen abaxially
opposite the lp lobes to accommodate a pair of small bacculae. Median occipital tubercle present. Surface,
excluding lurrows, is mostly granulose but the median lobe of the glabella carries five equispaced pairs of
tubercles that become progressively larger from front to rear. Glabella generally resembles that of Neseuretus
but the preglabellar field and anterior border are clearly defined, quite apart from the striking development
of the anterior spine. The large, paired glabellar tubercles are particularly distinctive and the rearmost pair
is visible also on the internal mould.

Paratype left librigena is ot typical calymenid form but lateral border is very wide and well defined. The
specimen shows the eye surface, though incomplete, to be short (exsag.), bounded by poorly defined eye
socle.

634

PALAEONTOLOGY, VOLUME 31

The pygidium is of calymenid type, with seven axial rings and five or six pairs of furrowed ribs. Outline
of axis is slightly constricted behind sixth axial ring, and postaxial ridge is apparently parallel-sided and
convex as in Neseuretus. Pleural furrows become progressively less well defined from front to rear and do
not quite attain the lateral margin. All the ribs are divided by faint interpleural furrows into two unequal
bands, the anterior twice as wide (exsag.) as the posterior. About midway between axial furrow and lateral
margin is a faint depression that corresponds to what Campbell (1967) termed a cincture, a coaptative
structure commonly developed in calymenids.

Discussion. C. tingi (Sun 1931, p. 29, pi. 3, fig. 9 a-g only, non 9 h) was redescribed by Zhou et al.
(1984, p. 29, fig. la-g, i,j) and differs from the new species in several respects: the glabella widens
considerably across the lp lobes, and the axial furrows, which contain bacculae, are strongly
curved, abaxially concave, in a manner recalling that in Vietnamia Kobayashi (1960, p. 43); the
anterior branches of the facial suture are curved; the anterior border is less distinctly defined and
forms a process that extends to produce a slim spine as long as the remainder of the cranidium;
the eyes are sited opposite the 2p lobes and furrows; the surface is finely granulose with no
tubercles.

C. granulosa Lu (1975, pp. 238, 450, pi. 47, figs. 1-5), from the top of the Linhsiang Formation
(lowest Ashgill, Nankinolithus Zone) at Chikangpo, Ichang district, west Flupei, China, has a
proportionately shorter, broader (esp. basally) glabella than the new species; the anterior border
and furrow are almost undefined; and the lateral border furrow is absent. In these respects C.
granulosa is more comparable with C. tingi.

In C. yinganensis Zhang (1981, p. 21 1, pi. 78, figs. 3-5), a species previously referred to Neseuretus
(Zhang et al. 1982, p. 72, table 10), from the Qilang Formation (Caradoc) of Kanling, Keping,
Xinjiang Province, China, the preglabellar field and anterior border, though incomplete, appear
less well defined than in the new species; they and the anterior branches of the facial suture are
more comparable with those of C. tingi, though the latter is readily recognized by the distinctly
large basal breadth of the glabella. Small, sparse tubercles with a suggestion of arrangement in
transverse rows ornament the glabella of C. yinganensis, and the pleural regions of the pygidium
show well-developed cinctures like those of C. tingi, but unlike the new species. C. zhejiangensis
Ju in Qiu et al. (1983, p. 250, pi. 87, figs. 1 1 and 12), from the Huangnekhan Formation (Ashgill,
Nankinolithus Zone) of Jiangshan, west Zhejiang Province, China, has distinct bacculae and the
uniformly tapered glabellar outline is more like that of C. altinasuta than that of C. tingi, though
the latter’s less well-defined preglabellar field is more comparable. The pygidium of C. zhejiangensis
is very different from that of the new species in having a broad (exsag.), deep cincture that divides
the pleural regions into small, coarsely ribbed proximal and weakly ribbed distal portions, and
coincides with the junction of terminal piece and post-axial ridge.

EXPLANATION OF PLATE 60

Figs. 1 3, 5, 6, 8 10, 12, 13. Calymenesun longinasuta sp. nov. Loc. Z.34. I and 2, cranidium. It. 19523,
x 3. 3, pygidium. It. 19524, x 4. 5, left librigena. It. 19525, x 3. 6 (internal mould), 10, 12, 13 (latex
cast), cranidium, It. 19526, x 3. 8 and 9, cranidium. It. 19527, x4. 8 and 9, holotype; remainder are
paratypes.

Figs. 4 and 24. Phorocephala sp. Loc. Z.34. 4, pygidium, It. 19528, x 9. 24, cranidium. It. 19529, x 5.

Figs. 7, 17-21, 23. Paraphillipsinella pilula sp. nov. 19 from Loc. Z.36; remainder from Loc. Z.34. 7,
pygidium. It. 19530, x 8. 17, cranidium. It 19531, x 6. 18, cranidium. It. 19532, x 6. 19, pygidium. It.
19533, x 8. 20, cranidium. It. 19534, x 6. 21, cranidium. It. 19535, x 6. 23, cranidium, It. 19536, x 6.
17 is holotype; remainder are paratypes.

Figs. 11, 14 16. Prionocheilus cf. obtusus (M‘Coy 1846). Loc. Z.34. 11, pygidium. It. 19537, x 5. 14,

cranidium. It. 19538, x 5. 15, cranidium. It. 19539, x 6. 16, pygidium and two, possibly three, attached
thoracic segments. It. 19540, x4.

Fig. 22. Harpidella sp. Loc. Z.34. Cranidium, It. 19541, x 6.

PLATE 60

%. J«* •' -‘-Mi j Srt '

'MyM

m* M

k iMj

DEAN and ZHOU, Turkish Ordovician trilobites

636

PALAEONTOLOGY, VOLUME 31

Subfamily pharostomatinae Hupe 1953
Genus prionocheilus Rouault 1847

Type species. Prionocheilus verneuili Rouault 1847, from an unnamed formation of Middle Ordovician age
at Poligne, Brittany, France.

Prionocheilus cf. ohtusus (IVTCoy 1846)

Plate 60, figs. 11, 14 16

Figured specimens. It. 19537 (PI. 60, fig. 11), It. 19538 (PI. 60, fig. 14), It. 19539 (PI. 60, fie. 15), It. 19540
(PI. 60, fig. 16).

Locality. §ort Dere, Z.34.

Description and discussion. M’Coy’s (1846, p. 54, pi. 4, fig. 6) holotype cranidium from the Chair of Kildare
Limestone in eastern Ireland was redescribed by Whittington (1965, p. 55, pi. 16, figs. 1-3, 6) and by Dean
(1971, p. 42, pi. 18, figs. 10, 12, 13), who figured additional topotype material both of the species and of
Calymene leptaenarum Tornquist 1884, placed in synonymy with it. The age of the type material is Ashgill,
probably Rawtheyan Stage. Both Turkish cranidia are slightly distorted and incomplete but generally resemble
the Irish material, and the anterior border and preglabellar field are similar. In both specimens the eye ridges
meet the axial furrows just in front of the 2p furrows as in the holotype of P. obtusus , but whether the
palpebral lobes are sited opposite the outer ends of the same furrows is less clear. Plate 60, fig. 14 shows the
characteristic widening of the axial furrows opposite the lp lobes to accommodate structures that have been
termed, variously, Pharostoma- Flecke (Opik 1937, p. 23) or paraglabellar areas (Harrington et al. in Moore
1959, p. 0123).

In Plate 60, fig. 11 the pygidiunr is slightly shortened by compression but the anterior half of the axis
carries four axial rings, followed by traces of a fifth; the remaining pleural region has five well-defined ribs
with faint interpleural furrows, and part of a sixth rib in addition to the anterior half rib. Pygidia of P.
obtusus from Ireland (Dean 1971, pi. 18, figs. 4, 5, 14; pi. 19, figs. 5, 10, 12) and Sweden (Warburg 1925,
p. 157, pi. 4, figs. 2 4) have five or six axial rings, the last poorly defined, and five pairs of ribs. The apparently
rounded tips of the ribs in the Turkish specimen are due to weathering, and are not an original feature.

The distribution of P. obtusus is not known in detail. Although the type material is probably of Rawtheyan
age, and P. cf. obtusus from the Rhiwlas Limestone of North Wales is likely to be of similar age, it is clear
that broadly comparable forms have an extended stratigraphic range and specimens from the Caradoc of
Norway described by Owen and Bruton (1980, p. 32, pi. 9, figs. 10, 11, 13-15) differ only in details of
ornamentation, length of preglabellar field, and the slightly more posterior position of palpebral lobes.

Xuanenia Zhou in Zhou et al. 1977, type species X. granulosa Zhou in Zhou et al. (1977, p. 263, pi. 79,
figs. 5-7) from the Linhsiang Formation (Ashgill) of Gaoluo, Xuanen, west Hubei, China, apparently differs
little from Prionocheilus. The anterior border is slightly less sharply defined, the front of the glabella is less
broadly rounded, and the lp and 2p lobes appear to coalesce to form composite structures bounded adaxially
by longitudinal furrows. The eyes and eye ridges are situated opposite the 2p lobes and furrows, and the
librigena shows a row of slim, ventrally directed spines. A possible trace of bacculae is visible in Zhou’s
pj. 79, fig. 5 (the holotype) but not in his pi. 79, fig. 6; similar structures are seen also in X. splendida Ju in
Qui et al. (1983, p. 251, pi. 87, figs. 9 and 10), from the Huangnekang Formation (low Ashgill) of Jiande,
west Zhejiang, China.

Family phillipsinellidae Whittington 1950
Genus paraphillipsinella Lu in Lu and Chang 1974

Type species. Paraphillipsinella globosa Lu in Lu and Chang 1974, Pagoda Formation (Caradoc), Chenkou,
Sichuan Province, China.

Junior subjective synonym. Protophillipsinella Chen in Li et al. (1975, p. 155).

Discussion. A translation of the original generic diagnosis, together with minor emendations, was
given by Zhou and Dean (1986, p. 766), who followed Lu and Zhou (1981, p. 14) in considering
Paraphillipsinella to include, at that time, only two species; P. globosa Lu in Lu and Chang (1974,
p. 133, pi. 53, figs. 8 and 9) and P. nanjiangensis Lu in Lu and Chang (1974, p. 133, pi. 53,

DEAN AND ZHOU: ORDOVICIAN TRILOBITES FROM TURKEY

637

fig. 10). Of the two, only P. globosa is regarded as sufficiently close to the new species to merit
discussion, the glabella of P. nanjiangensis being easily recognized by its transversely suboval
anterior lobe, short, squat posterior lobe, and wide fixigenae. To the previous generic criteria we
now add the presence, sometimes weakly developed, of a narrow anterior border and border
furrow immediately adjacent to the axial furrows. A median occipital tubercle is at least sometimes,
and possibly always, present.

Paraphillipsinella pilula sp. nov.

Plate 60, figs. 7, 17-21, 23

Diagnosis. Paraphillipsinella species characterized by: narrow (exsag.) anterior border and border
furrow well developed laterally, die out frontally; interocular portion of each fixigena narrow, little
more than half width of adjacent part of posterior lobe of glabella; well-defined palpebral lobes
sited opposite 2p and posterior half of 3p glabellar lobes.

Holotype. It. 19531 (PI. 60, fig. 17).

Paratypes. It. 19530 (PI. 60, fig. 7), It. 19532 (PI. 60, fig. 18), It. 19533 (PI. 60, fig. 19), It. 19534 (PI. 60,
fig. 20), It. 19535 (PI. 60, fig. 20, It. 19536 (PI. 60, fig. 23).

Localities. §ort Dere, Z.34 and Z.36.

Description and discussion. Rather than give a detailed description of the species, comments are confined
mainly to features relevant to its recognition. Lu’s illustrations of the holotype of P. globosa show the
subspherical anterior lobe and subcylindrical posterior lobe of the glabella occupying, respectively, 0-55 and
0-28 of the overall length of the cranidium. In P. pilula the corresponding figures are 0-52 and 0-39 in the
largest apparently undistorted cranidium but 0-47 and 0-42 in the smallest. In most published illustrations
of Paraphillipsinella the boundary between anterior and posterior lobes often appears as a sharp, transverse
furrow; but this is the result of compression and is not seen in specimens preserved in limestone (Zhou and
Dean 1986, pi. 62, figs. 13-15). The anterior border and border furrow of the new species are particularly
striking, with each end extending adaxially from the shallow axial furrows around the abaxial quarter of the
anterior lobe. The posterior lobe is gently tapered and there are four pairs of glabellar lobes, separated by
pit-like glabellar furrows; lp lobes are slightly larger than 2p to 4p pairs and form part of weakly defined
basal glabellar segment. Straight eye ridges run from anterior ends of palpebral lobes to meet axial furrows
opposite 4p glabellar lobes. The position of the palpebral lobes in P. globosa is difficult to distinguish in
published illustrations (Lu in Lu and Chang 1974, pi. 53, figs. 8 and 9) but is probably opposite the 2p
glabellar lobes, as in the new species. The same illustration of P. globosa showed no anterior extension of
the fixigenae beside, and overhung by, the anterior lobe. A whole exoskeleton identified by Ju (in Qiu et al.
1983) as P. hubeiensis Zhou ( 1974, p. 228, pi. 76, fig. 9) has since been put in synonymy with P. nanjiangensis
Lu in Lu and Chang 1974 by Zhou and Dean (1986, p. 767). The specimen lacks the anterior lobe and shows
both the large rostral plate and a forwards extension of the fixigenae, which are as wide as the posterior
lobe. A median occipital tubercle occurs in at least one paratype of P. pilula and may be a general feature
of Paraphillipsinella , though not clearly visible in all illustrations.

A subconcentric pattern of anastomosing, fine ridges on the anterior lobe of the glabella extends as
subparallel, longitudinal ridges on the posterior lobe. On the holotype there is a suggestion of pits in some
of the intervening grooves, an ornamentation generally resembling that found in Phillipsinella (Bruton 1976).
Evidence for similar ridges in other species of Paraphillipsinella is equivocal or absent, but the holotype of
P. globosa Lu in Lu and Chang (1974, pi. 53, figs. 8 and 9) shows rows of fine granules, apparently
concentrically arranged. The slightly greater width (tr.) of the glabella at the lp lobes in P. globosa also
recalls that in some species of Phillipsinella (see Bruton 1976 for various illustrations).

Two incomplete pygidia show a general resemblance to the type species of Phillipsinella , P. parabola
(Barrande 1846) from the Ashgill of Bohemia, redescribed by Whittington (1950, p. 559, pi. 75, figs. 4 and
7). One, a latex cast (PI. 60, fig. 19), has the axis slightly abraded but there are traces of three axial rings;
the pleural regions show three segments, separated by distinct rib furrows and carrying well-defined pleural
furrows. The other, an internal mould (PI. 60, fig. 7), has four segments and there are traces of ornamentation,
comprising oblique, anastomosing ridges, similar to that figured by Bruton (1976, pi. 106, fig. 1; pi. 108,
figs. 1, 5, 10, 12). Both Turkish pygidia have the posterior margin slightly indented medially and a marginal
rim is weakly developed.

638

PALAEONTOLOGY, VOLUME 31

Family asaphidae Burmeister 1843
Subfamily asaphinae Burmeister 1843
Genus birmanites Sheng 1934

Type species. Ogvgiles birmanicus Reed 1915, from the Hwe Mawng Beds (Lower Ordovician), Hwe Mawng
and Hpakhi. northern Shan States, Burma. The synonymy of the genus was discussed by Zhou et al. (1984,
p. 17).

Birmanites latus (Angelin 1851)

Plate 61, figs. 3-7; Plate 62, figs. 1, 2, 4, 8

1851 Niobe lata Angelin, p. 14, pi. 10.

1960 Opsimasaphus latus (Angelin); Kielan, p. 78, pi. 6, figs. 1 and 2; pi. 7, fig. 3; pi. 8, fig. 4; text-
fig. 20.

1981 Opsimasaphus', Dean, Monod and Peringek, p. 277 .

Figured specimens. It. 19545 (PI. 61, fig. 3), It. 19546 (PI. 61, fig. 4), It. 19547 (PI. 61, fig. 5), It. 19548
(PI. 61, fig. 6), It. 19549 (PI. 61, fig. 7), It. 19552 (PI. 62, fig. 1), It. 19553 (PI. 62, fig. 2), It. 19554 (PI. 62,
fig. 4), It. 19555 (PI. 62, fig. 8).

Localities. §ort Tepe, Z.33-4; §ort Dere, Z.34.

Description and discussion. Birmanites latus is easily the most abundant trilobite in the collections from the
§ort Tepe Formation. The material agrees closely with the lectotype and other specimens from the Red
Tretaspis Mudstones (Ashgill) of Vastergotland, Sweden, described by Kielan (see synonymy above) and
provides little additional information.

Some compressed cranidia (PI. 61, figs. 4 and 7) appear to show strong sigmoidal ridges extending from
the rear ends of the palpebral lobes, subparallel to the posterior branches of the facial suture, and ending
about half-way to the posterior margin. These structures are the result of crushing and are not invariably
developed. As in the lectotype cranidium, a large, low, median tubercle is sited just behind a line joining the
rear ends of the palpebral lobes. The flanks of the tubercle carry about five or six narrow, subconcentric
ridges and the apex has a trace of a small median perforation.

All the pygidia are dorsally compressed with median length slightly more or less than 0-6 of the breadth.
The almost straight-sided axis has a frontal breadth about 0-2 that of the pygidium and occupies about 0-8
of its length, though the terminal piece is not well defined. Largest examples (PI. 61, fig. 3) have at least
eight, low, transversely straight axial rings, separated by shallow ring furrows, in the anterior four-fifths of
the axis, the remainder being indiscernible. This matches closely Kielan’s illustration (1960, pi. 6, fig. 2),
where a further two rings and a tiny terminal piece are visible. The pleural fields show, in addition to the
large anterior half-ribs, five pairs of ribs clearly defined and a sixth less so; this agrees with the original of
Kielan 1960, pi. 6, fig. 1, though a better-preserved Swedish example (Kielan 1960, pi. 6, fig. 2) has six pairs
of ribs and a less well-defined seventh pair.

Kielan did not describe the hypostoma of B. latus but two associated Turkish specimens (PI. 62, figs. 1,
2) of asaphinid type are assigned to the species. Maximum breadth (including anterior wings) is about three-
quarters the overall length, and the posterior margin is deeply indented to form a narrow, median notch
with subparallel sides. Middle body is longitudinally subelliptical with length two-thirds that of hypostoma,
and with curved posterior margin concave rearwards, subparallel to median notch; posterolateral extremities

EXPLANATION OF PLATE 61

Figs. 1 and 2. Diacanthaspis sp. Loc. Z.34. 1, cranidium, It. 19542, x 6. 2, left librigena. It. 19543, x6.
Figs. 3-7. Birmanites latus (Angelin 1851). Loc. Z.34. 3, pygidium. It. 19545, x L5. 4, cephalon. It. 19546,
x2. 5, ventral surface of pygidial doublure. It. 19547, x 5. 6, exfoliated pygidium showing doublure. It.
19548, x2. 7, cranidium showing median tubercle. It. 19549, x2-5.

Figs. 8 10. Amphitryon ? sp. 8 and 9, loc. Z.34; 10, loc. Z.36. 8, cranidium. It. 19550, x4. 9, cranidium, It.
19551, x 5. 10, front of cranidium showing anterior border. It. 19476, x8.

PLATE 61

mm

mmt

MtQil

10

DEAN and ZHOU, Turkish Ordovician trilobites

640

PALAEONTOLOGY, VOLUME 31

formed by pair of large maculae, bounded anterolaterally by deep, triangular furrows. Posterior borders are
large, their lateral margins strongly curved, abaxially convex, linked by low ridges to anterior half of middle
body. Frontal portion of hypostoma formed by flat anterior border that circumscribes middle body and
widens (exsag.) distally to end in pair of short (tr. ), obtusely angular anterior wings. Overall breadth across
anterior wings slightly less than that across posterior borders, and the two structures are separated by broad
(exsag.) lateral notches. Except for a few terrace lines around margin of posterior notch and on front of
middle body, the surface is smooth.

The hypostoma of B. birmanicus has not been illustrated but that of B. hupeiensis Yi 1957, from the
Shih tzupu Formation (Llandeilo) of Guizhou Province, China, was redescribed by Zhou et al. (1984, p. 17,
fig. 3f). It differs from those attributed here to B. latus in having the middle body proportionately shorter
and less elliptical in outline; posterior wings are longer and more pointed; median notch is conspicuously
wider and longer (0-36 versus 0-25 of overall length of hypostoma), its sides converging forwards at 45°
instead of being subparallel; and the anterior border, though not clearly visible, appears to be proportionately
shorter.

Family aulacopleuridae Angelin 1854
Subfamily aulacopleurinae Angelin 1854
Genus harpidella IVTCoy 1 849

Type species. Harpesl megalops M‘Coy 1846, Upper Llandovery of Boocaun, Cong, County Galway, Ireland.

Harpidella sp.

Plate 60, fig. 22; Plate 62, fig. 6
Figured specimens. It. 19541 (PI. 60, fig. 22), It. 19559 (PI. 62, fig. 6).

Locality. §ort Dere, Z.34.

Description and discussion. This form is assigned to Harpidella on account of the large, posteriorly situated
palpebral lobes, considered by Thomas and Owens (1978, p. 71) as an important character in distinguishing
Harpidella from Otarion Zenker 1833. However, it also resembles the type species of Otarion , 0. diffraction
Zenker 1833 (see Thomas and Owens 1978, pi. 7, figs. 1-3, 5, 6), from the Kopanina Formation (Ludlow)
of Dlouha Hora, Czechoslovakia, in the narrow glabella, narrow (sag.) anterior border, faint palpebral
furrows, and vaulted preglabellar field. The Turkish specimens are inadequate for satisfactory comparison
but are generally similar to undetermined species of Otarion figured by Ingham (1970, pi. 5, fig. 12) from the
Ashgill, Cautleyan Stage, of northern England and by Dean (1974, pi. 26, fig. 9) from the Chair of Kildare
Limestone (Ashgill, Rawtheyan) in eastern Ireland, especially in the outline and size of the glabella, and in
the small lp glabellar lobes. Both these British and Irish species have large, backwardly placed palpebral
lobes and are probably better referred to Harpidella.

explanation of plate 62

Figs. 1, 2, 4, 8. Binnanites latus (Angelin 1851). 1, 2, 4, loc. Z.34; 8, loc. Z.33-4. 1, hypostoma, It. 19552,
x3. 2, hypostoma, It. 19553, x 3. 4, right librigena. It. 19554, x2. 8, dorsal exoskeleton. It 19555,
x 3.

Fig. 3. Miraspis sp. Loc. Z.34. Cranidium, It. 19556, x 7-5.

Lig. 5. Stenopareia sp. Loc. Z.36. Cranidium, It. 19558, x 5.

Lig. 6. Harpidella sp. Loc. Z.34. Cranidium, It. 19559, x9.

Lig. 7. Genus and species undetermined. Loc. Z.34. Pygidium, It. 16062, x 3-5.

Ligs. 9? and 13. Dicranopeltis sp. Loc. Z.34. 9, right side of cranidium. It. 19471, x 3 5. 13, pygidium. It.
1 9472, x 4.

Fig. 10. Diacanthaspis sp. Loc. Z.34. Pygidium, It. 19557, x 7.

Figs. 1 17, 12, 14. Lichas aff. laciniatus (Wahlenberg 1821). Loc. Z.34. 1 1, fragment of cranidium. It. 19473,
x4. 12, cranidium, It. 19474, x 3. 14, cranidium. It. 19475, x3.

PLATE 62

Ws.iSHk

■ ■Uv • ■'A

DEAN and ZHOU, Turkish Ordovician trilobites

642

PALAEONTOLOGY, VOLUME 31

Family komaspididae Kobayashi 1935
Genus phorocephala Lu in Lu et al. 1965

Type species. Phorocephala typa Lu in Lu et al. 1965, Siliangssu Formation (upper Arenig), Laingshan, south
Shaanxi Province, China.

Phorocephala sp.

Plate 60, figs. 4 and 24

Figured specimens. It. 19528 (PI. 60, fig. 4), It. 19529 (PI. 60, fig. 24).

Locality. §ort Dere, Z.34.

Description. Cranidium about twice as long as wide. Glabella well defined by deeply incised axial furrows,
its length 0-65 that of cranidium, slightly longer than wide, gently tapered forwards, rounded anteriorly.
Occipital ring one-sixth the cranidial length (sag.), wider (tr.) than base of glabella, with posterior margin
arched backwards; abaxial portions are narrower (exsag.) and curve forwards slightly to axial furrows.
Fixigenae narrow, with width one-sixth that of cranidium as measured across mid-length of palpebral lobes.
Palpebral lobes gently curved in plan and run slightly inwards anteriorly; their length is 0-45 that of cranidium
and they extend almost to posterior border furrow. Anterior branches of facial suture subparallel; preglabellar
area short, equal to one-sixth the cranidial length. Anterior border is dorsally convex, widens adaxially, and
is defined by distinct, though shallow, anterior border furrow. Preglabellar field depressed, as long (sag.) as
anterior border.

The pygidium is about twice as broad as long, its outline approximately lozenge shaped. The large, strongly
tapered axis has a frontal breadth half that of the pygidium. There are two large, curved axial rings, convex
forwards, with traces of a third; the pleural regions have a very thin, marginal rim and show two pairs of
deep pleural furrows and two pairs of shallow rib furrows.

Discussion. The cranidium, though incomplete, is comparable with that of the type species, P. typa
Lu (in Lu et al. 1965, p. 587, pi. 123, fig. 14; see also Lu 1975, pi. 34, fig. 13), in the shape of the
glabella and the size and location of the palpebral lobes. According to Zhou and Dean (1986,
p. 751) the preglabellar field of Phorocephala is absent in adult cranidia, though present in
immature cranidia, of all known species of Caradoc and Ashgill age. The present specimen has a
median length of only 2-3 mm and probably represents a juvenile individual.

The pygidium of the type species was not described by Lu but that of P. shizipuensis Yin (in
Yin and Lee 1978) from the Llandeilo of Guizhou Province, China, figured by Zhou et al. (1984,
fig. 5.x, z ) has an outline resembling that of the present specimen. Both have a large, triangular
axis but in the Turkish species the pleural regions are proportionately smaller, with straighter
margins, and there are only two pairs of pleural and interpleural furrows, compared with four.

Family remopleurididae Hawle and Corda 1847
Genus Amphitryon Flawle and Corda 1847

Type species. Caphyra radians Barrande 1846, p. 32 (a senior subjective synonym of Caphyra murchisonii
Flawle and Corda 1847), from the Kraluv Dvur Formation (Ashgill), Kraluv Dvur, Czechoslovakia.

Amphitryon ? sp.

Plate 61, figs. 8 10

Figured specimens. It. 19550 (PI. 61, fig. 8), It. 19551 (PI. 61, fig. 9), It. 19476 (PI. 61, fig. 10).

Localities. §ort Dere, Z.34 and Z.36.

Description and discussion. The most complete cranidium, though dorsally compressed, is apparently of low
convexity with overall breadth of 9 0 mm and median length (including preglabellar field) of 9 8 mm. Cranidial
length, excluding anterior tongue of glabella, is 7-2 mm, and basal breadth of glabellar tongue is 0.3 of
maximum glabellar breadth. Glabellar outline closely resembles that of Amphitryon radians and the most

DEAN AND ZHOU: ORDOVICIAN TR I LO B ITES F ROM TU R K E Y

643

obvious difference is the total absence of glabellar furrows in the Turkish material, though the value of this
as a generic character is unknown. Whittington’s (1966, p. 72, text-fig. 4a g) illustrations of A. radians from
Bohemia show three incised pairs of glabellar furrows, but the occipital ring and palpebral lobes are virtually
indistinguishable from those of the Turkish specimen, and the glabellar tongue ends behind a triangular
preglabellar field, the apex of which probably coincided with a median suture. The breadth (tr.) of the
glabellar tongue as shown by Whittington equals only about 014 the glabellar breadth, but in material from
Bohemia and Poland assigned to A. radians by Kielan (1960, pi. 2, figs. 3, 5, 6) the corresponding figure
varies from 0-2 to 0-25, though a small Polish cranidium identified as Amphitryon sp. (Kielan 1960, pi. 3, fig.
12) has a very narrow glabellar tongue (only 013 estimated). It is possible that the relative breadth of the
glabellar tongue changed during ontogeny, and for present purposes more importance is attached to the
triangular preglabellar field, which readily distinguishes the Turkish specimens from Remopleurides (see
account of type species R. colbii Portlock 1843 in Whittington 1950, p. 540).

Other species previously assigned to Remopleurides that have a triangular preglabellar field include the
Chinese forms R. nasutus Lu 1957 (see Lu 1975, pp. 109, 299, pi. 3, fig. 15; pi. 4, figs. 5, 8, 9), from the
Arenig of south Shensi and West Hupeh, and R. shihtzupuensis Lu 1957 (see Lu 1975, pp. Ill, 301, pi. 4,
fig. 15), from the Arenig to Llandeilo of North Kueichou and West Hupeh. R. nasutus has a wide glabellar
tongue with rounded frontal margin and the triangular preglabellar field is shorter and less distinct than in
the Turkish material; R. shihtzupuensis is closer to the latter in the shape of the preglabellar field but has a
much longer glabellar tongue. Sculptel/a and Sculptaspis from the Middle Ordovician of Norway (Nikolaisen
1982, pp. 265, 276) bear a superficial resemblance to the present material but their cranidia lack the triangular
preglabellar field.

Family illaenidae Hawle and Corda 1847
Genus stenopareia Holm 1886

Type species. I/laenus Linnarssoni Holm 1882, Boda Limestone (Ashgill Series), Dalarne, Sweden.

Stenopareia sp.

Plate 62, fig. 5

Figured specimen. It. 19558.

Locality. §ort Dere, Z.36.

Description and discussion. The Turkish cranidium is too compressed for specific identification, but the form
of the axial furrows and the width of the fixigenae are consistent with those of Stenopareia linnarssoni ,
redescribed by Warburg (1925, p. 117, pi. 2, figs. 14 18), though the position of the eyes is not clear.

Family lichidae Hawle and Corda 1847
Subfamily lichinae Hawle and Corda 1847
Genus lichas Dalman 1827

Type species. Entomostracites laciniatus Wahlenberg 1821, from the Dalmanitina Beds (Ashgill) of Bestorp,
Mosseberg, Sweden.

Lichas aff. laciniatus (Wahlenberg 1821)

Plate 62, figs. II?, 12, 14

Figured specimens. It. 19473 (PI. 62, fig. 1 1), It. 19474 (PI. 62, fig. 12), It. 19475 (PI. 62, fig. 14).

Locality. $ort Dere Z.34.

Description and discussion. Two cranidia, slightly distorted by dorsal compression, show clearly the composite
lateral lobes incompletely defined posterolaterally as in Lichas laciniatus , redescribed by Warburg (1925,
p. 295, pi. 8, figs. 16, 17, 20; 1939, p. 15, pi. 9, fig. 3a, h). The occipital lobes are also similar and the palpebral
lobes correspond in size and location. The long axes of the composite lateral lobes diverge forwards at about
45 , comparable with Warburg’s illustrations, but the median lobe narrows to half the breadth of the
composite lobes; this contrasts with L. laciniatus , where the median and composite lobes are of equal breadth

644

PALAEONTOLOGY, VOLUME 31

in larger specimens, though one small cranidium (Warburg 1925, pi. 8, fig. 17) has the median lobe slightly
narrower. The narrowest part of the median lobe is also much narrower than that of the neotype of L. affinis
Angelin, 1854, from the Ashgill of Sweden (Warburg 1939, pi. 9, fig. 13). The glabella in the Turkish material
is relatively shorter than that of L. laciniatus (length : breadth = 34 : 34 versus 38 : 34), the occipital ring is
proportionately narrower (exsag.), especially distally, and the occipital lobes are notably larger. L. laciniatus
ranges from the Ashgill into the Llandovery Series and cranidia from northern England described by Temple
(1969) correspond to the Swedish material.

Genus dicranopeltis Hawle and Corda 1847

Type species. Lidias scabra Beyrich 1845, upper part of the Liten Formation (Wenlock), Svaty Jan, near
Beroun, Czechoslovakia.

Dicranopeltis sp.

Plate 62, figs. 9? and 1 3

Figured specimens. It. 19471 (PL 62, fig. 9), It. 19472 (PI. 62, fig. 13).

Locality. §ort Dere, Z.34.

Description and discussion. The Turkish pygidiunr generally resembles material from the Ashgill (Boda
Limestone and Dalmanitina Beds) of Sweden described by Warburg, first as Dicranopeltis elegans (Tornquist
1884) (Warburg 1925, p. 291, pi. 7, figs. 27 and 31; pi. 8, figs. 9 and 10), and later put in synonymy with D.
polytomus (Angelin 1854) (Warburg 1939, p. 134, pi. 11, figs. 4 6). D. sp. differs from the pygidium of D.
polytomus in having a less sharply constricted terminal piece, and the axial furrows are moderately convergent
to what is essentially a low, slightly tapered post-axial ridge, separated by a pair of very shallow grooves
from the extremities of the third pleurae, which end in short free points separated by a very small median
notch. The third pleurae of the Turkish pygidium occupy a relatively smaller area and the equisized pleural
bands end in rounded tips at the broad border. The entire surface, except furrows, is covered with closely
spaced, small tubercles that become slightly smaller on the posterior border. The rearmost part of the axis,
immediately in front of the post-axial ridge, is bulbous as in the Swedish material.

The sole, possibly associated cranidial fragment is crushed and the position of some furrows apparently
displaced, so that the composite lateral lobes appear more divergent forwards than originally, and the
palpebral lobe has been displaced towards the glabella. The occipital ring and right occipital lobe resemble
those of D. polytomus , but the rearmost part of the axial furrow is almost indiscernible, in marked contrast
to the well-defined furrow in the Swedish species.

Family odontopleuridae Burmeister 1843
Subfamily odontopleurinae Burmeister 1 843
Genus diacanthaspis Whittington 1941

Type species. Diacanthaspis cooperi Whittington 1941, Lower Martinsburg Formation (Caradoc), Virginia,
USA.

Diacanthaspis sp.

Plate 61, figs. 1 and 2; Plate 62, fig. 10

Figured specimens. It. 19542 (PI. 61, fig. 1), It- 19543 (PI. 61, fig. 2), It. 19557 (PI. 62, fig. 10).

Locality. §ort Dere, Z.34.

Description and discussion. Cranidium has parallel-sided glabella with wide, subparallel-sided median glabellar
lobe and narrow, elongated Ip and 2p lateral glabellar lobes. It agrees well with the cranidium of D. sp. of
Lu and Zhou (1981, p. 20, pi. 3, fig. 10) from the Tangtou Formation (low Ashgill) of the Nanjing Hills,
China. In each case the cranidium is incomplete, and no associated pygidium is available for the Chinese
form, so that specific identity cannot be established. Occipital spines are not preserved on the Turkish
specimen, though a prominent median tubercle is present close to the occipital furrow. Another species of
Diacanthaspis with parallel-sided glabella is D. laokuangshanensis Lu and Chang (1974, p. 136, pi. 56, figs.

DEAN AND ZHOU: ORDOVICIAN TRILOBITES FROM TURKEY

645

5-7), from the Wufeng Formation (Ashgill) of western Sichuan, China, but the lateral glabellar lobes are
wider (tr.) and the median lobe expands forwards.

A fragmentary Turkish librigena shows a Hat, coarsely granulate genal field and small, raised eye socle.
Lateral border is narrow, ridge-like, with marginal spines that become successively shorter (tr.) anteriorly.

The pygidium has a semicircular outline and a convex axis comprises two axial rings and short terminal
piece. Pleural field narrow (tr.), weakly defined by narrow border of low convexity. There are five equispaced
pairs of slender, radiating border spines, whereas in all other known species of Diacanthaspis there are six
or more pairs.

Subfamily miraspidinae R. and E. Richter 1917
Genus miraspis R. and E. Richter 1917

Type species. Odontopleura mira Barrande 1846, from the Liten Formation (Wenlock), Lodenice, Czecho-
slovakia.

Miraspis sp.

Plate 62, fig. 3

Figured specimen. It. 19556.

Locality. §ort Dere, Z.34.

Description. Glabella tapers gently forwards, its basal breadth half that of cranidium. It is transversely convex
with elongated, subrectangular median lobe that expands abruptly to form very short frontal glabellar lobe.
Three pairs of lateral glabellar lobes well defined by deep longitudinal furrows; Ip lobes oval in outline, their
width equal almost to that of median glabellar lobe and to 0-4 of glabellar length; 2p lobes sub-square in
plan, slightly narrower than lp lobes; 3p lobes tiny, transverse, with length (exsag.) one-quarter that of 2p
lobes. Three pairs of almost transverse lateral glabellar furrows present; Ip and 2p furrows deeply incised,
3p pair shallow. Occipital furrow shallow, broad; anterior part of occipital ring with median tubercle and
pair of long, broadly based spines that extend upwards posterolaterally; posterior part of occipital ring not
preserved. Axial furrows distinct posteriorly, shallow anteriorly. Fixigenae narrower posteriorly than lp
lobes, and become still narrower further forwards. Palpebral lobes situated opposite mid-points of lp lobes.
Sutural and palpebral ridges slightly convex abaxially and converge forwards. Anterior border as narrow as
palpebral ridges, and defined by shallow preglabellar furrow. Cranidial surface covered with densely spaced
tubercles of different sizes.

Discussion. The cranidium, though imperfectly preserved, is compatible with that of the lectotype
of Miraspis mira (Barrande 1852, pi. 39, fig. 3, selected by Pribyl in Horny and Basil 1970, p. 203)
and of a well-preserved dorsal exoskeleton of that species figured by Prantl and Vanek (in Horny
et al. 1958, pi. 5, fig. 1) but differs in having the median glabellar lobe narrower, with straighter
sides, while the lp and 2p glabellar furrows are more distinct abaxially.

Ordovician species of Miraspis have been recorded from Sweden (Whittington and Bohlin 1958;
Bruton 1966), Norway (Owen and Bruton 1980), eastern Ireland (Dean 1974), North Wales and
Scotland (Whittington and Bohlin 1958, p. 43). In the shape of the cranidium and median glabellar
lobe and in the pattern of surface granulation the Turkish specimen resembles M. sp. of Owen
and Bruton (1980, p. 36, pi. 10, figs. 18 and 20), from the uppermost Solvang Formation (probably
low Ashgill) of Ringerike, Norway, but the latter has shallower Ip and 2p glabellar furrows and
the 2p glabellar lobes are proportionately much smaller.

Genus and species undetermined

Plate 62, fig. 7

Figured specimen. It. 16062.

Locality. §ort Dcre, Z.34.

Description and discussion. A fragmentary internal mould, interpreted tentatively as part of a pygidium, has
the surface, excluding furrows, covered with coarse, closely spaced tubercles. The ornamentation bears some

646

PALAEONTOLOGY, VOLUME 31

resemblance to that of the odontopleurids, and there is some evidence of one of a pair of subparallel ridges
ending in line with the first axial ring; in the odontopleurids similar ridges link the corresponding ring with
a pair of large marginal spines. Three markedly unequal axial rings, separated by transversely straight ring
furrows that deepen abaxially, are followed by a minute, weakly defined terminal piece and gently declined
postaxial field. No satisfactory comparison was made.

Acknowledgements. Dean’s field-work in the region south of Hakkari, supported in part by the Natural
Environment Research Council and the Royal Society, would have been impossible without the assistance
of T.P.A.O. and its geologists, especially Dogan Perii^ek. The help of Olivier Monod, Universite de Paris-
Sud, Orsay, is also gratefully acknowledged. Professor H. B. Whittington read the manuscript and made
suggestions for its improvement.

REFERENCES

altinli, i. e. 1963. Explanatory text of the geological map of Turkey, Cizre. Maden Tetkik Arama Enstit.
Haritasi, 103 pp.

Angelin, n. p. 1851 -1878. Palaeontologia Scandinavica: Academiae Regiae Scientiarum Suecanae (Holmiae);
Pars I. Crustacea formationis transitionis, pp. 1 24, pis. 1-24 [ 1 85 1 ); Pars II, pp. i-ix, 25-92, pis. 25-41
[1854]; republished in revised and combined form (ed. G. Lindstrom), pp. x -I- 96, pis. 1-42 [1878].
barrande, J. 1846. Notice preliminaire sur le Systeme silurien et les Trilobites de Boheme, vi + 97 pp. Leipzig.

— 1852. Systeme silurien du centre de la Boheme. lere partie. Recherches paleontologiques. Vol. 1 . Crustaces,
Trilobites, xxx + 935 pp. Prague and Paris.

BEYRICH, E. 1845. Ueber einige bohmischen Trilobiten , 47 pp. Reimer, Berlin.

bruton, d. l. 1966. A revision of the Swedish Ordovician Odontopleuridae (Trilobita). Bull. geol. Instn Univ.
Uppsala , 43, 1 40.

— 1968. A revision of the Odontopleuridae (Trilobita) from the Palaeozoic of Bohemia. Skr. norske
Vidensk-Akad. 25, 73 pp.

— 1976. The trilobite genus Phillipsinella from the Ordovician of Scandinavia and Great Britain.
Palaeontology , 19, 699-718.

burmeister, H. 1843. Die Organisation der Trilobiten , 147 pp. Berlin.

Campbell, k. s. w. 1967. Trilobites of the Henryhouse Formation (Silurian) in Oklahoma. Bull. Okla. geol.
Surv. 123, I 227.

chugaeva, m. N. 1958. The Ordovician of Kazakhstan. Ill: The Ordovician trilobites of the Chu-Ili Mountains.
Trudy geol. Inst., Leningr. 9, 5-138. [In Russian.]

dalman, J. w. 1827. Onr Palaeaderna, eller de sa kalladc Trilobiterna. Bill. K. Svenska VetenskAkad. Handl.

1, 113-152, 226-294.

dean, w. t. 1959. Duftonia , a new trilobite genus from the Ordovician of England and Wales. Palaeontology ,

2, 143-149.

— 1966. The Lower Ordovician stratigraphy and trilobites of the Landeyran Valley and the neighbouring
district of the Monlagne Noire, south-western France. Bull. Br. Mus. nat. Hist. (Geol.), 12, 245-353.

— 1967m The distribution of Ordovician shelly faunas in the Tethyan region. In adams, c. j. and ager,
d. v. (eds.). Aspects of Tethyan biogeography. Systematics Assoc. Spec. Publ. 7, 11-44.

— 19676. The correlation and trilobite fauna of the Bedinan Formation (Ordovician) in south-eastern
Turkey. Bull. Br. Mus. nat. Hist. (Geol.), 15, 81 123.

— 1971. The trilobites of the Chair of Kildare Limestone (LJpper Ordovician) of eastern Ireland.
Palaeontogr. Soc. [Monogr.], part 1, 1-60.

— 1974. Ibid. 2, 61 98.

— 1980. The Ordovician System in the Near and Middle East. Correlation chart and explanatory notes.
Publ. int. Union geol. Sci. 2, 1-22.

— monod, o. and perin^ek, d. 1981. Correlation of Cambrian and Ordovician rocks in southeastern
Turkey. Petroleum Activities at the 100th Year (100 Yilda Petrol Faaliyeti). Tiirkiye Cumhuriyet Petrol
Isleri Genel Mudurliigu Dergisi, 25, 269 291 (English), 292-399 (Turkish).

glimberg, c. f. 1961. Middle and Upper Ordovician strata at Lindegard in the Fagelsang district, Scania, S.
Sweden. Geol. For. Stockh. Fork. 83, 79 85.

hawle, J. and corda, A. J. C. 1847. Prodrom einer Monographic der bohmischen Trilobiten , 176 pp. J. G.
Calve, Prague.

DEAN AND ZHOU: ORDOVICIAN T R 1 LO B ITES F ROM TU R KE Y

647

holliday, s. 1942. Ordovician trilobites from Nevada. J. Paleont. 16, 471-478.

holm, G. 1882. De Svenska artena af Trilobitslagel Illaenus (Dalman). Bih. K. Svenska VelenskAkad. Hand/.
7, xiv + 148 pp.

— 1886. Illaeniden. Revision der Ostbaltischen silurischen Trilobiten von. Fr. Schmidt. Abt. III. Mem.
Acad. imp. Sci. St.-Petersb. (7) 33, 1-173.

horny, R. and bastl, F. 1970. Type specimens of fossils in the National Museum , Prague. Volume 1. Trilobita.
354 pp. Museum of Natural History, Prague.

— prantl, f. and vanek, j. 1958. K. otazce hranice mezi wenlockem a ludlowem v Barrandienu; Sb.
Ustfed. Ust. geol. 24, 217-256 (in Czech), 257-278 (in English).
hupe, p. 1955. Classification des Trilobites. Annls Paleont. 41, 91 325 (1 1 1-345).

ingham, j. k. 1970. The Upper Ordovician trilobites from the Cautley and Dent districts of Westmorland
and Yorkshire. Palaeontogr. Soc. [ Monogr .], part 1,1 58.

— 1974. Ibid. 2, 59-87.

jaanusson, v. 1963. Classification of the Harjuan (Upper Ordovician) rocks of the mainland of Sweden.
Geol. For. Stockh. Fork. 85, 110 144.

— 1982. Introduction to the Ordovician of Sweden. In bruton, d. l. and williams, s. h. (eds.). Field
excursion Guide. IV. International Symposium on the Ordovician System. Paleont. Contr. Univ. Oslo , 279,
1-217.

janvier, p., lethiers, f., monod, o. and balka§, o. 1984. Discovery of a vertebrate fauna at the Devonian
Carboniferous boundary in SE Turkey (Hakkari Province). Jl Petrol. Geol. 7, 147-168.
jones, t. r. and woodward, h. 1898. A monograph of the British Palaeozoic Phyllopoda (Phyllocarida,
Packard). Palaeontogr. Soc. [Monogr.]. part 3, 125-176.
kielan, z. 1960. Upper Ordovician trilobites from Poland and some related forms from Bohemia and
Scandinavia. Palaeont. pol. 11, I -198.

kobayashi, t. 1935. The Cambro-Ordovician formations and faunas of South Chosen. Palaeontology, Pi. 3,
Cambrian faunas of South Chosen with special study on the Cambrian trilobite genera and families. J.
Fac. Sci. Tokyo Univ. 4, 49 344.

1951. On the Ordovician trilobites in Central China. Ibid. Sect. II, 8, I 87.

— 1960. Some Ordovician fossils from east Tonkin, Viet Nam. Jap. J. Geol. Geogr. 31, 39 48.
koroleva, m. n. 1959. New species of trilobites from the Middle and Upper Ordovician of northern

Kazakhstan. Doki Akad. Nauk SSSR , 124, 1313-1316. [In Russian.]
li yaoxi, song lisheng, zhou zhiqiang and yang jingyao. 1975. Stratigraphical Gazetteer of Lower
Palaeozoic , western Dabashan , 372 pp. Geological Publishing House, Beijing. [In Chinese.]
linnarsson, j. g. o. 1869. Om Vestergotlands Cambriska och Siluriska Aflagringar. Bih. K. Svenska
VetenskAkad. Hand!. 8 (2), 1 89.

lu yanhao. 1975. Ordovician trilobite faunas of central and southwestern China. Palaeont. sin. 11,
1-484.

— and chang wentang. 1974. Ordovician trilobites, 124-136. In A handbook of Stratigraphy and
Palaeontology in southwest China. Science Press, Beijing. [In Chinese.]

— chu chaoling, chien yiyuan and hsiang leewen. 1965. Trilobites of China , 766 pp. Science Press,
Beijing. [In Chinese.]

— and zhou zhiyi. 1979. Systematic position and phylogeny of Hammatocnemis (Trilobita). Acta palaeont.
sin. 18, 415 434. [In Chinese with English summary.]

— 1981. Early Upper Ordovician trilobites from the Nanjing Hills. Bull. Nanjing Inst. Geol. Palaeont.
3, 1-27. [In Chinese with English abstract ]

m‘coy, f. 1846. A synopsis of the Silurian fossils of Ireland , 72 pp. McGlashand and Gill, Dublin.

1849. On the classification of some British fossil Crustacea, with notices of some new forms in the
Lhiiversity Collection at Cambridge. Ann. Mag. nat. Hist. (2) 4, 161-179, 330-335, 392-414.
moore, r. c. (ed.). 1959. Treatise on Invertebrate Paleontology , Pt. O, Arthropoda 7, xix + 560 pp. Geological
Society of America and University of Kansas Press, Lawrence, Kansas.
nicholson, h. a. and etheridge, r. 1878. A monograph of the Silurian fossils of the Girvan District in Ayrshire.
1. ix+135. Edinburgh and London.

nikolaisen, f. 1982. The Middle Ordovician of the Oslo Region, Norway, 32. Trilobites of the Family
Remopleurididae. Norsk geol. Tidsskr. 62, 231 -329.

olin, e. 1906. Om de Chasmopskalken och Trinucleusskiffern motsvarande bildningarne i Skane. Acta Univ.
lund. nf [2], 2 (3), 1-79.

opik, a. a. 1937. Trilobiten aus Estland. Acta Comment. Univ. tartu. (A) 32 (3), 1-163.

648 PALAEONTOLOGY, VOLUME 3 !

opik, A. A. 1967. The Mindyallan Fauna of north-western Queensland. Bull. Bur. Miner. Resour. Geol.
Geophys. Aust. 74, I 404.

Owen, a. and bruton, d. l. 1980. Late Caradoc-early Ashgill trilobites of the central Oslo region, Norway.
Paleont. Contr. Univ. Oslo , 245, 63 pp.

portlock, j. E. 1843. Report on the geology of the county of Londonderry , and of parts of Tyrone and
Fermanagh , xxi + 784 pp. Dublin (Milliken. Hodges and Smith) and London (Longman, Brown, Green,
and Longmans).

QIU HONGAN, LU YANHAO, ZHU ZHAOLING, BI DECHANG, LIN TIANRUI, ZHOU ZHIYI, ZHANG QUANZHONG, QIAN
yiyuan, ju TIANYIN, han nairen and WEI XIUZHE 1983. Trilobita. In Palaeontological Atlas of east China.
1, 28-254. Geological Publishing House, Beijing. [In Chinese.]
reed, f. r. c. 1915. Supplementary memoir on new Ordovician and Silurian fossils from the Northern Shan
States. Mem. geol. Surv. India Palaeont. indie a, NS 6 (1), 1-123.

Reynolds, s. h. 1894. Woodwardian Museum Notes. Certain fossils from the Lower Palaeozoic rocks of
Yorkshire. Geol. Mag. 31, 108- 111

richter, r. and richter, e. 1917. Uber die Einteilung der Familie Acidaspidae und iiber einige ihrer
devomschen Vertreter. Zentbl. Miner. Geol. Palaont. (1917), 462 472.
rouault, m. 1847. Extrait du Memoire sur les Trilobites du Departement d'Hle-et-Vilaine. Bull. Soc. geol.
Fr. (2) 4, 309 328.

sars, m. 1835. Ueber einige neue oder unvollstandig bekannte Trilobiten. Isis, Jena (1835), 4.
shaw, f. c. and ormiston, a. r. 1964. The eye socle of trilobites. J. Paleont. 38, 1001-1002.
sheng, s. f. 1934. Lower Ordovician trilobite fauna of Chekiang. Palaeont. sin. B 3 (1), I 19.

1974. Ordovician trilobites from western Yunnan and its stratigraphical significance. In Subdivision and
correlation of the Ordovician System in China , T- 1 53, 96 140. Geological Publishing House, Beijing. [In
Chinese.]

sun, y. c. 1931. Ordovician trilobites of Central and Southern China. Palaeont. sin. B 7, 1 47.
temple, j. t. 1954. The hypostome of Encrinurus variolaris and its relation to the cephalon. Geol. Mag. 91,
315-318.

1969. Lower Llandovery (Silurian) trilobites from Keisley, Westmorland. Bull. Br. Mus. nat. Hist.
(Geol.), 18, 197-230.

THOMAS, a. t. 1978. British Wenlock trilobites. Palaeontogr. Soc. [ Monogr .], 1-56.

— and owens, r. m. 1978. A review of the trilobite family Aulacopleuridae. Palaeontology , 21, 65 81.
tornquist, s. l. 1884. Undersokningar ofver Siljansomradets trilobitfauna. Sver. geol. Unders. Afh. C 66,

1 101.

tripp, r. p. 1958. Stratigraphical and geographical distribution of the named species of the trilobite superfamily
Lichacea. J. Paleont. 32, 574 582.

vogdes, a. w. 1890. A bibliography of Palaeozoic Crustacea from 1698 to 1889, including a list of North
American species and a systematic arrangement of genera. Bull. US geol. Surv. 63, 1 177.
wahlenberg, G. 1821. Petrificata Telluris Svecanae. Nova Acta R. Soc. Scient. upsala , 8, 1-116, 293-297.
warburg, E. 1925. The trilobites of the Leptaena Limestone in Dalarne. Bull. geol. Instn Univ. Uppsala , 17,

1 446.

1939. The Swedish Ordovician and Lower Silurian Lichidae. Bill. K. svensk VetenskAkad. Hand!. 17 (4),

I 162.

weber, v. n. 1932. Trilobites of Turkestan. Izd. Uses. Geol.-Razv. Ob'ed. NKTP , iv+157 pp. [In Russian
with English summary.]

1948. Trilobites of the Silurian beds. No. 1. Lower Silurian trilobites. Monogr. Palaeont. USSR, 69 ( I ),

1 110. [In Russian.]

whittard, w. f. 1955. The Ordovician trilobites of the Shelve inlier, west Shropshire. Palaeontogr. Soc.
[Monogr.], part 1,1 40.

Whittington, H. B. 1941. Silicified Trenton trilobites. J. Paleont. 15, 492-522.

1950. Sixteen Ordovician genotype trilobites. Ibid. 24, 531-565.

1965. A monograph of the Ordovician trilobites from the Bala area, Merioneth. Palaeontogr. Soc.
[Monogr.], part 2, 33-62.

- 1966. Ibid. 3, 63 92.

and bohlin, B. 1958. New Lower Ordovician Odontopleuridae (Trilobita) from Oland. Bull. geol. Instn
Univ. Uppsala , 38, 37-45.

WILLIAMS, A., STRACHAN, I., BASSETT, D. A., DEAN, W. T., INGHAM, J. K.., WRIGHT, A. D. and WHITTINGTON,

ii. b. 1972. A correlation of Ordovician rocks in the British Isles. Geol. Soc. Loud., Spec. Rep. 3, I 74.

DEAN AND ZHOU: ORDOVICIAN TRILOBITES FROM TURKEY

649

yi yongen. 1957. The Caradocian trilobate fauna from the Yangtze Gorges. Acta Palaeont. sin. 5 (4), 527
560. [In Chinese with English summary.]

yin gongzheng and lee SHANJi. 1978. Trilobita. In Palaeontological Atlas of Southwest China. Guizhou
Province (1), 385 595. Geological Publishing House, Beijing. [In Chinese.]
zenker, J. c. 1833. Beitrage zur Naturgeschichte der Utwelt , Organische Reste ( Petrefacten ) aus der Altenburger
Braunkohlen-Formation, dem Blankenburger Quadersandstein, jenaischen bunten Sandstein und bohmischen
Uebergangsgebirge, viii + 67 pp. Jena.

zhang tairong. 1981. Trilobita. In Atlas of palaeontology of NW China'. Xinjiang Volume 1, 134-213.
Geological Publishing House, Beijing. [In Chinese.]

zhang wentang. 1979. On the Miomera and Polymera (Trilobita). Scientia sin. 10, 996 1004. [In Chinese.]
li jijin, ge meiyu and chen junyuan. 1982. Subdivision and correlation of the Ordovician in China
Correlation chart and its explanation for the Chinese Ordovician System. In Correlation charts and then-
explanations for Chinese strata , 55 72. Science Press, Beijing.
zhou tianmei, liu yiren, meng xiansong and sun zhenhua. 1977. Trilobita. In Atlas of Palaeontology of
central and south China , 140-266. Geological Publishing House, Beijing. [In Chinese.]
zhou zhiyi and dean, w. t. 1986. Ordovician tri lobites from Chedao, Gansu Province, north-west China.
Palaeontology , 29, 743 786.

— yin gongzheng and tripp, r. p. 1984. Trilobites from the Ordovician Shihtzupu Formation, Zunyi,
Guizhou Province, China. Trans. R. Soc. Edinb. 75, 13-36.

w. T. DEAN

Department of Geology
University College

Cardiff CF1 1XL
South Wales

ZHOU ZHIYI

Institute of Geology and Palaeontology
Academia Sinica

Typescript received 19 March 1987
Revised typescript received 4 June 1987

Chi-Ming-Ssu
Nanjing, China

A SILURIAN CEPHALOPOD GENUS WITH
A REINFORCED FRILLED SHELL

by SVEN STRIDSBERG

Abstract. A new cephalopod genus, Torquatoceras, comprising two new species T. undulation and T. auritum ,
is described from the Silurian of Gotland. Torquatoceras is unique in that transverse crenulated frills have
been secreted during the entire growth of the shell. These frills, mainly consisting of prismatic layers, might
have served as a reinforcement of the shell. In T. undulation sexual dimorphism based on size variations is
demonstrated. In T. attrition there are two vertical septa inside the body-chamber, partly separating the
hyponomic sinus from the apertural opening.

The fossil record from the Baltic island of Gotland demonstrates very well that the Silurian
cephalopod fauna in the area was rich. The shallow, tropical Silurian sea favoured the establishment
of various genera and species, and thus far more than eighty species in fifteen genera have been
described from the island.

Mostly the cephalopods on Gotland are found in large thanatocoenoses and the new species
described herein are both collected at such a place, the Samsugns quarry in the Wenlock Slite Beds
(Laufeld 1974). This quarry is moderate in size, only 75-100 m across and about 10 m deep, but
is unique regarding the cephalopod fauna. No other locality on Gotland shows such a variety of
species with heterogeneous shell morphology, and altogether twenty species in ten genera have
been identified (Angelin and Lindstrom 1880, 1 species; Lindstrom 1890, 5 species; Hedstrom 1917,
8 species; Stridsberg 1985a, 4 species and herein 2 species). Still more taxa of cephalopods from
Storugns have been collected and are waiting description, but in contrast to those already described
they are mainly orthocones.

There are reasons for believing that all twenty species did not actually live in the Storugns area,
since from their shell morphology a number of them appear to have had the same mode of life.
However, the floating chambers of the cephalopod shells certainly contributed to post-mortem
drifting, and obviously Samsugns was a kind of meeting place for the Silurian drifters. Similar
drifting of extant Nautilus is well documented, especially from the south-western Pacific (Toriyama
et al. 1965; Hamada 1964; Saunders and Spinosa 1979).

In the varied cephalopod fauna from Samsugns there is a genus with a most unusual shell
surface, consisting of crenulated transverse frills (text-fig. 1). Various kinds of shell ornamentation
are well known from other cephalopod species, but this genus, Torquatoceras gen. nov., has an
exceptional protruding system of frills around the shell. Crenulate transverse frills also occur in
the Ordovician genus Zitteloceras Hyatt 1884, and according to Foerste (1916, p. 51) these frills,
or rather lamellae, ‘may have extended for a distance of about half a millimeter from the general
surface of the cyrtoceracone’. The frills in Zitteloceras appear to be strongly similar with those on
Pentameroceras facula Stridsberg 1985a, although the latter are not crenulated, and as discussed
in the description of Torquatoceras herein, the frills on P. facula are probably not constructed in
the same way as those in Torquatoceras.

The shell of the Bohemian species Corbuloceras corbulatum (Barrande 1866), is covered by
crenulated frills, extending a few millimetres from the shell wall (Barrande 1867, pp. 586-587;
Horny 1965, pp. 132-136, Tab. 1-2), however, Corbuloceras has longitudinal ribs on the shell and
when crossing these ribs the frills have distinct protrusions (text-fig. 2). Similar protrusions are not
preserved on any of the specimens of Torquatoceras.

| Palaeontology, Vol. 31, Part 3, 1988, pp. 651-663, pis. 63-64.|

© The Palaeontological Association

652

PALAEONTOLOGY, VOLUME 31

text-fig. 1 . Close-up of photographs of frills in Torquatoceras undulatum. a and b show the rhythmic pattern
found in some mature specimens with worn down (a) and well-preserved (b) shells, x6-5. c, enlarged
crenulations from d, to show the outline of the frills. RM Mo 57307, x4-5 and x 2-2.

In some Silurian cephalopods, for example Dawsonoceras , regularly repeated bulges surround
the shell. However, in these cases the bulges are the result of a repeated temporary widening of
the aperture, and the thickness of the shell is thus the same in the bulges as in the adjacent parts
of the shell (text-fig. 3). In Torquatoceras , however, the protruding frills are almost untraceable
from the inside of the shell. Only slight depressions indicate occasionally where the frills are
situated (text-fig. 4).

CONSTRUCTION OF FRILLS

The presence of frills on the shell surface of Torquatoceras , makes the cephalopod resemble a
rugose coral. In some rugose corals rhythmic shell growth is very common and frills similar to
those of Torquatoceras are found in various species. The shell growth, however, is far from similar

STRIDSBERG: SILURIAN CEPHALOPOD WITH FRILLED SHELL

653

text-fig. 2. Three views (a-c) from various angles of the same area of Corbuloceras corbulatum (Barrande
1866), specimen L 6561, Narodni Muzeum, Prague, x 1-75. a, lateral view with apical end upwards, b, view
towards the apical end showing the protruding parts of the frills on top of the longitudinal ribs. In c it is
evident that the protrusions are more striking than the ribs, and thus not just a reflection of the underlying

surface.

text-fig. 3. Cross-section of a specimen of Dawsonoceras, illustrating the regularly
repeated bulges surrounding the shell. The dark thin lines are the three last septa.

x 2-5.

654

PALAEONTOLOGY, VOLUME 31

text-fig. 4. Cross-sections of the outermost part of the shell in Torquatoceras
undulatum, RM Mo 57245. In a the reinforcement in the apertural area is
clearly visible. In b the beginning of the reinforcement can be seen on the
lowermost part of the shell, x 15.

as the frills on corals are interpreted as having been built up during extreme stretching out of the
secreting ectoderm. Such stretchings are supposed to have taken place when the polyps were
extended to release the planula larvae when the Moon was in a specific position. Similar rhythmic
shell growth has been suggested for the cephalopods (Kahn and Pompea 1978) but as explained
by Saunders and Ward (1979), comparisons can not be made.

In molluscs the mantle would hardly act in the same way as the ectoderm in the corals, and
furthermore, the frills on Torquatoceras are present from the most juvenile stage to the fully adult
specimen.

The construction of the frills in Torquatoceras is a procedure which is unusual among nautiloids,
as the mantle growth must have frequently changed directions. Instead of continuing shell secretion
along the apertural edge, the mantle must have turned round to secrete the apertural side of the
frill after the secreting of the apical side of each frill (text-fig. 5a-b). After the deposition of a frill
the shell-secreting epithelium inside the frill must have reduced its length, and during this phase
shell deposits eventually filled up the interior of the frill with nacreous layers (text-fig. 5c-d).

After the completing of a frill the shell-secreting epithelial cells from the interior of the frill must
have been reduced, as the inside of the body-chamber has a smooth surface (text-figs. 4 and 5).

STRIDSBERG: SILURIAN CEPHALOPOD WITH FRILLED SHELL

655

text-fig. 5. Four hypothetical phases (a d) in the secretion of a frill in
Torquatoceras, drawn from the frills in text-fig. 4b. a, the secretion of the
shell wall between two frills and the apical side of a partly secreted frill, b,
the mantle has turned round and secretes the apertural side of the new frill,
c, the frill is completed and the mantle continues the secretion of the
phragmocone wall, d, the intermediate space inside the frill is ‘filled up’ with,
most probably, nacreous layer and the mantle follows the phragmocone shell
wall. Abbreviations: ma = mantle; pe = periostracum; pr = outer prismatic
layer; na = nacreous layer.

Cross-sections of the shell do not show any space for extensions of the mantle into the frills (text-
fig. 4), and thus any damage on the frills could not be repaired after the withdrawal of the mantle.

Due to the recrystallization of the shell material, no details of the various shell layers can be
observed on any of the specimens examined. If, however, the shell of Torquatoceras was constructed
in the same way as the shell of Nautilus (Mutvei 1964), it can be assumed that the frills did not
have any semi-prismatic layer inside the nacreous layer, as the secretion of this semi-prismatic
layer took place long after the ‘closing’ of the frills when the epithelial cells turned to ‘muscle-
cells’. Only the periostracum the outer prismatic layer and the nacreous layer can have been
represented in the crenulated frills.

REGULARITY OF FRILL GROWTH

As septa and frills in Torquatoceras were constructed at intervals, there is reason to ask if secretion
of frills was in any way correlated with the secretion of septa. Frills, as well as septa, are very
closely spaced in the juvenile part of the phragmocone and considerably more widely spaced in
the mature part. During the final growth stage, however, the frills were more closely spaced again
but this was caused by the shape of the aperture. The area around the hyponomic sinus has very
limited space between the last frills, while on the other hand, the frills on the ventrolateral lobes
are fairly widely spaced.

Due to the apical end of the phragmocone always being missing, it is not possible to compare
the total number of septa and frills in a complete shell. However, by counting septa and frills
backwards, from the aperture towards the apical end, it is possible to make an hypothetical
reconstruction of various growth stages. If the growth of septum and frills were synchronized the
various reconstructed growth stages must show a shell with roughly the same proportions of body-
chamber and the chambered section of the phragmocone as in a mature specimen.

A specimen with an unusually large number of preserved septa has been used for reconstructions
of three different growth stages (text-fig. 6). In each case, the same number of frills and septa have
been removed, and the proportions of body-chamber and chambered part of the shell can be
compared. Due to the difficulties in observing the frills in the juvenile part of the shell,
reconstructions in this part of the shell have been omitted.

The comparisons of the reconstructed growth stages, including specimens not illustrated, show

656

PALAEONTOLOGY, VOLUME 31

text-fig. 6. A photograph and three drawings of a cut shell of Torquatoceras undulation , RM Mo 56353,
showing different growth stages, a, the cut shell, x2. b, the mature shell as shown in a. c, the same shell as
in b but with six frills and six septa removed, d, the same shell as in b but with twelve frills and twelve septa
removed. The missing apical end of the shell is reconstructed.

that the hypothesis of a synchronized secretion of septa and frills is realistic, although the body-
chamber proportionally decreases in relative volume from the juvenile reconstructions to the adult
specimens. This might be explained by the fact that the adult specimen has many more developed
frills, proportionally thicker as well as wider than those on the juvenile shells, and thus needed
more floating capacity. Furthermore, the possible negative buoyancy on the juvenile shell can also
be explained by an hypothesis that the juvenile Torquatoceras was benthic.

If the number of frills and of septa are the same in Torquatoceras , the secretion of these items
must have been parallel. Whether or not secretion was simultaneous is impossible to decide on
shell studies alone. Most probably, however, the apertural shell growth, and thus the frills,
continued all the time, and were not influenced by the repeated movements of the soft parts when
the mantle reorganized for secretion of a new septum.

FUNCTION OF FRILLS

The advantage of the frills, decorating Torquatoceras , is difficult to understand, and only
speculations can be made about their function. The presence of this ornamentation must add extra
weight to the shell, although the frills, as well as the shell, are rather thin. This conclusion was
reached after comparisons with other cephalopods of the same size from the same locality, but
unfortunately recrystallization precludes an exact comparison. On average cephalopods with no
frills had 50-60 % thicker shells. Since the shell was thinner in Torquatoceras than in other
comparable shells, the function of the frills might have been to reinforce the shell. For cephalo-
pods, at least those swimming forms, it is a question of priorities if the shell must be strong but
heavy or light but fragile. The combination light and strong is difficult to accomplish with the

STRIDSBERG: SILURIAN CEPHALOPOD WITH FRILLED SHELL

657

text-fig. 7. a, b, drawings of the apertural opening in Ihe two specimens
of Torquatoceras auritum, illustrating the vertical walls in the body-chamber.
See also Plate 64, figs. 2 and 5. c, d, illustrate the size difference between
the two morphs of T. undulatum. See also Plate 63, figs. 2 and 10.

usual mode of shell construction, as an increase in thickness of the shell to achieve strength,
necessarily causes an increase in weight. However, a thin shell supported by an external framework,
consisting of crenulated frills, can achieve considerably better strength than an ordinary shell of
the same size and the same weight, especially as the frills are always perpendicular to the shell
surface. Furthermore, the crenulations of the shell between the frills contribute further to a strong
and light construction. In the diagnosis of C. corbulatum, another species with a frilled shell. Horny
(1965, p. 136) stated that the shell wall was rather thin.

Other qualities in Torquatoceras , favouring the hypothesis of a thin but reinforced shell, are the
reinforced apertural edge and the vertical septa in T. auritum (text-fig. 7 and PI. 64). The reinforced
apertural edge is by no means a peculiarity of Torquatoceras , as this is common in almost all
oncocerid cephalopods from Gotland (Stridsberg 1985r/). Presumably this is an advantage for most
brevicones as they very often have a complex apertural opening with lobes and sinuses. Vertical
septa, however, have thus far only been observed in T. auritum. As discussed in the taxonomic
part herein, our knowledge is very limited about the outlines of these septa, but naturally any kind
of extra shell construction must support the total strength of the shell. Possibly, the vertical septa
not only strengthened the shell, but also assisted in protecting the soft parts from external danger.
The lateral view shows clearly that the outer parts of the vertical septa protrude outside the
apertural opening (PI. 64, fig. 3).

If the hypothesis that the crenulated frills served to strengthen the shell is correct, then there
must be a reason for a reinforced shell. Above all, the function of the shell was to protect the
inhabitant from any kind of danger. Mostly this danger comes from predators, but occasionally
from the energy of the environment. Regarding Torquatoceras , the only known predators capable
of destroying such a strong shell, were the eurypterids. These animals are well known from Gotland
and in the Hogklint Beds, lower Sheinwoodian, where the occurrence of eurypterids resulted in
the Pterygotus beds.

The possibility that a reinforced shell developed to withstand physical damage in a turbulent
reef environment, might explain the presence of the frills. One must keep in mind that the Palaeozoic
reef fauna probably consisted considerably more of shell carrying organisms than is the case today.
Not only the cephalopods but also all the fishes and many gastropods have abandoned external
skeletons or shells, and at least the two first groups rely on speed in favour of armour, while some
gastropods are bad tasting or poisonous. The use of a reinforced shell like that of Torquatoceras ,
might have been an attempt to survive in a tough environment.

SYSTEMATIC PALAEONTOLOGY

Order oncocerida Flower, 1950
Family trimeroceratidae Hyatt, 1900
Genus Torquatoceras gen. nov.

Derivation of name. Latin torquatus , adorned with a collar.

658

PALAEONTOLOGY, VOLUME 31

Type species. Torquatoceras undulatum sp. nov.

Diagnosis. Circular, exogastric (convex ventral side) brevicone. Contracted aperture with two
ventrolateral lobes on each side of the hyponomic sinus. Shell surface covered with transverse
undulating frills, which on the ventral side are V-shaped due to the growth of the hyponomic sinus.
Slender empty siphuncle.

Discussion. The genus Torquatoceras is placed in the family Trimeroceratidae due to the outline
of the aperture and the slender empty siphuncle. Furthermore, Torquatoceras appears to be closely
related to Pentameroceras facula, another species of the Trimeroceratidae.

Species. T. undulatum sp. nov. and T. auritum sp. nov.

Torquatoceras undulatum sp. nov.

Plate 63, figs. 1-10; text-figs. 6 and 7c, d
Derivation of name. Latin undulatus, undulating, referring to the undulating frills.

Holotype. RM Mo 56365.

Type stratum. Slite Beds, unit g, upper Sheinwoodian.

Type locality. Samsugns 1, Gotland, Sweden.

Material. Eleven macroconchs and forty to fifty more or less well-preserved microconchs from Gotland,
Sweden. Macroconchs RM Mo 56274-56275, 56842-56845, 57305, 57307, 155957 155960. Microconchs RM
Mo 56353, 56356, 56358, 56365-56366, 56368-56369, 56373-56374, 56377, 56379, 56383, 56392, 57244-57249,
152776, 157719 and twenty to thirty less well-preserved specimens, all at Naturhistoriska Riksmuseet,
Stockholm, Sweden. All well-preserved specimens are mature.

Diagnosis. A species of Torquatoceras with a circular exogastric brevicone, having crenulate
transverse frills. In mature specimens a contracted aperture with two ventrolateral lobes.

Description. Slightly exogastric, circular phragmocone with a straight body-chamber. The shell surface consists
of crenulate transverse frills, 1 -5-2-5 mm wide, perpendicular to the shell wall. The crenulation on the frills
continues on the shell surface between the frills. In some of the bigger specimens the crenulation shows a
specific rhythm (text-fig. 1), which is repeated on all frills. The distance between the frills increases during
growth to reach a maximum at about mid body-chamber. Close to the aperture the distance between the
frills decreases and the last frills can only be observed on the ventrolateral lobes. On the ventral side of the
phragmocone the frills are V-shaped, due to the position of the hyponomic sinus during the shell growth.

Because of a distinct size dimorphism the distance between the frills varies considerably. As an average
the maximum distance on macroconchs is 5 mm and on microconchs 2 mm. The slender empty siphuncle,
located close to the ventral shell wall, has an average thickness of about 2 mm in macroconchs and about
1 mm m microconchs (PI. 63, fig. 3). The ratio of the length and width of the body-chamber is 3 to 2.
Inside the apertural rim the shell is reinforced (text-fig. 4) and this reinforcement is further developed by
the crowding of the last frills around the apertural opening. Altogether this shell growth forms a ridge along
the edge of the aperture.

Dimorphism. The material is divided into two very distinct size groups, the macroconchs being about twice
the length and width of the microconchs, and this size dimorphism is interpreted as sexual dimorphism (text-
fig. 7c, d; PI. 63, figs. 2 and 10). Dimorphism is fairly common among Silurian oncocerid cephalopods

EXPLANATION OF PLATE 63

Figs. 1 10. Torquatoceras undulatum sp. nov. Slite beds, unit g. Samsugns 1. 1 and 2, ventral and apertural
views of RM Mo 56842 with worn down frills, x 1-5. 3, an enlargement of the siphuncle of the holotype
RM Mo 56365, x 6. 4-7, dorsal, lateral, ventral, and apertural views of the holotype RM Mo 56365,
x 1-5. 8 10, dorsal, lateral, and apertural views of RM Mo 152776, x 1-5.

PLATE 63

STRIDSBERG, Torquatoceras

660

PALAEONTOLOGY, VOLUME 31

(Stridsberg 1985a, b). The apertural shape in combination with the shell morphology in the macroconchs as
well as in the microconchs, indicates a specific mode of life, and it is most unlikely that two species with the
same mode of life lived in the same place. Although after the death of the animals the empty shells are known
to drift around in the ocean, and perhaps become concentrated in some places, it must be pointed out that
all specimens of T. undulation are found in one locality only, macroconchs as well as microconchs.

Discussion. The condition of the preserved material varies widely and the most complete specimens
are the microconchs. Some of these have a well-preserved apertural region and an almost complete
phragmocone, although the apical end is always broken off. The crenulations, however, are best
observed on some of the macroconchs, and here the rhythmic pattern can easily be followed from
frill to frill (text-fig. 1a, b). These patterns are not so well pronounced on the microconchs but the
state of preservation of these does not permit any accurate measurements.

A similar rhythmic pattern is also found in C. corbulatum (text-fig. 2) but in this species the
frills regularly protrude. On worn-down specimens, however, the pattern is very similar to that of
Torquatoceras.

Comparison. The outline of T. undulation is very close to that of T. auritum, and the only known
difference between the two species is the presence of two ventrolateral vertical septa, partly enclosing
the hyponomic sinus, in T. attrition. The size of the two species is almost the same, as is the
configuration of the frills. Due to the very limited material of T. attrition, only two incomplete
specimens, further comparison is not possible.

In P.facula Stridsberg 1985a the shell surface has transverse surficial annulations, although they
are not crenulated as in Torquatoceras. Furthermore, the annulations of P. facula do not seem to
be constructed in the same way as in T. undulatum, in which the frills were secreted during an
extraordinary position of the mantle. The apertural constrictions on P. facula and T. undulation
are totally different in regard to the lobes and sinuses, although both species have a distinct
hyponomic sinus and very pronounced ventrolateral lobes. Probably the apertural shape in the
two species is the result of convergent evolution. The size of P. facula is about the same as the
microconchs of T. undulation.

Torquatoceras auritum sp. nov.

Plate 64, figs. 1-9; text-fig. 7a, b

Derivation of name. Latin auritum, referring to something with ears.

Holotype. RM Mo 56284.

Type stratum. Slite beds, unit g, upper Sheinwoodian.

Type locality. Samsugns 1, Gotland, Sweden.

Material. Two specimens from Gotland; RM Mo 56277 and RM Mo 56284 in the Naturhistoriska Riksmuseet,
Stockholm, Sweden. Both specimens are mature.

Diagnosis. A species of Torquatoceras with a circular, probably exogastric breviconic phragmocone,
having crenulate transverse frills. In mature specimens a contracted aperture with two ventrolateral
lobes, each with a vertical septum partly enclosing the hyponomic sinus.

EXPLANATION OF PLATE 64

Figs. 1-9. Torquatoceras auritum sp. nov. Slite beds, unit g, Samsugns 1. 1-3, ventral, apertural, and lateral
views of RM Mo 56277. In the lateral view (3) the protrusion of the vertical septa outside the apertural
opening is shown, x 1-5. 4, 5, 7, 8, dorsal, apertural, ventral, and lateral views of the holotype RM Mo
56284 with worn down frills, x I -5. 6, enlarged detail of the apertural area of specimen RM Mo 56277,
x4-5. 9, enlarged detail of the apertural area of the specimen RM Mo 56284, x4-5.

PLATE 64

STRIDSBERG, Torquatoceras

662

PALAEONTOLOGY, VOLUME 31

Description. Circular straight body-chamber on a probably exogastric phragmocone. The shell surface consists
of crenulate transverse frills, which reach about T5 mm perpendicularly out from the shell surface. Between
the frills the crenulations can be followed on the shell surface as longitudinal furrows (PI. 64, figs. 4 and 7).
The distance between the frills on the body-chamber (the only preserved part of the shells) varies between
2-5-3-5 mm. In the apertural ridge, however, the frills are piled up on each other, forming an edge about 2 0
mm wide. A few of the frills in the apertural ridge are exposed on the ventrolateral lobes, situated on both
sides of the hyponomic sinus (PI. 64, fig. 2).

From the ventrolateral lobes, two vertical septa protrude almost 5 mm towards the centre of the apertural
opening. The pronounced shape of the hyponomic sinus can be followed on all older frills by a V-shaped
bend towards the apical end of the shell (PI. 64, fig. 7). The empty slender siphuncle is almost 2 mm wide in
the last chambers. The ratio of the length and width of the body-chamber is estimated to be 3 to 2.

Discussion. As the only known specimens of T. auritum are both incomplete, knowledge of the
phragmocone is very limited. However, the shape of the body-chamber of the holotype indicates
an exogastric curvature, as is also the case with T. undulatum. The two vertical septa, one from
each ventrolateral lobe, are well developed in the apertural area and are firmly connected to the
apertural rim. Since the specimens are recrystallized, as is the sediment in the body-chamber, it is
not possible to document the extensions of the vertical septa along the inside of the body-chamber.
Sections made inside the body-chamber in specimen RM Mo 56277 (PI. 64, figs. 13), show no
details at all of the continuation of the vertical septa. Due to the recrystalization it is not known
if the two protrusions really are the outer part of two septa or not, and naturally they might as
well have been two spines, secreted during the build-up of the apertural rim. Anyhow, the exposed
remains of the construction favour the interpretation of two minor septa, perhaps ending a few
millimetres behind the apertural edge.

The function of the vertical septa is hard to understand, primarily because we have incomplete
knowledge of their shape, but presumably they supported the hyponome in one way or another.
The area left for the hyponome, restricted by the apertural rim and the two vertical septa, would
still allow a fairly flexible hyponome. As the only likely means of navigation was to alter the
direction of the hyponome this was essential for any kind of swimming. Naturally any distension
of soft parts must influence the swimming direction but as this would produce a notable drag for
the animal, this method is unlikely.

Comparison. T. auritum is in many ways identical with T. undulatum , and actually the two vertical
septa are the only distinguishing characteristics of T. auritum. The size of the macroconchs of T.
undulatum is the same as the conchs of T. auritum. Due to these facts it could be questioned
whether the two morphs are one species or not, and in case they were the same species, the vertical
septa could be some kind of sexual dimorphism. However, in T. undulatum there is a most distinct
size dimorphism and there are strong reasons to believe that a specific type of aperture would
favour a specific mode of life. It must be assumed that the identity of the apertural shape in the
two size groups of T. undulatum is a better argument for sexual dimorphism, than the external
similarities with T. auritum. The apertural shape of the latter might possess other qualities and
thus another mode of life. The apertural rim is notably thicker on T. auritum than on T. undulatum ,
but due to the limited material of T. auritum the thick ridge might as well be the result of better
preserved specimens of this species.

Acknowledgements. For valuable comments and improvements of the manuscript I would express my thanks
to Charles H. Holland, Dublin, and Lennart Jeppsson, Lund. Thanks also go to Christin Andreasson for
great help in preparing the drawings and Vojtech Turek, Prague, for assistance with the Bohemian material.
Grants from Statens Naturvetenskapliga Forskningsrad (NFR) made it possible for me to study collections
in Stockholm and Prague.

STRIDSBERG: SILURIAN CEPHALOPOD WITH FRILLED SHELL

663

REFERENCES

angelin, n. p. and lindstrom, G. 1890. Fragmenta silurica e dono Caroli Henrici Wegelin , 60 pp. Stockholm.
barrande, j. 1866. Systeme silwien du centre de la Boheme 2, Cephalopodes, pis. 108-244. Prague, Paris.

— 1867. Systeme silwien du centre de la Boheme 2, Cephalopodes , texte, pt. I. 712 pp. Prague, Paris.
foerste, a. f. 1916. Notes on Richmond and related fossils. J. Cincinn. Soc. nat. Hist. 22, 42- 55, 3 pis.
hamada, T. 1964. Notes on the drifted Nautilus in Thailand. Contributions to the Geology and Palaeontology

of Southeast Asia, XXIV. Scient. Pap. Coll. gen. Educ. Tokyo. 14, 255-278, pis. 1 5.
hedstrom, H. 1917. Uber die Gattung Phragmoceras in der Obersilurformation Gotlands. Sver. geol. Unders.
Afh. 15, 35 pp., 27 pis.

hyatt, a. 1884. Genera of fossil cephalopods. Proc. Boston Soc. nat. Hist. 22, 253-338.
horny, R. 1965. Corbuloceras gen. n., novy onkoceridni hlavonozec (Cephalopoda, Oncocerida) z ceskeho
siluru. Cas. ndrod. Muz. 134, 132-137, tab. 1-2.

kahn, p. G. k. and pompea, s. m. 1978. Nautiloid growth rhythms and dynamical evolution of the Earth
Moon system. Nature , Land. 275, 606-61 1.

laufeld, s. 1974. Reference localities for palaeontology and geology in the Silurian of Gotland. Sver. geol.
Unders. Afh. 705, 172 pp.

lindstrom, G. 1890. The Ascoceratidae and the Lituitidae of the upper Silurian formation of Gotland. Kungl.
svenska VetenskAkad. Hand!. 23-12, 54 pp., 7 pis.

mutvei, H. 1964. On the shells of Nautilus and Spirula with notes on the shell secretion in non-cephalopod
molluscs. Ark. Zool. 16, 221-278, 22 pis.

saunders, w. b. and spinosa, c. 1979. Nautilus movement and distribution in Palau, Western Caroline
Islands. Science, NY, 204, 1199 1201.

and ward, p. d. 1979. Nautiloid growth and lunar dynamics. Lethaia, 12, 172.

stridsberg, s. 1985a. Silurian oncocerid cephalopods from Gotland. Fossils and strata, 18, 65 pp.

— 19856. Functional morphology of Silurian oncocerids from Gotland. Lund Pubis Geol. 34, 24 pp.
toriyama, r., sato, t., hamada, t. and komalarjun, p. 1965. Nautilus pompilius drifts on the West Coast

of Thailand. Contributions to the Geology and Palaeontology of Southeast Asia, XX. Jap. J. Geol. Geogr.
36, 149-161, pi. 3.

Typescript received II May 1987
Revised typescript 10 August 1987

SVEN STRIDSBERG

Institute for Historical Geology and Palaeontology
University of Lund
Solvegatan 13

S-223 63 Lund
Sweden

PALAEOCORYNID-TYPE APPENDAGES IN
UPPER PALAEOZOIC FENESTELLID BRYOZOA

by ADRIAN J. BANCROFT

Abstract. Palaeocorynid-type structures (Family Palaeocorynidae Duncan and Jenkins 1869), currently
regarded as being of uncertain zoological affinities, are here interpreted as being a specialized form-appendage
of Upper Palaeozoic fenestellid Bryozoa. Palaeocorynid-type appendages are morphologically complex, and
consist of a short stem developed at right angles from the branch of the bryozoan, terminating in a cone-
shaped body from whose lateral margins a variable number of long slender spines or branchlets emanate at
high angles. Spines form simple, distally tapering structures; branchlets are much longer and repeatedly
bifurcate, converge and fuse to develop an anastomosing reticulate meshwork. The external ornament and
internal microstructure of these structures is identical and continuous with that of the branch of the bryozoan
on which they occur. Up to five developments have been found in situ on a colony, occurring anywhere over
the colony surface, and nearly all are developed from the obverse surface of branches. They are interpreted
as having a defensive function, giving a protective covering to feeding autozooecial polypides beneath by
providing a surface deterrent to predatory organisms.

Calcified appendages are commonly developed on Fenestella s.l. and other fenestrate bryozoan
genera from Upper Palaeozoic strata. They generally form slender, distally tapering, cylindrical,
unbarbed, or barbed stem-like structures up to several centimetres in length, and can diverge from
the lateral margins, obverse or reverse surface of branches in a colony (text-fig. 1a, b). They are
particularly abundant in the proximal parts of colonies and are interpreted as supporting struts
that acted in association with the heavily calcified holdfast (King 1850; Young and Young 1874;
Vine 1879c/, b\ Cumings 1906; Ferguson 1963; Tavener-Smith 1969).

During ongoing revision of British Carboniferous fenestrate Bryozoa, large numbers of another
type of structure occurring on Fenestella s.l. have been examined. Morphologically they consist of
a short cylindrical stem, attached to the underside of a cone-shaped body from whose lateral
margins a number of long slender spines or branchlets project. The base of the stem is directly
connected at right angles to the branch of the bryozoan and they nearly always occur on the
obverse surface of branches, being developed anywhere over the colony surface.

Although these curious and morphologically complex structures have been the subject of several
detailed studies, their zoological affinities and functional significance have remained somewhat
enigmatic. They were first described by Duncan and Jenkins (1869), who suggested that they were
hollow and represented the trophosomes of a hydroid that attached itself to Fenestella. Duncan
and Jenkins erected the genus Palaeocoryne with two species, within the new family Palaeocorynidae,
which they classified within the Order Tubulariidae. In a subsequent paper, Duncan (1873)
reiterated the zoological affinities of the Palaeocorynidae. Allman (1872) refuted Duncan and
Jenkins’s interpretation, and suggested that the group had foraminiferal affinities. Young and
Young (1874) stated that the Palaeocorynidae were merely outgrowths of a bryozoan colony,
and were solid structures directly connected to the skeletal tissues of the branch on which they
occur. Vine (1879//, b ) agreed with Young and Young’s observations and suggested that these
structures had a combined supportive and reproductive function. Barnes (1903) described the body
and spines of two specimens of Palaeocorynidae, and assigned them to the phylum Polyzoa under
the genus Evactinopora Meek and Worthen. Elias and Condra (1957) discarded evidence suggested
by G. F. Papenfuss of a relationship between Palaeocoryne and the living red alga Asparagopsis
armata, and regarded the structures as appendages of Fenestella. Ferguson (1961) erected the

| Palaeontology, Vol. 31, Part 3, 1988, pp. 665-675, pi. 65.|

© The Palaeontological Association

666

PALAEONTOLOGY, VOLUME 31

text-fig. 1 . a, b, morphology of stem-like appendages diverging from branches, a, BOM 25-09-238, Fenestella
plebeia M‘Coy (Visean), Halkyn, Clwyd, x 7 0. b. BM(NH) PD. 7794, F. bicellulata Etheridge Jun., Fifth
Limestone (Asbian), Alston Group, Penruddock, near Penrith, Cumbria; showing occurrence of barbs on
stems, x 16 0. c-F, morphology of in situ palaeocorynid appendages, c, BH(NH) PD. 7795, single spinose
development on F. multispinosa Ulrich, shales in Upper Fell Top Limestone (Pendleian), Haltwhistle,
Northumberland, x 2-9. D, BM(NH) PD. 2371, four spinose developments on F. multispinosa colony. Car-
boniferous (Visean), Halkyn, Clwyd, x2T. e, detail of one development shown in d, x8-6. f, BM(NH)
PD. 2609, reticulate development on F. plebeia , Carboniferous, locality and horizon unknown, x2 l.

BANCROFT: PALAEOCORYNID-TYPE APPENDAGES IN BRYOZOA

667

new palaeocorynid genus Claviradix , and concluded that the Palaeocorynidae were separate
organisms from the host bryozoan, using it only as support, and in a following paper (1963) he
stated that they probably had bryozoan affinities.

Since Ferguson (1963), no systematic studies have been undertaken on the Palaeocorynidae. The
discovery of abundant, well-preserved fragmented and in situ material, including significantly larger
and more complex developments than hitherto recognized has prompted the present work. This
study incorporates a detailed re-examination of the morphology of palaeocorynid-type structures
and reassessment of their zoological affinities and functional significance. External and internal
details of morphology have been examined under the SEM. Cited material is located in the
collections of Bolton Museum (abbreviated BOM) and the British Museum (Natural History),
London (abbreviated BM(NH)).

MORPHOLOGY OF PALAEOCORYNID-TYPE STRUCTURES

External

Palaeocorynid-type structures almost exclusively occur on the obverse surface of branches, and
may be developed anywhere over the colony surface. Their occurrence has been documented on
the reverse surface of branches (Ferguson 1963, p. 156), and one example was found on the reverse
side of a colony of F. frutex M‘Coy during the course of the present study. Palaeocorynid-type
developments have been found in situ on the obverse surface of the following taxa: F. plebeia
M'Coy and F. multispinosa Ulrich. F. K. McKinney (pers. comm.) has reported their occurrence
in the fenestellid genus Archimedes Hall.

Stems range between 0-50 mm and 1-80 mm in length, and may be barrel-shaped, expand distally,
or be of uniform diameter (PI. 65, figs. 1, 2, 4). They generally arise at right angles from branches,
and their external ornament is continuous with that of the branch on which they occur (PI. 65,
figs. 3 and 4). The disposition, shape, and size of autozooecial apertures is usually not affected by
the development of palaeocorynid-type structures (PI. 65, fig. 4), except where buttress-like features
are developed at their bases when apertural shape may be distorted (PI. 65, fig. 1). These buttress-
like structures were interpreted as root-like processes by Ferguson (1961), who established the
palaeocorynid genus Claviradix on the basis of their occurrence, the taxon being distinguished
from Palaeocoryne which apparently does not possess them. Stems are longitudinally striate, with
a single row of closely spaced, small, pustules situated on ridges (PI. 65, fig. 4).

In all the described species of Claviradix and Palaeocoryne , with one exception, stems are single
cylindrical structures. The form C. bifurcate/ Ferguson (1961) is apparently unique in that the stem
bifurcates. However, only one incomplete fragment of this taxon is known, of which only the bifid
stem is preserved, and the recognition of this form as a palaeocorynid-type of development cannot
be qualified.

The body of palaeocorynid-type developments varies significantly in shape and size, from small
box-like structures, 0-20 mm in diameter, to large high-angle cones, 0-60 mm in diameter (PI. 65,
figs. 5-8). The centre of the bodies upper surface is most commonly depressed or flat, but is
occasionally slightly elevated into a dome-like structure and may rarely be developed into a
prominent spine up to 0-40 mm in length (PI. 65, figs. 8-11). The external ornament of the body
is continuous with that developed on the stem, with striae being radially arranged (PI. 65, figs. 4,
7, 1 1).

Spines are regularly developed and geometrically arranged around the lateral margins of the
body, and display considerable variation in their number, shape, and size. Between four and fifteen
spines may be developed around the body, and they most commonly project slightly upwards away
from it (PI. 65, figs. 2, 4-6). Spines generally form long, straight, cylindrical, distally tapering
structures and are longitudinally striate, their ornamentation being continuous with that of the
stem and body (PI. 65, figs. 4, 6, 7, 11). In all the described species of Palaeocoryne and Claviradix ,
spines are equally developed around the body (PI. 65, figs. 6-9, 1 1 ). However, in several specimens
recently discovered one spine is significantly more robust and appears to have been longer than

668

PALAEONTOLOGY, VOLUME 31

any of the others (PI. 65, figs. 12 and 13). Spines range from 010 mm to 0-20 mm in diameter
(measured at their proximal extremities), and the largest spine examined in the present study was
6 0 mm in length (an incomplete example) (text-fig. Id, e).

Considerable morphological variation exists in the spinose developments occurring on F.
multispinosa (incorrectly identified as F. nodulosa (Phillips) by Ferguson 1963); the number of
spines ranges between seven and ten, and the open cone-shaped body ranges between 0-25 mm
and 0-40 mm in diameter. While only one palaeocorynid development is usually found preserved
on colonies of F. multispinosa examined, up to four may be present (text-fig. lc, D). In one
colony where four do occur, some of the spines from individual structures converge and overlap
(text-fig. Id).

Two species of Claviradix described by Ferguson (1963) are unusual in that each of the four
spines developed from the body bifurcate, once in the case of C. ashfellensis and twice in C.
cruciformis. Flowever, several recently discovered colonies of F. plebeia IVTCoy exhibit significantly
larger and more complex developments of C. cruciformis than hitherto described. It is apparent
that Ferguson (1963) had only examined incomplete specimens of this particular growth form
developed on F. plebeia (incorrectly identified by Ferguson as Parafenestella formosa (M‘Coy)). In
these larger developments, individual spines, more appropriately termed branchlets, repeatedly
bifurcate at high angles. Individual branchlets also converge and fuse, so that an anastomosing
reticulate meshwork is developed around the body (text-figs. If, 2a-f, 4a). The largest recorded
single development is 40 mm in diameter (measured on an incomplete structure). Extremely thin
lateral offsets commonly diverge from branchlets at right angles and are of variable morphology.
They may be straight bars that extend fully across the gap between adjacent branchlets or else
form short barb-like structures projecting laterally into the gap (text-fig. 2c, e). Branchlets appear
to taper distally and range between 015 mm and 0-27 mm in diameter (measured away from
points of bifurcation and convergence), and are longitudinally striate with an oval cross-section
(text-fig. 2f).

In one colony of F. plebeia five such developments are preserved in situ , and they overlap and
appear to fuse irregularly together. The ‘superstructure’ is only partially preserved and is somewhat
covered by matrix, but it possibly covered the entire obverse surface of the colony fragment
(measuring 70 mm x 45 mm), and was developed parallel to it (text-fig. 4a).

Internal

SEM investigations undertaken on the internal ultrastructure of palaeocorynid-type developments
have shown that they are structurally continuous with the branch of the bryozoan on which they
occur, as originally suggested by Young and Young (1874). The granular primary skeleton
surrounding autozooecial chambers on branches also forms the core of the stem, body, spines,
and branchlets of palaeocorynid structures (text-figs. 3a-f and 4b-d). This observation contrasts
with those made by Elias and Condra (1957) and Ferguson (1963), who concluded that the granular
(axial) core of the stem did not join that of the branch but terminated at the base of the stem. The
granular primary skeleton in the stem, body, spines, and branchlets is surrounded by laminated
secondary skeleton continuous with that surrounding the granular primary skeleton on branches

EXPLANATION OF PLATE 65

Figs. 1 13. Morphology of palaeocorynid appendages. Material from shales above the Main Limestone
(Namurian, Pendleian), Hurst, North Yorkshire Moors. 1, BM(NH) PD. 7796, x 30. 2, BM(NH) PD. 7797,
x 24. 3, BM(NH) PD.7798, x 24. 4, BM(NH) PD.7799, x 30. 5, BM(NH) PD.7800, x 21. 6, BM(NH)
PD. 7802, x 24. 7, BM(NH) PD.7802, x 18. 8, BM(NH) PD. 7803, x 24. 9, BM(NH) PD. 7804, x 42.
10, BM(NH) PD. 7805, x 24. 11, BM(NH) PD. 7806, x48. 12, BM(NH) PD. 7807, x 18. 13, BM(NH)
PD. 7808, x21. All are SEM photographs.

PLATE 65

BANCROFT, palaeocorynid Bryozoa

670

PALAEONTOLOGY, VOLUME 31

text-fig. 2. a-f, morphology of reticulate palaeocorynid appendages on Fenestella plebeia M'Coy. a, BM(NH)
PD. 7809, Hardrow shales (Visean, Brigantian), Middle Limestone Group, Mill Gill, Askrigg, North Yorkshire;
body of structure and proximal extremities of diverging branchlets with initial bifurcations, x 7. B, c, BM(NH)
PD. 7810, Carboniferous, Cambeck, locality and horizon unknown. B, x 9; c, detail of branchlets, x 19. d-
f, BM(NH) PD. 2609, Carboniferous, locality and horizon unknown, d, proximal portion of structure, x4-3.
E, curved barb-like structures developed from lateral margins of branchlets, x 42. F, striated ornamentation

of branchlets, x 32. All are SEM photographs.

below (text-figs. 3a d and 4b-d). Although Ferguson (1963) also observed this fact he suggested
that the laminated skeleton of Palaeocoryne developed after that of the branch of Fenestella on
which the structure occurs.

The granular core divides in the body of palaeocorynid structures, and the resultant cores

BANCROFT: PALAEOCORYNID-TYPE APPENDAGES IN BRYOZOA 671

text-fig. 3. a-f, ultrastructure of palaeocorynid appendages; a, b, e, f of Fenestella multispinosa Ulrich and
other spinose developments and c-d of reticulate developments in F. plebeia M'Coy. Material from shales
above the Main Limestone (Namurian, Pendleian), Hurst, North Yorkshire Moors. A, BM(NH) PD. 7811,
continuation of laminated secondary skeleton around autozooecial chamber and proximal portion of
development (right), transverse section through normal branch to left, x90. b, BM(NH) PD.7812, transverse
section through body showing several granular cores diverging away from its centre, x 180. c, BM(NH)
PD. 781 3, slight oblique section of linear granular skeleton with lateral offsets in branchlet, x420. d, BM(NH)
PD. 78 14, transverse section through stem showing linear arrangement of granular skeleton and zone of
poorly defined laminated skeleton between ridges, x 300. e, BM(NH) PD.7815, transverse section through
stem showing stellate granular core and radiating stellate arrangement of surrounding laminated skeleton,
x 360. f, BM(NH) PD. 7818, oblique section through spine, showing granular core (bottom left), granular

ridge, and stylets in laminated skeleton, x 420.

developed form the axial cores of spines and branchlets emanating from the lateral extremities of
the body (text-figs. 3b and 4b). The morphology of the granular skeleton in the spinose developments
of F. multispinosa and fragmented specimens of other spinose developments of unknown provenance,
is significantly different from that of reticulate developments on F. plebeia. The morphology of the
granular skeleton in transverse section in the stems and branchlets of F. plebeia has a linear
structure and it possesses lateral offsets that may bifurcate (text-figs. 3c, d and 4c). The main axis
of the granular skeleton in branchlets lies parallel to the plane of the development of the branchlets
(text-fig. 4c). In F. multispinosa , and fragments of other spinose developments, the granular core
in stems and spines has a stellate appearance in transverse section (text-figs. 3e and 4d), and is
identical in most respects to the morphology of the granular skeleton in stem-like appendages and
dissepiments that interconnect branches in fenestellid colonies.

672

PALAEONTOLOGY, VOLUME 31

Hi

text-fig. 4. a, BM(NH) PD. 7819, Fenestella
plebeia M‘Coy, Hardrow shales (Visean, Brigan-
tian), Mill Gill, Askrigg, North Yorkshire; show-
ing the occurrence of five, possibly six, in situ
reticulate palaeocorynid developments on one
colony. Thick solid lines indicate location of
branchlets, dashed lines indicate areas where
obverse surface of colony is visible, and orien-
tation of branches. Scale bar for size, b-d,
ultrastructure of palaeocorynid appendages, b,
longitudinal section through palaeocorynid de-
velopment and branch of fenestellid showing
arrangement of various skeletal elements, x 36.
c, d, transverse sections through stems of palaeo-
corynid developments, c, linear arrangement of
granular skeleton in F. plebeia M'Coy, x 78; D,
stellate arrangement of granular skeleton in F.
multispinosa Ulrich, x 78.

The junction between the granular primary skeleton and the laminated secondary skeleton is
well defined (text-fig. 3d). The morphology of the laminated secondary skeleton is variable, with
no difference in morphology occurring between spinose and reticulate developments. Laminae
within the inner portion of this unit are often poorly defined and pass gradationally into an outer
region where they become well defined (text-figs. 3d and 4c, d). The poorly laminated inner zone
probably represents the additional granular layer recognized by Ferguson (1963) between the
central granular core and the outer laminated skeleton. Ferguson (1963) distinguished the skeletal
structure of Palaeocoryne and Claviradix from Fenestella on the presence of this additional granular
layer, and used this feature to support his suggestion that palaeocorynid-type structures were
separate organisms.

The laminated secondary skeleton is typically arranged in orally flexed ridges around the ridges
of the granular skeleton, with additional ridges commonly developed in between, and has a well-
defined radiating stellate appearance in transverse section (text-figs. 3d, e and 4c, d). Close to the
outer surface, laminae forming the ridges are commonly arranged in closely spaced orally flexed
nests (termed stylets), forming papillae or small pustules on the outer surface (text-fig. 3f; PI. 65,
fig. 11).

The granular core of spines and branchlets continues along their length in all the material
examined, and has an identical stellate or linear appearance to that developed on stems, with the

BANCROFT: PALAEOCORYNID-TYPE APPENDAGES IN BRYOZOA

673

granular core and ridges appearing to thin distally (text-fig. 4b). No bifurcations have been observed
in the lateral offsets of the granular skeleton in branchlets (text-fig. 3c).

ZOOLOGICAL AFFINITIES

The preceding morphological assessment of palaeocorynid-type structures unequivocally proves
that they are not a distinct group of organisms which were parasitic on fenestellid bryozoans, as
suggested by some previous workers (Duncan and Jenkins 1869; Duncan 1873; Ferguson 1961,
1963), nor are they of algal origin (Elias and Condra 1957). The fact that their external ornament
and internal microstructure is continuous with that of the bryozoan on which they occur proves
that palaeocorynid-type developments are merely a form of appendage. In accordance with this
conclusion, the generic and specific names applied to individual morphotypes by Duncan and
Jenkins (1869), Duncan (1873), and Ferguson (1961, 1963) should perhaps best be regarded as
invalid.

Skeletal secretion in fenestrate bryozoans is inferred to have been undertaken by an external
epithelial tissue common to the whole colony, comparable to that of some living Bryozoa (Elias
and Condra 1957; Tavener-Smith 1969; Gautier 1973). Accordingly, palaeocorynid-type appendages
must have been secreted by an epithelium continuous with that covering the rest of the colony.

FUNCTIONAL SIGNIFICANCE

The diverse and complex morphology of palaeocorynid-type appendages, coupled with the fact
that they may occur anywhere over the colony surface, suggests that they did not have a supportive
function, akin to that interpreted for unbarbed or barbed long stem-like appendages commonly
present in the proximal parts of fenestellid colonies.

The discovery of large, anastomosing reticulate meshworks on F. plebeici is particularly interesting,
and is reminiscent of superstructures developed above the obverse surface of colonies in certain
other fenestellid genera, such as Cyclopelta Bornemann, Unitrypa Hall, and Hemitrypa Phillips.
These three genera possess colony-wide superstructures that are developed as outgrowths of carinal
nodes or the median carina on the obverse surface of branches. Hemitrypa possesses the most
complex type of superstructure that is developed as geometrically arranged lateral bar-like
outgrowths of the crests of elongate carinal nodes, and forms an intricate interlocking, perforate,
hexagonal latticework situated at a uniform distance above the main reticulate meshwork below
(text-fig. 5a, b). The superstructure in Hemitrypa is interpreted to have acted as a protective screen

text-fig. 5. a, b, Hemitrypa hibernica IVLCoy. BM(NH) PD. 6642, High Glencar Limestone (Visean, Asbian),
Carrick Lough, County Fermanagh, Northern Ireland. Silicified colony fragment with the superstructure
in situ, a, with almost the entire superstructure intact, x 26. b, showing an area where the superstructure
is broken away revealing the obverse surface of the main meshwork below, x 26.

674

PALAEONTOLOGY, VOLUME 31

for feeding autozooecial polypides functioning between the branch surface and the superstructure,
by providing a surface deterrent to predatory organisms (Tavener-Smith 1973; Bancroft 1986).
Such a function may also be inferred for other fenestellid taxa (e.g. Cyclopelta) with different and
less intricate superstructures, in which the superstructure consists of a vertical extension of the
median carina that bifurcates into two lateral wedges at a uniform distance above the meshwork
(see McKinney and Kriz 1986).

The reticulate meshworks preserved on F. plebeia appear to have covered a relatively large area
of the colony surface, and in one large colony fragment where five such developments occur, they
may have completely covered it. These facts possibly suggest that these structures had a function
analogous to that inferred for the superstructure in Hemitrypa. The radiating spine-like structures
observed on F. multispinosa may also have had a comparable function. Although only one
development is usually found on colonies, up to four have been observed (text-fig. Id).

Although palaeocorynid-type appendages have only been found in situ on the obverse surface
of two fenestellid taxa, the variety of morphotypes found in fragmented specimens examined that
cannot be attributed to either F. multispinosa or F. plebeia suggests their occurrence in several
other taxa. The presence of a palaeocorynid development on the reverse surface of F. frutex
suggests that this taxon was capable of growing such appendages, but its occurrence on the reverse
surface of branches cannot be explained other than as a growth enigma in the light of the preceding
discussion.

The rare in situ occurrence of palaeocorynid-type appendages, and their apparent intracolonial
sparsity in taxa known to possess them, is possibly accounted for by their low preservation potential
as they are delicate structures. Abundant fragments of spines, bodies, and branchlets have been
found at several horizons in association with fenestellid bryozoans, with none being found in situ.
Their disposition is such that they would have readily broken away on the death of the colony
and its subsequent post-mortem fragmentation.

However, the occurrence of palaeocorynid-type appendages does appear to be spatially and
temporally intermittent, and at many horizons where fenestellids are abundant (including F.
multispinosa and F. plebeia ), no fragments of palaeocorynid developments have been found.
Laboratory experiments on the living cheilostome bryozoan Membranipora membranacea have
shown that colonies exposed to direct predation by slow feeding nudibranch molluscans have the
ability to grow protective chitinous and membranous spines around autozooecia to defend them
from attack (Harvell 1984). These spines grow rapidly, during the course of predation, and are
fully developed within a day or two. They serve to control effectively the pattern of predation,
reduce the extent of intracolonial mortality and to slow down significantly the rate of predation.
The development of palaeocorynid-type appendages in fenestellid bryozoans may also have been
predator-induced, their spatial and temporally intermittent occurrence reflecting that of possible
molluscan predators.

CONCLUSIONS

1. The external ornament and internal microstructure of palaeocorynid-type structures developed
on Fenestella s.l. is continuous with that of the branch of the bryozoan colony on which they
occur.

2. Palaeocorynid-type structures are almost exclusively developed on the obverse surface of
branches and may occur anywhere over the colony surface.

3. They are a specialized form of appendage, and possibly had a defensive function, in that the
extensive array of spines or branchlets developed laterally from the distal extremity of stems served
to give a protective covering to feeding autozooecial polypides beneath, by providing a surface
deterrent to predatory organisms.

Acknowledgements. I would like to thank Dr P. D. Taylor (British Museum (Natural History)) for the loan
of material and for help with SEM work. This study was undertaken during the tenure of a Department of
Education Post-doctoral Fellowship, at the Department of Geology, Trinity College, Dublin. I extend my

BANCROFT: PALAEOCORYNID-TYPE APPENDAGES IN BRYOZOA

675

gratitude to Professor C. H. Holland for allowing me use of the departments tehnical facilities, and to the
technical staff for photographic assistance. Text-fig. 5 is reproduced here by kind permission of the editors
of the Irish Journal of Earth Sciences.

REFERENCES

allman, G. J. 1872. A monograph of Gymnoblastic or Tubularian hydroids. Part II. R. Soc. Publ. 47, 155
450.

Bancroft, a. j. 1986. Revision of the Carboniferous fenestrate bryozoan Hemitrypa hibernica M‘Coy. Ir. J.
earth Sci. 7, 111-124.

barnes, j. 1903. On a fossil Polyzoa from the Mountain Limestone District, Castleton. Manchr geol. Min.
Soc. Trans. 28, 243-245.

cumings, e. r. 1906. Description of the Bryozoa of the Salem Limestone of Southern Indiana. Ann. Rept.
Dept. Geol. nat. Res., Indiana, 30, 1274-1296.

duncan, p. M. 1873. On the genus Palaeocoryne , Duncan and Jenkins, and its affinities. Q. Jl geol. Soc. Load.
29, 412-417.

and jenkins, h. m. 1869. On Palcieocoryne, a genus of Tubularine Hydrozoa from the Carboniferous
Formation. Phil. Trans. R. Soc. 159, 693-699.

elias, m. k. and condra, G. E. 1957. Fenestella from the Permian of West Texas. Mem. geol. Soc. Am. 70,
1-158.

ferguson, j. 1961. Claviradix, a new genus of the Family Palaeocorynidae from the Carboniferous rocks of
County Durham. Proc. Yorks, geol. Soc. 33, 135 148.

— 1963. British Carboniferous Palaeocorynidae. Trans, nat. Hist. Soc. Nortlmmb. 14, 141-162.

Gautier, t. 1973. Growth in bryozoans of the Order Fenestrata. In larwood, g. p. (ed. ). Living and Fossil

Bryozoa, 271-274. Academic Press, London.

harvell, c. d. 1984. Predator-induced defense in a marine Bryozoan. Science , NY, 224, 1357-1359.
king, w. 1850. A monograph of the Permian fossils of England. Palaeontogr. Soc. [Monogr.], 258 pp.
mckinney, F. k. and kriz, J. 1986. Lower Devonian Fenestrata (Bryozoa) of the Prague Basin, Barrandian
Area, Bohemia, Czechoslovakia. Fieldiana Geol. 15, 1-90.
tavener-smith, r. 1969. Skeletal ultrastructure and growth in the Fenestellidae. Palaeontology, 12, 281 309.

- 1973. Fenestrate Bryozoa from the Visean of County Fermanagh, Ireland. Bull. Br. Mus. nat. Hist.
(Geol.), 23, 389-493.

vine, G. R. 1879«. Physiological character of Fenestella. Sci. Gossip. 15, 50-54.

1879 b. On Palaeocoryne, and the development of Fenestella. Ibid. 225-229, 247-249.

young, J. and young, J. 1874. On Palaeocoryne and other polyzoal appendages. Q. Jl geol. Soc. Loud. 30,
684-687.

ADRIAN J. BANCROFT

Department of Geology
Trinity College

Typescript received 20 May 1987 Dublin 7

Revised typescript received 2 July 1987 Ireland

TREMADOC TRILOBITES FROM THE SKIDDAW
GROUP IN THE ENGLISH LAKE DISTRICT

by A. W. A. RUSHTON

Abstract. The Tremadoc trilobite fauna from the Skiddaw Group exposed in the river Calder, western Lake
District, consists of ten species and is referred to the upper Tremadoc Angelina sedgwickii Biozone. Some of
the constituent genera are of wide geographical range in outer shelf environments. Two species, Pareuloma
expansion and Prospectatrix brevior, are new.

Elles (1898) studied the graptolite fauna of the Skiddaw Slates Group in northern England and
concluded that part of the succession was of Tremadoc age. Subsequent revision did not uphold
her claim (Rose 1954; Jackson 1962), and for many years it was believed that the Skiddaw Group
was no older than the Arenig Series. Recently, however, Molyneux and Rushton (1985) demonstrated
the presence of the Tremadoc Series by means of acritarchs and trilobites collected from a small
area in the valley of the river Calder in Cumbria (a locality apparently unknown to Elles). Since
that discovery prolonged collecting has added to the number of trilobites found, and the locality
has now yielded the richest trilobite fauna (in specimens and species) so far known in the Skiddaw
Group, which is as a rule notoriously barren. Detailed examination of the collection confirms the
late Tremadoc age of the fauna and indicates a correlation with the Angelina sedgwickii Biozone
of the Welsh Tremadoc succession and with the Triarthrus tetragonalis-Shumardia minutula Biozone
or the Notopeltis orthometopa Biozone of Argentina. Further investigations have subsequently
revealed Tremadoc rocks in other parts of the Lake District: Tremadoc acritarch floras have been
found in the Buttermere area (S. G. Molyneux, pers. comm.), and evidence from graptolites and
acritarchs demonstrates the presence of Lancefieldian (late Tremadoc?) strata in the Uldale Fells,
in the northern Lake District (Rushton 1985).

LOCALITIES AND STRATIGRAPHY

The trilobite locality is on the east bank of the river Calder, 1 120 m at 297 from the summit of
Latter Barrow hill, 6 km east of Egremont, Cumbria (text-fig. 1 ). At grid reference NY 0687 1 178
a meander of the river cuts into the bank and exposes a flat-lying slump fold of Skiddaw Slate.
Downstream are grey mudstones, siltstones, and thin beds of sandstone, generally dipping to the
west or north-west at 20°-30°. Allen and Cooper (1986) mapped the base of the overlying
Latterbarrow Sandstone Formation; it crosses the river 200 m downstream from the meander
(Allen and Cooper 1986, figs. 2 and 4).

Shackleton (1975, p. 35) mentioned the discovery of trilobites at this place, but of those specimens
the only example I have been able to locate is the ‘ Cvclopyge ’ preserved in the collections of the
British Geological Survey (no. GSM 87362). It is an undeterminable fragment of a thorax, and
was collected from one of the small outcrops on the west bank of the river.

More recently fossils have been collected at three places on the east bank of the river (text-
fig. 1). Locality 1, about 80 m downstream from the meander, yielded Geragnostus callavei (Raw
in Lake, 1906), Shumardia ( Conophrys ) sp., Pareuloma expansion sp. nov., ParabolineUa triarthroides
Harrington, 1938, Peltocare olenoides (Salter, 1866), Bohemilla sp., Niobina davidis Lake, 1946,
Nileid spp. 1 and 2, Prospectatrix brevior sp. nov., sponge spicules (cruciform), acrotretid
brachiopod indet., gastropod indet. Locality 2, about 15 m upstream from Locality 1, yielded

| Palaeontology, Vol. 31, Part 3, 1988, pp. 677-698, pis. 66-68.|

© The Palaeontological Association

678

PALAEONTOLOGY, VOLUME 31

text-fig. I Map to show the fossil localities on the river Calder, western Lake District. Geology from Allen
and Cooper (1986). Numbered grid lines relate to National Grid square NY.

Peltocare olenoides, Prospectatrix brevior , and acrotretids. Locality 3, about 30 m upstream from
Locality 1, yielded Peltocare olenoides and acrotretids.

Macrofossils are scarce at the river Calder localities. They have suffered from sedimentary
compaction but are only slightly deformed tectonically. Many of the specimens, especially those
in fresh rock, are preserved as ‘ghosts’ (i.e. the rock splits a fraction of a millimetre above or below
the bedding-plane with the fossil, as in the case of the thorax of the specimen in PI. 67, fig. 3); it
is at best laborious, and more generally impossible, to develop them. When the mudstone is
weathered, preservation may be fairly good. The trilobite remains appear to be exuviae. Many
specimens consist of partial exoskeletons such as axial shields (i.e. without free cheeks) and there
are few complete specimens that could be construed as dead individuals. One specimen of P.
olenoides has some of the pleurae shortened, suggestive of a healed injury (PI. 67, fig. 10; see Owen
1984).

Of the species in the collections from the river Calder, only four are known elsewhere. N. davidis
occurs in the Upper Tremadoc Penmorfa Beds and Garth Hill Beds in North Wales, and is reported
from New Brunswick (see below). The Penmorfa Beds are referred to the S. pusilla Biozone and
the Garth Hill Beds represent the best development of the A. sedgwickii Biozone (Cowie et al.
1972). Parabolinella triarthroides, as restricted below, occurs in the upper Tremadoc of northern
Argentina: it is rare in the T. tetragonalis-S. minutula Biozone and common in the Notopeltis
orthometopa Biozone there (Harrington and Leanza 1957, pp. 28, 107). Peltocare olenoides is

RUSHTON: TREMADOC TRILOBITES

679

text-fig. 2. A global palaeogeographic reconstruction for the Tremadoc Series (modified from Scotese et al.
1979) showing the distribution of Parabolinella and Peltocare (crosses) and euolomids (rings).

known only from the Garth Hill Beds. If P. glabrum is a synonym, the species also occurs in the
T. tetragonalis-S. minutula Biozone in Argentina but is not known in the N. orthometopa Biozone.
G. callavei is recorded from the S. pusilla Biozone in Shropshire and North Wales, but the
Shumardia ( Conophrys ) is a slightly different form from S. (C.) pusilla that characterizes the
S. pusilla Biozone in the same areas. The Pareuloma and the Prospectatrix are new species, and
these are of uncertain correlative value.

The Tremadoc age of the fauna is significant as it contributed to Allen and Cooper’s (1986,
p. 70) interpretation of the unconformable contact at the base of the Latterbarrow Sandstone.
Furthermore, the fauna is an example of an outer-shelf or slope trilobite assemblage in the latest
Tremadoc. Outer-shelf faunas tend to be rare during periods of world-wide marine regression
(Fortey 1984), and the latest Tremadoc has been interpreted as such a regressive interval.

Taken at generic level the fauna from the river Calder is referable to the Ceratopygid Province
of Whittington and Hughes (1974). All the genera present are assigned to families (for example,
Asaphidae, Cyclopygidae) typical of the Province, though Parabolinella , Geragnostus , and Shumar-
dia ( Conophrys ) were regarded as cosmopolitan genera. Parabolinella is widely distributed in the
Ceratopygid Province; Peltocare, although less widely distributed, generally occurs at localities
where Parabolinella is known, an exception being at Digermul in Finnmark (Nikolaisen and Hen-
ningsmoen 1985). Niobina is recorded from Britain, New Brunswick (see systematic section below),
Sweden (Tjernvik 1956), and Argentina (Harrington and Leanza 1957); Prospectatrix is known
from Britain, Turkey, and Kazakhstan (see systematic section below). When the occurrences of
genera known from the river Calder locality are plotted on a palaeogeographic reconstruction,
they are seen to lie on the margins of cratonic areas from the circum-equatorial belt to near the
Antarctic circle (text-fig. 2), though Parabolinella may occur as a rarity in shallow shelf deposits
(e.g. Winston and Nicholls 1967, p. 76). The distribution somewhat resembles that of the lower
Ordovician Isograptid biofacies shown by Fortey and Cocks (1986, fig. 3). Therefore, if one adapts
the interpretation offered by Fortey and Owens (1978, p. 239), the Ceratopygid Province seems

680

PALAEONTOLOGY, VOLUME 31

better regarded as an outer-shelf benthic association, inhabiting cool water over a wide latitudinal
range (Cocks and Fortey 1982), rather than a geographically defined province.

SYSTEMATIC PALAEONTOLOGY

The terminology used generally follows that of Henningsmoen (1957). The glabella excludes the occipital
ring and the glabellar lobes and furrows are labelled LI, L2 . . . and SI, S2 . . . forwards from the back. All
the new material is in the Type and Stratigraphical Collection of the British Geological Survey, Keyworth,
Nottinghamshire, UK (specimen numbers prefixed by GSM and RX).

Family metagnostidae Jaekel, 1909
Subfamily metagnostinae Jaekel, 1909
Genus geragnostus Howell, 1935

Type species. By original designation, Agnostus sidenbladhi Linnarsson (see Tjernvik 1956, p. 188). Fortey
(1980, p. 24) discussed the applicability of the family name Metagnostidae.

Geragnostus callavei ( Raw in Lake, 1 906)

Plate 66, figs. 1, 2, 4, 5, II

1906 Agnostus callavei Raw MS in Lake, p. 25, pi. 2, fig. 20.

Material. Four exoskeletons, one cephalon, and three pygidia, mostly in counterpart (RX 318-328, 1 523
1524a, b). All are from river Calder, Loc. 1.

Description. All cephala poorly preserved. Glabella about one-third of cephalic width, bluntly pointed in
front. Furrow marking off anterior lobe faint, form uncertain. Elongate median node immediately behind
position of this furrow. Cephalic border furrow wide, border rather fiat. Thorax of usual agnostid type.

Pygidial axis occupies two-thirds of length of pygidium and two-fifths of its width. Two anterior lobes of
axis (Ml +M2 of Robison 1982, p. 134) together only little more than half as long as posteroaxis. M2 slightly
longer than and less wide than M 1 . Elongate median node extends along M 1 and M2 and appears to be
divided by transverse furrow. Posteroaxis rounded behind, with small terminal node. Pleural regions subequal
in width beside and behind axis. Border furrow broad and shallow; border flat with pair of small posterolateral
marginal spines.

Discussion. In 1985 I recorded this species as Micragnostus (Molyneux and Rushton 1985) but the
discovery of better specimens (e.g. PI. 66, figs. 1 and 2) showed that the cephalic border, and
probably the glabella also, are unlike those in species of Micragnostus , as restricted by Fortey
(1980, p. 20). Geragnostus , after the exclusion of several species not referable to the genus (Fortey
1980, p. 27), offers better forms for comparison.

explanation of plate 66

Except where otherwise stated the specimens are from the Upper Tremadoc of the river Calder, Cumbria
(see text-fig. 1). All are internal moulds except where indicated, and were whitened before photography.

Figs. 1, 2, 4, 5, 11. Geragnostus callavei (Raw in Lake, 1906). 1, 2, 5, 11, from Loc. 1; 4, from Shineton

Shales, pusilla Zone, Sheinton Brook, Shropshire (NGR SJ 608 037). 1 and 2, latex cast of external mould
(RX 1524a) and counterpart (RX 1523a). 4, GSM 48670, lectotype. 5, RX 1523b. 11, RX 325. All x 6.

Fig. 3. Geragnostus sp. Loc. 1. RX 2549, with longer Ml + M2 lobes on the pygidial axis, x 6.

Figs. 6-10. Shumardia ( Conophrys ) sp. All from Loc. 1 . 6 and 7, RX 292 and RX 920, latex casts of cranidia.
8, RX 1548, internal mould. 9 and 10, RX 921 and RX 922, fragmentary thorax and pygidium; latex cast
of external mould and counterpart. All approx, x 16.

Figs. 12 15. Pareuloma expansion sp. nov. All from Loc. 1 . 12 and 1 5, RX 1 543a and 1 543b, small cranidium,
with latex cast of counterpart, both x 8. 13, RX 1546, holotype, x 3. 14, RX 520, damaged specimen,
but shows fragmentary free cheek and pygidium (see text-fig. 4a), x 3.

PLATE 66

RUSHTON, Geragnostus , Shumardia ( Conophrys ), Pareulonta

682

PALAEONTOLOGY, VOLUME 31

The species from the river Calder is distinguishable from most other species of Geragnostus by
the relative shortness of Ml +M2 compared with the posteroaxis, and in this feature is most
similar to G. callavei , from the upper Tremadoc S. pusilla Biozone of the Shineton Shales. The
pygidium is identical in detail, even to the presence of a terminal node which, though not seen in
Lake’s original figure, is present on the lectotype (selected Morris 1988, refigured here on PI. 66,
fig. 4). The cephalon of the river Calder specimen appears to differ from that of the lectotype of
G. callavei only in having a more pointed glabellar front.

Another species in which Ml +M2 is relatively short (though not as short as in G. callavei ) is
G. mediterraneus Howell, 1935, from the Arenig of the Montagne Noire (figured by Dean 1966,
pi. 2, fig. 8, and Capera et al. 1978, pi. 5, fig. 4), but that species differs also in having a relatively
wider axis. Some specimens of G. nesossii Harrington and Leanza (1957, p. 65, figs. 9.2 and 9.5),
from lower Tremadoc strata in Argentina, have a glabella shape like that of the specimens from
the river Calder, but in G. nesossii the pygidial axis is relatively short and Ml +M2 form a greater
proportion of the axis. In G. sidenbladhi M1+M2 are together more than half as long as the
posteroaxis and M2 is considerably longer than Ml (Tjernvik 1956, pi. 1, fig. 6). One specimen
from the river Calder locality, collected and kindly donated by Mr M. J. N. Cullen (PI. 66, fig. 3),
differs from other specimens from the same locality but resembles G. sidenbladhi in having M 1 + M2
about two-thirds as long as the posteroaxis. Although the pygidium resembles that of G. sidenbladhi ,
the cephalon does not show the relatively narrow cephalic border of that species.

Family shumardiidae Lake, 1907
Genus shumardia Billings, 1862

Type species. Shumardia granulosa Billings, 1862.

Subgenus shumardia (conophrys) Callaway, 1877

Type species. Conophrys salopiensis Callaway, 1 877, from the Shineton Shales (upper Tremadoc) of Shropshire.
This species has long been treated as a junior synonym of S. pusilla (Sars) from the Ceratopyge Shale (upper
Tremadoc) of Norway, and it may be so; but there has been no revision of S. (C.) pusilla since Stormer’s of
1940, and the pygidium in particular is not well known. The synonymy of pusilla and salopiensis remains
therefore ‘not yet definitely settled’ (Stubblefield 1926, p. 347).

Discussion. Fortey (1980, p. 33) discussed Shumardia and listed many species assigned to the genus
as conceived in a broad sense. Subsequently Fortey and Owens (1987, p. 119) have argued for the
use of subgenera within Shumardia , as recognized by combinations of cephalic and pygidial
characters. The material considered below has small anterolateral lobes to the glabella, a
macropleural thoracic segment, and a transverse pygidium, and is referred to Shumardia
( Conophrys ), using Fortey and Owens’s criteria.

Shumardia ( Conophrys ) sp.

Plate 66, figs. 6 10

Material. Four cranidia and one pygidium with fragment of thorax in counterpart; two poorly preserved
fragmentary individuals (RX 292, 523-524, 921-923, 1547, 1548). All from river Calder, Loc. 1.

Description. Cranidium of typical Conophrys form, with bluntly pointed front of glabella outlined by shallow
furrow, and sides of glabella defined by deep axial furrows. Anterolateral glabellar (‘eye-like’) lobes small
and faintly delimited, less tumid than those of S. (C.) salopiensis. No basal glabellar (SI) furrows seen.

Only one interpretable thorax preserved, but if macropleural segment assumed to be fourth (as in other
species), thorax composed of six segments altogether.

Pygidial axis nearly one-third of width and two-thirds of length of pygidium, and composed of three rings
and terminal part. There are two oblique pleural grooves. Border flat and margin entire.

Discussion. The present form differs from most species referred to S. ( Conophrys ) because the
pygidium lacks a raised marginal rim, such as is well shown by S. (C.) salopiensis (Fortey and

RUSHTON: TREMADOC TRILOBITES

683

Rushton 1980, figs. 11 and 16). Among those species which lack a raised pygidial rim are three-
s'. botinica Wiman (1902, pi. 3, figs. 35-38), S. curta Stubblefield in Stubblefield and Bulman (1927,
pi. 4, figs. 4 and 5), and S. ctenata Robison and Pantoja-Alor (1968, pi. 99, figs. 19 and 20) —
which differ from the river Calder species because the pygidial pleurae curve backwards to become
subparallel with the axial line distally and remain separated by the interpleural grooves to the very
margin. The present species appears to be more like S. (C.) oelandica Moberg (1900, pi. 14, figs.
4-6) in pygidial features, but the glabella in Moberg’s figure is wider in proportion. In that respect
the river Calder form is more like the specimen of S. oelandica figured by Balashova (1961, pi. 4,
fig. 15). In S. (C.) oelandica , however, the thorax has only one segment between the macropleural
segment and the pygidium, whereas there appear to be two such segments in the present species.
The pygidium of S. (C.) liantangensis Lu and Lin (1984, pi. 6, figs. 15 and 16) is similar to that
from the river Calder but differs because it has an additional pygidial segment.

Family eulomidae Kobayashi, 1955

Discussion. The Eulomidae have been discussed by Courtessole and Pillet (1975), Shergold (1980),
and Shergold and Sdzuy (1984, p. 80). They are ptychoparioid trilobites showing conservative
features such as a conical glabella with simple furrows, distinct ocular ridges, and a well-developed
preglabellar field. Distinctive of Euloma and several other eulomids are the deep glabellar furrows
(SI and S2) that run into the axial furrows; SI is strongly oblique inwards and backwards,
separating a subtriangular LI and giving the glabella a calymenid appearance. The palpebral lobe
is separated from the ocular ridge by a furrow. Most species have pits in the anterior border
furrow (but Pareuloma impunctatum Rasetti does not). The pygidium in eulomids is wide and
short, composed of few segments, and has a narrow border; the pleural areas have one or two
pairs of pleural furrows but the interpleural grooves are faint or absent. None of the above features
is especially distinctive, and all could be matched in genera that are not regarded as eulomids. The
plesiomorphic nature of the group makes it difficult to diagnose, and this is made more difficult
by including in the Eulomidae genera with weak or no glabellar furrows, as has been done in
recent years. Yet Apollonov and Chugaeva (1983, text-figs. 3-14; pis. 7 and 8) have demonstrated
a morphological gradation between Ketyna , some of which have no or only very weak glabellar
furrows, and various forms of Euloma with deep furrows; although the SI glabellar furrows in K.
venusta Apollonov and Chugaeva (1983, pi. 7, figs. 14 and 15) are weak, they are oblique and the
species has pits in the anterior border furrow. Shergold and Sdzuy (1984, p. 81) also considered
that some eulomid genera might be descended from species of Ketyna.

Included genera. Numerous genera have been referred to the Eulomidae. These are listed below, with their
type species. ( Euloma , and names derived from it by the addition of a prefix, are neuter in gender, but some
specific names such as brachymetopa are nouns in apposition and therefore do not decline. I have treated
abunda as an invariate arbitrary combination of letters.)

Euloma Angelin, 1854 (type species E. laeve , for which see Tjernvik 1956, p. 274).

Pareuloma Rasetti, 1954 (P. brachymetopa).

Euloma (Proteuloma) Sdzuy, 1958 (Conocephalites geinitzi Barrande, for which see Sdzuy 1955).

Eulomina Ruzicka, 1931 (Euloma initiation Riizicka, 1926) appears to have glabellar furrows that do not
reach the axial furrow, and is here excluded from the Eulomidae.

Eulomella Kobayashi, 1955 (E. mckayensis ) has weak and rather transversely directed glabellar furrows,
and is probably not a eulomid.

Ketyna Rosova, 1963 (K. ketiensis ) has weak glabellar furrows but it was placed in the Eulomidae by
Apollonov and Chugaeva (1983).

Dolgeuloma Rosova, 1963. In 1963 Rosova described two species, D. dolganense (originally doiganensis)
and D. abunda , and designated D. dolganense as type species (Rosova 1963, p. 17). In 1968 Rosova (footnote
on p. 131) stated that this designation was a mistake and sought to alter the type species to D. abunda ; but
this is inadmissable without resort to the plenary powers of the ICZN. In 1968 Rosova also proposed the
Subgenus D. ( Pseudoacrocephalites ), with D. dolganense as type species. As D. (Pseudoacrocephalites) is a
junior homonym of Pseudoacrocephalites Maximova, 1962, Courtessole and Pillet (1975, footnote on p. 253)

684

PALAEONTOLOGY, VOLUME 31

proposed the replacement name D. ( Rosovaspis ). In consequence Pseudoacrocephalites Rosova (not Maximova)
and Rosovaspis are both objective synonyms of Dolgeuloma as they all have the same type species. In
Dolgeuloma the glabellar furrows are weak and do not connect with the axial furrow, and the ocular ridge
joins the palpebral lobe without an intervening furrow. These features indicate that Dolgeuloma is not to be
placed in the Eulonridae.

Lopeuloma Rosova, 1968 (L. loparense).

Duplora Shergold, 1972 (D. clara).

Euloma ( Lateuloma ) Dean, 1973 (E. (L.) latigena).

E. ( Plecteuloma ) Shergold, 1975 (E. (P.) strix).

E. ( Mioeuloma ) Lu and Qian, 1977 (E. ( M .) subquadratum). See Lu and Qian (1983, p. 39), who give
differences in proportion that are intended to distinguish E. (M.) from E. ( Proteuloma ). Peng ( 1984) justifiably
regarded these subgenera as synonyms.

E. (Archaeuloma) Lee in Yin and Lee, 1978 (E. (A.) guizhouense ) resembles Ketyna except that the palpebral
lobe appears to be confluent with the ocular ridge, so it may not be a eulomid.

Iveria Shergold, 1980 (I. iverensis).

Karataspis Ergaliev, 1983. The type species, K. blednovi Ergaliev (1983, pi. 4, fig. 7) is difficult to interpret,
but K. peculiaris Apollonov and Chugaeva (1983, pi. 8, figs. 4-9) resembles Ketyna.

E. ( Spineuloma ) Lu and Lin, 1984 (E. (S.) spinosum).

Duplora ( Euduplora ) Zhou and Zhang, 1984 (D. (E.) ambigua).

Some of the above taxa, for example Duplora and Iveria , are characterized by distinctive features, but others
depend on such doubtful features as the strength of the glabellar furrows or minor variations in the length
or position of the eyes. The presence of pits in the anterior border furrow is a distinctive feature but is not
treated as of generic value in, for example, the olenid Parabolinella. If the presence of curved and strongly
oblique SI furrows is taken as a unifying feature of the Eulomidae, forms in which this furrow is absent
( Arclieuloma , Spineuloma) should not be referred to the family. Among the genera assigned to this restricted
view of the Eulomidae, the distinctions between the genera (or subgenera) Pareuloma , Proteuloma , and
Lateuloma remain arbitrary, as implied by Dean (1973, p. 300).

Genus pareuloma Rasetti, 1954

In Pareuloma species the palpebral lobes are small, and the frontal area and pleural regions are
broad so that the glabella occupies a correspondingly small proportion of the cranidium. This
morphology resembles that of other trilobites of the atheloptic community of Fortey and Owens
(1987, p. 105), that they interpreted as inhabitants of outer-shelf or slope environments. The
supposed olenid Pie slop arabolina proparia Harrington and Leanza (1957, p. 87) has a very similar
morphology, but I regard the short, pit-like glabellar furrows as evidence that Plesioparabolina is
neither a eulomid nor an olenid. Its affinities are uncertain.

Pareuloma expansion sp. nov.

Plate 66, figs. 12-15; text-fig. 4a
Name. Latin, expanded, referring to the frontal area.

Material. Holotype cranidium with part of thorax (RX 1546; PI. 1, fig. 13). Paratypes: a damaged cephalon
with fragment of thorax and pygidium (RX 520), a small cranidium (RX 1523a, b), and some other
fragmentary specimens (RX 311, 1529, 1538-1540). There are other fragments, and some ‘ghosts’ (e.g. RX
309, 518) that proved impossible to develop, and these suggest that the species is not rare. All are from the
river Calder, Loc. I .

Diagnosis. Pareuloma with small palpebral lobes, anterior border nearly flat, broad, as wide (sag.)
as preglabellar field, with pits in the anterior border furrow. Preocular sutures strongly divergent
forwards.

Description. Glabella (excluding occipital ring) about as long as wide across LI (longer in proportion in
smaller specimens). Glabellar furrows deep: SI oblique backwards, S2 similar but shorter, S3 short, indents
side of glabella. Occipital ring as wide as base LI, with faint median node. SO composite. Frontal area nearly

RUSHTON: TREMADOC TRILOBITES

685

as long as preoccipital glabella. Preglabellar field slightly inflated in front of glabella. Anterior border broad,
flat, at least as long (sag.) as preglabellar field. About sixteen pits lie along anterior border furrow. Palpebral
lobes small (about one-tenth as long as cranidium), centred opposite anterior ends of SI. Ocular ridge thin,
distinct, oblique. Interocular cheeks about three-quarters as wide as width of glabella across L2. Preocular
sutures divergent forwards, making broad inward curve across anterior border. Postocular cheeks about as
wide as occipital ring, pleuroccipital furrow curved forwards at its outer end. Postocular sutures oblique,
curving back distally.

Free cheeks not well preserved. Hypostome and ventral features not known. Anterior thoracic segments
have pleurae wider than the axis. Posterior segments not known. A fragmentary pygidium (PI. 66, fig. 14;
text-fig. 4a) is subtriangular in outline, with border. Axis about one-third of total width but badly preserved.
Pleural fields appear to have two pairs of pleural furrows. Surface of convex parts of exoskeleton finely
granulose.

Discussion. P. expansum is unusual among eulomids in having an almost flat anterior border, but
the small truncate, conical glabella and the small eyes are typical features of Pareuloma. P.
expansum differs in many details from P. brachymetopa , from beds of supposedly Trempealeau
(early Tremadoc) age at Cap des Rosiers, Gaspe, and at Broom Point, western Newfoundland.
The anterior border is much broader and is practically flat; the arc of pits in the anterior border
furrow extends behind a transverse line through the front of the glabella, which is further back
than in P. brachymetopa. The preglabellar swelling is weaker than in P. brachymetopa and the
palpebral lobes are a little shorter, judging from Rasetti’s reconstruction (1954, text-fig. 2).

P. expansum resembles P. impunctatum Rasetti ( 1954, pi. 61, figs. 1 and 2) in having an expanded
and not very convex border, but differs because the border is much broader and is arcuate in plan.
The border furrow has well-marked pits, whereas there are none in P. impunctatum , and the
preocular sutures are strongly divergent forwards. P. insuetum Apollonov and Chugaeva (1983,
pi. 8, fig. 15) has narrower interocular cheeks and a narrower frontal border. P. spinosum Palmer
(1968, p. 76, pi. 11, figs. 1-9) differs in having an occipital spine, longer eyes, and a narrower,
straighter frontal border.

The arcuate anterior border of P. expansum recalls those of E. ( Lateu/oma ) latigena Dean (1973,
pi. 3, figs. 5, 6, 8-1 1) and E. (L.) kasachstanicum Balashova (1961, pi. 4, figs. 11-13), although it
differs in being much broader (sag.); the wide interocular cheeks and posteriorly placed eyes are
also points of similarity. In Lateu/oma species the glabella is contracted in front of L2 and is
rounded in front, whereas in Pareuloma the glabella is truncate and conical. P. expansum differs
from Lateu/oma also in having smaller palpebral lobes. The broad, arcuate border and the small
eyes distinguish P. expansum from all species of Proteuloma.

Family olenidae Burmeister, 1843
Subfamily oleninae Burmeister, 1843
Genus parabolinella Brogger, 1882

Type species. Parabolinella limitis Brogger, 1 882.

Discussion. The predominantly Tremadoc genus Parabolinella is in some respects a conservative
member of the Subfamily Oleninae, and retains features of its supposed ancestral stock, for example
the well-developed preglabellar field, the position of the eye, and the forms of the postocular cheek,
free cheek, and hypostome. However, advanced characters that could be used to characterize
Parabolinella include: 1, the geniculate and bifurcate SI furrows; 2, the composite occipital furrow
(SO); 3, the accessory lateral glabellar furrow between SO and SI (seen also in Hypermecaspis)\ 4,
the inflated preglabellar field; and 5, pits in the anterior border furrow. Compared with earlier
olenine genera there are many thoracic segments and a small pygidium. In the type species, P.
limitis , features 1 and 2 are well developed, 3 and 4 are faint, and 5 is present (Henningsmoen
1957, pi. 12, figs. 2 and 3). Among other species such as P. triarthra (Callaway) and P. argentinensis
Kobayashi, all the cited features are shown, though the accessory furrow (no. 3) is generally
indistinct.

686

PALAEONTOLOGY, VOLUME 31

Included species. Henningsmoen (1957, p. 132) reviewed the species of Parabolinella then known. Several
further species have since been referred to Parabolinella'.

Moxomia liecuba Walcott, referred to Parabolinella by Harrington and Leanza (1957, p. 107).

P. coelatifrons Harrington and Leanza (1957, p. 109) (transferred to Angelina by Robison and Pantoja-
Alor 1968, p. 787).

P. chilienensis Chang and Fan, 1960.

PI fortunata Lazarenko, 1966.

P. prolata Robison and Pantoja-Alor, 1968.

P. tumifrons Robison and Pantoja-Alor, 1968 (referred to P. hecuba by Ludvigsen 1982, p. 63).

P. variabilis Robison and Pantoja-Alor, 1968.

P. latilimbata Lu and Chien in Yin and Lee, 1978 (see Lu and Qian 1983, p. 49, pi. 6, fig. 2).

P. contracta Lu and Zhou in Lu, Zhou and Zhou, 1981.

P. panosa Ludvigsen, 1982.

Remizites bolati Ergaliev, 1983, pi. 3, figs. 12 and 13, is referable to Parabolinella as considered here.

P. sayramensis Xiang and Zhang, 1984.

P. lata Xiang and Zhang, 1984 (not P. lata Henningsmoen, 1957).

P. jiangnanensis Lu and Lin, 1984.

P. ocellata Lu and Lin, 1984.

P. borohoroensis Xiang and Zhang, 1985 (possibly better referred to Parabolinites).

Parabolinella xinjiangensis Xiang and Zhang, 1985.

Some of the above names are likely to be synonyms. P. contracta is distinguished by small eyes, ocular
ridges that slope outwards and slightly forwards, and postocular cheeks that are as wide as the occipital ring;
an accessory glabellar furrow is inserted close to the axial furrow between SO and the geniculate SI, and the
surface is finely granulose. Through the kindness of Dr Zhou Zhiyi I have examined latex casts of P. contracta ,
and I consider that the fragmentary Parabolinella ? figured by Rushlon (1982, pi. 3, figs. 23, 247, 25), from
the Acerocare Zone, just below the base of the Tremadoc Series as defined in the section at Bryn-llin-fawr
in North Wales, is referable to the same species. P. sayramensis, from the lower part of the Sayram Formation
in north Tianshan (assigned to the lower Tremadoc) seems to be identical with P. contracta. P. lata Xiang
and Zhang (not Henningsmoen) is based on distorted material from the same formation as P. sayramensis,
but at a different locality. It shows the same features as P. contracta and is probably a synonym. P. ocellata
Lu and Lin, from the basal Tremadoc part of the Yinchupu Formation in Zhejiang is also similar to P.
contracta, but the ocular ridges do not slope forwards from the eyes. P. xinjiangensis differs from P. contracta
only in the weakness of the accessory glabellar furrow, and may also be a synonym. Finally, P. bolati
(Ergaliev) resembles P. contracta in most features but the figured examples have a proportionally longer
preglabellar field.

Parabolinella triarthroides Harrington, 1938
Text-fig. 3c

1938 Parabolinella triarthroides n. sp., Harrington, p. 194, pi. 7, figs. 10 and 1 1.

non 1951 Parabolinella triarthroides Harrington; Shaw, p. 102, pi. 22, figs. 1-10.

1957 Parabolinella triarthroides Harrington; Harrington and Leanza, p. 105, fig. 39.1.

non 1967 Parabolinella triarthroides Harrington; Winston and Nicholls, p. 76, pi. 13, fig. 14.

New material. A small cephalon (RX 913 914), a fragmentary free cheek (RX 260), and fragments of thorax
(RX 921, 922, 1534, 1535), all from river Calder, Loc. I

Description. Cranidium 5-2 mm long. Glabella plus occipital ring three-quarters of cranidial length,
subquadrate, widens slightly forwards from occipital ring to L3, anterolateral corners rounded. Occipital
ring simple, with small median node. Preoccipital glabella as long as its greatest width. SI oblique, scarcely
geniculate or forked. S2 simple, slightly oblique. S3 faint, S4 not discerned. Preglabellar field somewhat
down-sloping, but shows signs of having been originally steeper. Anterior border not well preserved, pits
may be present in border furrow, but this is not certain. Preocular sutures slightly divergent forwards.
Palpebral lobes small, opposite L3, interocular cheeks about one-third as wide as glabella; ocular ridges
extend outwards and slightly forwards. Postocular cheeks a little wider than occipital ring; postocular sutures
oblique, slightly sinuous, curving backwards across posterolateral border furrow.

RUSHTON: TREMADOC TRILOBITES

687

text-fig. 3. A, b, Peltocare modestum Henningsmoen, 1957, BM(NH) It 12903, Ceratopyge Limestone,
Bjerkasholmen, Oslo region (Coll. R. A. Fortey), anterior and dorsal views, x 6. c, Parabolinella triarthroides
Harrington, 1938, RX 913, river Calder, Loc. 1, x 6.

The thoracic fragments are insufficient for description but indicate specimens of large size, bigger than the
large specimen of P. triarthra figured by Lake (1913, pi. 7, fig. 4).

Discussion. P. triarthroides , originally described from the upper Tremadoc rocks of Argentina, is
based on small cranidia, the figured examples being convex specimens about 3 mm long. Compared
with the larger but flattened specimen from the river Calder, the glabellar furrows, the position of
the eyes, the width of the fixed cheeks, and the course of the facial sutures are alike. The river
Calder specimen is more like Harrington’s paratype (1938, pi. 7, fig. 11) than his holotype (fig. 10)
in the squarish anterolateral corners of the glabella and the length of the occipital ring.

Shaw (1951, pi. 22, figs. 1-10) referred several specimens from the Gorge Formation of Vermont
(a horizon near to or just below the base of the Tremadoc Series) to P. triarthroides , since when
P. triarthroides has been mentioned in discussions of the correlation of the Cambrian-Ordovician
(or, rather, the Trempealeauan-Canadian) boundary. However, all Shaw’s specimens differ from
P. triarthroides , for example, in having postocular cheeks that are narrower than the occipital ring
(see also Fortey in Fortey et al. 1982, p. 112, and Harrington and Leanza 1957, p. 107). A fragment
figured as P. triarthroides by Winston and Nicholls ( 1967, pi. 13, fig. 14) is also unlike Harrington’s
type material as the eye is well back opposite L2 and is close to the glabella.

P. lata Henningsmoen (1957, pi. 12, fig. 8) from the upper Tremadoc Ceratopyge Limestone of
Royken, Norway, is very like P. triarthroides , but differs in having wider fixed cheeks. P. limitis ,
from the upper Tremadoc Ceratopyge Shale of Norway, has narrower fixed cheeks than P.
triarthroides, a longer palpebral lobe, and a strongly geniculate SI (Henningsmoen 1957, pi. 12,
figs. 1-3). P. triarthra from the Shineton Shales of Shropshire has narrower fixed cheeks than
P. triarthroides in both large and small specimens (Lake 1913, pi. 7, figs. 4-12); furthermore, the
preoccipital glabella tends to be wider than long and S3 is often distinct. The same is true of P.
argentinensis (Harrington and Leanza 1957, figs. 37 and 38) which is like P. triarthra but
distinguished by the more widely divergent preocular sutures. P. latilimbata Lu and Chien (see Lu
and Qian 1983, pi. 6, fig. 2) differs slightly in several respects; the preglabellar field is longer and

688

PALAEONTOLOGY, VOLUME 31

the fixed cheeks wider than in P. triarthroides , the glabellar furrows are deeper, the occipital ring
is more strongly composite, and the occipital node is stronger.

Subfamily pelturinae Hawle and Corda, 1847
Genus peltocare Henningsmoen, 1957

Type species. Acerocare norvegicum Moberg and Moller, 1898.

Discussion. Henningsmoen (1957) assigned three other species to Peltocare , namely P. olenoides
(Salter), P. rotundifrons (Matthew), and P. glabrum (Harrington). Since then two further species
have been described: P. modestum Henningsmoen (1959, p. 158, pi. 1, figs. 9 and 10; text-fig. 3a,
b herein), and P. compaction Nikolaisen and Henningsmoen (1985, p. 21, figs. 8 and 16a-r).
Nikolaisen and Henningsmoen (1985) considered that the specimens described by Robison and
Pantoja-Alor (1968, p. 793, pi. 103, figs. 14-23) as P. norvegicum represent an independent species.
Alimbetaspis kelleri Balashova (1961, pi. 3, figs. 15-19) resembles Peltocare species, but has larger
palpebral lobes and more transverse postocular sutures. Nikolaisen and Henningsmoen (1985,
p. 27) treated it as a synonym of Jujuyaspis.

Peltocare olenoides (Salter, 1866)

Plate 67; text-fig. 4b

1866 Conocoryphe olenoides n. sp., Salter, p. 308, pi. 8, fig. 6.

1919 Peltura olenoides (Salter); Lake, p. 100, pi. 12, figs. 4 and 5.

1938 Cyclognathus glaber sp. nov., Harrington, p. 212, pi. 9, figs. 1, 5, 12.

1957 Acrocarina glaber Harrington; Harrington and Leanza, p. 93, fig. 323a-d.

1957 Peltocare olenoides (Salter); Henningsmoen, p. 249.
non 1968 Peltura olenoides (Salter); Curtis, pi. 9a.

1985 Peltocare olenoides (Salter); Molyneux and Rushton, fig. 1.14.

Type material. Salter’s monotype is a distorted cephalon (GSM 10846; PI. 67, fig. 4) from the Garth Hill
Beds (upper Tremadoc Series) of Garth Hill, near Minfordd, North Wales. 1 here interpret the species by
reference to a previously unfigured topotype (BGS Zi 1668, 1669; PI. 67, figs. 6 and 7). This is smaller than
the type but (if correctly referred to P. olenoides ) gives a much better idea of the species.

New material. Twenty specimens and fragments of cephala, axial shields, and fragments of thorax and
pygidium (including RX 279, 302-305, 312, 313, 329, 330, 490, 491, 517, 918, I530M534). Most are from
Loc. I; one or two specimens from each of Loc. 2 and 3.

Description. Glabella plus occipital ring nearly parallel-sided, bluntly rounded in front. Glabellar furrows
not seen, occipital furrow distinct. Occipital ring less than one-quarter of length of cephalic axis, with faint
median node. Palpebral lobes short, inconspicuous, placed well forward such that their anterior ends are
nearly in line with anterior end of glabella. Faint ocular ridges seen in one small specimen (PI. 67, fig. 3).
Frontal area about one-tenth of cranidial length, not differentiated into border and preglabellar field.
Preocular sutures short. Postocular cheeks about 0-7 times as wide as occipital ring. Postocular sutures long,

explanation of plate 67

Figs. 1-11 .Peltocare olenoides (Salter, 1866). 1-3, 5, 8-1 1, all Loc. 1; 4, 6, 7, Upper Tremadoc, Garth Hill,
near Minfordd, Gwynedd, North Wales (NGR SH 593 393). I, RX 1531, latex cast of external mould,
x4. 2, RX 517, cranidium with left free cheek; white pointer indicates position of left eye shown in text-
fig. 4b, x 2. 3, RX 312, small cranidium with part of thorax exposed, x 8. 4, GSM 10846, holotype, x4.
5, RX 297, latex cast of cranidium of specimen figured by Molyneux and Rushton (1985, fig. 1.14), x4.
6 and 7, BGS Zi 1669 and 1668, internal mould and latex cast of counterpart, both x 4. 8 and 11, RX
330 and 329, internal mould of fragmentary cephalon showing the fixed cheek and latex cast of counterpart
prepared so as to show the free cheek, both x4. 9, RX 303, thorax and pygidium, flattened, x4. 10,

RX 302, part of thorax and pygidium; note terrace lines on the pygidium and the truncated pleurae, x4.

PLATE 67

RUSHTON, Peltocare

690

PALAEONTOLOGY, VOLUME 31

convexly curved. No sutural ridge seen, except for vestige seen at posterolateral corner of one small cranidium
(PI. 67, fig. 3). Pleuroccipital furrow curved forwards somewhat at its outer end.

Free cheek semicircular in outline, with broad, faint border furrow and narrow border, and no genal spine.
Posterior end produced into long pointed process directed adaxially (PI. 67, fig. 11). Specimen in Plate 67,
fig. 2 shows eye with holochroal facets, visible when moistened with alcohol (see text-fig. 4b). Hypostome,
known from one indistinct impression (PI. 67, figs. 8 and 11), of normal pelturoid type.

Thorax of twelve segments. Anterior pleura about two-thirds as wide as axial ring. Thorax widens slightly
backwards to fifth segment, behind which it narrows. The last pleura is nearly as wide as axis.

Pygidium more than twice as wide as long. Pleural field as wide as axis. Axis with three rings and terminal
part. Pleural fields marked by one or two pleural furrows. Margin entire with narrow border. Dorsal surface
of pygidium marked by terrace lines subparallel with posterior margin, as in other pelturines (PI. 67, fig. 10).
Remainder of exoskeleton smooth.

Discussion. As discussed by Henningsmoen (1957), the species of Peltocare are all rather similar.
P. compaction has narrower pleural regions than the other species, and in P. modestum (text-fig.
3a, b) the glabella is comparatively sharply rounded in front. P. rotundifrons has a smaller number
of pygidial axial segments than other species.

Henningsmoen (1957) suggested that P. glabnon might be a synonym of P. norvegicum. At that
time P. olenoides was poorly known, but the new material shows that it, too, is very like P.
norvegicum. I can see nothing to distinguish P. glabnon from P. olenoides , and accordingly regard
it as a synonym; but I hesitate to synonymize P. norvegicum because there appear to be slight
differences: the pleuroccipital furrow in P. norvegicum seems to curve forward more strongly than
that of P. olenoides ; the anterior pleura in P. norvegicum appears to be only half as wide as the
axial ring, and is thus shorter (tr.) than in P. olenoides (note, however, that this observation is
based only on Brogger’s figure (1882, pi. 1, fig. 4) of an imperfect specimen and on the external
mould figured by Henningsmoen (1957, pi. 27, fig. 8) in which the full extent of the pleurae may
not be shown).

The specimen that Curtis (1968) figured as P. olenoides differs from the species as revised here
because the eyes are further back and further from the glabella, and the postocular cheeks are
narrower with a more transverse posterolateral margin. It may be referable to Leptoplastides.

Family bohemillidae Barrande, 1872
Genus bohemilla Barrande, 1872

Type species. By monotypy, Bohemilla stupenda Barrande, 1872.

Bohemilla ( Bohemilla ) sp.

Plate 68, figs. 4 and 7; text-fig. 4c

Material. One small fragmentary cranidium, associated with two free cheeks and some thoracic fragments
(RX 31 6a, b) from river Calder, Loc. I

explanation of plate 68

Figs. 1-3. Prospectatrix brevior sp. nov. 1, RX 916, Loc. 1, holotype, latex cast of external mould, x4.
2, RX 928, Loc. 2, cranidium; note faint glabellar furrows, x4. 3, RX 1542, Loc. 1, free cheek, latex cast
of external mould, x 4.

Figs. 4 and 7. Bohemilla (Bohemilla) sp. RX 316b, Loc. 1. 4, right free cheek with fragment of left cheek.

7, cranidium with fragment of thoracic axis; both x 12. For reconstruction, see text-fig. 4c.

Fig. 5. Nileid sp. I . RX 317, Loc. I, x 3.

Fig. 6. Niobina davidis Lake, 1946. RX 259, Loc. 1, latex cast of thorax and pygidium (figured by Molyneux
and Rushton 1985, fig. 1.11), x 2.

Figs. 8 and 9. Nileid sp. 2. RX 521 and 522, Loc. I, fragmentary cranidium and free cheeks, internal mould
and latex cast of counterpart, x 2.

PLATE 68

RUSHTON, Prospectatrix , Bohemilla ( Bohemilla ), Niobina, nileid

692

PALAEONTOLOGY, VOLUME 31

text-fig. 4. a, Pareuloma expansum sp. nov., RX 520, reconstruction of pygidium (see PI. 66, fig. 14),
x 6. b, Pehocare olenoides (Salter, 1866), RX 517 (sketch), oblique anterolateral view of cephalon to show
holochroal eye (see PI. 67, fig. 2). c, Bohemilla ( Bohemilla ) sp., based on RX 316b, reconstruction of

cephalon (PI. 68, figs. 4 and 7).

Discussion. The cranidial fragment shows the SI furrow to be simple and not hooked proximally
as in B. stupenda and other Bohemian species (Marek 1966). The glabella is contracted in front of
S2, but gradually, and not abruptly as in B. pragensis Marek (1966, pi. 2, figs. 10 and 11) or B.
tridens Rushton and Hughes (1981, pi. 5, figs. 11, 15, 16). Only a fragment of the postocular cheek
is visible, but it shows that the base of the border furrow is directed unusually transversely for a
Bohemilla , making an angle of 50°-60° to the sagittal line. The free cheek associated with the
cranidium has an acute inner spine angle; the ocular incisure shows clearly that the eye was short.
If it extended forwards from S2, as in other species of Bohemilla , it did not reach as far forward
as S3.

The free cheek also shows that the postocular suture was comparatively long (about three times
as long as the ocular incisure). In other species the postocular suture is no longer than the ocular
incisure (Marek 1966, pi. 1, fig. 8, pi. 2, fig. 4; Rushton and Hughes 1981, pi. 5, fig. 12).

This fragmentary specimen is the oldest described bohemillid. Compared with the upper Arenig
species B. ( Fenniops ) sabulon Fortey and Owens (1987, p. 129) the present form has a contracted
glabella, which is considered a comparatively advanced character, but shares the more primitive
form of the glabellar furrows and postocular cheeks.

Family asaphidae Burmeister, 1843
Subfamily niobinae Jaanusson, 1959
Genus niobina Lake, 1946

Type species. By original designation, Niobina davidis Lake, 1946.

Niobina davidis Lake, 1946
Plate 68, fig. 6

1946 Niobina davidis Lake, p. 334, pi. 47, figs. 1-5 (synonymy).

1985 Niobina davidis Lake (?); Molyneux and Rushton, p. 126, fig. 1.11.

Material. Thorax and pygidium in counterpart (RX 258, 259), and two poorly preserved free cheeks (RX
289, 292). All from river Calder, Loc. I .

RUSHTON: TREMADOC TRILOBITES

693

Discussion. The figured thorax and pygidium agree in all details with Lake’s figured specimens and
with other material from the upper Tremadoc series in North Wales. The pygidium has about
eight axial rings, and about seven pairs of pleurae marked by interpleural grooves. It thereby
differs from N. taurina Harrington and Leanza (1957, p. 180, fig. 91.1) from the lower Tremadoc
of Argentina which has twelve axial rings and eleven pairs of pleural furrows. Tjernvik’s Niobina
sp. (1956, p. 234, pi. 5, fig. 17), from the Apatokephalus serratus Biozone (upper Tremadoc) of
Sweden has about ten axial rings and nine or ten pairs of pleural furrows. The free cheeks collected
from the river Calder locality are imperfect but show the blunt genal angle and the anterior
extension of the doublure cut off by the median suture. No example of the cranidium has been
collected, so although the pygidium agrees precisely with Lake’s N. davidis , there remains an
element of doubt about specific determination.

N. davidis is recorded from the Shumardia pusilla Biozone and Angelina sedgwickii Biozone in
the upper Tremadoc Series in North Wales, and in the US National Museum there is a specimen
from the Upper Tremadoc of New Brunswick (Dr R. A. Fortey, pers. comm.).

Family nileidae Angelin, 1854
Nileid sp. 1

Plate 68, fig. 5

Material. One axial shield lacking the anterior part of the cranidium (RX 317), from river Calder, Loc. 1.

Description. Postocular facial suture practically straight and directed outwards and backwards at about 30
to sagittal line. Postocular cheek about half as wide as occipital ring. Thorax of seven segments: anterior
pleura two-thirds width of axis; posterior pleura as wide as axis. Pleural geniculation close to axial furrow
throughout. Pygidium semi-elliptical, length two-thirds of width. Axis occupies one-third of width and 0-6
of length of pygidium and lacks ring furrows. Pleural regions unfurrowed. Doublure wide, extending inwards
to fulcral line.

Discussion. The present specimen may be referable to Barrandia M’Coy, as discussed by Hughes
(1979, p. 154), but as the thorax has only seven segments and the pygidium has a better-marked
flattened marginal rim, I hesitate to include this form in Barrandia. The postocular suture resembles
that of certain other nileids such as Peraspis omega Fortey (1975, pi. 20, fig. 1) from the Arenig
Series in Spitsbergen, though that species has narrower postocular cheeks. The poorly known
Hemibarrandia holoubkovensis Ruzicka (1926, pi. 2, figs. 5 and 6), from the lower Tremadoc of
Bohemia, differs from the present form in having a more transverse pygidium with a narrower
doublure (Ruzicka 1931, pi. 1, fig. 8). As the cranidium is fragmentary, closer comparison is
impossible, but the present form is clearly distinct from most nileids (e.g. Nileus , Symphysurus,
Platypeltoides) in the length and straightness of the postocular suture. The pygidium is peculiar in
being elongate and semi-elliptical.

Nileid sp. 2
Plate 68, figs. 8 and 9

Material. Fragmentary cranidium and conjoined free cheeks (RX 521, 522), from river Calder, Loc. 1.

Discussion. The long postocular suture, curved outwards and backwards, clearly distinguishes this
form from Nileid sp. 1, above. The form of the free cheek (PI. 68, fig. 9) shows that the eye was
comparatively far forward and rather small compared with most nileid genera. Illaenopsis Salter
has very small eyes but in most species the glabella is fairly well marked where it expands at its
forward end, e.g. I. thomsoni Salter (Whittard 1961, pi. 31, fig. 3); I. gaspensis (Rasetti 1954,
pi. 60, figs. 9 and 10); I. griffei Courtessole and Pillet (1975, pi. 27, figs. 5-1 1 and presumably also
pi. 26, fig. 21). There is no sign of this expansion in the present fragment (PI. 68, fig. 8). In the
early Tremadoc species Psilocephalinella innotata (Salter) the axial furrow is nearly effaced (Lake

694 PALAEONTOLOGY, VOLUME 31

1942, pi. 44, figs. 2-7), but the eye is further back than in the present form, so that the free cheek
is of a different shape.

Family cyclopygidae Raymond, 1925
Genus prospectatrix Fortey, 1981

Type species. By original designation, Cyclopyge genatenta Stubblefield in Stubblefield and Bulman, 1927.

Prospectatrix brevior sp. nov.

Plate 68, figs. 1 -3

Name. Latin, shorter (than the type), the thoracic axis having six rather than seven segments.

Material. Holotype, an axial shield in counterpart (RX 915, 916; PI. 68, fig. I) from river Calder, Loc. 1.
Paratypes, a cranidium (RX 928) from Loc. 2, and a visual surface (RX 1542) from Loc. I. One poorly
preserved but nearly complete specimen (RX 2550), collected by Mr M. J. N. Cullen from Loc. 1, is thought
to belong to this species.

Diagnosis. A species of Prospectatrix with relatively broad postocular cheeks and very narrow
interocular cheeks. Thorax of six segments. Pygidial axis divided into three distinct rings and a
terminal part.

Discussion. A full description is unnecessary here because this material resembles that of P.
genatenta , as described by Stubblefield (in Stubblefield and Bulman 1927, p. 138) and redescribed
by Fortey (1981, p. 611). Compared with P. genatenta , the cephalic axis and glabellar furrows of
P. brevior have the same form (PI. 68, fig. 2) and the postocular cheeks are of similar size (about
0-3 of the basal width of the cephalic axis). A significant difference lies in the reduced interocular
cheeks, which in P. brevior constitute only a very narrow rim close to the glabella, whereas in P.
genatenta the interocular cheek is about one-sixth as wide as the glabella. The visual surface
appears to widen posteriorly, as in P. genatenta (Fortey 1981, pi. lc). There are only six thoracic
segments in P. brevior , whereas P. genatenta has seven. The pygidium is longer than that of P.
genatenta , the axis has three rings and a terminal part, rather than two rings as in P. genatenta.
The doublure in the two species is similar, with a median cusp extending forward to the tip of the
pygidial axis. These differences— the reduction in the fixed cheeks and the reduction in number of
thoracic segments, with a concomitant increase in the size of the pygidium— appear to be advances
from more primitive character-states in P. genatenta , and P. brevior thus lies morphologically
between Prospectatrix and other Cyclopygidae.

Fortey (1981, p. 612) suggested that Pricyclopyge super ciliata Dean (1973, p. 314), from beds
tentatively ascribed to the lower Arenig in Turkey (Dean 1973, p. 343), might be a species of
Prospectatrix , and Fortey and Owens (1987, p. 176) subsequently compared a specimen of
Prospectatrix from the Fennian (Upper Arenig) of South Wales with Pricyclopyge superciliata.
The cranidium of P. superciliata (Dean 1973, pi. 6, fig. 6) resembles that of the new species,
although the cephalic axis is shorter relative to its length. Compared with Prospectatrix brevior ,
Pricyclopyge superciliata has broader interocular cheeks, whereas the postocular cheeks are both
narrower and shorter (exsag.), indicating that P. superciliata had larger eyes than Prospectatrix
brevior.

A cranidium illustrated by Apollonov et al. (1984, pi. 23, fig. 11) may also be referable to
Prospectatrix. It resembles Pricyclopyge superciliata rather than Prospectatrix brevior in its short,
wide cephalic axis and small postocular cheeks.

Acknowledgements. 1 thank the many friends who helped with fossil collecting, especially Dr R. D. Hutchison
who found some of the best specimens (text-fig. 3c; PI. 68, figs. I and 2), and my colleagues, notably E. P.
Smith and S. P. Tunnicliff, for assistance in the field. I thank Dr S. G. Molyneux (who initiated the study
by finding the specimen in PI. 68, fig. 6) and Dr R. A. Fortey for much helpful discussion. Sir James
Stubblefield read the manuscript and offered helpful criticism. This paper is a contribution to the BGS Lake

RUSHTON: TREMADOC TRILOBITES

695

District Regional Geological Survey, and is published by permission of the Director, British Geological
Survey (NERC).

REFERENCES

allen, p. m. and cooper, d. c. 1986. The stratigraphy and composition of the Latterbarrow and Redmain
sandstones. Lake District, England. Geol. Jl , 21, 59 76.

ANGELIN, N. p. 1854. Palaeontologici Scandinavica , Pars 1. Crustacea formationis transitionis. Fasc. II,
i-ix + 21 92, pis. 25-41. Lund.

APOLLONOV, M. K. and CHUGAEVA, M. N. 1983. Pp. 66-90, pis. 7-10. In APOLLONOV, M. K., BANDALETOV, S. M.
and ivshin, n. k. (eds.). The Lower Paleozoic stratigraphy and palaeontology of Kazakhstan, 176 pp.,
39 pis. Acad. Sci. Kazakh SSR, ‘NAUKA’, Alma-Ata. [In Russian.]

— and dubinina, c. V. 1984. Trilobites and conodonts from the Batyrbay Section (uppermost Cambrian-
Lower Ordovician ) in Malyi Karatau Range (atlas of the palaeontological plates), 48 pp., 32 pis. Acad. Sci.
Kazakh SSR, ‘NAUKA’, Alma-Ata. [In Russian.]

balashova, e. a. 1961. Some Tremadoc trilobites of the Aktyubin region. Trudy geol. Inst. Acad. Nauk SSSR
[Moscow], 18, 102-145, 4 pis. [In Russian.]

barrande, j. 1872. Systeme Silurien du centre de la Boheme. lere partie. Recherches Paleontologiques ,
supplement au Vol. 1. Trilobites , Crustaces diverse et Poissons , xxx + 647 pp., 35 pis. Prague, Paris.
billings, e. 1861-1865. Palaeozoic Fossils , Vol. I, 427 pp. Geological Survey of Canada, Montreal.
brogger, w. c. 1882. Die silurichen Etagen 2 und 3 im Kristianiagebiet, etc., vni + 376 pp., 12 pis. Kristiana.
burmeister, H. 1843. Die Organisation der Trilobiten , etc., xii+ 148 pp., 6 pis. Berlin.

Callaway, c. 1877. On a new area of Upper Cambrian rocks in south Shropshire, with a description of a
new fauna. Q. Jl geol. Soc. Lond. 33, 652-672, pi. 24.

capera, J. c., courtessole, r. and pillet, j. 1978. Contribution a l’etude de l’Ordovicien inferieur de la
Montagne Noire. Biostratigraphie et revision des Agnostida. Annls Soc. geol. N. 98, 67-88, pis. 5-7.
chang went’ang and fan chiasung. 1960. Pp. 83-147, pis. 1 10. In yin tsanhsun (ed.). Geological
gazetteer of the Chi-Lien Mountains , 4 (1, Palaeontological descriptions . . .). Science Press, Beijing.
[In Chinese.]

cocks, l. r. m. and fortey, r. a. 1982. Faunal evidence for oceanic separations in the Palaeozoic of Britain.
Jl geol. Soc. Lond. 139, 465 -478.

courtessole, r. and pillet, j. 1975. Contribution a l’etude des faunes trilobitique de l’Ordovicien inferieur
de la Montagne Noire. Les Eulominae et les Nileidae. Annls Soc. geol. N. 95, 251-271, pis. 24-27.
cowie, j. w., rushton, a. w. a. and Stubblefield, c. J. 1972. A correlation of the Cambrian rocks in the
British Isles. Spec. Rep. geol. Soc. Lond. 2, 42 pp.

Curtis, M. L. K. 1968. The Tremadoc rocks of the Tortworth Inlier, Gloucestershire. Proc. Geol. Ass. 79,
349-362.

dean, w. t. 1966. The Lower Ordovician stratigraphy and trilobites of the Landeyran Valley and the
neighbouring district of the Montagne Noire, south-west France. Bull. Br. Mus. nat. Hist. (Geol.), 12, 245
353, 21 pis.

1973. The Lower Palaeozoic stratigraphy and faunas of the Taurus Mountains near Bey§ehir, Turkey.
III. The trilobites of the Sobova Formation (Lower Ordovician). Ibid. 24, 279-348, 12 pis.
elles, G. L. 1898. The graptolite-fauna of the Skiddaw Slates. Q. Jl geol. Soc. Lond. 54, 463-539.
ergaliev, g. ch. 1983. Pp. 35-66, pis. I 6. In apollonov, m. k., bandaletov, s. m. and ivshin, n. k. (eds.).
The Lower Paleozoic stratigraphy and palaeontology of Kazakhstan, 176 pp., 39 pis. Acad. Sci. Kazakh
SSR, ‘NAUKA’, Alma-Ata. [In Russian.]

fortey, r. a. 1975. The Ordovician trilobites of Spitsbergen. II. Asaphidae, Nileidae, Raphiophoridae and
Telephinidae of the Valhallfonna Formation. Skr. norsk Polarinst. 162, 207 pp., 42 pis.

— 1980. The Ordovician trilobites of Spitsbergen. III. Remaining trilobites of the Valhallfonna Formation.
Ibid. 171, 163 pp., 25 pis.

— 1981. Prospectatrix genatenta (Stubblefield) and the trilobite superfamily Cyclopygacea. Geol. Mag. 118,
603-614, 1 pi.

1984. Global earlier Ordovician transgressions and regressions and their biological implications, 37-50.
In bruton, d. l. (ed.). Aspects of the Ordovician System. Palaeont. Contr. Univ. Oslo, 295.

— and cocks, l. r. m. 1986. Marginal faunal belts and their structural implications, with examples from
the Lower Palaeozoic. Jl geol. Soc. Lond. 143, 151-160.

696

PALAEONTOLOGY, VOLUME 31

fortey, r. a., landing, e. and skevington, d. 1982. Cambrian Ordovician boundary sections in the Cow
Head Group, western Newfoundland, 95-129. In bassett, m. g. and dean, w. t. (eds.). The Cambrian-
Ordovician boundary: sections, fossil distributions, and correlations. Geol. Ser. nat. Mus. Wales, 3, 227 pp.

— and Owens, R. m. 1978. Early Ordovician (Arenig) stratigraphy and faunas of the Carmarthen district,
south-west Wales. Bull. Br. Mus. nat. Hist. (Geol.), 30, 225-294, 1 1 pis.

1987. The Arenig Series in South Wales. Ibid. 41, 69-307, 146 figs,
and rushton, a. w. a. 1980. Acanthopleurella Groom 1902: origin and life-habits of a miniature trilobite.
Ibid. 33, 79-89.

Harrington, H. J. 1938. Sobre las faunas del Ordoviciano Inferior del Norte Argcntino. Revta Mus. La Plata,
Seccion. Paleont., ns, 1, 109-289, pis. 114.

— and leanza, A. 1957. Ordovician trilobites of Argentina. Spec. Pubis Dep. Geol. Univ. Kansas, 1,
276 pp., 140 figs.

hawle, i. and corda, a. j. c. 1847. Prodrom einer Monographie der bohmischen Trilobiten, 176 pp., 7 pis.
Prague.

henningsmoen, g. 1957. The trilobite family Olenidae. Skr. norske Vidensk.-Akad. mat.-nat. Kl. 1957, no. 1,
303 pp., 31 pis.

— 1959. Rare Tremadocian trilobites from Norway. Norsk geol Tidsskr. 39, 153 174, 2 pis.

howell, b. f. 1935. Cambrian and Ordovician trilobites from Herault, southern France. J! Paleont. 9, 222-
238, pis. 22 and 23.

hughes, c. p. 1979. The Ordovician trilobite faunas of the Builth-Llandrindod Inlier, central Wales. Part
III. Bull. Br. Mus. nat. Hist. (Geol.), 32, 109 181.

jaanusson, v. 1959. In Harrington, h. j. et al. Trilobita, 38-540. In moore, r. c. (ed.). Treatise on invertebrate
paleontology. Part O, Arthropoda 1, xix-t-560 pp. Geological Society of America and University of Kansas
Press, New York and Lawrence, Kansas.

jackson, d. e. 1962. Graptolite zones in the Skiddaw Group in Cumberland, England. Jl Paleont. 36, 300-
313.

jaekel, o. 1909. Uber die Agnostiden. Z. dt. geol. Ges. 61, 380-401.

kobayashi, t. 1955. The Ordovician fossils from the McKay Group in British Columbia, western Canada,
with a note on the early Ordovician palaeogeography. J. Fac. Sci. Tokyo Univ. sect. 2, 9 (3), 355-493,
pis. I -9.

lake p. 1906 1946. British Cambrian trilobites. Palaeontogr. Soc. [ Monogr .], Part 1 (1906), 1-28, pis. 1 and
2; Part 2 (1907), 29 48, pis. 3 and 4; Part 4 (1913), 65 88, pis. 7- 10; Part 5 (1919), 89 120, pis. 11 14; Part
13 (1942), 307-332, pis. 44 46; Part 14 (1946), 333-350, pi. 47.

Lazarenko, n. p. 1966. Biostratigraphy and some new trilobites from the Upper Cambrian of the Olenek
Uplift and the Kharaulakh Mountains. Uchen. Zap. nauchno-issled. Inst. geol. arkt., Paleont. Biostratigr.
11, 33-78, 8 pis. [In Russian ]

lu yan-hao and lin huan-ling. 1984. Late late Cambrian and earliest Ordovician trilobites of Jiangshan-
Changshan area, Zhejiang, 45 144, 19 pis. In Stratigraphy and palaeontology of systemic boundaries in
China. Cambrian- Ordovician boundary (7). Nanjing Inst. Geol. Palaeont., Acad. Sinica.

— and qian yi-yuan. 1977. In zhou tien-mei et al. (eds.). Trilobita, 104 266, pis. 33-81. Atlas of fossils of
central and south China. 1. Lower Palaeozoic. Geological Publishing House, Beijing. [In Chinese.]

— 1983. Cambro Ordovician trilobites from eastern Guizhou. Palaeont. Cathay. 1, 1- 105, 12 pis.

— zhou zhiyi and zhou zhiqiang. 1981. Cambrian Ordovician boundary and their related trilobites in
the Hangula region, W. Nei Monggol. Bull. chin. Acad. geol. Sci., ser. 10, 2(1), I 22, pis. 1-3. [In Chinese,
English summary.]

ludvigsen, r. 1982. Upper Cambrian and Lower Ordovician trilobite biostratigraphy of the Rabbitkettle
Formation, western District of Mackenzie. Contr. Life Sci. Div. R. Ont. Mus. 134, 188 pp., 70 figs.
marek, l. 1966. The trilobite family Bohemillacea Barrande, 1872 in the Ordovician of Bohemia. Cas. narod.
Mus. 135, 143 153, pis. 1 and 2 ( = 11 and 12).

maximova, z. a. [maksimova]. 1962. Biostratigraphy of the Palaeozoic of the Siberian Platform, part 5.
Trilobites of the Ordovician and Silurian of the Siberian Platform. Trudy vses. nauchno-issled. geol. Inst.
ns, 76, 215 pp., 18 pis. [In Russian.]

moberg, J. c. 1900. Nya bidrag till utredning af fragan om gransen mallan undersilur och kambrium. Geol.
For. Stockh. Forh. 22, 523-540, pi. 14.

— and moller, h. 1898. Om Acerocarezonen. Ibid. 20, 198-290, pis. 10-14.

molyneux, s. G. and rushton, a. w. a. 1985. Discovery of Tremadoc rocks in the Lake District. Proc. Yorks,
geol. Soc. 45, 123 127 (dated 1984).

RUSHTON: TREMADOC TRILOBITES

697

morris, s. F. 1988. A review of British trilobites, including a synoptic revision of Salter’s monograph.
Palaeontogr. Soc. [Monogr.], 316 pp.

nikolaisen, f. and henningsmoen, G. 1985. Upper Cambrian and lower Tremadoc olenid trilobites from the
Digermul Peninsula, Finnmark, northern Norway. Bull. Norsk geol. Unders. 400, 1 49, 18 figs.
owen, a. w. 1985. Trilobite abnormalities. Trans. R. Soc. Edinb ., Earth Sci., 76, 255-272.
palmer, a. r. 1968. Cambrian trilobites of east-central Alaska. Prof. Pap. US geol. Surv. 559-B, 115 pp.,
15 pis.

peng sanchi. 1984. Cambrian Ordovician boundary in the Cili Taoyuan border area, northwestern Hunan
with descriptions of relative trilobites, 285-405, 18 pis. In Stratigraphy and palaeontology of systemic
boundaries in China. Cambrian - Ordovician boundary (1). Nanjing Inst. Geol. Palaeont., Acad. Sinica.
rasetti, f. 1954. Early Ordovician trilobite faunules from Quebec and Newfoundland. .// Paleont. 28, 581
587, pis. 60 and 61.

Raymond, p. e. 1925. Some trilobites of the lower Middle Ordovician of eastern North America. Bull. Mus.
comp. Zool. Harv. 67, 1-180, pis. 1 10.

robison, r. a. 1982. Some Middle Cambrian agnostoid trilobites from western North America. .// Paleont.
56, 132- 160.

and pantoja-alor, j. 1968. Tremadocian trilobites from the Nochixtlan region, Oaxaca, Mexico. Ibid.
42, 767-800, pis. 97 104.

rose, w. c. c. 1954. The sequence and structure of the Skiddaw Slates in the Keswick- Buttermere area. Proc.
Geol. Tss. 65, 403-406.

rosova, a. v. [rozova] 1963. Biostratigraphic scheme of upper parts of Middle Cambrian and of Upper
Cambrian of north-western part of Siberian Platform, and new Upper Cambrian trilobites of the River
Kulyumbe area. Geol. Geofiz. Novosibirsk. 9, 1-19, pis. I and 2. [In Russian.]

1968. Biostratigraphy and trilobites of the Upper Cambrian and Lower Ordovician of the northwestern
Siberian Platform. Trudy Inst. Geol. Geofiz. sib. Otd. 36, 196 pp., 17 pis. [In Russian.]
rushton, a. w. a. 1982. The biostratigraphy and correlation of the Merioneth Tremadoc Series boundary
in North Wales, 41-59. In bassett, m. g. and dean, w. t. (eds.). The Cambrian Ordovician boundary:
sections, fossil distributions, and correlations. Geol. Ser. nat. Mus. Wales , 3, 221 pp.

1985. A Lancefieldian graptolite from the Lake District. Geol. Mag. 122, 329 333.

— and hughes, c. p. 1981. The Ordovician trilobite fauna of the Great Paxton Borehole, Cambridgeshire.
118, 623-646, 6 pis.

ruzicka, r. 1926. Faune des couches a Euloma du gisement ferrugineux pres de Holoubkov (a Ouzky). Bull,
int. Acad. Sci. Bo he me. for 1926, 1-24, pis. 1-3.

1931. Supplement a la faune des couches a Euloma du gisement metallifere pres de Holoubkov (a

Ouzky). Vest. st. geol. Ust. csl. Repub. 7, cislo 4-5, 1 20, pi. I (18).

salter, j. w. 1866. On the fossils of North Wales. Appendix, pp. 239-363. In ramsay, a. c. (ed. ). The geology
of North Wales. Mem. geol. Surv. Gt Br. 3, viii + 381 pp., 28 pis.

scotese, c. R., bambach, r. K.., barton, c., van der voo, r. and ziegler, a. m. 1979. Paleozoic base maps.

Jl Geol. 87, 217-277.

sdzuy, k. 1955. Die Fauna der Leimitz-Schiefer (Tremadoc). Abb. senckenb. naturforsch. Ges. 492, 1-74,
pis. 1-8.

— 1958. Fossilien aus den Tremadoc des Montagne Noire. Senckenberg. letli. 39, 255-258, pis. 1 3.
shackleton, E. h. 1975. Geological excursions in Lakeland , 127 pp. Dalesman, Nelson.

shaw, a. b. 1951. The paleontology of northwestern Vermont. 1. New Upper Cambrian trilobites. J! Paleont.
25, 97-114, pis. 21 24.

shergold, j. h. 1972. Late Upper Cambrian trilobites from the Gola Beds, western Queensland. Bull. Bur.
Miner. Resour. Geol. Geophys. Aust. 112, 127 pp., 19 pis.

— 1975. Late Cambrian and early Ordovician trilobites from the Burke River structural bell, western
Queensland. Ibid. 153, 251 pp., 58 pis. (2 vols.).

1980. Late Cambrian trilobites from the Chatsworth Limestone, western Queensland. Ibid. 186,
1 1 1 pp., 35 pis.

and sdzuy, k. 1984. Cambrian and early Tremadoc trilobites from Sultan Dag, central Turkey.
Senckenberg. leth. 65, 51 135, pis. I 8.

stormer, l. 1940. Early descriptions of Norwegian trilobites. The type specimens of C. Boeck, M. Sars and
M. Esmark. Norsk geol. Tidsskr. 20, 113-151, 3 pis.

Stubblefield, c. J. 1926. Notes on the development of a trilobite, Shumardia pusilla (Sars). Jl Linn. Soc..
Zool. 36, 345-372, pis. 14 16.

698

PALAEONTOLOGY, VOLUME 31

Stubblefield, c. J. and bulman, o. m. b. 1927. The Shineton Shales of the Wrekin district. Q. Jl geol. Soc.
Lond. 83, 96-146, pis. 3-5.

tjernvik, t. e. 1956. On the early Ordovician of Sweden. Bull. geol. Instn Univ. Uppsala , 36, 109-284,
1 1 pis.

whittard, w. f. 1961. The Ordovician trilobites of the Shelve Inlier, west Shropshire. Pa/aeontogr. Soc.
[Monogr.], Part 6, 197-228, pis. 26-33.

Whittington, h. b. and hughes, c. p. 1974. Geography and faunal provinces in the Tremadoc epoch. Spec.
Pubis Soc. econ. Paleont. Miner. 21, 203-218.

wiman, c. 1902. Studien fiber das Nordbaltische Silurgebiet. Bull. geol. Instn Univ. Uppsala , 6, 10-76, 4 pis.,
map.

winston, d. and nicholls, h. 1967. Late Cambrian and early Ordovician faunas from the Wilberns Formation
of central Texas. J! Paleont. 41, 66-96, pis. 9 13.

xiang li-wen and zhang tai-rong. 1984. Tremadocian trilobites from western part of northern Tianshan,
Xinjiang. Acta palaeont. sin. 23, 399-410, 3 pis. [In Chinese, English summary.]

— 1985. Stratigraphy and trilobite faunas of the Cambrian in the western part of northern Tianshan,
Xinjiang. People's Republic of China , Ministry of Geology and Mineral Resources. Geological Memoirs , ser.
2, 4, ix + 243 pp., 52 pis. [In Chinese, English summary. |
yin gong-zheng and lee shan-ji. 1978. Trilobita, 385-594, pis. 144 192. Atlas of fossils of southwest China.

Guizhou province. 1 . Cambrian to Devonian. Geological Publishing House, Beijing. [In Chinese.]
zhou zhi-yi and zhang jin-lin. 1984. Uppermost Cambrian and lowest Ordovician trilobites of north and
north-east China, 63-163, 29 pis. In Stratigraphy and palaeontology of systemic boundaries in China.
Cambrian-Ordovician Boundary (2). Nanjing Inst. Geol. Palaeont., Acad., Sinica.

Typescript received 29 May 1987

Revised typescript received 7 September 1987

a. w. a. rushton

British Geological Survey
Keyworth

Nottingham NG12 5GG

NEW MATERIAL OF THE EARLY TETRAPOD
ACANTHOSTEGA FROM THE UPPER
DEVONIAN OF EAST GREENLAND

by J. A. CLACK

Abstract. New material of one of the oldest known tetrapods, Acanthostega gunnari , is described: three
skulls, together in one block, in association with postcranial material. This is the hrst postcranial material
to be described for Acanthostega. The skulls show an animal with a broad, closed, denticulated palate in
which the pterygoids meet in the mid-line as in loxommatids and Ichthyostega. The ventrally grooved
parasphenoid resembles that of some osteolepiform hsh rather than that of tetrapods. The basal articulation
is tetrapod-like with well-developed basipterygoid processes. The otic capsules appear to be well ossified and
the braincase tits flat under the skull table, in contrast to the complex facets in Ichthyostega. No synapomorphies
with any particular tetrapod group have been discovered, but one additional character defining all tetrapods
(large ornamented interclavicle) and two defining all neotetrapods (presplenial-anterior coronoid suture,
surangular contributes significantly to margin of adductor fossa) have been identified. The latter two can be
used to establish whether isolated lower jaws belong to fishes or to tetrapods.

The earliest tetrapods yet known have been found in rocks of Upper Devonian (Famennian) age.
They have now been recorded from several continents, including Australasia (Campbell and Bell
1977; Warren el al. 1986), South America (Leonardi 1983), and Eurasia (Lebedev 1984), but by
far the largest number and best-preserved specimens derive from East Greenland. Tetrapods were
first recognized there in 1931 during a series of expeditions led by Lauge Koch. The majority of
described specimens from these expeditions pertain to the genus Ichthyostega , one has been placed
in a second, related, genus Ichthyostegopsis (Save-Soderbergh 1932), while two pertain to a third
genus, Acanthostega (Jarvik 1952).

Ichthyostega and Ichthyostegopsis were first described in a preliminary report by Save-Soderbergh
(1932), who unfortunately died before being able to carry out the work more completely. His
report gave basic descriptions of the skull roofs of several specimens, to many of which he gave
separate specific names. Further information about Ichthyostega was published by Jarvik (1952),
including details of the fish-like tail, the vertebral column, the hindlimb, the unique overlapping
ribs, and new reconstructions of the skull and of the whole animal. The skull was shown to have
many unusual features including apparently advanced ones such as the lack of an intertemporal
and fused postparietals, and primitive ones such as a braincase retaining the ventral cranial fissure
with the otic capsule not underlain by the parasphenoid. Jarvik (1965) gave more information on
the limbs, with reconstructions of the pelvic girdle and the pectoral limb following in 1980. A.
gunnari is known so far only from the skull roof in two specimens. A possible third specimen
mentioned by Jarvik (1952) is not now included in this genus (Jarvik, pers. comm.).

The material to be described here was collected in 1970 during one of a series of expeditions led
by Dr Peter Friend, then of the Scott Polar Institute, now of the Department of Earth Sciences,
University of Cambridge (Friend et al. 1983). The fossils were collected by John Nicholson (Friend
et al. 1976), as a secondary casual activity, the main objective being to draw up stratographic
sections. Fossils from each collecting site were grouped under one ‘lot’ number prefixed G, and
each item was also numbered separately.

Tetrapods were found at three sites during the series of expeditions: G656, G680, and G920.
The latter site, visited on 1 1 August 1970, yielded by far the bulk of the tetrapod remains, consisting

IPalaeontology, Vol. 31, Part 3, 1988, pp. 699-724.|

© The Palaeontological Association

700

PALAEONTOLOGY, VOLUME 31

of many isolated elements, gathered as it was from scree on the mountainside. However, much of
the great value of this material comes from the fact, which I have subsequently discovered, that
many of the items from site G920 fit together to form one composite block. The cranial material
is identifiable as belonging to the poorly known A. gunnari. The associated postcranial elements
in the block may be attributed to this form with reasonable confidence, though not those on
isolated blocks. This material therefore more than doubles the known specimens of this form,
substantially increases our knowledge of its anatomy, and indicates a new locality.

MATERIALS AND METHODS

Nicholson’s site G920 is located on the south-east slope of Stensios Bjerg, and derives from the top of the
Britta Dal Formation. Material from this site consists of both isolated and associated cranial and postcranial
elements, most of which are preserved in a weathered reddish-grey micaceous sandy siltstone, which is
irregularly bedded. A few specimens are from a harder and more finely laminated greyer, but still micaceous,
sandy siltstone, and are clearly from a different bedding plane. Bands of calcite are found both in this and
in the redder rock, several fragments having a calcite lining along one edge.

In most instances, the bone is heavily weathered and preservation is often poor. Dermal bone is usually
split through the middle spongy layer and the outer layer of dermal ornament often lost. Where endochondral
bone has been exposed to weathering, the inner spongy bone is often reduced to a soft caramel-like substance.
In other places it appears that chemical interchange has occurred between the bone and matrix, areas of
apparently rotted bone having become coarse and crystalline, some of which has subsequently weathered to
a powder. The matrix formed by the reddish-grey sandy siltstone is highly variable in character. The outer
layer of weathered rock is usually soft and easily removed mechanically. In other places, the bone is covered
by a thin layer of very fine- and even-grained red haematitic matrix which is so soft as to be removable with
a stiff brush or fine pin. In other places the matrix is hard, coarse, and crystalline with much pink calcite,
which differs little in colour from the outer layer of bone which is slightly browner in tone. The calcite

text-fig. 1. Acanthostega gunnari Jarvik. Diagram of composite block UMZC TI300 to show distribution

of elements. Scale bar, 10 mm.

CLACK: DEVONIAN TETRAPOD FROM GREENLAND

701

crystals often adhere firmly to the bone, making preparation extremely difficult in these places. The bulk of
this matrix had to be removed by careful use of a pneumatic pen or dental mallet, but removal of the final
layer required the use of a very fine, frequently sharpened mounted needle, individual crystals being picked
or scraped off to avoid damage to the bone. The matrix contains many mica flakes, sometimes lying over
the bone, and here they help separation of bone and matrix. Many fragments of broken bone and scutes add
to the difficulty of preparing and interpreting this material.

The material consists of a composite block (text-fig. 1 ) about 280 mm in length containing remains of
three skulls, an isolated premaxilla, a lower jaw, two clavicles, an interclavicle, and a scapulocoracoid. One
skull (skull A, University Museum of Zoology, Cambridge (UMZC) number T1300u-c) (text-fig. 2) consists
of the skull table with both tabular horns complete, part of the interorbital region and portions of the
squamosals. Most of it is exposed in dorsal view, but the surface ornament has been eroded away except on
the tabular horns which were exposed by mechanical preparation. The second skull (skull B, UMZC T1300</
h) (text-fig. 3) is essentially complete except for the suspensorium on each side. A section through the skull
can be seen posteriorly, where the specimen is broken obliquely. The posterior part of the skull roof is
preserved in ventral view on the counterpart of the specimen, while some of the snout region has been
exposed in dorsal view by mechanical preparation. The lower jaws have remained in situ and the skull has
been little disturbed except for flattening.

The third skull (skull C, UMZC T 1 300/) (text-figs. 4 and 5) cannot be certainly identified as Accmthostega ,
but is attributed to that genus on the grounds of its association with the other material in the same block,
and it also has the posteriorly convex margin to the postparietals seen in Accmthostega. It provides an unusual
view: the right side including the cheek and lower jaw has been folded underneath and most of the skull roof
except for the posterior part of the postparietals has been lost. This has exposed what remains of the braincase
and palatoquadrate in dorsal view. Both lower jaws remain attached to their respective quadrates, the left
being more or less completely exposed in external view. The right lower jaw has lost its lateral (external) face
so that the bones of the mesial face are exposed in lateral (internal) view. A section through the anterior part
of the skull is visible where the snout has been lost (text-fig. 5c). This skull is associated with cervical elements
and three ribs.

Other recognizable elements on isolated blocks include three skull table fragments, a frontal prefrontal
unit, part of an articular, portions of dentary and maxilla, an isolated lower jaw and humerus, a clavicle,
two interclavicles, and a pelvic girdle. These cannot be assigned taxonomically at present.

Before preparation, the specimens were photographed and the more important cast in silastic (Silastomer
RTV 9161 ), which was also used to provide a backing during mechanical preparation. Sections were provided
where the calcite lining of the composite block could be removed mechanically and the section polished using
fine-grade carborundum paper. The specimens are now registered as UMZC T 1 29 1 T 1 302.

In addition to the material discovered by Nicholson, I have been able to examine specimens of Ichtliyostega ,
the holotype of A. gunnari (GM A33), the second specimen (GM A85), and two other unidentified skull
specimens (GM A88, GM A90), collected by the Danish Swedish expeditions.

Abbreviations used for institutions: GM, Geologisk Museum, Copenhagen; NRS, Naturhistoriska
Riksmuseet, Stockholm; UMZC, University Museum of Zoology, Cambridge.

STRATIGRAPHY

The Upper Devonian in East Greenland outcrops in a number of localities surrounding Kejser
Franz Josephs Fjord, an area about 500 miles north of the Arctic Circle on the East coast. Outcrops
occur on Yrners 0 on the slopes of Celsius Bjerg, along the slopes of Sederholms Bjerg in
Paralleldal, and around the mountains of Gauss Halvo, including Smith Woodwards Bjerg, Stensios
Bjerg, and Wimans Bjerg. The stratigraphy was described by Save-Soderbergh (1932, 1933, 1934),
Jarvik (Johansson) (1935), Butler (1961), and has been amplified by the work of Friend et al.
(1983). Nomenclature in this paper follows the latter work (Table 1).

The Upper Devonian System in East Greenland can be divided into three major groups. The
Kap Kolthof Group underlies the Kap Graah Group, dating from the Frasnian through to the
Fower Middle Famennian, the latter being equivalent to the Phyllolepis Series of Save-Soderbergh.
These are overlain by the Mount Celsius Supergroup which completes the Upper Devonian strata
of the area. The Mount Celsius Supergroup in turn is divided into two groups, the lower Remigolepis
Group which is equivalent to Save-Soderbergh and Jarvik’s Remigolepis Series, and the upper

702

PALAEONTOLOGY, VOLUME 31
TABLE 1

Friend et al. (1983)

Save-Soderbergh - Jarvik (1935)

CL

C3

O

Gronlandaspis Group

Arthrodire Sandstone Series

600 m

cn

IS)

s_

CL

CD

<u

CL

*Britta Dal Formation

O

L

Upper Reddish Division

o

i/1

CD

in

LO

IS)

IS)

IS)

Wimans Bjerg Formation

CL

CD

CL

CD

Middle Grey Division

o

o

CNJ

CD

C_>

cn

cn

4->

*Aina Dal Formation

E

E

Lower Red Division

O ;

CSL

CSL

E

Kap Graah Group

Phyllolepis Series

o

o

T —

*Tetrapods

Gronlandaspis Group equivalent to Save-Soderbergh and Jarvik’s Arthrodire Sandstone Series. It
is from the former Group that the tetrapods derive.

The Remigo/epis Group consists of three distinct formations which can be recognized over the
whole area, though the three vary in thickness. The lower Aina Dal Formation, equivalent to
Save-Soderbergh and Jarvik’s Lower Reddish Division, consists of red coarse- and medium-grained
siltstones and has yielded a rich fauna including many specimens of Ichthyostega. It reaches a
maximum thickness of 80 m on Gauss Halve, where it passes smoothly into the grey siltstones of
the Wimans Bjerg Formation, equivalent to Save-Soderbergh and Jarvik’s Middle Grey Division.
This is essentially unfossiliferous. Its maximum thickness is 200 m and it passes into the upper
Britta Dal Formation, equivalent to Save-Soderbergh and Jarvik’s Upper Red Division, which
reaches its maximum thickness of 550 m on Stensios Bjerg. This consists of red and grey siltstones
and some red sandstones, and is interpreted by Nicholson and Friend (1976) as representing
dominantly fluviatile channel and floodplain sedimentation. It has also yielded a rich fauna
including Ichthyostega and Accmthostega. The Upper Devonian sequence is terminated by the grey
fine- and medium-grained sandstones of the Gronlandaspis Group, which reaches a maximum
thickness of 600 m. As Friend et al. interpret it this sequence was originally of much greater
thickness but was eroded during the Carboniferous Period. Friend et al. (1983) accept Jarvik’s
(1961) dating of the whole sequence based on the vertebrate fauna, and place the Remigolepis
Group firmly within the Famennian. Spore analysis of rocks from several parts from this sequence
was attempted by Friend et al. (1983) but all samples proved unproductive.

SYSTEMATIC PALAEONTOLOGY

Family acanthostegidae Jarvik, 1952
Diagnosis of family. As for Acanthostega.

Type species. A. gunnari Jarvik, 1952.

CLACK: DEVONIAN TETRAPOD FROM GREENLAND

703

Key to textures used in figures (unless otherwise indicated)

true bone surface

split dermal bone )

) sometimes not separable
natural mould )

matrix

broken endochondral bone
eroded bone

text-fig. 2. Acanthostega gunnari Jarvik. UMZC T1300a c, skull A, dorsal view, with mterorbital region
(exposed in ventral view) reversed and shown as transparent. Scale bar, 10 mm.

Diagnosis. Devonian tetrapod with skull table lacking intertemporal and with cheek-skull table
junction spanned by arrowhead-shaped supratemporal. Tabular with deep embayment and long
laterally developed horn; tabular-squamosal junction smooth. Postparietals relatively long, with
convex posterior margin. Narrow interorbital region. Prefrontal large, excluding lachyrmal from
orbit. Nasals broad anteriorly; ?internasal present. Palate broad, closed, denticulate, small but
evident interpterygoid vacuities, pterygoids meet anterior to cultriform process. Marginal palatal
bones narrow, bearing numerous small teeth but ?no tusks. Parasphenoid grooved in mid-line;
groove broadens between basipterygoid processes. Basipterygoid processes well developed. Otic
capsules heavily ossified; ?roof of braincase closed. Simple abutment of braincase roof on to skull
table; only small facet on tabular for attachment. Ornament groove and ridge, with some tubercular
development; grooves often elongated near bone margins, in regions of growth, though this not
invariable. Lateral-line canals in tubes through bone. Orbits circular to oval. Dentary teeth about
seventy or more; maxillary dentition about forty-six; premaxillary dentition ?about twenty.

DESCRIPTION

Skull

Dermal Skull Roof. The new material substantially confirms and reinforces much of the information published
by Jarvik (1952), but gives little further knowledge of areas such as the snout which were missing from the
original material. It is unfortunate that the suspensorial region, difficult to interpret in the original specimens,
is not represented in the new material, so that the presence or absence of a preopercular cannot be confirmed.
Lacking also is any evidence about the shape and position of the external naris.

The unique horn and embayment, described by Jarvik (1952) in the original material, are major
autapomorphies used to identify the new material as Acanthostega. In skull A the horns have both been
exposed by mechanical preparation and show the unweathered bone surface to be ornamented dorsally (text-
fig. 2). They are more substantial than those in either of the original specimens, both of the latter having
suffered a certain amount of erosion. The holotype tabular horn shows a smooth mesial edge which was
presumably embedded in soft tissue in life as Jarvik suggests, but this is not evident in skull A. Where the
tabular meets the supratemporal and squamosal, it is thickened and is a substantial ellipse in cross-section,
but further distally, where it becomes the tabular horn, it is flattened. The tabular-squamosal suture is simple
and lacks inlerdigitations, the sutural surface of the tabular at this point is seen on the left horn of skull A,
where there is no overlap surface at all for another bone. The lateral margin of the tabular turns mesially
where it would have lost contact with the squamosal to become free tabular horn, but there is no evidence
that it was embayed to correspond to the squamosal tabular embayment of the holotype.

704

PALAEONTOLOGY, VOLUME 31

The question arises as to which of the two embayments of Acanthostega is the homologue of the ‘otic’ or
‘spiracular’ notch of other early fossil amphibians, which lies between the junction of the skull table and
cheek regions. It is usually bounded by the tabular, and sometimes the supratemporal, dorsally, and the
squamosal ventrally. At first sight, the lower of the two embayments in Acanthostega seems to fulfill these
criteria. However, the state of the sutures bounding the tabular and contacting the squamosal and
supratemporal suggest an alternative hypothesis. It is possible that the tabular has in effect ‘grown around’
the site of the original embaymenl, sealing the primitive kinetism found at this point in fishes. Thus the
tabular embayment encloses the notch which may have housed a persistent spiracle, and a second embayment
was produced where the tabular has ‘sprung away’ from the margin of the squamosal to form the horn. The
suture of the tabular and squamosal remained uninterdigitated, betraying its history as part of the kinetic
mechanism, though it is not suggested that there was any movement here. This hypothesis requires more
information on the nature of the tabular-squamosal embayment.

At the anteromesial corner of the embayment the tabular bears a tiny process on the ventral surface, seen
in the counterpart of skull B (text-fig. 3c), which may have been a facet attaching to the braincase. Also in
this specimen, it is clear that the tabular is penetrated by a canal running almost from the posterior margin,
anteriorly, parallel to the mesial edge of the embayment. It can be seen both in section and in ventral view
where some of the underlying dermal layer has been lost (text-fig. 7b). The canal can also be identified on
the left side of the holotype, whereas on its right side, because of the way the bone is preserved, a partial
section through the canal gives the deceptive appearance of a downwardly curving flange.

One of the most striking features in the skull table is the arrowhead-shape of each supratemporal,
manifested particularly in the posterolateral and posteromesial corners, and seen best in an isolated skull
table (text-fig. 6). This character is not as obvious in the original material since the sutures are difficult to
trace, but it is consistent among the new skull table specimens. So characteristic is it that it can be used as
a means of identification of incomplete skull table fragments. The posterolateral corner of the supratemporal
is drawn out into a diminishing process ‘squeezed’ between the tabular and squamosal, until the latter meet
in a butt-joint. This is particularly well seen in skull A, where the lateral margin of the tabular is well
preserved.

The course of the squamosal postorbital suture is rather difficult to establish in the new specimens of
Acanthostega , resulting, apparently, from a substantial overlap on the inner surface between adjacent bones.
Thus, internal and external views give a very different picture from one another and, where the bone is split
horizontally, conclusions about the course of a suture can be quite contradictory. In the isolated skull table
the postorbital appears to be a large bone, with an interdigitating suture with the squamosal at about the
level of the apex of the tabular embayment. The specimen is exposed in internal view, but the bone is split,
and the pattern it reveals is probably that of the external surface. On re-examination, the holotype shows a
similar pattern. In the counterpart of skull B (text-fig. 3c), however, exposed also in internal view but with
the bone here complete, what is apparently a good squamosal postorbital suture defines a much smaller
postorbital, the suture being positioned much further anteriorly than in the isolated example. Sutural overlap
can be seen in the section through the counterpart of skull B, at the tabular postparietal suture. Though
quite clear in ventral view, in section a very thin lamina of bone from the postparietal lies on the ventral
surface of the skull table, and it is the margin of this which is taken for the suture in ventral view (text-
fig. 7a). The margin so formed follows the course which the suture would be expected to take, though no
other evidence of the suture can be seen in the section. The same situation applies to the squamosal-
postorbital suture of this specimen.

In all but one of the new specimens, the skull table is exposed in ventral view. Apart from that on the
tabular, no other facets for support of the braincase have been identified, although the posteriormost parts
of the postparietals are not preserved in ventral view in any specimen. In this respect, Acanthostega resembles
Eusthenopteron and contrasts with Ichthyostega , in which there are complex facets under the whole of the
postparietal. As in many other tetrapods the skull table is thickened in the region of the mid-line of the
postparietals and parietals. Anterior to the parietal foramen in Acanthostega , the growth lines within the bone
form a strongly transverse pattern, manifested as a thickened ridge in complete specimens.

Acanthostega resembled most other early tetrapods in the relatively small size of the otic region, judging
from the proportions of the postparietals and parietals. Ichthyostega , with its apparently rather large otic
region, was much more fish-like in this respect, as noted by Jarvik (1980).

The maxilla is preserved in skull C where it has remained in contact with bones of the palate, even though
the dermal roofing bones are missing. This contrasts with the holotype, in which the maxilla lies apparently
a little detached from the roofing bones. Jarvik (1980) interprets this to mean that it was independent from
the roof. However, a more likely explanation is that it was sutured to them by a flat butt-joint similar to

CLACK: DEVONIAN TETRAPOD FROM GREENLAND

705

text-fig. 3. Acanthostega gunnari Jarvik. a, UMZC T 1 300/', skull B. dorsal view (ornamented part of
squamosal from T1300f). b, 71300/; skull B, ventral view, c, T1300g, counterpart of skull B, skull roof in
ventral view, d, T1300g, isolated maxilla on reverse of specimen, e, T1300/i skull B, section through posterior

part of skull. Scale bars, 10 mm.

706

PALAEONTOLOGY, VOLUME 31

text-fig. 4. Acanthostega gunnari Jarvik. UMZC T 1 300/, skull C, dorsal view
showing braincase, palate, left maxilla and lower jaw, and cervical elements. Scale

bar, 10 mm.

that in embolomeres (Clack 1987), a structure which resists vertical forces during biting. The suture of
the dentary to other bones of the lower jaw was of a similar form and can be seen in the section through
skull C (text-fig. 5c).

An isolated premaxilla is preserved on the same block as the counterpart of skull B (text-fig. 3d).
Surprisingly, it is narrow anteriorly and broadens towards the posterior end which is blunt and rounded.
The anterior end shows an embayment presumably for accommodation of an internasal bone, found also in
Ichthyostega , the loxommatids, and predicted for Acanthostega by Jarvik (1952) from the shape of the
preserved fronlals.

Dermal ornament is only preserved where it has not been exposed to weathering and has been prepared
out mechanically. This includes areas on the frontals, nasals, and squamosal of skull B, and on the
postparietals of skull C. Here it shows some difference from that of the second original specimen (GM A85)
which is the only other specimen in which it is preserved. In the former, as in Ichthyostega , there are strongly
radiating grooves and ridges present, the ridges often bearing raised tubercles, in contrast to the more
‘honeycomb’-like arrangement of pits in A85.

Specimen A85 also differs in other ways from the majority of Acanthostega specimens. The posterior
margin of the skull table between the tabular embayments is less markedly convex than in the holotype, and
has a ‘squared off’ appearance. In those new specimens in which the posterior margin of the skull table is
complete, the corners of the tabulars are gently rounded, and the posterior convexity is less marked than
in the holotype. In A85 the postparietals appear relatively shorter than in other specimens altering the
proportions of the skull table. Though the skull table width between the tabular embayments is roughly
similar, the distance from the apex of the tabular embayment to the orbit is a little shorter in A85. It is also
broader between the orbits. The differences cannot be taken to be taxonomically significant at this stage,
since both possess the tabular horn and embayment definitive of A. gunnari. They must be regarded as
individual variation unless discovery of further specimens shows otherwise.

CLACK: DEVONIAN TETRAPOD FROM GREENLAND

707

max tooth

text-fig. 5. Acanthostega gunnari Jarvik. UMZC T1300/. Skull C. a, reverse of
specimen, showing right side of cheek, right lower jaw. b, sections through anterior
part of skull, c, sections through left lower jaw. Scale bars, 10 nun.

The new specimens show some size variation. The composite block contains two skull tables of identical
size. These are significantly smaller than the holotype. An isolated specimen in which the skull tabic and
horns are complete is intermediate in size between the holotype and A85. It has relatively short postparietals
but resembles the holotype in interorbital width. Of two further isolated specimens, indentified on supra-
temporal shape, one is similar to the holotype and the other representative of by far the largest individual.
As in other amphibians ( Proterogyrinus Holmes 1984; Archeria , pers. obs.) the size of the parietal foramen
varies unpredictably in different individuals.

Lateral-line canals are occasionally discernible in Acanthostega , as in Ichthyostega , running in tubes through
the bones. They are difficult to detect in complete specimens, but are often more obvious in eroded ones
where they can be seen in section, or as substantial canals infilled with matrix, or as a series of pores (text-
figs. 3a, 4, 5a). They have been traced on the nasal, frontal, postfrontal, jugal, squamosal, and lower jaw.
The canals and pores are difficult to distinguish from a second system which also leaves evidence of superficial
foramina.

The dermal bones of Acanthostega have a middle layer penetrated by a complex interconnecting system
of canals and tubules, which is responsible for the poor preservation of the bone. The bone usually splits
through this weak layer, leaving the denser inner and outer layers on part and counterpart. Seen in section,
the system of tubes and canals produces a network, in places so cavernous as to appear more space than
bone. It may indicate that the bone was highly vascularized. The tubes are linked in places to pores on
the outer and inner bone surfaces (text-fig. 7c). Where the outer ornamented layer has been removed, the

708

PALAEONTOLOGY, VOLUME 31

text-fig. 6. Acanthostega gunnari Jarvik. UMZC T1299. Isolated
skull table. Stipple, matrix; dermal bone, split. Scale bar, 10 mm.

remnants of large vacuities can be seen, often in consistent places in different skulls, for example in the
supratemporals of skulls A and B and the isolated skull table. The canal in the tabular appears to be part
of this system. Without more and thinner sections of better preserved bone, it is not possible to elucidate the
relationships of this pore system to that of the lateral-line systems. At first sight it resembles that described
by Bystrow (1947) for Benthosuchus, though with a much more complex tube and pore system and without
a rete vasculosum. Since the clavicle also shows the canals and pores, some connection with the vascular
system is more likely than with the lateral line system.

Skull A shows a steep angle between the cheek and the skull table, rather greater than that in the holotype,
but which is the more natural is hard to assess.

The right side of skull B shows an almost undistorted orbit which is effectively circular (text-fig. 3a). The
right orbit of the holotype, by contrast, is somewhat elongated anteroposteriorly. Whether this difference is
the result of the larger size of the holotype, or to its being compressed, is not certain.

Palate. The palate is visible in ventral aspect in skull B (text-fig. 3b), in dorsal aspect in skull C (text-fig. 4),
in section through the posterior part (skull B) (text-fig. 3e), and anterior part (skull C) (text-fig. 5b).

There is a broad, almost closed, palate as in Ichthyostega , but with clear though narrow interpterygoid
vacuities, bordered by the thickened mesial margins of the pterygoids, lying on either side of the parasphenoid.
There was clearly no contact between the pterygoids and the parasphenoid at this point. This contrasts with
the description which Jarvik (1980) gives of Ichthyostega , in which there are only tiny vacuities rather
anteriorly placed at the front of the parasphenoid. Elsewhere, he figures the pterygoids as meeting the
parasphenoid. However, Save-Soderbergh (1932, pis. 4 and 8; pers. obs.) shows clearly that at least in some
specimens of Ichthyostega , narrow interpterygoid vacuities did exist beside the cultriform process. Beyond
the anterior end of the parasphenoid, the pterygoids met in Acanthostega , and may either have sutured or
simply abutted each other. Lateral to the thickened mesial margins, the pterygoids are grooved in ventral
view, especially posteriorly. Both the groove and the ridge fade as they pass anteriorly.

It is not possible to distinguish between the pterygoid and epipterygoid either around the basal articulation
or on the quadrate ramus, though it is presumably the epipterygoid portion which forms the region
accommodating the basipterygoid process. This can be seen in ventral view in skull B, and is in essence like
that of other early tetrapods with a peg and socket arrangement (text-fig. 3b). Just anterior to the basipterygoid
processes, the mesial margin of the pterygoid turns laterally through almost a right angle to form a posteriorly
facing ledge. It is against this which the basipterygoid processes appear to articulate, but this could result
from compression having forced the pterygoids somewhat apart. The margin is then scooped out into a
socket to accommodate the tip of the basipterygoid process. The whole area surrounding the socket is
thickened and the socket itself is bordered by a lip.

Posterior to the basal articulation, the quadrate ramus produces its ascending ramus, seen in section in
skull B and in dorsal view in skull C. This was a thin sheet, much crushed in skull C, though in skull B, on
the left side where the section is more anterior, the ascending ramus has remained intact. It reaches almost
to the skull roof, where its dorsal margin is somewhat thickened. On the left it has been folded over and lies
at a narrow angle to the horizontal (text-fig. 3e).

CLACK: DEVONIAN TETRAPOD FROM GREENLAND

709

text-fig. 7. Acanthostega gunnari Jarvik. UMZC T1300/! Sections through
dermal skull roof, a, through postparietal/tabular junction to show sutural
overlap; b, through tabular to show canal; c, through tabular to show tube
system. Scale bar, 10 mm.

At the level of the basal articulation, in dorsal view in skull C, the thickened mesial margins of the pterygoid
rise smoothly into vertical buttresses, where presumably they incorporate the epipterygoids and form the
columellae cranii (text-fig. 4, ‘col cran’)- That on the left shows a smooth, rounded tip. On the right side of
skull C, the columella cranii has been pushed laterally so that its mesial face is exposed, and a patch of
unfinished bone at its base may represent part of the recess accommodating the basipterygoid process, though
it provides no useful detail.

The lateral margin of the subtemporal fossa has been exposed in skull C, and is robust and thickened. No
muscle scars are apparent. The rounded margins of the fossa strongly suggests that the quadrate ramus of
the pterygoid did not project below the level of the jaw margin as it does in some anthracosaurs such as
Palaeoherpeton (Panchen 1964) and Proterogyrinus (Holmes 1984). In the section provided by skull B, this
region of the pterygoid lies almost horizontal, though this skull is much compressed.

Most of the visible palate is formed by the pterygoids, dcnliculated on the ventral surface as in most other
primitive tetrapods, but not described for Ichthyostega. The apparent absence of denticulation in Ichthyostega

710

PALAEONTOLOGY, VOLUME 31

may be simply a result of the type of preservation in which the true bone surface is rarely exposed. Small
denticles would easily be missed when the bone splits through the spongy layer.

The marginal palatal bones have not been exposed in skull B, and only broken remnants remain in skull C.
The latter does, however, show them in section at about the level of the posterior part of the palatine. This
shows that at least the palatine overlapped the pterygoid internally to a marked degree, but that little would
have been exposed in ventral view (text-fig. 5b). This contrasts with Ichthyostega , in which the marginal
palatal bones are broad elements in ventral view.

The left quadrate of skull C is visible in section and is a substantial element with a considerable dorsal
component. The posteroventral margin bears an embayment lying above the retroarticular process of the
lower jaw, and was perhaps the site of attachment of a joint-stabilizing ligament (see below).

Parasphenoid and braincase. In skull B the braincase is visible in ventral view and in the oblique section which
passes through the otic region (text-fig. 3b, e). A dorsal view of the much disrupted braincase of skull C is
available where the dermal roofing bones have disappeared (text-fig. 4).

The parasphenoid is a long tapering element reaching anteriorly to a point about level with the front of
the orbit. It does not contact the pterygoids, nor does it appear to continue above the point at which the
latter meet each other. It is strongly ridged in the mid-line anteriorly, except for the first few millimetres, but
as it passes back the ridge divides into two, enclosing a deep groove. The ridges diverge posteriorly for most
of their length, but just anterior to the basal articulation they converge, and meet in a smooth curve just
posterior to the basal articulation.

This form of parasphenoid has not been described in any other tetrapod. It most closely resembles that in
some specimens of Eusthenopteron (e.g. NRS P6849a, pers. obs.) and Megalichthys (S. M. Andrews, pers.
comm.). In these, however, the region between the ridges is denticulated, and pierced by a persistent
hypophyseal foramen. In Acanthostega , the floor of the groove does not appear to be lined with periosteal
bone, and is extremely difficult to prepare. Thus not all the matrix lying between the ridges has been removed.
However, as far as it has, there is no evidence of either denticulation, or of a foramen. It is possible that the
ridges represent the margins of a large gap in the dermal parasphenoid, with the floor of the chondrocranium
visible above it. The hypophyseal fenestra appears to have closed, whereas the ossification of the parasphenoid
was still incomplete. This could represent the retention of an embryonic condition, if the parasphenoid ossifies
from paired centres as it does in Sphenodon and Lepidosiren (de Beer 1937). In other early tetrapods the
parasphenoid is convex, usually with the strong mid-line ridge of the processus cultriformis in the hypophyseal
region, and nothing is known about its development.

The parasphenoid sheathes the basipterygoid processes as in other tetrapods, clearly separated from the
more medial regions by smooth periosteal bone, but not by conspicuous carotid grooves as they are for
example in anthracosaurs ( Palaeoherpetron , Panchen 1964; Eoherpeton, Panchen 1975), and runs back from
the basal articulation on either side. Just posterior to the point where the ridges converge, however, the bone
is strongly depressed into a median concavity, apparently natural, but with the periosteal bone having a
broken edge. If periosteal bone were present covering this concavity in life, it must have been very thin and
thus not preserved. Alternatively, it was missing altogether. I am sufficiently confident of my preparation
technique to believe that had it been preserved, it would have been found. Only further specimens could
confirm the condition, but the implication of this specimen is that in Acanthostega , like Ichthyostega (Jarvik
1980; pers. obs.), the parasphenoid did not grow back to underlie the whole of the otic region. Thus
Acanthostega would be only the second tetrapod to display this feature, otherwise only seen in primitive or
paedomorphic fish.

Among the tetrapods, Crassigyrinus (Panchen 1985) appears most similar to Acanthostega in this region.
In this animal, there was a large triangular concavity between and posterior to the basipterygoid processes.
It is in rather a different position relative to that of both the groove on the mid-line of the parasphenoid and
the more posterior concavity of Acanthostega , and it is not clear to which of these that in Crassigyrinus
would be homologus.

The basipterygoid processes are tetrapod-like in being relatively large structures projecting laterally from
the side-walls of the braincase. The articular faces lie with their anteroventral margins at approximately right
angles to the parasagittal plane, but the shape of the articular surfaces is not known.

Both skulls B and C indicate that the otic region of the braincase was well ossified. Although skull C is
much disturbed, there are clearly solidly ossified units which are best explained as otic capsules, though they
are not interpretable in detail.

From skull B the section shows endochondral bone lying beneath the dermal bones, forming an ossified
roof to the braincase. The underside of the skull table shows no significant facets attaching to the braincase,

CLACK: DEVONIAN TETRAPOD FROM GREENLAND

711

so that the roof of the braincase would have made full but unsutured contact with the dermal skull roof over
its whole surface. The situation is directly comparable to that in fishes such as Eusthenopteron (Jarvik 1980).
It is in direct contrast to that in Ichthyostega , in which complex facets lie beneath the postparietal for
attachment of the otic region, though the otic region itself is poorly ossified and difficult to interpret in the
conventional pattern of either fishes or tetrapods (Jarvik 1980, pcrs. comm.; pers. obs.). Most other tetrapods
in which the otic capsule is known to have an ossified roof, such as the loxommatids (Beaumont 1977),
Eoherpeton (Smithson 1985), Pholiderpeton (Clack 1987) have more or less well-developed facets, especially
on the tabular, for attachment of the braincase, in addition to smooth contact between the surfaces of
braincase and skull table.

Laterally, the endochondral bone of the braincase roof descends to form the side wall, presumably of the
otic capsule, with periosteal bone lining both lateral, ventral, and some of the mesial surface, seen on the left
side (text-fig. 3e). This separates the upper part of the braincase wall clearly from the more ventral parts,
presumably formed from the basioccipital, and indicates the presence of a fenestra of some kind at this point.
There is not enough evidence to describe this as a fenestra ovalis, though it is in about the expected position
for one.

It has been suggested (Jarvik 1952) that the tabular embayment might represent an excavation of the skull
roof lying above the equivalent of the fossa bridgei in the braincase. In Eusthenopteron , the fossa bridgei
perforates the posterior wall of the otic-occipital unit, separating the paroccipital processes from the body
of the braincase. Laterally the paroccipital processes contract the skull roof under the tabulars (terminology
of Westoll 1943). Therefore, if the tabular embayment of Acanthostega is a dorsally open fossa bridgei, some
contact between tabular and braincase would be expected lateral to the embayment. However, judging from
the section afforded by skull B there appears to be none, with the embayments purely a character of the
dermal skull roof. Other possible explanations for them are either that they were the site of attachment of
axial musculature, developed in association with the elaboration of the tabular horn, or that they housed a
persistent spiracle, as has been postulated for the 'otic notch’ of Crassigyrinus (Panchen 1985).

Beneath the otic region, the basioccipital region can be seen as paired convex areas of endochondral bone
with periosteal lining present laterally but fading to disappear in the mid-line. As described above, it is
uncertain whether its total absence was natural or not. There appears to be no certain endochondral bone
at this point in the mid-line, though it is difficult to distinguish from matrix, but its absence would accord
with the presence of a persistent notochord running through the basioccipital as in Ichthyostega.

Lower jaw

Two skulls from the composite block have lower jaws in articulation. A further lower jaw specimen is
associated with a humerus (see below) but cannot be attributed to Acanthostega. It is poorly preserved and
offers little significant detail.

The left side of skull C provides the best-preserved lateral face of the Acanthostega jaws, though it is
incomplete and the bones a little disarticulated anteriorly. The pattern of bones is that typical of a primitive
tetrapod as far as can be ascertained. In one respect, however, it differs from the published account of
Ichthyostega. In this form, Jarvik (1980) figures the dentary as running back to contract the articular, as it
does in Eusthenopteron , but in no other described tetrapod. In Acanthostega , and also in the isolated lower
jaw, the dentary terminates at about the mid-point of the adductor fossa, so that the surangular contributes
to the margin of the fossa (text-fig. 4).

The dentary suture with the underlying bones (presumably coronoids, though none is well enough preserved
to merit description) takes the form of a smooth shelf, a narrow flange descending laterally to meet the
splenials (text-fig. 5c).

The lower jaw is not exposed in mesial view in any specimen, but the mesial components are exposed in
lateral view on the right side of skull C. This shows clearly that the prearticular is a very large bone, as it is
in Ichthyostega , and it passes as far anteriorly as the jaw is preserved (text-fig. 5a). It has a thickened ridge
around the adductor fossa presumably for insertion of adductor musculature. The lower border is missing,
precluding description of the suture with the splenials and the state of any Meckelian fossae. Portions of the
disrupted coronoids lie along the dorsal border of this element.

The articular is exposed where the lateral components of the lower jaw are missing and it passes anteriorly
to about the level of the middle of the adductor fossa. The articular surface is not exposed in any specimen.
Posteriorly the surangular wrapped around the articular leaving none exposed dorsally as far as preserved.
Both the left lower jaw of skull C and the isolated specimen show a small retroarticular process on the
surangular, which may well have attached by a ligament to the quadrate to stabilize the jaw-joint as in
Proterogyrinus (Holmes 1984).

712

PALAEONTOLOGY, VOLUME 31

Dentition

Marginal teeth are preserved best in skull C where they have been exposed by preparation. As in the holotype,
they are almost even in size, though diminishing towards the rear of the row. They are simple cones, slightly
recurved at the tips, and of oval cross-section with the long axis orientated bucco-lingually. Sections show
that there was infolding of the enamel at the root of each tooth, but not in the exposed crown. Maxillary
and dentary teeth show few differences, except for the slightly larger size of dentary teeth seen in skull C.

Tooth counts are difficult to estimate since the dentigerous bones of skull C are incomplete, and in skull
B the maxillae are missing while many of the dentary teeth are missing or obscured by matrix.

The teeth in the maxilla of skull C, as exposed by preparation, apparently alternate regularly with spaces,
while those in the dentary are in places closely spaced. In skull B, where visible, the teeth are also very closely
packed, with ten to thirteen teeth per centimetre. A conservative estimate of the dentry tooth count, given a
dentary length of 7 cm, would be about seventy. This is rather more than the maxillary count of the larger
holotype, which is about forty-six (including spaces). The significance of this must await the discovery of
further specimens.

The isolated premaxilla (admittedly only tentatively assigned to Acanthostega ), shows remains of nine teeth
with spaces for a further seven or eight. A premaxilla with a total of around twenty teeth would account for
the difference between the dentary count of skull B and the maxillary count of the holotype.

Coronoids are not well represented in the specimens from site G920 and there is no firm evidence of
coronoid teeth. Skull C shows a section of the left lower jaw in which a possible coronoid tooth is preserved
(text-fig. 5c), but this could be a broken and displaced fragment of dentary tooth.

There are dentigerous fragments among the isolated specimens from G920, of which some show closely
spaced teeth and some in which teeth alternate regularly with spaces. These and the identified specimens are
in accord with the studies of Rocek (1986), in which both replacement patterns can occur in both
Eusthenopteron and Ichthyostega.

max

text-fig. 8. Greenland Geological Survey specimen GM
A88, section through right dentition, in dorsal view.
Scale bar, 10 mm.

Palatal teeth are not exposed in skull B, but are visible in skull C on the right side and in section. They
are uniformly small, much smaller than the maxillary teeth, and on the exposed length there are about
twenty-seven. This arrangement of palatal teeth is unusual among early tetrapods. Typically, the vomer,
palatine, and ectopterygoids carry large tusks, often occurring in a pair in which one tusk is functional, the
other being represented by a replacement pit. Loxommatids (Beaumont 1977), and the early anthracosaurs
Eoherpeton (Panchen 1975) and Greererpeton (Smithson 1982) all show this pattern, and it is also found in
osteolepiform fishes such as Eusthenopteron (Jarvik 1980). However, in the latter case, the palatal bones also
carry a row of small toothlets lateral to the tusks, similar in number and arrangement to the toothlets seen
in Acanthostega. What cannot be stated with certainty at this stage is that Acanthostega did not also carry
a more mesial tusk-row. There is no evidence of it in the section, but it remains possible that the section
failed to pass through such teeth on either side. However, the small ventral exposure of the lateral palatal
bones which the section reveals suggests that the small toothlets were the only teeth present.

Jarvik (1980) states that Ichthyostega also lacked palatal tusks, and in the figures given by Save-Soderbergh,
only the vomer consistently shows teeth at all. Jarvik’s reconstructions show a row of small teeth running
the length of the marginal palatal bones, in this case about six on the ectopterygoid, seven on the palatine,
and four on the vomer. Clearly this is different from the pattern in Acanthostega. However, among the
specimens from G920 is an isolated tooth-bearing element in which one large tusk and a tusk pit is followed
by four smaller teeth. On current evidence it belongs neither to Acanthostega nor to Ichthyostega.

Specimen GM A88, collected in 1947 by the Danish Swedish expeditions, from the south side of Celsius
Bjerg, shows the natural mould of a denticulated palate in which the marginal dentition is still present,
exposed in dorsal view sectioned across the tooth roots (text-fig. 8). On the reverse side of the specimen, the

CLACK: DEVONIAN TETRAPOD FROM GREENLAND

713

A

B

C

text-fig. 9. Interclavicles, a, UMZC T1293, isolated specimen, b, T1300<7, b,
associated specimen, c, TI292, isolated specimen. Scale bar, 10 mm.

lower jaws are almost in life position. This clearly shows a palatal formula in which there are both tusks and
smaller teeth on ectopterygoid, palatine, and vomer, though the vomerine teeth are not well preserved. The
palatal tooth formula would be expressed thus: vomer2(2 + ) palatine(2)2 ectopterygoid(2)2 + (6) in which
bracketed numbers indicate small teeth, unbracketed, tusks.

The maxillary teeth of this specimen are likewise exposed in section across the roots, and in the whole
length of the maxilla there are nine teeth preserved with spaces for a maximum of twelve. The anterior teeth
are much larger than the posterior ones. The tooth row is about 7 cm, about the same as that of skull B.
There are perhaps a maximum of seven teeth in the premaxilla. Eighteen teeth are exposed in the dentary,
with spaces for a further eight, unless the teeth are actually alternating with space, which does not seem to
be the case. Thus the complete marginal tooth count for this specimen would be about twenty-six to twenty-
eight per side.

In summary, specimen A88 is quite different in tooth formula from Acanthostega , and also from Ichthyostega
as described by Jarvik. It is possible that this unknown form is also present at site G920, and contributed
the isolated palatal clement described above. It represents a third, as yet unnamed and undescribed species
of tetrapod from the Upper Devonian of East Greenland.

Specimen GM A90 from Wimans Bjerg appears to have a similar dentary tooth count to skull B in a tooth
row of comparable size, and might be attributable to Acanthostega , though it is associated with an
ichthyostegan type of clavicle.

Pectoral girdle

Interclavicle. Three interclavicles are preserved (text-fig. 9). One is closely associated with skull A in the
composite block and may confirm the identity of the two isolated elements from the same site. All
three interclavicles are kite-shaped and resemble those of the anthracosaurs Pholiderpeton (Clack 1987),
Proterogyrinus (Holmes 1984), and the temnospondyl Dendrerpeton (Carroll 1967). They are quite different
from that of Ichthyostega which has a long parallel-sided posterior stem very like that of Seymouria (White
1939). This suggests a different adaptation of the pectoral girdle from that in Ichthyostega. Kite-shaped
interclavicles are more often found in aquatically adapted animals and long-stemmed ones in more terrestrially
adapted ones, though the correlation is not invariable (Clack 1987). Unfortunately, none of the Acanthostega
interclavicles has an adequately preserved external (ventral) surface, so that neither the form of the ornament
nor the region of clavicular overlap can be ascertained. It has been assumed that the broader portion of the bone
would have been anteriorly placed as in embolomeres, rather than the more tapering portion as in colosteids. The
largest specimen is preserved with its internal (dorsal) surface moderately well preserved and this is smooth
and featureless.

714

PALAEONTOLOGY, VOLUME 31

text-fig. 10. Clavicles, a, UMZC T1300r/, e, associated specimen, b,
T 1294a, b , isolated specimen. Sections through stems figured to right
of specimen, mesial surface figured uppermost. Scale bar, 10 mm.

Clavicle. Three clavicles are preserved, two associated with the composite block. One of these has the blade
preserved chiefly in section, with a little of the stem visible, but it supplies little useful information. The
second shows most of the blade and a little of the stem. The third is on an isolated block and is complete
except for the tip of the stem. These two clavicles are rather different from one another (text-fig. 10).

In that associated with the composite block (text-fig. 10a), the angle between the anterior and posterior
margins is about 60°. The base of the stem is supported by a stout buttress internally, with a smooth groove
running up the anterior margin, and the section available reveals that the posterior margin was also grooved.
If the true mesial edge is as preserved, the blade would have been a triangle with its posteromesial edge a
right angle. The angle between the anterior and posterior margins of the isolated example (text-fig. I 0b) is
about 40 , giving the blade the shape of an isosceles triangle. Its stem appears rather slender, judging from
the section available and though it is in hard crystalline matrix and difficult to prepare, no evidence of a
groove along the anterior margin can be found.

This isolated element compares closely with that illustrated by Jarvik (1980) for Ichthyostega, and may
indicate that this genus was also present at site G920. The associated example may be assignable to
Acanthostega.

Cleithrum. A cleithrum has not been positively identified, but a bone associated with skull C (text-fig. 4) may
represent one. A long narrow bone lies along the preserved margin of the right quadrate ramus of the
pterygoid, its free end eroded, the other obscured by possible braincase elements. The bone preservation
suggests endochondral rather than dermal bone, but if it is not a cleithrum, the bone is not indentifiable at
present. The bone is an almost parallel-sided strut, with a deep groove along the dorsally exposed face, which
tapers out as the bone runs forward beneath other parts of the skull.

Scapulocoracoid. This is preserved in association with skull C, exposed in lateral view, the anterior margin
obscured by overlying bones (text-fig. 11). Given the lack of disturbance of other postcranial elements
associated with skull C, this bone can be assigned to Accmthostega with moderate confidence. The ventral
margin, having been poorly ossified in life, becomes increasingly difficult to distinguish from matrix and has
not been completely exposed. A section passing through the scapular region and the posterior part of the
glenoid shows that the bone below the glenoid is very thin.

Like those of most other early tetrapods, this scapulocoracoid shows a substantial ossification of the
scapular region, though it is narrower than most. It contrasts with that of Ichthyostega in which no
endochondral scapular region is found, its place being occupied by the large dermal cleithrum. The posterior
margin of the scapulocoracoid curves strongly and smoothly to form almost a full semicircle, similar to that
seen in the embolomere Pholiderpeton (Clack 1987). It is thickened especially in the supraglenoid region, but
no supraglenoid foramen, such as is usually present in early tetrapods, has been found in the exposed part.
There is a very small foramen situated beneath a curving ridge running anteroposteriorly across the bone at
approximately the level where the scapular region merges into the coracoid region (text-fig. 1 1). It is unlikely
to be equivalent to the supraglenoid foramen of other tetrapods.

CLACK: DEVONIAN TETRAPOD FROM GREENLAND

715

text-fig. 11. Scapulocoracoid, UMZC T1300/, with
section through bone, associated with skull C, orien-
tation uncertain. Scale bar, 10 mm.

tubercle

?supracor for

The orientation of the glenoid is unknown, as is its shape and surface form. At its anterior end, the glenoid
is supported by the stout supraglenoid buttress which forms a tubercle at the anterodorsal corner of the
glenoid. A thin flange of bone runs above the dorsal margin of the glenoid as far as preserved. Just
anterodorsal to this tubercle there is another small foramen, possibly equivalent to the supracoracoid foramen
of other early tetrapods (for example, Archeria , Romer 1957), but in a rather different relative position (text-
fig. 11).

In summary, though there are differences in detail between this scapulocoracoid and that of other tetrapods,
as far as preserved it is much more typical of the tetrapod pattern than is that of Ichthyostega.

Other postcranial elements associated with skull C

Fragments of three ribs lie in association with skull C, approximately in life position, but very little information
can be gained from them. One shows a flange developed on the anterodorsal margin, but it is very different
from the massive overlapping ribs developed even in the cervical region of Ichthyostega (Jarvik 1952).

There are cervical elements associated with skull C, again more or less in life position, but the preservation
makes interpretation very difficult (text-fig. 4). One element may be an atlas arch, another a pro-atlas (or
perhaps a disarticulated exoccipital). Two slender spines (probably a pair) were present (one now removed
and preserved separately), one on each side of the vertebral column, which may have been atlantal ribs.
Atlantal ribs are not usually found in early tetrapods, and these would represent a primitive feature.

Beneath skull B lies a very thin curved bone. It has blunt ends and is featureless. It cannot be identified
as belonging to any known fish, and may be interpreted as part of the hyoid apparatus or other parts of a
vestigial gill support system.

Numerous scutes lie in the composite block, particularly associated with skull C. They are narrowly oval,
with a pronounced ridge along one edge which varies in height among the scutes.

Isolated humerus

A poorly preserved humerus (text-fig. 12a) is associated with a lower jaw from site G920, but attribution of
either to Acanthostega cannot be made at this stage. However, it will be described because it shows some
differences from that described by Jarvik (1980) for Ichthyostega. All that remains of the bone substance is
the internal surface of its thin perichondral lining seen in ventral view. The rest of the outline is preserved
as a natural mould which renders little detail. There is no evidence on the surface of the radial condyle,
situated ventrally in Ichthyostega , though since the outer layer of bone is gone, this is not conclusive evidence
of its absence here. Nothing useful remains of the other articular surfaces.

The bone is kidney-shaped, with the entepicondyle arising in a gentle curve from the shaft of the bone, at
an even more obtuse angle than in Ichthyostega. There may have been some distortion during diagenesis,
since the entepicondyle lies almost in the same plane as the shaft of the bone. Some degree of torsion between
the two would normally be expected in a primitive tetrapod humerus, as in Ichthyostega. It has an anterior
flange, as in the humeri of primitive tetrapods such as Proterogyrinus and Greererpeton (Holmes 1980), in

716

PALAEONTOLOGY, VOLUME 31

?obtur for

text-fig. 12. a, humerus, UMZC T1295, untextured portions represented by thin
shell of eroded endochondral bone, b, pelvic girdle, T 1 29 1 . Scale bar, 10 mm.

that of Ichthyostega as reinterpreted by Panchen (1985), and apparently in that of Tulerpeton (Lebedev 1984).
The ectepicondyle is unfortunately not visible.

The humerus possesses an entepicondylar foramen situated in the usual place for tetrapod humeri, and it
also shows accessory foramina. There are two foramina equivalent to those labelled ‘d’ by Jarvik (1980) in
the humerus of Ichthyostega and also in that of Crassigyrinus (Panchen 1985), but otherwise unknown in
tetrapods, and one equivalent to the ‘c’ foramen in Ichthyostega which is not found in Crassigyrinus. In
Ichthyostega , the ‘d’ foramina lie either side of a ridge which runs obliquely across the bone from Jarvik's
‘process 6’ about half-way along the length of the bone, to terminate at the posteromedial corner of the
entepicondyle. In the humerus from G920, a ridge, which is probably equivalent, runs down from the head
of the bone parallel with the shaft, and merges into the margin of the entepicondyle. This appears more
similar to the position of the ‘d’ foramina in Crassigyrinus than in Ichthyostega.

This humerus, though unidentified and poorly preserved, is significant for two reasons. First, it shows the
humerus of a second genus of tetrapod from the Upper Devonian of East Greenland, other than Ichthyostega.
Jarvik (1952) mentioned the existence of an 'Eryops- like’ humerus in the material collected from East
Greenland during the Danish Swedish expeditions, but he does not now believe this to be so (pers. comm.).
Secondly, this humerus shows that the primitive foramina found in the humerus of Eusthenopteron (Andrews
and Westoll 1970) are now known in at least three species of primitive tetrapod.

Isolated pelvic girdle

Like the humerus, this element (text-fig. 12b) cannot be attributed to Acanthostega. but will be described
because it too shows substantial differences from that of Ichthyostega (Jarvik 1980). The left half of the girdle
is exposed in lateral view, and is preserved more or less intact. The anterior and ventral margins are incomplete
and were probably poorly ossified in life. The tip of the postiliac process has been broken off. It is not
possible to be sure whether the element was ossified as a unit or as three separate ossifications, since there
are breaks across the regions where these sutures might be expected.

The ilium was well ossified and has a substantial postiliac process directed posteriorly, with its dorsal
margin at an angle of approximately 25° to the ventral margin of the element. This contrasts with Ichthyostega
(Jarvik 1980) in which these two margins are almost parallel. The process broadens distally, and the section
available is a narrow oval orientated dorsoventrally. A more significant difference from the ilium of
Ichthyostega is in the complete absence of an iliac crest. Instead, the dorsal margin slopes anteroventrally,
and two very slight processes arise above the base of the postiliac process. These may indicate where the
sacral rib attached, though without an internal view, it is impossible to be sure. In this respect this pelvic
girdle resembles that of temnospondyls, such as that attributed to Dendrerpeton (Carroll 1967), and those of
Amphibamus (Carroll 1964) and an as yet undescribed specimen from the Lower Carboniferous of Scotland
(UMZC T 1 26 1 ). The girdle attributed to Baphetes (the ‘Pictou Girdle’: Watson 1926; Panchen 1970) also
apparently lacked an iliac crest. Microsaur pelvic girdles vary greatly, some with iliac crests (e.g. Ricnodon)
and some without (e.g. Hyloplesion) (Carroll and Gaskill 1978). All known anthracosaurs, such as

CLACK: DEVONIAN TETRAPOD FROM GREENLAND

717

Proterogyrinus (Holmes 1984) and Eoherpeton (Smithson 1985), have a large iliac blade arising dorsally and
in this respect resemble Ichthyostega.

The body of the ilium is thickened to support the acetabulum, with an anteroventrally directed buttress
above it which terminates in unfinished bone. A more complex region lies posterior to the acetabulum, where
an almost hemispherical depression imparts a lobed shape to its posterior margin. As preserved, therefore,
the acetabulum is essentially heart-shaped. The lobed region may be equivalent to that in Eoherpeton
(Smithson 1985) where a supra-acetabular notch is interpreted as the site of a ligament attaching to the
femur.

The posterovenlral portion of the acetabulum is supported on what appears to be a thickened horizontal
buttress, but this could well be an artefact caused by compression. The surface of the acetabulum is not
visible; as in other parts of this material, unlined endochondral bone is almost impossible to distinguish from
matrix. The acetabulum lies much further anteriorly in the ilium than it does in most other tetrapod pelvic
girdles. Typically, the acetabulum lies directly beneath the point at which the postiliac process arises.

The ischium is relatively thin, but quite well ossified except at the margins. The posterior margin has a
similar hatchet shape to that of Ichthyostega. The pubic region is similarly preserved, but the anterior margin
is incomplete. It is not obvious what, if any, contribution the pubis made to the acetabulum. It is possible
that the whole unit was continued more anteriorly in cartilage. Only one small foramen pierces the pubic
region of this pelvic girdle, which is difficult to interpret as an obturator foramen.

In Ichthyostega , the pubic region appears very truncated as illustrated by Jarvik (1980), though he notes
that the anterior margin was cartilage-finished. In examining the specimens of the pelvic girdle of Ichthyostega.
I found one which appears to show a long, rather narrow and poorly ossified pubis, with large obturator
foramina, in articulation anteriorly. It seems as though the whole pubis remained largely cartilaginous and
was only rarely preserved. This could well have been the case in the ‘Pictou Girdle’, in which the pubic region
appears to be even more truncated than in Ichthyostega.

While the pubis in early tetrapods was apparently the last element of the pelvic girdle to ossify, and is
often not preserved, in the pelvic girdles of osteolepiform fishes (Andrew and Westoll 1970; Jarvik 1980),
there is a single ossification which is generally homologized with the pubis of tetrapods because it is anteriorly
directed. The contrast suggests that close homologies between the two elements may not be possible. It seems
more likely that the element in osteolepiform fish is homologous with those in other fish groups, where
homologies with the tetrapod girdle are not evident.

DISCUSSION

The new material of Acanthostega reveals, as Jarvik (1952) suspected, an animal quite different
from the better-known Ichthyostega , and if the postcranial elements are correctly assigned to
Acanthostega , the differences are known to extend to the postcranium. This serves to emphasize
what has become apparent from more recent finds of Devonian tetrapods, that by the late
Devonian, tetrapods had radiated widely both in space and ecologically, and that the emergence
of tetrapods occurred much earlier than the late Devonian.

Although the new specimens of Acanthostega are so incomplete, they nevertheless provide
evidence of both similarities and differences between it and Ichthyostega which contribute to the
debate, not so much about the origin of tetrapods or their relationships to any fish group, but of
what primitive tetrapods were actually like, in other words, what were the primitive characters
of tetrapods, and which of them were tetrapod autapomorphies. Most of these, like the majority of
those cited by Gaffney (1979), are directly related to overcoming the problems of life on land.

Historically, since the work of D. M. S. Watson (especially 1926), the embolomeres (in which
group Watson included the loxommatids), were considered to be the most primitive tetrapods,
both because they were the earliest tetrapods known at the time, and because they showed
resemblances to the osteolepiform fishes from which they were considered to have emerged. These
tetrapods were all late Carboniferous in age, by which time it is now known that the group had
undergone a considerable radiation, possibly explosive in character. As Devonian tetrapods become
better known, it may become clearer which characters shown by Carboniferous forms were actually
primitive, thus which characters may legitimately be taken to represent tetrapods as a whole in
the debate about their closest relatives. In searching for the true primitive state of a character.

718

PALAEONTOLOGY, VOLUME 31

evidence from neither stratigraphy nor functional morphology can be ignored. Panchen and
Smithson (1987) and Schultze (1987) have recently used a combination of both these lines of
evidence in a debate about which characters are true autapomorphies of lungfishes and can be
used to represent them in a cladistic analysis, as distinct from those which characterize a subgroup
(albeit the majority) which arose subsequently.

The differences between Acanthostega and Ichthyostega , as shown by the new evidence, include
the ossification of the otic region and its relationship to the skull roof, a character of the lower
jaw, and those seen in the postcranial skeleton. Similarities include the broad, closed palate, and
the lack of any skull table-cheek kinetism, though the pattern of skull table bones is quite different
in each. Possibly similar also is the presence of an internasal bone and a persistent ventral otic
fissure and notochordal basioccipital, though the evidence for these is less certain.

Among the similarities between them, none yet discovered can be considered as indicating any
special relationship, that is, a synapomorphy which unites them more closely to each other than
to other tetrapods. By the same token, neither shows any synapomorphies which could unite it
with any other early tetrapod group. The material is still too imperfectly known to warrant any
more detailed discussion of the possible relationships of Acanthostega to other tetrapods.

The closed, plate-like palate of Ichthyostega, in which the parasphenoid sutured to the pterygoids
laterally, has been considered a unique feature of the genus (Jarvik 1980), though this has also
been seen as a character uniting tetrapods with lungfishes by Rosen et al. (1981). They saw it as
similar to the palate in lungfishes, where a short broad parasphenoid sutures along its length to
the pterygoids. In some respects, however, the palate of Ichthyostega shows primitive characters,
and one of these is the suture between the pterygoids anterior to the parasphenoid. This character
has been considered primitive for tetrapods since Watson (1919, 1926).

My examination of the palate of Ichthyostega convinces me that the parasphenoid was separated
from the pterygoids by narrow but distinct interpterygoid vacuities, as in other primitive tetrapods.
In Acanthostega, narrow interpterygoid vacuities were certainly present, and again the pterygoids
met anteriorly. The isolated specimen from Celsius Bjerg, A88, clearly neither Acanthostega nor
Ichthyostega, also shows a broad, closed, and somewhat dorsally convex palate, though there is
no evidence concerning the relationship of the parasphenoid to the pterygoids.

Among the better known Carboniferous groups, the pattern in these Devonian forms is most
closely matched by that in the loxommatids (Beaumont 1977). In other forms, interpterygoid
vacuities, though still narrow, are nevertheless significantly larger, and the anterior suture between
the pterygoids more restricted, allowing the parasphenoid a longer ventral exposure. Anteriorly,
the pterygoids are also generally narrower. These features can be seen in the colosteid Greererpeton
(Smithson 1982), Crassigyrinus (Panchen 1985), and the embolomeres Proterogyrinus (Holmes
1984) and Pholiderpeton (as ‘ Eogyrinus Panchen 1972; Clack 1987). It is this form, rather than
the closed loxommatid palate, which has usually been considered primitive for tetrapods, primarily
because of its apparent similarity to that of osteolepiform fishes, in particular that of Eusthenopteron
(text-fig. 13).

The presence of the broad, closed plate-like palate in each of three Devonian forms and in the
loxommatids presents a prima facie case for consideration of this pattern, rather than that of
embolomeres, as primitive for tetrapods. What are the implications of this?

Seen in ventral view, the area about the mid-line of the palate seems very similar in embolomeres
and osteolepiforms, with narrow pterygoids, long narrow interpterygoid vacuities, and a long
exposure of the parasphenoid, but the similarities may be more apparent than real. In Eusthenop-
teron, on either side of the parasphenoid, the pterygoids descend to form a strongly vertical
component. This creates the illusion of narrow pterygoids and narrow, but real, interpterygoid
vacuities, similar to those of embolomeres. In fact there is only a very small gap between the
parasphenoid and the pterygoids. The vertical component of the pterygoids can be seen clearly in
section (Jarvik 1980), and this results from the fact that in primitive osteichthyan fish both head
and body are laterally compressed, consequent upon their streamlined fusiform shape, an adaptation
for aquatic locomotion. It remains true among recent forms that, in general, tetrapods are

CLACK: DEVONIAN TETRAPOD FROM GREENLAND

719

text-fig. 13. Palates of fishes and early tetrapods (marginal dentition omitted), a, Osteolepis macrolepidotus
(anterior part only); b, Eusthenopteron foordv, c, Crassigyrinus scoticus ; d, Pholiderpeton scutigerum\ e, Ichthyo-
stega sp.; f, Megalocephalus pachycephalus\ G, Acanthostega gunnari. (a, b, e, after Jarvik (1980); c, after
Panchen (1985); d, after Clack (1987); f, after Beaumont (1977); G, original.)

dorsoventrally compressed as compared with the lateral compression common in fish. Thus the
broad, closed palate of these Devonian forms could result from dorsoventral flattening of a palate
like that of an osteolepiform.

The resemblance between the palate of the embolomeres, Crassigyrinus , and Eusthenopteron ,
may be associated with a secondary adaptation to aquatic locomotion and subsequent deepening
of their skulls.

At the anterior end of the palate, the resemblances between any early tetrapod and osteolepiform
fishes, in particular Eusthenopteron , are less obvious (text-fig. 13). Two character differences are
of interest here. In all the earliest tetrapods so far discussed, the pterygoids meet anteriorly, whether
it be in a sutural contact or simple abutment. In no osteolepiform is this so. In Eusthenopteron ,
the pterygoids are separated along their length by the parasphenoid, and this seems to have been
true of all osteichthyans except lungfishes. One of the characteristic differences between fish and
tetrapods is the elongation of the snout in the latter. This not only influenced the bones of the
dorsal part of the skull around the naris and the orbit, but also, it seems, of the underlying palate,
causing the pterygoids, but not the parasphenoid, to lengthen anteriorly and meet in the mid-line.
Though lungfishes exhibit the same pattern, it was clearly not derived in association with elongation
of the snout, since it is also present in short-snouted forms (Miles 1977).

The second character to be considered is the relationship between the pterygoids, vomers, and
parasphenoid. In the early tetrapods discussed so far, the vomers meet in the mid-line anteriorly.

720

PALAEONTOLOGY, VOLUME 31

though in most, with the exception of the loxommatids Megalocephalus and some specimens of
Baphetes (Beaumont 1977), and in Acanthostega in which the condition is not known, they are
separated posteriorly by anterior extensions of the pterygoids. In Eusthenopteron, by contrast, the
vomers are separated throughout most of their length by the parasphenoid, while the pterygoids
lie lateral to both. It is difficult to see how the tetrapod pattern could be derived from this rather
specialized condition. The osteolepidids, however, show a condition closer to the primitive
sarcopterygian pattern in having vomers which barely meet in the mid-line, their common junction
meeting the anterior tip of the parasphenoid. Elongation of the snout could more easily have
produced the tetrapod pattern from this than from the eusthenopterid condition (text-fig. 13).
Panchen and Smithson (1987) have recently argued that eusthenopterids rather than osteolepids
form the sister-group of tetrapods.

In neither Ichthyostega , nor Acanthostega , nor the loxommatids, all of which show the broad
closed palate, is there any sign of the skull table-cheek kinetism found in fish, associated with
movements of the cheek and opercular region during ventilation and feeding, which is usually
assumed to have its homologue in the straight, unconsolidated suture found in this region in, for
example, embolomeres, Eoherpeton (Smithson 1985), and Crassigyrinus (Panchen 1985). On the
same basis used for consideration of the palate, is it justifiable to consider the consolidated skull
as primitive for tetrapods?

Movement between the skull table and cheek bones in osteichthyan fish is necessary to
accommodate the expansion of the gill chamber during the ventilatory cycle. However, it is
characteristic of tetrapods that the opercular series is all but lost; when gill breathing was superseded
by other methods of ventilation it became unnecessary. Gill breathing in adults would have been
eliminated at an early stage in tetrapod evolution. In the dorsoventrally flattened skulls of Devonian
tetrapods, the appropriate movements of the cheek would have been difficult to achieve. However,
particularly in a dorsoventrally flattened skull, there would have been some benefit to eliminating
the weakness at the skull table-cheek junction. It is significant in this context that in Ichthyostega ,
Acanthostega , and the loxommatids, the result has been achieved in different ways, and so
presumably by convergent evolution. Only the loxommatids retain the pattern of bones in the skull
table which comparison with osteolepiforms suggests to be primitive, retaining the intertemporal
at least in early members of the group.

Why then did embolomeres, Eopherpeton, and Crassigyrinus apparently have a "kinetic’ skull
roof reminiscent of that of osteolepiform fish? It has been suggested (see Clack 1987) that the
‘kinetism’ in these forms was rather the result of development of a butt-joint between the horizontal
skull table and the steeply sloping cheek, which enhanced resistance to compressive forces during
jaw closure. Perhaps, like the palate, the similarities to osteolepiforms are associated with secondary
deepening and lateral compression in the skulls of these animals. Embolomeres and Crassigyrinus
were secondarily aquatic, though apparently Eoherpeton was not. The condition is derivable from
that of an early loxommatid, and it is the latter, rather than the embolomere pattern, which may
represent the true primitive condition for tetrapods. This hypothesis would be supported if further
finds of Devonian tetrapods show dorsoventrally flattened skulls with broad palates, and would
be more satisfactorily refuted by the discovery of an early tetrapod with an undeniably flattened
skull which was nevertheless ‘kinetic’, rather than a steep-sided skull with no ‘kinetism’.

Consideration of the differences between Ichthyostega , Acanthostega , and other tetrapods, has
highlighted three characters of which one is a true tetrapod autapomorphy, and two may be
autapomorphies of all tetrapods other than Ichthyostega (‘Neotetrapoda’, Gaffney 1979).

1. Differences between the interclavicles of Acanthostega and Ichthyostega. The differences may
well be caused by differences in the functional morphology of the rest of the skeleton, and how
well adapted it was for terrestrial locomotion, but this will be hard to assess until more of the
postcranium of Acanthostega is known. However, the possession of a large dermal interclavicle
exposed ventrally between the clavicles, and bearing ornament, appears to be characteristic of
early tetrapods. It is probably associated with both protection of the thorax and elaboration

CLACK: DEVONIAN TETRAPOD FROM GREENLAND

721

A

B

C

Meek bone

D

artic

add foss suranq artic

surang

pospl P°SP'

F

H

text-fig. 14. Lower jaws of fishes and early tetrapods (dentition omitted). Eusthenopteron foordi: a, lateral
view; b, section through anterior end; c, mesial view. Ichthyostega sp.: d, lateral view; e, mesial view.
Megalocephalus pachycephalus : F, lateral view. G, mesial view. Eoherpeton watsoni : h, lateral view; J, mesial
view, (a-e, after Jarvik (1980); f, g, after Beaumont (1977); h, j, after Smithson (1985)).

of the pectoral musculature in terrestrial locomotion. It contrasts with the small interclavicle of
Eusthenopteron , a form in which the interclavicle is known. In most early sarcopterygian groups,
the interclavicle is not known, suggesting that it was also small or absent altogether. It was present
as a small element in primitive actinopterygians and could represent an apomorphy of osteichthyans
(Gardiner 1984). However, an interclavicle bearing dermal ornament and large with respect to the
clavicle, is found only in tetrapods and may be cited as a tetrapod autapomorphy, resulting directly
from adaptation to terrestrial locomotion.

2. Differences in the relationships of the dentary to the articular between Ichthyostega and
Acanthostega. As figured by Jarvik (1980), the dentary of Ichthyostega runs along the whole of the
dorsal margin of the lower jaw, to contact the articular. This pattern is found in Eusthenopteron
and many other sarcopterygian fishes. It differs from that in Acanthostega and in all other described
tetrapods, where the dentary is excluded from most of the dorsal margin of the adductor fossa
by the surangular (text-fig. 14). Assuming Jarvik’s description to be accurate, this represents an
autapomorphy of all tetrapods other than Ichthyostega , and on this evidence the lower jaw of
Metaxygnathus (Campbell and Bell 1977) appears to be a true tetrapod. Loss of contact between
the dentary and articular could have been associated with elongation of the snout, characteristic
of tetrapods, and in this respect it is surprising to find that Ichthyostega retains the fish-like
condition.

3. A suture between the anterior coronoid and the presplenial on the mesial surface of the lower
jaw, at the anterior end (text-fig. 14). Ichthyostega differs from all other described tetrapods in
lacking this feature, although, unfortunately, Acanthostega yields no information on this. The
presplenial curves round under the ventral margin of the jaw ramus to meet the anterior coronoid,
forming a tube in cross-section enclosing the Meckelian space. Although the jaw associated with
skull C appears tubular in cross-section (text-fig. 5c), the bones are disturbed and broken and the
elements difficult to interpret.

722

PALAEONTOLOGY, VOLUME 31

In Eusthenopteron, and in other primitive sarcopterygian fishes, the presplenial (‘anterior
infradentary’ in fish terminology) is essentially a flat bone in cross-section. Beneath the anterior
coronoid lies a convex ridge formed by the Meckelian bone (seen in section in Jarvik’s 1980, fig.
76 and reproduced here in text-fig. 14e), which may or may not be overlain on the mesial surface
by the prearticular. It is difficult to be sure from his figure where the anterior suture of the
prearticular lies. In Ichthyostega , the prearticular appears from his figure to pass along the complete
length of the jaw ramus to the symphysis. In neither case, however, is there any contact between
the presplenial and the anterior coronoid (text. -fig. 14).

The typical tetrapod condition could have arisen by reduction of the Meckelian bone, a process
that certainly occurred in tetrapods, where as a rule the only ossification of Meckel’s cartilage to
survive is the articular. Formation of a tubular cross-section at the anterior end of the lower jaw
would have conferred greater stiffness to this element, and so would be more resistant to bending
or twisting forces than the fish jaw. It would represent a more economical use of materials: a
tetrapod jaw of this design would be stiffer than a fish jaw of the same mass, or the same stiffness
could be achieved for less mass. The difference could represent fundamental differences in the
musculature of the jaws in the two groups in which there may have been lateral forces produced
by the jaw muscles of tetrapods which were not experienced by fish.

A presplenial anterior coronoid suture may thus be cited as a further apomorphy of neotetrapods,
again explicable in terms of the demands of terrestrial life. As described by Campbell and Bell
(1977), Metaxygnathus is a true neotetrapod on this character, though the specimen is very poorly
preserved (A. L. Panchen, pers. comm.).

It would be of great interest to know the state in Elpistostege (Schultze and Arsenault 1985) of
each of these three characters and also to know the pattern of the palatal bones. The second lower
jaw character could be confirmed quite easily by a section across the anterior end of the skull,
which might also yield some information about the relations of the pterygoids, vomers, and
parasphenoid.

Acknowledgements. My thanks go first to Dr John Nicholson for finding these specimens in the first place,
and to Dr Peter Friend for giving me access to them and for all his subsequent help, both with the stratigraphy
and in being instrumental in bringing about a further expedition to John Nicholson's collecting site. 1 also
record my appreciation of the generous access which 1 was allowed to the material of Acanthostega and
Ichthyostega by Professor Eric Jarvik and the staff of the Natural History Museum, Stockholm, who made
my visit there such a happy one. Dr Svend Eric Bendix-Almgreen of the Geologisk Museum, Copenhagen,
kindly sanctioned the loan of the holotype of A. gunnari, and three other Devonian tetrapod specimens, and
1 also thank him for his co-operation and efforts in making arrangements for our joint expedition to
Greenland. I am grateful to my colleagues Drs Andrew and Angela Milner, Alec Panchen, and Tim Smithson
for reading and commenting on the manuscript and for helpful discussion during the early stages of this
work. I also thank Dr Colin Patterson for helpful suggestions and for eliminating a major howler. Mr Peter
Whybrow and Mr Ronald Croucher of the British Museum (Natural History) gave invaluable advice on
preparing the material.

REFERENCES

Andrews, s. M. and westoll, T. s. 1970. The postcranial skeleton of Eusthenopteron foordi Whiteaves. Trans.
R. Soc. Edinb. 68, 207-329.

beaumont, E. 1977. Cranial morphology of the Loxommatidae (Amphibia: Labrinthodontia). Phil. Trans. R.
Soc. B 280, 29-101.

butler, h. 1961. Devonian deposits of central East Greenland. In raasch, g. o. (ed.). Geology of the Artie,
188-196. University of Toronto Press, Toronto.
bystrow, a. p. 1947. Hydrophilous and xerophilous labyrinthodonts. Acta Zool. 28, 137 164.

Campbell, K. s. w. and bell, m. w. 1977. A primitive amphibian from the late Devonian of New South
Wales. Alcheringa , 10, 369-381.

Carroll, R. L. 1964. Early evolution of dissorophid amphibians. Bull. Mus. comp. Zool. Harv. 131, 163-250.
1967. Labyrinthodonts from the Joggins fauna. J. Paleont. 41, 111 142.

CLACK: DEVONIAN TETRAPOD FROM GREENLAND

723

and gaskill, p. 1978. The order Microsauria. Mem. Am. phil. Soc. 126, 1 211.

clack, J. a. 1987. Pholiderpeton scutigerum Huxley, an amphibian from the Yorkshire Coal Measures. Phil.
Trans. R. Soc. B 318, 1 107.

de beer, g. R. 1937. The development of the vertebrate skull, 552 pp. Clarendon Press, Oxford.
friend, p. f., alexander-marrak, p. d., allen, k. c., nicholson, j. and yeats, a. k. 1983. Devonian sediments
of East Greenland, VI. Review of results. Meddr Gronland , 206 (6), 1-96.

-nicholson, j. and yeats, a. k. 1976. Devonian sediments of East Greenland, II. Sedimentary
structures and fossils. Ibid. 206, (2), 191.

gaffney, E. s. 1979. Tetrapod monophyly: a phylogenetic analysis. Bull. Carneg. Mus. nat. Hist. 13,
92 105.

Gardiner, b. G. 1984. The relationship of the palaeoniscid fishes, a review based on new specimens of Mimia
and Moythomasia from the Upper Devonian of Western Australia. Bull. Br. Mus. nat. Hist. ( Geol .), 37,
173-428.

holmes, R. 1980. Proterogyinus scheeli and the early evolution of the labyrinthodont pectoral limb. In
panchen, A. L. (ed. ). The Terrestrial Environment and the Origin of Land Vertebrates , 351 376. Academic
Press, London.

— 1984. The Carboniferous amphibian Proterogyrinus scheeli Romer and the early evolution of tetrapods.
Phil. Trans. R. Soc. B 306, 431 527.

jarvik, e. 1952. On the fish-like tail in the ichthyostegid stegocephalians. Meddr Gronland , 114, 1 90.

— 1961. Devonian vertebrates. In raasch, g. o. (ed.). Geology of the Arctic, 197-204. University of Toronto
Press, Toronto.

1965. Specialisations in early vertebrates. Annls Soc. r. zool. Belg. 94, 11 95.

— 1980. Basic Structure and Evolution of Vertebrates (vol. 1), 575 pp. Academic Press, London.
johansson (jarvik), a. e. v. 1935. Upper Devonian fossiliferous localities in Parallel Valley on Gauss

Peninsula, East Greenland. Meddr Gronland, 96, 1 96.
lebedev, A. O. 1984. The first find of a Devonian tetrapod vertebrate in the U.S.S.R. Dokl. Akad. Nauk
SSSR, Palaeont. 278, 1470 1473.

leonardi, g. 1983. Notopus petri nov. gen., nov. sp. : une empreinte d’amphibien de Devonian au Parana
(Bresil). Geobios, 16, 233-239.

miles, r. s. 1977. Dipnoan (lungfish) skulls and the relationships of the group: a study based on new species
from the Devonian of Australia. Zook J. Linn. Soc. 61, 1-328.
nicholson, j. and friend, p. f. 1976. Devonian sediments of East Greenland, V. The central sequence, Kap
Graah Group and Mount Celsius supergroup. Meddr Gronland, 206, I 117.
panchen, a. l. 1964. The cranial anatomy of two Coal Measure anthracosaurs. Phil. Trans. R. Soc. B 242,
207-281.

— 1970, Anthracosauria. In kuhn, o. (ed.). Handbuch der Palaoherpetologie, Teil 5a, 84 pp. Gustav Fischer,
Stuttgart.

1972. The skull and skeleton of Eogyrinus attheyi Watson (Amphibia: Labyrinthodontia). Phil. Trans.
R. Soc. B 263, 279-326.

1975. A new genus and species of anthracosaur amphibian from the Lower Carboniferous of Scotland
and the status of Pholidogaster pisciformis Huxley. Ibid. 269, 581-640.

1985. On the amphibian Crassigyrinus scoticus Watson from the Carboniferous of Scotland. Ibid. 309,
461-568.

and smithson, t. r. 1987. Character diagnosis, fossils and the origin of tetrapods. Biol. Rev. 62, 341
438.

rocek, z. 1986. Tooth replacement in Eusthenopteron and Iclithyostega. In duncker, h.-r. and Fleischer, g.

(eds.). Vertebrate Morphology. Gustav Fischer, Stuttgart, Fortschritte der Zoologie, 30, 249 252.
romer, a. s. 1957. The appendicular skeleton of the Permian embolomerous amphibian Archeria. Univ. Mich.
Contr. paleont. Mus. 13, 103 159.

rosen, d. e., forey, p. l., Gardiner, b. g. and Patterson, c. 1981. Lungfishes, tetrapods, paleontology and
plesionrorphy. Bull. Am. Mus. nat. Hist. 167, 154-276.
save-soderbergh, g. 1932. Preliminary note on Devonian stegocephalians from East Greenland. Meddr
Gronland, 94 (7), 1-107.

1933. Further contributions to the Devonian stratigraphy of East Greenland, I. Results from the summer
expedition, 1932. Ibid. 96 (1), 1 40.

1934. Further contributions to the Devonian stratigraphy of East Greenland, II. Investigations on Gauss
Peninsula during the summer of 1933. Ibid. 96 (2), 1-74.

724

PALAEONTOLOGY, VOLUME 31

schultze, h. p. 1987. Dipnoans as sarcopterygians. In bemis, w., burggren, w. w. and kemp, n. (eds.). The
Biology and evolution of lungfishes. J. Morph. Supp. 1, 39-74.

— and arsenault, m. 1985. The panderichthyid fish Elpistostege: a close relative of tetrapods? Palaeontology ,
28, 293-310.

smithson, T. R. 1982. The cranial morphology of Greererpeton burkemorani Rorner (Amphibia: Temnospon-
dyli). Zool. J. Linn. Soc. 76, 29-90.

— 1985. The morphology and relationships of the Carboniferous amphibian Eoherpeton watsoni Panchen.
Ibid. 85, 317 410.

warren, a., jupp, r. and bolton, b. 1986. Earliest tetrapod trackway. Alcheringa , 10, 183-186.
watson, d. m. s. 1919. The structure, evolution and origin of the Amphibia— the 'Orders’ Rachitomi and
Stereospondyli. Phil. Trans. R. Soc. B 209, 1-73.

1926. Croonian Lecture. The evolution and origin of the Amphibia. Ibid. 214, 189-257.
westoll, t. s. 1943. The origin of tetrapods. Biol. Rev. 18, 78-98.

white, t. e. 1939. Osteology of Seymouria baylorensis Broili. Bull. Mus. Comp. Zool. Harv. 85, 325-409.

j. a. clack

University Museum of Zoology

Typescript received 10 June 1987 Downing Street

Revised typescript received 12 September 1987 Cambridge CB2 3EJ, UK

ABBREVIATIONS

acet

acetabulum

pmx

premaxillary

add foss

adductor fossa

po

postorbital

ang

angular

pofr

postfrontal

artic

articular

pospl

postsplenial

bocc

basioccipital

PP

postparietal

bptpr

basipterygoid process

prearlic

prearticular

br/case

braincase

prefr

prefrontal

clav

clavicle

prespl

presplenial

cleith

cleithrum

proatl/exocc

proatlas or exoccipital

col cran

columella cranii

psph

parasphenoid

cor

coronoid

psph (pr cult)

processus cultriformis of para-

dent

dentary

sphenoid

ect

ectopterygoid

Pi

pterygoid

entep for

entepicondylar foramen

qj

quadratojugal

epipt

epipterygoid

qu

quadrate

fr

frontal

qu ram pt

quadrate ramus of pterygoid

i/clav

interclavicle

r a pr

retroarticular process of lower

i pt vac

interpterygoid vacuity

jaw

jug

jugal

rt pt

mesial margin of right pterygoid

11c

lateral-line canal

scapcor

scapulocoracoid

lr jaw

lower jaw

sphet

sphenethmoid

max

maxilla

st

supratemporal

Meek bone

Meckelian bone

squ

squamosa]

obtur for

obturator foramen

supracor for

supracoracoid foramen

otic caps

otic capsule

surang

surangular

pal

palatine

tab

tabular

pal/ect

palatine or ectopterygoid

tab emb

tabular embayment

pal tooth/teeth

palatal tooth/teeth

tab h

tabular horn

par

parietal

vom

vomer

AN EXTINCT ‘SWAN-GOOSE’ FROM
THE PLEISTOCENE OF MALTA

by E. MARJORIE NORTHCOTE

Abstract. Qualitative and quantitative studies on extinct Cygnus equitum/ Anser equitum from the Ipswichian
(Eemian) Interglacial of Malta (c. 125 000 b.p.) show it was a broad-bodied, dwarf swan with some goose-
like features. It was closer to Whooper and Bewick’s C. cygnus than Mute Swans C. olor though the relative
shortness of the chief hand bones resembles the latter. Feathered wing span was c. 15 m. The wings were
probably more ‘elliptical’ than in other swans; ‘stouter’ carpometacarpus and ulna(?) suggest higher camber
and relatively shorter hand bones suggest lower aspect ratio ( length : width) than in Whooper Swans. There
is no evidence to support assertions that it was flightless. The wings were fully feathered, it was light enough
(c. 3-5-4 0 kg) to fly and the flight apparatus was not reduced. The femur was comparatively ‘stout’. Abundant
on the island, C. equitum may have swum on fresh and brackish water, walked well and. unlike other swans,
have habitually taken off and alighted on land. It probably ate highly calorific plant food in enclosed, rather
terrestrial habitats. Morphological differentiation facilitated coexistence with Whooper Swan and the giant,
flightless, extinct swan C. falconeri. The two extinct, more advanced swans probably arose from the same
fully flighted stock as Whooper Swans.

Bate (1916) based Cygnus equitum on fossils of what she considered to be a small extinct swan
and Lambrecht (1933) and Howard (1964) agreed but Brodkorb (1964) named them Anser equitum
(Bate), an extinct goose. Bate (1916) briefly described and figured the holotype (a carpometacarpus)
and paratypes (a proximal humerus and a coracoideum) of equitum and mentioned fragments,
now lost, of two ulnae and a radius, all from Pleistocene deposits at Ghar Dalam, Malta.

This, the first detailed study on equitum , aims to ascertain the genus and affinities of the bird,
to suggest its probable size, form, and habitat and to investigate its habit, particularly with respect
to Bate’s (1916) claim that equitum was flightless.

AGE OF THE FOSSILS

In Ghar Dalam cavern the equitum type series lay in red earth matrix (Bate 1916), characteristic of the bone-
bearing stratum of Maltese caves and fissures. The stratum is thin so Adams (1870) and de Bruijn (1966)
considered all the bones were deposited in a short time span and represent one faunal sample. The matrix is
highly calciferous. No countable pollen for dating has been found (Zammit-Maempel 1982; Northcote 1982a);
indeed no precise dates are available for the sediments or fauna (Pedley 1981, p. 71). At times during the
Pleistocene, Sicily and Malta were connected by an isthmus or island chain with sea-level lower than at
present (Zammit-Maempel 1977; Sondaar and Boekschoten 1967). Bones of equitum were associated with
extinct pygmy elephant Palaeoloxodon melitensis (Falconer, 1862), that flourished on Siculo-Malta in a period
equivalent to the Ipswichian (Eemian) Interglacial Stage of more northern countries (Sondaar 1971),
114000 135 000 years ago (Gascoyne et al. 1983). This then, may also be taken as the date of equitum.

SPECIMENS, METHODS, AND TERMINOLOGY

The type series of C. equitum Bate, 1916 is in the National Museum of Natural History, Malta (Specimens
NMM 20 and 21). Casts, catalogued A. equitum (Bate) are in the British Museum (Natural History), London
(Specimens BMNHL A 161 3, 1614, 1615). From Maltese Pleistocene anseriform fossils, unidentified or
identified as C. falconeri Parker, 1865 or C. equitum , in those museums and the University Museum of
Zoology, Cambridge (UMZC) I chose specimens consistent with the equitum types. Reference skeletons

(Palaeontology, Vol. 31, Part 3, 1988, pp. 725 740, pis. 69-70.1

© The Palaeontological Association

726

PALAEONTOLOGY, VOLUME 31

include Greylag A. anser , White-fronted A. albifrons. Barnacle Branta leucopsis and Brent Geese B. bernicla,
and Whooper and Bewick’s C. cygnus and Mute Swans C. olor from the following: University Museum of
Zoology, Cambridge, Sedgwick Museum, Cambridge (SMC), BMNH, Tring (BMNHT), Royal Scottish
Museum, Edinburgh (RSM), Glasgow Museum (GM), Leicester Museum (LM), and Colchester and Essex
Museum (CEM).

I chiefly use Anser (less specialized than Branta , Johnsgard 1965), in particular Greylag Goose (the largest
western Palaearctic goose. Cramp and Simmons 1977) for comparisons with geese. I follow Johnsgard (1974)
in treating Whooper and Bewick’s Swans as elastically similar Eurasian subspecies of C. cygnus. Because of
their more southern Palaearctic distribution, I chiefly use Whooper C. c. cygnus and Mute Swans for
comparison with swans.

I follow Verheyen (1953, 1955), Simpson et al. (1960), and Woolfenden (1961) who used ratios for
mensurational comparison. For comparing ‘stoutness’, where accurate measurements are obtainable, viz.
humerus and carpometacarpus, I follow Kuhry and Marcus (1977) and compare logarithms of ratios. Weight
predictions are made using scaling formulae. Following Scott (1983) they are based on several parameters
within similar morphological groups. For estimating equitum weight, I use the humerus and femur (the bones
least likely to be modified by habit, Bellairs and Jenkin 1960).

Methods of preparation and measurement are given elsewhere (Northcote 1979a, b , 1982a). Many of the
distinctions between Anserini cited follow Woolfenden (1961). Taxonomy follows Delacour (1954) and
Johnsgard (1974). Anatomical nomenclature follows Baumel (1979) and Vanden Berge (1979).

QUALITATIVE CHARACTERS

Cranium

Specimen BMNHL A3267 (text-fig. 1) comprises the frontal area with right supraorbital margin,
postorbital region and occipital plane with condyle, foramen magnum, and alae tympanicae.
Specimen UMZC 252 a comprises a braincase infill with a posterior frontal bone fragment attached
to an occipital plane with dorsoventrally compressed condyle and foramen magnum.

The sulcus gl. nasalis in equitum resembles certain geese and extant northern swans in being
comparatively extensive (text-fig. 1 a-c). However, the equitum cranium differs from geese, but is
like swans, as follows: 1, the foramen n. olfactorii et sulcus olfactorius are overarched with bone
(Shufeldt 1909); 2, the proc. postorbitalis is enlarged rostrally and directed more ventrally (text-
fig. 1 d-f); 3, the crista temporalis forms a distinct ridge, and the fossa temporalis is large and
distinct (text-fig. 1 h-j); 4, the crista nuchalis transversa forms a distinct ridge demarcating the
occipital plane (text-fig. lg-/); and, 5, occipital fontanelles are absent (Stejneger 1882) (text-fig.
1 g-I). There is no indication in the extinct bird of the bony frontal bill knob diagnostic of Mute
Swans (text-fig. 1 b).

In equitum the large sulcus for the glandula nasalis (salt gland) suggests that it could live near
estuaries or the sea (Holmes and Phillips 1985). The other cranial characters indicate the
comparatively larger ligaments and muscles of a longer swan-like beak.

EXPLANATION OF PLATE 69

Figs. 1, 4, 7, goose; 2, 5, 8 a-c, equitum (BMNHL A5218, 5221, 5222, 5186, respectively); 3, 6, 9, Whooper
Swan.

Figs. U3. Scapula, lateral surface of cranial extremity showing acromion (F).

Figs. 4-6. Coracoideum, dorsal aspect of cranial extremity showing area (G) between proc. procoracoideus
and acrocoracoideus.

Figs. 7-9. Humerus, a, caudal surface of proximal extremity showing caput humeri (H), tuberculum ventrale
(I), fossa pneumotricipitalis (J), margo caudalis (K), incipient second fossa pneumotricipitalis (L), crista
pectoralis (m), impressio m. supracoracoidei (N), tuberculum dorsale (O). b, caudal surface of distal
extremity showing fossa olecrani and sulci m. humerotricipitis and scapulotricipitis (P). c, cranial surface
of distal extremity showing fossa m. brachialis (Q).

All magn. x 1 .

PLATE 69

1

1

/ *

8$£-

Nk:

4

F

NORTHCOTE, goose, swan, and swan-goose

728

PALAEONTOLOGY, VOLUME 31

text-fig. 1. Cranium, a, d, g, h, goose; b , e, i, k , equitum BMNHL
A3267; c,f, j, /, Whooper Swan, a-c, sulcus gl. nasalis (A); d-f, proc.
postorbitalis (B); h-j, crista temporalis (C) and lossa temporalis (D); g,
k, /, crista nuchalis transversa (E) and occipital plane. All magn. x 1.

Scapula and coracoideum

The equitum scapula differs from geese, but resembles swans, in lacking a pneumatic foramen
laterally between the acromion (that is cranially attenuated) and the facies artic. humeralis (PI. 69,
figs. 1-3). The equitum coracoid differs from geese but resembles swans as follows: 1, the area
between the proc. procoracoideus and acrocoracoideus is flat (PI. 69, figs. 4-6); and, 2, numerous
small pneumatic foramina occur under the entire edge of the facies artic. clavicularis; in geese there
is only one large hole.

NORTHCOTE: PLEISTOCENE ‘SWAN-GOOSE’

729

In swans, absence of a scapular air sac may facilitate upending, and in equitum its similar absence
may indicate a similar habit.

Humerus

Proximally, the equitum humerus differs from geese but resembles swans as follows: 1, the
tuberculum ventrale is less attenuated; and 2, there is an incipient second (dorsal) fossa
pneumotricipitalis (an advanced character, Bock 1962) bordered by a ridged margo caudalis (PI.
69, figs, la-9a) (Bate, 1916 considered that equitum had a single deep fossa). In one character, the
equitum humerus resembles geese; at the cranial end of the crista pectoralis there is an impressio
m. supracoracoidei forming a caudal lip on the tuberculum dorsale. As Bate (1916, p. 430) observed
of equitum , ‘the general outline is squarer’ than in swans (PI. 69, figs. la-9a). Distally, the equitum
humerus differs from extant swans and geese: 1, the fossa olecrani is much shallower, and the sulci
m. humerotricipitalis and scapulotricipitalis are much deeper (PI. 69, figs. 76-96); and, 2, the fossa
m. brachialis is deeply excavated and oval-shaped (PI. 69, figs. lc-9c).

The equitum humerus differs from Mute, but resembles Whooper, Swans: 1, the m. latissimi
dorsi insertion is clearly marked on the caudal shaft surface and turns ventrally below the caput
humeri; in Mute Swans the line is indefinite and straight; and, 2 insertion of the m. scapulohumeralis
in the fossa pneumotricipitalis is poorly marked and lacks a raised border; in Mute Swans it is
clearly marked and bordered.

The supracoracoideus muscle that inserts on the crista pectoralis and on the tuberculum dorsale
and impressio m. supracoracoidei (when present) (Baumel 1979) is essential for take-off from level
ground (Sy 1936), and most highly developed in birds specialized for slow flapping flight and jump
take-offs (Pennycuick 1972). Swans usually take off and land on water by pushing the water with
their feet (Cramp and Simmons 1977); their lack (atrophy?) of an impressio m. supracoracoidei
may be correlated with this habit. Greylag Geese, like other large geese, more frequently perform
jump take-offs and land on level ground using their wings (Cramp and Simmons 1977); the presence
of an impressio m. supracoracoidei in them may be correlated with this habit and the same may
apply to equitum. The differences between equitum and recent Anserini in both fossa olecrani and
fossa m. brachialis suggest differences in elbow flexion, and, therefore in lift mechanisms.

Antebrachium

I can find no difference in radius or ulna between equitum and recent Anserini. Bate (1916) stated
the equitum ulna lacked papillae remigiales caudales, but ulnae such as BMNHL A5225 (PI. 70,
fig. 1) bear papillae. Bate’s (1916) specimen may have been eroded. Contrary to Brodkorb (1964),
the equitum carpometacarpus resembles swans, rather than geese: 1, the proximal articulatory
surface is almost flat (PI. 70, figs. 2a-4a)\ Bate (1916) erroneously considered it even flatter than
in swans; 2, the proc. extensorius of the os metacarp, alulare is less attenuated and the angle
between this process and the trochlea carpalis is larger (PI. 70, figs. 26-46); and, 3, the dorsal rim
of the facies artic. dig. major forms an arc. According to Bate (1916, p. 429), in equitum the os
metacarpale minus and major separate ‘for a comparatively much shorter distance (than in a recent
swan) causing the articular ends to be more massive’. On the holotype (as on other specimens)
only the minus ends remain so there is no evidence for her statement. Like the ulna, the metacarpale
majus of equitum bears feather papillae (PI. 70, fig. 3c). The phalanx proximalis digiti majoris of
the equitum manus resembles swans in having a discrete proximodistal ridge between two grooves
(PI. 70, figs. 5-7).

Papillae remigiales caudales on ulna and carpometacarpus indicate that equitum had the chief
flight feathers. The flatter proximal surface and rounder rim of the facies artic. dig. major in
equitum and swans may be related to the shape and disposition of the proc. extensorius of the os
metacarp, alulare (concerned with muscles extending the hand and keeping taut the propatagial
skin fold, George and Berger 1966) and indicate greater rotation at wrist and major digit in equitum
and swans than in geese. Tendons of muscles that control wing-tip movement cross the proximal
phalanx and insert on the second phalanx of the dig. majoris (George and Berger 1966). In geese.

730

PALAEONTOLOGY, VOLUME 31

there is a certain amount of play of the tip, but in swans the tendons are constrained by the ridge
and its flanking grooves on the proximal phalanx with, consequently, less play. This must also
have been the condition in equitum. All these similarities in form of wrist and hand bone in equitum
and swans suggest similar use of the wing tip, e.g. during wing-tip reversal for fast speed (Brown
1963).

Hind limb bones

A femur shaft NMM F.22, No. 31 reported by Despott (1928/1929), combined with the extremity
BMNHL A5812, represents an equitum right femur. Compared to geese, the trochlea fibularis and
condylus lateralis flare less laterally in equitum and swans (PI. 70, figs. 8-10). A distal equitum
tibiotarsus NMM No. 26, reported by Despott (1928/1929) has pons supratendineus, canalis
extensorius, and incisura intercondylaris, but damaged condyles. An equitum tarsometatarsus
fragment (BMNHL A5810) is a distal shaft with trochlea of metatarsals III and IV enclosing the
incisura intertrochlearis lateralis and typical anserine bridge. In geese and equitum , but not swans,
the trochlear groove of metatarsal IV has a proximodistal swelling (PI. 70, figs. 11-13).

In resembling geese rather than swans, the equitum leg-bone characters suggest that, like the
former, the extinct bird walked efficiently and may contribute evidence that equitum habitually
took off and landed on level ground.

QUANTITATIVE CHARACTERS

‘ Stoutness ’

On the equitum coracoid dorsoventral width at the cotyla scapularis is 19-9-22-2 % of length; this
is above the range for geese (15 0-17-6%) but like that for swans ( 1 6-4-22- 1 %). For the equitum
coracoid shaft, range for ratio (width : length) is approx. 0- 1 52-0- 161; for Mute Swan UMZC 249
it is approx. 0-141 and for Whooper Swan UMZC 250 and Bewick’s Swan UMZC P6 approx.
0-158 and 01 59, respectively. Thus the equitum shaft, though relatively wider than in the Mute
Swan, is, contrary to Bate (1916), not wider than in the Whooper Swan. For Greylag Geese SMC
533-544 and BMNHT 1852.2.20.10, this ratio is approx. 0-138. Bate (1916) also stated the equitum
coracoid has greater mediolateral facies artic. clavicularis width than a swan. However, in the
fossils the facies edge is eroded.

Limb-bone measurements of equitum. Greylag Geese, and extant Palaearctic swans are given in
appendices 1 and 2 (lodged in the British Lending Library, no. 14035), means in Table 1. Log10
(ratio width : length) for equitum and Whooper Swan humeri do not significantly differ (95 % level;
P > 0-05). Whooper Swan humeri are significantly ‘stouter’ than Mute Swan humeri ( P < 0 001,
Northcote 1981), hence equitum humeri also are significantly ‘stouter’. However, equitum humeri are
significantly less ‘stout’ (P < 0-05) than those of Greylag Geese. The ratio (width : length) for two
equitum ulnae (approx. 0 042) is less than in Greylag Geese (0 051), but greater than in Whooper
and Mute Swans (0-039 and 0 038, respectively). Log10 (ratio width : length) comparisons for the

EXPLANATION OF PLATE 70

Figs. 2, 5, 8, 11, goose; 1, 3, 6, 9, 12, equitum. (1), BMNHL A5225; (3a, b), BMNHL A5216; (3c), NMM
Q.102.F25; (6) BMNHL A5219; (9a), NMM F.22; (9b), NMM F.22 (above), BMNHL A5812 (below);
(12), BMNHL A5810; 4, 7, 10, 13, Whooper Swan.

Fig. 1. Ulna, caudal aspect showing papillae remigiales caudales (R).

Figs. 2-4. Carpometacarpus. a, cranial aspect showing proximal articulatory surface (S). b, dorsal aspect
showing proc. extensorius of os metacarp, alulare (T) and trochlea carpalis (U). c, caudal aspect showing
papillae remigiales caudales (V)-

Figs. 5-7. Phalanx proximalis digiti majoris. 5, 6a, 7, dorsal surface; 6b, distal view, showing ridge (W).
Figs. 8-10. Femur, a, cranial, b, caudal surface showing trochlea fibularis (X) and condylus lateralis (Y).
Figs. 11-13. Tarsometatarsus, showing proximodistal swelling (Z) on the trochlear groove of metatarsal IV.
All magn. x 1 .

PLATE 70

NORTHCOTE, goose, swan, and swan-goose

732

PALAEONTOLOGY, VOLUME 31

table 1. Mean limb-bone measurements (mm) of equitum. Greylag Geese, and extant Palaearctic swans.

Measurements are given in appendices I and 2.

n

equitum

n

Greylag

Geese

n

Whooper

Swans

n

Bewick's

Swans

n

Mute

Swans

Humerus
Max. length

2

19715

6

169-37

28

275-5

8

233-3

33

290-9

Min. shaft width

2

9-60

6

9-38

28

12-30

8

10 91

33

12-29

Ulna

Max. length

2

c. 187

7

152-77

25

259-7

8

219-5

28

257-3

Min. shaft width

2

7-80

7

7-86

25

10-16

8

8-79

28

9-80

Carpomet.
Max. length

4

9118

5

96-44

17

137-47

2

118-90

9

133-36

Max. dorso-
ventral width
met. majus.

4

7-95

5

5-66

17

8 16

2

6-20

9

7-67

Phalanx

Max. length

9

33-42

2

43-40

16

58-29

2

51-15

5

51-42

Femur

Max. length

1

c. 79

5

80-16

26

108-78

8

94-33

34

104-67

Min. shaft width

1

9-90

6

7-52

26

10-46

8

9-39

34

10-20

Tarsomet.

Min. shaft width

1

6 61

3

5-80

20

8-24

5

7-60

23

8-40

table 2. Verheyen's (1955) osteometric indices applied to equitum, Greylag Geese, Whooper and Mute Swans.

Index

equitum 1

Greylag Geese2

Whooper Swans2

Mute Swans2

Humerus : ulna

c. 1 05

105-110

0-99- 1-09

1 00-117

Humerus : carpomet.
Wing index (ulna +

216

1-73-1-77

1-88-2-08

1 95 2-32

carpomet. : humerus)

c. 1 41

1 -47 1-51

1-411 -53

1-31-1 46

Femur : humerus

c. 0-40

0-47-0-49

0-38 0-42

0-34-0-37

1 From Table 1.

2 From appendix 2 and Verheyen (1955). These indices cannot be compared statistically since Verheyen published no
raw data.

carpometacarpus of equitum , Greylag Geese, Whooper and Mute Swans confirm Bate’s (1916)
opinion that the equitum carpometacarpus is, first, very much ’stouter’ than extant geese or swans —
significantly ‘stouter’ than Whooper Swans (P < 0 001) and therefore, of Greylags and Mute Swans
that are less ‘stout’ than Whooper Swans— and secondly, closer in proportion to Whooper than
Mute Swans. Ratio (width : length) of the composite equitum femur shows it is ‘stouter’ (ratio «
01 27) than in Greylag Geese (0 094) and Whooper and Mute Swans (0 096 and 0 097, respectively).

Ratios of limb-bone lengths

A ratio diagram (text-fig. 2) comparing bone lengths in Greylag Geese, Whooper and Mute Swans
with equitum shows that the ratios for the goose deviate from equitum more than for the swans.
Four osteometric indices used by Verheyen (1955) to characterize Greylag Geese, Whooper and
Mute Swans are applicable (Table 2). The index (humerus : ulna) for equitum is within the ranges
for Greylag Geese and the swans. The index (humerus : carpometacarpus) for equitum is greater
than ranges for Greylag Geese and Whooper Swans but within that for Mute Swans. (Bate,
1916, p. 427 considered the equitum carpometacarpus ‘relatively very much shorter’ than in recent
swans.) The wing index (ulna + carpometacarpus: humerus) for equitum is less than for Greylag
Geese, but within the ranges for the swans. The index (femur : humerus) for equitum is less than

NORTHCOTE: PLEISTOCENE ‘SWAN-GOOSE’ 733

Whooper Mute Greylag

Swan Swan equitum Goose

text-fig. 2. Simpson’s ratio diagram comparing mean lengths for six
bones of Greylag Geese, Whooper and Mute Swans, and equitum. The
horizontal scale represents the deviation from equitum (the standard) of
the logarithm of each dimension. No vertical scale is used. Though the
line for no recent species lies exactly parallel to the one for equitum ,
which is straight, those for the swans are straighter than that for the
goose. The relative proportions of equitum are, therefore, more like the
swans than the goose.

for Greylag Geese, but greater than for Mute Swans; it is within that for Whooper Swans. For
equitum , the index (chief phalanx: carpometacarpus) (0-37) is less than that for Greylag Geese,
Whooper and Mute Swans (0-45, 042, and 039, respectively). The phalanx proximalis digiti
majoris is significantly shorter in relation to the carpometacarpus in equitum than in Whooper
Swans (P < 0 001) and hence Greylag Geese but this ratio is not significantly different from Mute
Swans ( P = 0-6-0-7).

DISCUSSION AND CONCLUSIONS

Genus and Species

Bate (1916) was correct in assigning the Maltese fossils to Cygnus. The comparatively longer beak
and characteristic form of the scapula and coracoid, humerus head, carpometacarpus, and proximal
phalanx of the major wing digit, ‘stoutness’ of the limb bones, and ratios of their lengths to one
another all show equitum to be less like geese than swans. So far, there is little contrary evidence;
only one feature on the proximal humerus, and one each on distal femur and tarsometatarsus.
Brodkorb (1964) assigned the bird erroneously to Anser on account of the small size of the type
specimens relative to extant swans, and Bate’s (1916) figures of the proximal humerus (a paratype)
and carpometacarpus (holotype).

Greater affinity between the extinct swan and C. cygnus than C. olor is indicated by the absence
of a bony bill knob, two features proximally on the humerus, and perhaps the proportions of
coracoid and humerus, and the relationship between femur and humerus lengths. However, the
relative shortness of the chief wing phalanx and carpometacarpus is more similar to C. olor. This
last character, combined with greater ‘stoutness’ of carpometacarpus and femur (and perhaps

3. Estimation of weight (kg) of Cygnus equitum. Bone measurements from Table 1. Extant swan weights calculated from data given by Scott

734

PALAEONTOLOGY, VOLUME 31

s

0

C

<g

C/3

G

O

U

N

Os

CO

bO

G

Q,

O

O

X .

£

C/2

G

G

T3

<D

G

C/1

s

Lh

CD

'-G

G

G

W

s

0)

S

Ph

PCh

ds

•C *
rt X

.1 ^

0 ^ ^ ^ r t ^

'C ■£ 8 G’S-S’o’S'o

1 i<.a ^ IS

- w S 1 3 3

X ^
A OS

so

G-

I r-)

r-

os

l/~>

r-

Os

> m

-*-» r- ^ r-
£ Os X Os

, ^

— . X

U & m g ^ g

i, ‘C ° .. P

X5 II XS c3 ^ X

G , G ;

G qj <G £

X ^ X S _C3

<D <D CL) ‘ ’"H

_ .. X

G , G

e? o

G ^
O _]

CD ^

c/2 ^

o3 *— I

3

£

X

W OO

OS

c<) 4C

"^5

^5

too

.g

bD ”0
G G

CD C$
^ X
J-. <D

1<

• • ■ m

^ o os so

1J II II Jj,

% n <-! c

b- P-H

Actual \

1 9-72, r

2 6 05, r

3 10 63,

NORTHCOTE: PLEISTOCENE SWAN-GOOSE’

735

text-fig. 3. Estimation of distance between caputi humeri. AB joins articulation
points of the caputi humeri with the facies artic. humeralis of the coracoids. In
Anserini the anguli medialis of the coracoids meet (but do not overlap, Beddard
1898) at the spina externa of the sternum (C). D is the mid-point of AB. CD is
perpendicular to AB. CD = \ AB. AC = BC = the hypoteneuse of a right-angled
isoseles triangle. Therefore, AC2= BC2= CD2 + (iAB)2= 2 (^AB)2. For Cygnus
equitum mean AC = mean BC * 56 mm, n = 4. Thus AB, the width between the
caputi humeri = 2\/562/2 ss 79 mm.

ulna), shorter limb-bone lengths, and distal humerus structure are diagnostic of the extinct bird;
they justify Bate’s (1916) designation of it as sp. nov.

Shape and size

Though C. equitum had the general shape of a swan, especially Whooper and Bewick’s, its
wings were probably more ‘elliptical’, that is, more highly cambered, with a low aspect ratio
(length : width); greater ‘stoutness’ of the carpometacarpus, and probably of the ulna, suggests a
thicker leading edge, and relative shortness of the hand bones suggest a lower aspect ratio. Though
C. equitum was smaller than other swans (Table 1), its coracoids are comparatively long (length
= 74-7 mm for C. equitum and for Bewick’s Swan UMZC P6), so its body was comparatively
broad.

Weight

Humerus diameter : length (see earlier) and femur : humerus length (Table 2) are the same in the
extinct bird and in Whooper Swans, so weights of Bewick’s and Mute Swans and C. equitum are
estimated from scaling formulae using these measurements and constants calculated using Whooper
Swans (Table 3). The formulae for geometric similarity give the closest estimates for Bewick’s
Swans. (Estimated weights of Mute Swans, since they are a different species from Whooper, are
less close.) Using this formula, and femur and humerus lengths (the best parameters), mean weight
of C. equitum ss 3-55 kg. Though within ranges for the largest geese. Greylag Geese, and the
smallest swans, Bewick’s Swans, 2- 16-4-56 kg and 3-40-7-80 kg, respectively (Cramp and Simmons
1977), it is less than the mean weight for Bewick’s Swans (6-05 kg from data given by Scott et al.
1972). At approximately 3-5-4 0 kg, C. equitum is the smallest known swan.

Wing span

Feathered wing span for C. equitum was estimated using the wing-skeleton span and the weight.
1, wing-skeleton span = sum of wing-bone lengths (from Table 1) + width between caputi humeri
( ss 79 mm, text-fig. 3). Wing structure is similar in C. equitum to that in the extant swans (see
earlier) where the ratio (wing-skeleton span : feathered wing span) se 0-7-0-8 (Table 4). Therefore,

736

PALAEONTOLOGY, VOLUME 31

a

a

R

fee

£

c

cd

a

C/5

00

G

cd

3

Q

<G

a>

jd

Oh

G

cd

T3

G

cd

s

a

a

G

cd

Oh

C/D

G

O

W)

G

X)

cd

H

bJ)

G

(N

X

B

S

&Q

G

G

O

x>

it)

C

• • W)

, G G

w) S '3
.5 &.1
c "S

n £ <L) ^

.2 G c

5 j S 5^

^y/ i-X C) Oh

MH c/5 G— I C/5

6D

G

O, C

£ p

cd

W ,-H

£ §

Q-> *- r-<

^ O G
<D ^-v .G

h> c r-

J? cd ^ ^
£ Oh g ON
Mh co cd ’ — 1

G X
, O O

• G 0) cd Qh
> ^ a 03

£> C/5 <75 ' — '

X

G

jd

15

g:

Oh

J-H

o

cd*

6

C/5

G

‘5b

S

|?

cd G

O G
<u w-

O a .

G in

Hi

>R a> G

Q -a G

V©

o

G-

G\

G*

o\

o

o

G-

qs

G-

r-

<N

G-

G\

a q3

a

o ^

Oh P

o 5

GO

©

6\

v©

G"

<L>

G

cd

G

£

£>

'5b

o

-G

G

<

s

O 15

-H ^

J-H G
cd cd

3 15

G G3

Oh

6" <“
a 2
o & ,

O H=!

H-> *"H

C O

if

<D 5“

Eh c/3

3 G

£

CO

<L>

G

s

-G

a

15

6

cd jd

3 3

^ < TO to

2 Q u u

NORTHCOTE: PLEISTOCENE ‘SWAN-GOOSE’

737

wing span for C. equitum k 1 -44-1-65 m (Table 4). 2, for all birds, wing span = IT x weight0'33
(Tucker 1977). Using estimated weights of C. equitum (Table 3), its wing span = 1-50-1-78 m,
mean « 1-68 m. For swans, wing span = a constant x weight 0 39 (Alexander 1971). Wing span of
C. equitum = 1-30-1-88 m, 1-29-1-96 m, or 1-20-1-78 m, corresponding respectively to Whooper,
Bewick’s, or Mute Swans. The ratios (humerus : carpometacarpus) and (chief phalanx : carpometa-
carpus) in equitum are like Mute rather than Whooper Swans (see earlier), so the best estimate
may be 1-20-1-78 m. In summary, the feathered wing span of C. equitum k 1-44-1-65 m (using
wing skeleton), 1-50-1-78 m, mean 1 -68 m or 1-20-1 -78 m (using scaling formulae).

Habitat and habit

Today Malta is relatively arid and bare, but remains of pygmy elephants and hippopotami, giant
dormice, and land and freshwater turtles (Adams 1870, 1877), cranes (Lydekker 1890) of two
species (Northcote 19826, c, 1984, 1984-1985), geese (Parker 1865, 1869; Bate 1916), and two other
swan species (Parker 1865, 1869; Northcote 1982a, 1981-1983) besides C. equitum , suggest that
about 125 000 years ago there were stretches of fresh water and marshes besides that between Sicily
and Malta (Northcote 1982a) and luxuriant vegetation including deciduous forest. The climate
was probably warmer and moister than now as it was elsewhere in the Mediterranean according
to Van der Hammen et al. (1971). Parker (1865, 1869) and Bate (1916) thought that foxes preyed
on this fauna, though Falconer (1868) and Adams (1870) commented on the absence of carnivore
bones from their Maltese excavations and Sondaar and Boekschoten (1967) and Sondaar (1971)
considered that there were no large carnivores on Mediterranean islands in the Pleistocene. Zammit-
Maempel (1982, p. 254) listed occurrences of bear remains on Malta but noted their sparcity and
rarity.

No structural evidence supports Bate’s (1916) statement that C. equitum , like some other island
birds, was flightless. Its wings bore flight feathers, it was light enough to fly (the upper limit ss 12
kg, Pennycuick 1972) and there was no reduction of coracoid or wing bones in ‘stoutness’ or
relative lengths. In addition, the ratio (length of crista pectoralis : humerus (i.e. insertion of the
main flight muscles)) is similar in equitum (0-298) to Greylag Goose and Whooper Swan (approx.
0-300), indicating that it had fully formed flight muscles. (McGowan (1986), however, has shown
the wing musculature of the flightless rail Gallirallus australis to be indistinguishable from Fulica
americana , a fully flighted coot.) Characters of the proximal wing skeleton, as well as of the femur
and tarsometatarsus indicate that C. equitum may have habitually taken off and alighted on level
ground and was perhaps more terrestrial than extant swans. Taken together with the manoeuvrability
conferred by its smaller size, its more ‘elliptical’ wing shape, and perhaps its mode of elbow flexion,
these factors suggest that equitum could live in such enclosed habitats as marshes, reed beds, and
fen carr. C. equitum probably could not fly far because of its ‘elliptical’ wing shape and broad
body (that are associated with slower flight, McFarland et al. 1979), together with wing-bone
proportions less like the migratory Whooper and Bewick’s Swans (that have tapered ‘high speed’
wings) than the relatively sedentary Mutes.

C. equitum occurred centrally on the island as well as in brackish and marine deposits (Brodkorb
1964). Evidently its large salt gland allowed it to eat plants from different areas. C. equitum was
associated with the giant extinct swan C.falconeri Parker, 1865 and with Whooper Swan (Parker
1865, 1869; Bate 1916). Remains of the last named swan also occur in Devensian (Weichselian) as
well as Ipswichian (Eemian) Interglacial deposits elsewhere in Europe (Lydekker 1891; Northcote
19796), but the extinct dwarf and giant swans occur only in Interglacial deposits on Malta. Though
able to forage on land, Whooper Swans eat mainly leaves, stems, and roots in shallow water
(Cramp and Simmons 1977). Comparatively smaller herbivores tend toward a more selective
browsing diet of higher calorific value (Prothero and Sereno 1982), so C. equitum probably ate
mainly roots, shoots, flowers, fruits, and seeds on the water’s edge. Comparatively larger herbivores
nearly always eat food of lower calorific value (Prothero and Sereno 1982), so C. falconeri , an
inland grazer (Northcote 1982a, 1981-1983), probably consumed a higher ratio of fibre to protein
by cropping unselected grasses and whole plants on drier ground. Morphological differentiation.

738 PALAEONTOLOGY, VOLUME 31

by conferring ability to utilize different subniches, could thus have facilitated coexistence of the
three swan species.

Evolution and extinction

The Maltese islands and Sicily are remnants of the land that emerged from the early Pliocene
Mediterranean about five million years ago (Zammit-Maempel 1977). Ensuing Pleistocene climatic
fluctuations facilitated rapid speciation (McFarland et al. 1979). During the 21 000 years of the
last Interglacial, Siculo-Malta was isolated from mainland Italy by strong currents in the Straits
of Messina (Sondaar and Boekschoten 1967). In both C. equitum and C. falconeri the change in
size and assumption (or retention) of terrestrial, sedentary habits were probably adaptations to
isolation in a mild climate, with plentiful food and rarity or absence of large carnivores.

In overall structure, both C. equitum and C. falconeri (Northcote 1982a, 1981-1983) differ from
Mute, but resemble Whooper, Swans. Presumably, the actively flying Eurasian stock that, according
to Johnsgard (1974), gave rise to C. cygnus, also produced C. equitum and C. falconeri. Terrestrial
Anseriformes are more advanced than aquatic (Johnsgard 1965; Olson and Feduccia 1980), so
that the terrestrial Whooper are more advanced than Mute Swans. C. equitum and C. falconeri
were probably even more terrestrial than Whooper Swans. This characteristic, taken with their
respective nanism and gigantism, indicates that both were even more advanced than Whooper
Swans. However, the goose-like features and smaller size of C. equitum may be parallelisms or
they may be primitive retentions, and so may the Mute-like hand proportions. In addition, remains
of C. equitum chiefly represent fore-limbs, while those of C. falconeri are chiefly hind-limbs so that
it is not possible to propose a more specific hypothesis of interrelationships.

Dwarf and giant swans probably evolved from separate invasions (maybe at different times) of
ancestors derived on the mainland by allopatric speciation. It is unlikely that C. equitum and C.
falconeri evolved sympatrically from, or in parallel with, intermediately sized swans such as
Whooper Swans on Siculo-Malta because, as shown by Kondrashov and Mina (1986), an increased
proportion of intermediates resulting from breeding with marginal populations would have
prevented phenotypic separation of the marginals.

Rather rapid environmental changes accompanied the fall in temperature that terminated the
Interglacial (Charlesworth 1957; Starkel 1977). Tectonic disturbances (Zammit-Maempel 1977)
caused sea submergence of the area between Sicily and Malta and produced faulting, upthrowing,
and tilting further south (Pedley 1981). Habitats were lost as a result of the sea-level changes and
torrential rainfall eroded the sloping surfaces. These factors, combined with few, if any, large
predators, may have led to overcrowding, overgrazing, and starvation. The less specialized Whooper
Swans, migrants to Siculo-Malta, survived. The endemic C. falconeri and C. equitum , like many
island bird species (Diamond 1981; McGowan 1986) may have been reluctant, rather than unable,
to cross water.

Acknowledgements. I am grateful for the co-operation of the curators who gave me access to their collections.
I received invaluable help from Dr G. Zammit-Maempel of the National Museum of Natural History, Malta,
and from Mr C. A. Walker (London) and Mr G. S. Cowles (Tring) of the British Museum (Natural History)
for which I thank them. I am indebted to Messrs M. J. Ashby and J. W. Rodford for assistance with the
illustrations and to Mrs A. Maxwell for preparing the typescript. I thank Dr K. A. Joysey for help in other
ways. I am very grateful to Professor H. B. Whittington and the referees for helpful comments.

REFERENCES

adams, A. L. 1870. Notes of a naturalist in the Nile Valley and Malta , 117-187. Edmonton and Douglas,
Edinburgh.

— 1877. On gigantic land-tortoises and a small freshwater species from the ossiferous caverns of Malta,
together with a list of their fossil fauna; and a note on chelonian remains from the rock-cavities of
Gibraltar. Q. Jl geol. Soc., bond. 33, 177-191.

NORTHCOTE: PLEISTOCENE SWAN-GOOSE’

739

Alexander, R. M. 1971. Size and shape. The Institute of Biology’s Studies in Biology, 29. Edward Arnold,
London.

1983. Allometry of the leg bones of moas (Dinornithes) and other birds. J. Zool. Loud. 200, 215-231

bate, d. m. a. 1916. On a small collection of vertebrate remains from the Har Dalam cavern, Malta; with
note on a new species of the genus Cygnus. Proc. zool. Soc. Loud. 28, 421-430.
baumel, J. J. 1979. Osteologia and Myologia. In baumel, j. j., king, a. s., lucas, a. m., breazile, j. e. and
evans, h. E. (eds.). Nomina anatomica avium, 53-123, 123-173. Academic Press, London.
beddard, F. E. 1898. The structure and classification of birds. Longmans, Green and Co., London.
bellairs, a. and jenkin, c. r. 1960. The skeleton of birds. In marshall, a. j. (ed.). Biology and comparative
physiology of birds , 1, Academic Press, London.
bock, w. j. 1962. The pneumatic fossa of the humerus in the passeres. Auk, 79, 425 443.
brodkorb, p. 1964. Catalogue of fossil birds. Part 2. Anseriformes through Galliformes. Bull. Fla St. Mus.
biol. Sci. 8, 195-335.

brown, r. h. j. 1963. The flight of birds. Biol. Rev. 38, 460-489.

CHARLESWORTH, s. K. 1957. The Quaternary Era, with special reference to its glaciation. Arnold, London.
cramp, s. and Simmons, k. e. l. (eds.). 1977. The birds of the Western Palearctic, 1, University Press, Oxford.
de bruijn, h. 1966. On the Pleistocene Gliridae (Mammalia, Rodentia) from Malta and Mallorca. Proc. K.
ned. Akad. Wet. B 69, 480 496.

delacour j. 1954. Systematic list and Introduction. The waterfowl of the world , 1, 13-18. Country Life,
London.

despott, G. 1928/1929. Annual Report on the Working of the Museum Department during 1928-29, S.VII-
VIII. National Museum of Malta, Malta.

diamond, J. m. 1981. Llightlessness and fear of flying in island species. Nature, 293, 507-508.
falconer, h. 1868. IV. On the fossil remains of Elephas melitensis an extinct pigmy species of elephant; and
of other Mammalia, etc., from the ossiferous caves of Malta. Palaeont. Mem. Falconer , 2, 292 308.
Gascoyne, m., schwarcz, H. p. and ford, d. c. 1983. Uranium-series ages of speleothem from northwest
England; correlation with Quaternary climate. Phil. Trans. R. Soc. Lond. B 301, 143-164.
george, J. c. and berger, a. j. 1966. Avian myology. Academic Press, London and New York.
holmes, w. n. and Phillips, J. G. 1985. The avian salt gland. Biol. Rev. 60, 213-256.

Howard, h. 1964. Lossil Anseriformes. In delacour, j. (ed.). The waterfowl of the world, 4, 233 326. Country
Life, London.

johnsgard, p. a. 1965. Handbook of waterfowl behaviour. Constable and Co., Ltd, London.

1974. The taxonomy and relationships of the northern swans. Wildfowl, 25, 155 160.

kondrashov, a. s. and mina, m. v. 1986. Sympatric speciation: when is it possible? Biol. Jl Linn. Soc. 27,
201-223.

kuhry, b. and marcus, l. F. 1977. Bivariate linear models in biometry. Syst. Zool. 26, 201-209.
lambrecht, k. 1931. Cygnopterus und Cygnavus, zwei fossile schwane aus dem Tertiar Europas. Bull. Mus.
R. Hist. nat. Belg. 7, 1-6.

1933. Handbuch der Palaeornithologie. Borntraeger, Berlin.

lydekker, r. 1890. On the remains of some large extinct birds from the cavern-deposits of Malta. Proc. zool.
Soc. Lond. 28, 403-411.

1891. Catalogue of the fossil birds in the British Museum ( Natural History). British Museum (Natural

History), London.

mcfarland, w. N., pough, F. h., cade, t. j. and heiser, j. b. 1979. Vertebrate life. Macmillan, New York.
mcgowan, c. 1986. The wing musculature of the Weka (Gallir alius australis), a flightless rail endemic to New
Zealand. Jl Zool., Lond. 210, 305-346.
mcmahon, t. a. 1973. Size and shape in biology. Science, 179, 1201-1204.

— — 1975. Using body size to understand the structural design of animals; quadrupedal locomotion. J. appl.
Physiol. 39, 619 627.

northcote, e. m. 1979a. Determination of age and sex of long bones of Mute Swan Cygnus olor. Ibis, 121,
74-80.

19797). Comparative and historical studies of European Quaternary swans and other aquatic birds.

Ph.D. thesis (unpublished). University of Cambridge.

1981. Differences in weight and habit of Whooper Cygnus cygnus cygnus and Mute C. olor swans in

relation to differences in their long bones. Bull. Br. Orn. Club, 101, 266 267.

— 1982a. Size, form and habit of the extinct Maltese Swan Cygnus falconeri. Ibis , 124, 148-159.

1982 b. The extinct Maltese Crane Grus melitensis. Ibis, 124, 76-80.

740

PALAEONTOLOGY, VOLUME 31

northcote, e. m. 1982c. Sympatry of Common Cranes Grus grus with larger cranes in the last c. 125 000
years. Bull. Br. Orn. Club, 102, 141 142.

1981 1983. The giant Maltese Swan. Il-Merill, 22, 6-8.

1984. Crane Grus fossils from the Maltese Pleistocene. Palaeontology, 27, 729- 735.

1984-1985. The giant Maltese crane. Il-Merill, 23, 1-4.

olson, s. l. and feduccia, a. 1980. Presbyornis and the origin of the Anserifornres (Aves; Charadriomorphae).
Smithson Contr. Zool. 323, 1 24.

Parker, w. K. 1865. Preliminary notes on some fossil birds from the Zebbug Cave, Malta. Proc. zool. Soc.
Lond. 1865, 752-753.

— 1869. On some fossil birds from the Zebbug Cave, Malta. Trans, zool. Soc. Lond. 6, 119-124.
pedley, h. m. 1981. Quaternary sediments, Malta. In bosence, d. w. j., pedley, h. m. and rose, e. p. f. (eds.).

Field guide to the Mid-Tertiary carbonate facies of the Maltese Islands, 71-80. Palaeontological Association,
London.

pennycuick, c. J. 1972. Animal flight . Arnold, London.

prothero, d. r. and sereno, p. c. 1982. Allometry and paleoecology of medial Miocene dwarf rhinoceroses
from the Texas Gulf Coastal Plain. Paleobiology , 8, 16-30.
scott, k. m. 1983. Prediction of body weight of fossil Artiodactyla. Zool. Jl Linn. Soc. 77, 199-215.
scott, p. and the wildfowl trust, slimbridge (eds.). 1972. The swans. Michael Joseph, London.
shufeldt, r. w. 1909. Osteology of birds. Bull. NY Mus. 130, 1-38.

simpson, G. G., roe, A. and lewontin, R. c. 1960. Quantitative zoology. Harcourt, Brace and World, United
States.

sondaar, p. y. 1971. Paleozoogeography of the Pleistocene mammals from the Aegean. In strid, a. (ed. ).
Evolution in the Aegean. Opera Botanica, 30, 65-70.

— and boekschoten, G. J. 1967. Quaternary mammals in the South Aegean Island Arc; with notes on
other fossil mammals from the coastal regions of the Mediterranean. Proc. K. ned. Alcad. Wet. B 70, 565-
576.

starkel, l. 1977. The palaeogeography of mid- and east Europe during the last cold stage, with west
European comparisons. Phil. Trans. R. Soc. Lond. B 280, 351-372.
stejneger, l. 1882. Outlines of a monograph of the Cygninae. Proc. US natn. Mus. 5, 174-221.
sy, M. 1936. Funktionell-anatomische Untersuchungen am Vogeltliigel. Jl Orn., Lpz. 84, 199 296.
tucker, v. a. 1977. Scaling and avian flight. In pedley, t. j. (ed.). Scale effects in animal locomotion, 497
509. Academic Press, London.

vanden berge, j. c. 1979. Myologia. In baumel, j. j., king, a. s., lucas, a. m., breazile, j. e. and evans,
h. e. (eds.). Nomina anatomica avium, 175-219. Academic Press, London.
van der hammen, T., wiJMSTRA, T. a. and ZAGWIJN, w. h. 1971. The floral record of the late Cenozoic of
Europe. In turekian, k. k. (ed.). The Late Cenozoic Glacial Ages, 391-424. Yale University Press, Newhaven
and London.

verheyen, r. 1953. Bijdrage tot de Osteologie en de Systematik van der Anseriformes. Le Gerfaut, 43, 373-
497.

— 1955. La systematique des Anseriformes basee sur l’osteologie comparee. Bull. Inst. R. Sci. nat. Belg.
31 (35), 1-18; (36), 116; (37), 1-22; (38), 1-16.

woolfenden, G. E. 1961. Postcranial osteology of the waterfowl. Bull. Fla St. Mus. biol. Sci. 6, 1-129.
zammit-maempel, G. 1977. An outline of Maltese geology. Progress Press, Malta.

— 1982. A Maltese Pleistocene sequence capped by volcanic tufa. Atti. Soc. Tosc. Sci. nat. Mem. A 88,
243-260.

Appendices 1 and 2 have been deposited with the British Library, Boston Spa, Yorkshire, UK, as Supplemen-
tary Publication No. SUP BLL 14035 (7 pages).

E. MARJORIE NORTHCOTE

Department of Zoology
University of Cambridge

Typescript received 25 February 1987 Downing Street

Revised typescript received 25 September 1987 Cambridge CB2 3EJ

A NEW ALGA FROM THE CARBONIFEROUS
FROSTERLEY MARBLE OF NORTHERN ENGLAND

by GRAHAM F. ELLIOTT

Abstract. A new alga from the Frosterley Marble (Namurian; Carboniferous) of northern England is
reconstructed from fragmentary material. It is compared with the Carboniferous genera Kulikia and
Sphinctoporella , sharing with them the distinctive profusion of spherical cavities within the calcified axial
surround or 'sleeve'. It differs from them in being formed of successional separate calcified discs or verticil
units, which came apart after death, and is thus described as Frosterleyella diaspora gen. et sp. nov.

Frosterley Marble was much worked in medieval and later times. The word 'marble', as with
the better-known Purbeck Marble, is used in the popular and trade sense, not scientifically. It is
in fact a hard limestone taking a good polish, in which the light-grey fossils, largely corals, stand
out in section against the dark background to give an attractive pattern. Geologically it comes
from the uppermost of three fossiliferous bands in the local Great Limestone of Namurian
(Carboniferous) age in the northern Pennines of northern England (Johnson 1958). Mills and Hull
(1976, p. 31) suggest from the lithology and from the mode of occurrence of the corals in the rock,
that original deposition of the sediment was probably under the influence of strong currents or
waves.

This is confirmed by thin-section study of the rock, which, where detail is not destroyed by
diagenesis, shows a profusion of ill-sorted organic debris and microfossils. Debris of echinoderms
(mostly crinoids), Bryozoa, brachiopod test and spines, and whole small foraminiferidids are
common, and the original calcite is moderately well preserved. Less common are fragmentary
remains of an obvious dasyclad alga, frequently seen as rings with smooth inner surface and very
ragged exterior. The presumed original organic aragonite of the living plant has been converted
to white calcite usually with near-complete obliteration of fine organic detail. The present study
was made to see how much could be reconstructed from the unpromising abraded and diageneticized
remains. Other algae are rare; noted were INanopora (Wood 1964), the problematic Hypocaustella
(Elliott 1980), and Aphralysia which was described as an alga, but later interpreted as a
foraminiferidid (Garwood 1914; Belka 1981).

DESCRIPTION

Examination of the dasyclad rings reveals that those examples showing most of the calcite around the inner
axial cavity, i.e. those less worn and not too ill-preserved, show the calcite to be full of close-set spherical
cavities. Dimensions vary in random cut, but have been seen up to 73 /<m in diameter, possibly the original
maximum size. They correspond in appearance to 'sporangia' or reproductive cavities seen in other dasyclads.
No traces of connecting branches between them are to be seen. Along one random radius, from axial cavity
to outer worn edge, there is space for three cavities and also the serrated, broken-away edge of a fourth, so
permitting a diameter to be postulated. It is such broken spheres, forming the outer edge of abraded rings,
which give these their characteristic ragged appearance. The axial diameter can be measured; if a d/D ratio
of 25-33% is postulated, this would give an estimated outer diameter of 1.1 1.5 mm. No vertical sections of
dasyclad calcareous tube ('sleeve') have been observed, but two vertical sections of single rings indicate the
plant to have been built up in life from consecutive discs or whorls which came apart after death, as is known
in certain other dasyclads. Each disc (whorl, verticil) had a central axial cavity, keg-shaped, with maximum
diameter equatorially, narrowing a little above and below. Numerous random oblique sections confirm this

(Palaeontology, Vol. 31, Part 3, 1988, pp. 741-745.|

© The Palaeontological Association

742

PALAEONTOLOGY, VOLUME 31

when the plane of section traverses above or below the aperture. Two specimens suggest that in life a very
thin calcification may have joined the consecutive discs at the axial cavity surface.

The combined evidence suggests a plant built up of numerous successive calcareous whorls, each circular
in outline, flatfish, biconvex, and externally rounded, giving a deeply incised outline to the whole plant,
somewhat like that of Queenslandella (Mamet and Roux 1983). The calcareous skeleton of the Frosterley
alga contains numerous adjacent spherical cavities; the branch system connecting them is not known.

CLASSIFICATION

The key character in seeking to classify the Frosterley alga is the multitude of spherical cavities
within the calcified discs. Two known Carboniferous genera show this: Kulikia (Golubtsov 1961)
and Sphinctoporella (Mamet and Rudloff 1972). In the case of Kulikia , the author figured a selection
of thin-section specimens, but managed to emphasize the spherical nature of the whorls at the
expense of the continuous axial cavity, although several of his figures show this latter feature
clearly. This led to the confusion of later authors. Mamet and Rudloff (1972) and Mamet and
Roux (1975) in describing Sphinctoporella , and in describing a new species in a queried allocation
to this genus, did not compare it with Kulikia. Not until later (Mamet et al. 1981) was this
comparison made.

Skompski (1984) published an account of Kulikia from new, magnificently preserved pyritic
material in Poland. He was able to develop both solid, detached, three-dimensional specimens and
very clear thin-sections, better than any before. In particular, he described the branch system of
Kulikia , not clearly preserved in the previous limestone matrices. He showed that the spherical
cavities are occasioned by two kinds of branches, which he called active and passive. Active
branches divide and produce many connected spheres, which he suggests are the assimilatory
branches. Passive branches are single spheres or ovoids, immediately outside the axial cavity, which
he suggests were reproductive in function. He re-defined Kulikia , transferring the queried species
rozovskaiae from Sphinctoporella to Kulikia , because it has similar branch structures to the type-
species K. sphaerica. The new generic diagnosis is in one respect a little restrictive; it includes
tetragonal axial symmetry as a character. But, from the nature of dasyclad branching, polygonal
axial symmetry is also likely to occur in other species if found. However, this is a minor point.
Now that the Kulikia branch system is known, with important significance for our knowledge of
dasyclad reproductive evolution, as Skompski emphasizes, how should Sphinctoporella be regarded?
Kulikia and Sphinctoporella seem closely related; would the latter, known only in normal limestone
preservation, show Kulikia- type branching if found better preserved?

There is here more than one taxonomic possibility. Sphinctoporella spp. could be transferred to
Kulikia , assuming that they will be found eventually to have Kulikia-type branching. However,
this is a character of more than generic importance, and it seems right to leave Sphinctoporella as
a valid genus, for probable eventual ‘tribal’ grouping with Kulikia. It is against this background
that the Frosterley alga must be considered. It is the worst preserved of them all, but shows the
distinctive numerous calcified spheres of the other two. In one sense its separate whorl units are
the final stage of the incised outline of Sphinctoporella , but this development occurs also in other
dasyclad lineages.

I therefore describe it below as a new genus and species, clearly related to Kulikia and
Sphinctoporella , and I hope that the incompleteness of our knowledge will be rectified in the future.

SYSTEMATIC PALAEONTOLOGY
Frosterleyella gen. nov.

Type species. Frosterleyella diaspora sp. nov. Carboniferous (Namurian) of northern England.

Derivation of name. After the type-locality of Frosterley, Weardale, County Durham, England.

Diagnosis. Calcified successional dasyclad verticil discs, biconvex in vertical section and circular in

ELLIOTT: CARBONIFEROUS CALCIFIED ALGA

743

text-fig. 1. Frosterleyella diaspora gen. et sp. nov. Thin-sections of abraded examples from the Frosterley
Marble (Namurian); Shittlehope Quarry, between Stanhope and Frosterley, Weardale, County Durham,
England. British Museum (Natural Flistory), Dept. Palaeontology, registered nos. V. 62755c; (a, b), V.62755g
(c), V. 62755;; (d), all x 60. a, equatorial transverse section of uniformly worn example; biological detail poor
but partly recognizable, b, transverse section showing central (inner periaxial) calcification, with incomplete
area of ‘sleeve’ wall-calcite to right, showing clearly numerous spherical cavities, c, vertical section of single,
small, much worn example, all biological detail destroyed by diagenetic calcification, d, oblique section,
poorly preserved, but showing traces of cavities (ragged exterior, and interior of calcite), and possible thin

‘below verticil’ periaxial calcification.

outline, each with central keg-shaped axial cavity. Calcification full of spherical cavities, believed
to be similarly organized to those of Kulikia.

Frosterleyella diaspora sp. nov.

Text-fig. 1a d

Derivation of name. Diaspora or dispersal, a reference to the scattered occurrence of the broken fossil remains.

744

PALAEONTOLOGY, VOLUME 31

text-fig. 2. Reconstruction of living Frosterleyella , gen. nov. Basal regions
conjectural, x 17 approx.

Syntypes. The specimens shown in text-fig. 1a d. British Museum (Natural History), Department of
Palaeontology, registered numbers V. 62755a, g, n.

Horizon and Locality. Frosterley Band, in Great Limestone Formation (Namurian), Shittlehope Quarry,
between Stanhope and Frosterley, Weardale, County Durham, England.

Diagnosis. As for genus; thickness of verticil discs up to 0-5 mm., estimated external diameter 1-1-
1-5 mm.; diameter of axial cavity up to 0366 mm.; diameter of spherical cavities up to 0 073 mm.

Other material. Numerous examples in thin-sections, registered nos. V.62755 a-z inclusive.

Acknowledgements. My grateful thanks are due to Dr B. R. Rosen (B.M.(N.H.)), who very kindly provided
me with a large piece of Frosterley Marble for my study, and to Mr M. Crawley (B.M.(N.H.)), who prepared
text-fig. 2.

REFERENCES

belka, z. 1981. The alleged algal genus Aphralysia is a foraminifer. Neues Jb. Geol. Paldont. Mb. H5, 257-266.
elliott, g. f. 1980. A new, possibly algal, microproblematicum from the Lower Carboniferous of England.
Bull. Br. Mus. nat. Hist. (Geol.), 33 (2), 73 77.

garwood, e. j. 1914. Some new rock-building organisms from the lower Carboniferous beds of Westmorland.
Geol. Mag. 61, 265-271.

golubtsov, v. k. 1961. Kulikia', a new genus of calcareous alga from the Visean. Akad. Nauk. Biel. SSR,
Inst. Geol. Sci. Minsk , 3, 348-353. [in Russian.]

JOHNSON, G. a. l 1958. Biostromes in the Namurian Great Limestone of northern England. Palaeontology,
1, 147 157.

mamet, b., dejonghe, l. and roux, a. 1981. Sur la presence de Kulikia (Dasycladacee) dans le Viseen des
Grands Malades (Jambes). Bull. Soc. belg. Geol. 89, 291-295.

and roux, a. 1975. Dasycladacees devoniennes et carbonileres de la Tethys occidenlale. Revta Esp.
Micropaleont. 7, 245-295.

1983. Algues devono-carboniferes de l’Australie. Rev. Micropaleont. 26, 63-131.

ELLIOTT: CARBONIFEROUS CALCIFIED ALGA

745

mamet, b. and rudloff, b. 1972. Algues carboniferes de la partie septentrionale de l’Amerique du Nord. Rev.
Micropaleont. 15, 75-1 14.

mills, d. a. c. and hull, j. h. 1976. Geology of the country around Barnard Castle. Mem. geol. Surv. Gt
Britain , 365 pp.

skompski, s. 1984. The functional morphology of the Carboniferous dasycladacean genus Kulikia. Neues Jb.
Geol. Paldont. Mb. H.7, 427-436.

wood, a. 1964. A new dasycladacean alga, Nanopora , from the Lower Carboniferous of England and
Kazakhstan. Palaeontology ,7, 181 185.

Typescript received 26 March 1987
Revised typescript received 29 November 1987

G. F. ELLIOTT

Department of Palaeontology
British Museum (Natural History)
Cromwell Road, London SW7 5BD

Note added in proof. After submission and revision of this paper I learned of Dr S. Skompski’s description
of a poorly preserved dasyclad as Diploporeae gen. indet. Some random sections of his Form A of this Polish
material appear very similar if not identical to those of Frosterleyella. There are slight differences in the
features seen and in our interpretations, but we agree on general structure and on a close relationship to
Kulikia.

REFERENCE

skompski, s 1986. Upper Visean calcareous Algae from the Lublin Coal Basin. Acta geol pol. 36, 251 280.

THE MOSASAUR GO RO NYO SAURUS FROM
THE UPPER CRETACEOUS OF SOKOTO STATE,

NIGERIA

by T. SOLI AR

Abstract. New mosasaur material from the Upper Cretaceous of Sokoto State, Nigeria, was described in
detail by Azzaroli et al. (1972, 1975). Largely on the basis of what they interpreted as a highly unusual jugal
bone in the skull, they erected in 1972, the new genus Goronyosaurus , and subsequently (1975) placed it in a
new subfamily, the Goronyosaurinae, representing a totally unique development of a squamate skull.

A reassessment of the material suggests that the description of the jugal by Azzaroli et al. (1972, 1975) is
incorrect. However, other characters described by them, plus new information added here, vindicates the
erection of a new genus, which can be tentatively assigned to the Tylosaurinae.

Mosasaurs, large marine varanid lizards, were widely distributed in the Upper Cretaceous in
both the Old and New Worlds. The first mosasaur specimen was found in 1780 by Dr Hoffmann
in the district of Maastricht, Holland, in rocks of Upper Cretaceous age. Over fifty years elapsed
before the fossil was named Mosasaurus by Conybeare (in Parkinson 1822). Dollo (1890) divided
the family Mosasauridae into three groups based on the degree of development of the rostrum—
microrhynchous, mesorhynchous, and megarhynchous.

Williston (1897) placed Dollo’s groups in the subfamilies Mosasaurinae, Platecarpinae, and
Tylosaurinae. In Camp and Allison’s (1961) classification the Platecarpinae is replaced by the
Plioplatecarpinae.

MOSASAURS FROM WEST AFRICA

Mesozoic vertebrate remains in West Africa are relatively rare. In contrast Sokoto State, in north-
west Nigeria, yields quantities of marine vertebrates from the uppermost Cretaceous, Maastrichtian
Dukamaje Formation (Swinton 1930; Jones 1948; Reyment 1965; Kogbe 1973; Petters 1979a, b\
Halstead 1979c).

Vertebrate remains from the Sokoto region were first described by Nopcsa (1925) as including
dinosaurs from the Tertiary. Swinton (1930) subsequently demonstrated them to be crocodilian
and of Palaeocene age (see Halstead and Middleton 1976). However, Swinton (1930) described
further material from a Cretaceous horizon, the ‘Mosasaurus shales’. On the basis of postcranials
and jaw fragments Swinton (1930) erected the new species, M. nigeriensis. A new crocodilian from
the Cretaceous was described under the name Sokotosuchus (Halstead 1973, 1975; Buffetaut 1976,
1979) and a new turtle genus Sokotochelys (Walker 1979; Halstead 1979a, b) from the same
horizon.

Until 1970, the only work on African mosasaurs, besides Swinton’s in 1930, was on rather
isolated finds. Broom (1912) described Tylosaurus capensis from South Africa; Deperet and Russo
(1925) described Leiodon anceps from Morocco and Arambourg (1952) described further mosasaur
remains; Antunes (1964) described a species of mosasaur from Angola. Further mosasaur material
has been recorded from Libya (Quass 1902) and Egypt (Stromer and Weiler 1930; Zdansky 1935;
Leonardi and Malaroda 1946).

In December 1969 to January 1970 and February and March 1971 expeditions were mounted
by Professor A. Azzaroli of Florence University, Italy, to collect fossil vertebrates in Sokoto State,
Nigeria. New mosasaur material from the Goronyo district of Sokoto State was discovered by de

| Palaeontology, Vol. 31, Part 3, 1988, pp. 747-762. |

© The Palaeontological Association

748

PALAEONTOLOGY, VOLUME 31

Guili et al. (1970) and described in detail by Azzaroli et al. (1972, 1975). The remains comprised
a large number of vertebrae, two right humeri, a fragment of a premaxilla, eight fragments of
mandibles, a fragment of a pterygoid of a small size, and an almost complete skull, albeit badly
crushed.

The first mosasaur remains recognized from Nigeria were described by Swinton (1930). The two
associated dorsal vertebrae (BMNH R5674) were selected as the lectotype of M. nigeriensis, by
Halstead and Middleton (1982). The new materials discovered by the Italian expeditions and
figured by de Guili et al. (1970), Azzaroli et al. (1972, 1975), and Halstead and Middleton (1982),
were assigned to the same species.

A number of unusual features were described by Azzaroli et al. (1972, 1975) including the
following:

the maxillaries extended beyond the posterior margin of the orbits;

the jugal had a broad ascending ramus;

the inner surface of the frontal bore a fully closed canal housing the olfactory lobes;

m the pterygoids the roots of the ectopterygoidal processes and of the quadratic rami
were widely spaced and the ectopterygoids articulated with the posterior margins of the
ectopterygoidal processes.

On the basis of these characters Azzaroli et al. ( 1972) established the new genus Goronyosaurus.
However, as they did not consider it possible to place it in any of the existing subfamilies of
mosasaurs, Azzaroli et al. (1975) subsequently erected the new subfamily, the Goronyosaurinae
for its reception.

A key character and also by far the most contentious related to the features of the jugal. In the
normal squamate skull the lower temporal bar is missing and the jugal which forms the anterior
border of the infratemporal vacuity is greatly reduced. According to Azzaroli et al. (1975) in
Goronyosaurus the jugal was enormously expanded into a deep elongated sheet covering a large
proportion of the infratemporal vacuity. This would make it a unique feature in the squamate
reptile skull.

GEOLOGY

An early note by Raeburn and Tattam (1930) on the geology of the Sokoto region was later
followed by a more detailed account by Jones (1948).

The Cretaceous Dukamaje Formation and the Palaeocene Dange Formation yield abundant
vertebrate remains. Vertebrates of the Dukamaje Formation are concentrated in a bone bed about
40 cm thick and belong to the Maastrichtian period (Swinton 1930; White 1934; Jones 1948;
Reyment 1965; Kogbe 1973; Petters 1979a, b ; Halstead 1979c, 1980). See text-fig. 1.

The type localities listed by Reyment (1965) were simply the villages which gave their names to
the formations. Kogbe (1973) published detailed sections which were subsequently revised by
Petters (1979a, b) and Halstead (1979c, 1980). See text-fig. 2.

SYSTEMATIC PALAEONTOLOGY

Order squamata
Suborder sauria
Superfamily varanoidea
Family mosasauridae
Subfamily tylosaurinae
Genus goronyosaurus Azzaroli et al., 1972
Goronyosaurus nigeriensis (Swinton, 1930)

Text-figs. 3-6, 7c-f

Type species. Goronyosaurus nigeriensis.

SOLI AR: CRETACEOUS MOSASAUR FROM NIGERIA

749

text-fig. 1. Geological map of Sokoto region (after Parker and Carter 1965).

Lectotype. Two associated dorsal vertebrae, R5674, housed in the British Museum (Natural History), London,
figured Swinton 1930, pi. 10, fig. 2 a-c.

Type horizon and locality. Gypsiferous shale member, Dukamaje Formation, Maastrichtian, Cretaceous;
Benbow Hill, south-east of village of Gilbedi, Sokoto State, Nigeria.

Diagnosis. Small premaxillary rostrum, with wide internarial bar; closed canal on undersurface of
frontals; ectopterygoid process dorsoventrally flattened forming two fork-like processes; postorbital
maxillary teeth.

750

PALAEONTOLOGY, VOLUME 31

QJ

<

FORMATION

UJ

b

Z

<

<

<

h

Y

1

o

O

Q

CL

1-

UJ

<

<

Y

0

2

-J

<

b

DUKAMATE

TSO0A

OOO

OOO

“A —

—

1 1

1

1 1

1

1 1

1

1

1 1

1

1 1

Jill

.111.

— -

PALAEOCENE

6AM BA FM.

KALAM-

BAINA

FM.

DANGE FM.

WURNO

FM.

MAESTRICWTIAN

DUKAMAJE

FM.

TALOKA

FM.

TALOKA

o o o o

O O 0 0

“A — — "

II

1 1

1 1

A

i O 101

~ -r:

TTTTT"

JT-i-i

tz.

—

hnni

WURNO

_ovo_o/^

TT~T

1:1

-I..J-L

0

■-IOM

ooo

ooo

A

Later ite
Shal <z

Gypsiferous Shale
Limestone/ Marl

KEY

Siltstone

Sandstone

burrow's

Coal

text-fig. 2. Geological sections (after Petters 1979a, b ).

SOLI AR: CRETACEOUS MOSASAUR FROM NIGERIA

751

Associated material. Skull and associated postcranial remains, housed in Museum of Geological Survey of
Nigeria, Kaduna South, Nigeria, IGF 14750, figured Azzaroli et al. 1975, pis. I XI; complete set of casts
housed in Museum of Geology and Palaeontology, University of Florence, Italy; cast of skull in Cole
Museum, Department of Zoology, University of Reading.

Horizon and locality of associated material (IGF 14750). Gypsiferous shale member, Dukamaje Formation,
Maastrichtian, Cretaceous; summit of hill overlooking village of Taloka, Sokoto State, Nigeria.

DESCRIPTION OF MATERIAL

Skull

The largest fragment of the skull is crushed with considerably distorted parts. The crushing appears to
have taken place with the right and dorsal surfaces coming to lie uppermost and the left side underneath
(text-fig. 3).

The skull appears to be unusually long and narrow. The length of the skull is 655 mm (estimated
reconstructed length 710 mm) and the width of the widest part of the frontals between the orbits is 1 12 mm.
Table 1 gives a comparison of length/width of skull for several groups of mosasaurs (data from Russell
1967).

TABLE 1

Species

Skull length
in mm

Skull width
in mm

Ratio of skull
length : skull width

Platecarpus ictericus

431

97

4-4: 1

Plotosaurus tuckeri

588

148

4:1

Prognatlwdon overtoni

702

188

3-73: 1

Tylosaurus nepaeolicus

717

116

618: 1

T. proriger

585

118

5:1

T. proriger

600

113

5-3 : 1

Mosasaurus missouriensis

614

142

4-3 : 1

M. maximus

1091

262

416: 1

Plesiotylosaurus crassidens

880

156

5-64: 1

Goronyosaurus n igeriensis

710

112

6-31 : 1

Jugal. Because of the extreme crushing in this region particularly, the boundaries of the left jugal (ventrally
positioned in situ) are difficult to ascertain (see text-fig. 4a). The interpretation of a large postorbital lamina
of the jugal (Azzaroli et al. 1975) would make it a unique feature among mosasaurs and all diapsid reptiles.

In their discussion Azzaroli et al. (1972, 1975) suggest that these broad postorbital rami of the jugals are
superposed to similarly descending rami of the postorbitofrontals and that these two bones were possibly
united by ligamentary tissue making a certain movement possible. There is, however, no precedent for such
a function among mosasaurs (see Callison 1967).

Halstead and Middleton (1982) viewed the description of the jugals by Azzaroli et al. (1975) as doubtful.
I am of the opinion that the broad postorbital laminae of the jugals described by Azzaroli et al. (1975) is an
artefact of crushing. Careful examination reveals that the supposed ascending ramus of the jugal is composed
of at least three separate parts, with the actual ascending ramus of the jugal having been broken off at a
point just above its posteroventral process and displaced. A bone protruding from the left supratemporal
vacuity may represent a portion of the true ascending ramus of the jugal.

The upper part of the ‘lamina’ adjacent to the orbit is composed of a fragment of bone approximately
5 mm thick. The thickness is observed readily along the orbital and postorbitofrontal margins. The ventral
part of the lamina is made up of another thicker tongue-shaped bone approximately 10 mm thick anteriorly,
and flattening out posteriorly to approximately 5 mm. Dorsoposteriorly to this is a thin flange of bone which
has fused to the preceding two fragments of bone.

The horizontal axis of the jugal itself, is on average 19 mm thick and the great and abrupt difference in
thickness with the lamina makes it clear that the horizontal axis of the jugal is unrelated to the ascending

752

PALAEONTOLOGY, VOLUME 31

Premaxilla

A

external Prefrontal Frontals

POP

Parietal

Parietal

foramen

Maxilla

Jugal

Jugal

i IQQmm i
Internarial bar

C

D

text-fig. 3. Goronyosaurus nigeriensis skull, IGF 14750-1, lacking occipital unit, a, lett view, b, right and
dorsal view. Reconstruction of skull, c, left view, d, dorsal view. Abbreviation: POF = postorbitofrontal.

SOLI AR: CRETACEOUS MOSASAUR FROM NIGERIA

753

A

text-fig. 4. a, detail of jugal and surrounding area, b, I and II represent fragments of the coronoid making
up the ‘lamina’ of the jugal, c, jugal of Leiodon perlatus (BMNH R35637-9).

‘lamina’. Despite the absence, in the main, of discernible seams separating the three bones making up the
‘lamina’, demarcation is made clear by abrupt changes in thickness and in surface consistency, and including
seams along the anterior and dorsal boundaries.

With reasonable confidence it is possible to identify part of the lamina adjacent to the orbit as a portion
of the coronoid from the lower jaw. The thin posteriormost wall of the coronoid is in life buttressed
ventromedially to the surangular (see Russell 1967, p. 53). It is at this weak point that the coronoid would
have been likely to have broken otf. Following which the fragment of the coronoid (text-fig. 4bI) could have
been displaced to lie postorbitally at a slight angle forming a part of the ‘lamina’ adjacent to the orbit. The
fragment of bone forming the posteroventral part of the ‘lamina’ is identified with equal confidence as the
remaining part of the coronoid (text-fig. 4bII). The dimensions of this part of the ‘lamina’ are compatible
with this interpretation. It can, therefore, be assumed that the jugal is of typical mosasauroid dimensions,
perhaps comparing more closely with the more robust jugal of L. perlatus (BMNH R35637-9) possessing a
strong posteroventral process (text-fig. 4c). A similar large posteroventral process is also found in T.
nepaeolicus (BMNH R3627).

Premaxillae. The premaxilla has suffered considerable lateral compression. The tooth sockets and tooth bases
are large suggesting that the premaxilla possessed strong teeth. The foramina which mark the exits of the
ophthalmic ramus of the fifth cranial nerve appears to be situated in a cluster either on or very close to the
dorsal mid-line. In Clidastes, Mosasaurus, Plotosaurus , Platecarpus , and Prognat hodon , Russell (1967, p. 16)
states that the foramina are located on either side of the dorsal mid-line while in Tylosaurus they are
distributed randomly on the sides of the rostrum. Although in a giant premaxilla of the genus Prognathodon
sp. indet. (BMNH 49939) the foramina appears to be nearer the mid-line as in Goronyosaurus. The premaxilla
in Goronyosaurus ends abruptly in front of the anterior teeth as in Platecarpus , Plioplatecarpus, Prognathodon ,
and Plesiotylosaurus (Dollo 1889, p. 275 and Russell 1967, p. 16).

754

PALAEONTOLOGY, VOLUME 31

Maxillae. The maxillae are highly unusual as they extend beyond the posterior margin of the orbits and are
also toothed postorbitally (see text-tig. 3). There does, however, appear to be a slight exaggeration in the
length of the postorbital part of the maxillae due to a severe break in the left maxilla at a point just anterior
to the last four maxillary teeth. Consequently, during the crushing of the skull, the maxilla, from the point
of the break, apparently was forced slightly posteriorly along the horizontal axis of the jugal. It is probable
that the gap caused by this break may have been mistaken for a tooth socket by Azzaroli et al. (1972, 1975).
While it is possible that there is room for another tooth in this area, it nevertheless remains uncertain and
the number of maxillary teeth is more probably eleven.

Internarial Bar. This is a very robust bone with a maximum width of 29 mm and minimum width of
135 mm. There is very little narrowing of the internarial bar along most of its length including especially the
area between the external nares. The narrowest point being where the internarial bar joins the premaxillary
rostrum (see text-figs. 3 and 7). The ruggedness of the internarial bar, its wide and almost unconstricted
form, and its penetration of the frontals to a point far behind the posterior termination of the external narial
opening, resembles that of members of the Tylosaurinae subfamily, in particular T. proriger. In, for example,
Clidastes, Mosasaurus , Plotosaurus , Platecarpus , Ectenosaurus , and Plesiotylosaurus (Russell 1967, p. 17) the
internarial bar is constricted between the external nares and is a much more slender body. However, the
point of origin of the internarial bar is more triangularly shaped unlike in Tylosaurus where it is rectangular.
The nares themselves in Goronyosaurus appear unusually small and rather posteriorly situated. Both the
ruggedness of the internarial bar and the posterior position and smallness of the nares are unusual among
mosasaurs and the nearest comparison may be made among members of the subfamily Tylosaurinae, in
particular T. proriger.

Splenial. On the right side of the skull, embedded with the upper jaw material, there is a fragment of the
right splenial which had become displaced during the crushing of the skull. This bone is laterally compressed
and it is possible to locate, at its posterior end, the surface which articulated with the angular bone of the
lower jaw. In posterior outline the articulating surface is laterally compressed as in Prognathodon. A small
portion of the thin ala which is normally expanded to enclose the medial surface of the dentary, can be seen.

Frontal. The frontal is a triangularly shaped bone and links with the internarial process from the premaxilla.
The internarial process penetrates deeply into the anterior end of the frontal and at some distance from the
external nares as in Tylosaurus. Dorsally there is an anteromedian ridge present.

Prefrontal. Crushing and distortion have resulted in the left prefrontal being displaced slightly anteriorly and
ventrally. Azzaroli et al. (1975) have stated that the prefrontals emarginate with the external nares. A re-
examination demonstrates that this is not the case, even when disregarding the slight displacement. The
position of the prefrontals is similar to that found in Tylosaurus. The indications are that the slender
prominences from the prefrontals where they border the orbits, and the anterior tips of the postorbitofrontals
originally linked these two bones, thereby completing the emargination of the orbits. The frontals themselves
do not appear to emarginate with the orbits.

Parietal. The parietal, although a strong unit, is particularly narrow in G. nigeriensis. Because of a severe
break anteriorly, it is difficult to distinguish exactly the point of contact with the frontal bone, but the
indications are that the groove separating the frontal from the parietal was fine, indicating that mesokinetic
movement was consequently limited. The parietal foramen is long and narrow. It is widely separated from
the frontal suture as in T. nepaeolicus.

Pterygoid. Azzaroli et al. (1972, 1975) represented the pterygoid as a greatly shortened bone lacking
basisphenoid and palatine proceses, quite unlike the pterygoids of other known mosasaurs. A re-examination,
however, indicates that there is no reason to assume that the main body of the pterygoid is different from
that of a typical mosasaur. A distinct break can be observed along the posterior margin of the left pterygoid,
at a point which marks the origin of the basisphenoid process (text-fig. 5). Anteriorly, too, the indications
are that the process leading to the palatine has broken off.

Breaks in these regions of the palatine are not uncommon in mosasaurs as both the basisphenoid and
palatine processes are frequently quite flattened and slender. The palatine processes especially are invariably
lost during fossilization. Russell (1967, p. 43) mentions that anteriorly the pterygoid in mosasaurs must have
been firmly buttressed against the palatine but the contact is rarely preserved.

The ectopterygoidal process is, however, unusual. It is a dorsoventrally flattened structure consisting of
two fork-like processes, the longer of which ends in a broad expanded termination, similar to Tylosaurus

SOLI A R : CRETACEOUS MOSASAUR FROM NIGERIA

755

50 mm

A

B

C D

text-fig. 5. a, occlusal view of the left pterygoid, b, dorsal view of the left pterygoid, c, reconstruction of

the left pterygoid by Azzaroli et al.\ d, by Soliar.

(Russell 1967), which made contact with the ectopterygoid probably by means of a fibrous joint. The shorter
process may have served to prevent the pterygoid from slipping too far backwards by apposing with the
ectopterygoid and may have served also as a point of origin of the M. pterygoideus profundus or M.
pterygoideus superficialis muscles. A groove on the dorsal side of the quadratic ramus probably marked the
insertion point of the M. protractor pterygoid muscle.

Occipital Unit. Although the occipital unit is in a poor state of preservation some important features can be
observed. In general, the occipital unit is narrow and bears certain tylosaurine characteristics in addition to
an unusual supraoccipital bone (see text-fig. 6e).

During fossilization the supraoccipital must have been dislodged, and at present lies in a semi-vertical
position, but appears fairly undistorted. The bone is unusual when compared with descriptions of the
supraoccipital in other mosasaurs. Instead of the roof-shaped element described by Russell (1967, p. 40), the
dorsal surface of the supraoccipital has a deep groove along its mid-line with gently convex surfaces on either
side to the lateral edges. The ventral surfaces of the supraoccipital slopes inwards and fits over the paroccipital
process. As in other mosasaurs the supraoccipital is strongly grooved and ridged longitudinally along the
ventral surfaces. The ventral extremity was probably hollowed to fit over the posterior part of the brain stem.

The lateral wall of the basisphenoid is unusually steeply sulcate inwardly from the ventral edge. The vidian
canal which carries the internal carotid artery and the palatine branch of the seventh cranial nerve into the
basisphenoid, is uncovered. However, it is possible that the lateral wall of the vidian canal, usually a thin
sheet of bone, is lost. Russell (1967, p. 33) makes a similar observation in the basisphenoid of a specimen of
Plioplatecarpus. Features that compare with those of Tylosaurus are the elongate basisphenoid and slender
basipterygoid process together with a narrow alar process of the prootic bone. The basal tuber of the
basioccipital appears to be intermediate in size as in the tylosaurs.

Teeth. Remains of the teeth can be seen on both the left and right maxillae (see text-fig. 3a). Besides four
small complete postorbital teeth and one partial tooth crown at the anterior end of the left maxilla, all the
remains are of tooth bases only.

756

PALAEONTOLOGY, VOLUME 31

Basal tuber

Basipterygoid

Paraoccipital

Part of
parietal

Disarticulated

Opisthotic

50 mm

text-fig. 6. Goronyosaurus nigeriensis occipital unit, IGF 14750-2. a, right view, b, posterior view, c,
reconstruction, view of left side, d, reconstruction, posterior view, e, reconstruction, left view of supraoccipital.

The posterior maxillary teeth are small, bicarinate, with recurved tips. The teeth appear to be smooth and
enamelled. The tooth bases in the remaining part of the maxillae are large, indicating strong teeth.

The partial tooth crown is covered at the base by extraneous material. Despite this and the tip being
broken off much of the tooth can still be seen. The tooth is elongated and particularly straight with slight
lateral compression. Fore and aft carinae are present. The surface appears to have been smooth but it is
damaged and it is hard to tell with any certainty whether vertical striations may have been present. In cross-
section the lingual and buccal surfaces appear to be similar. This tooth appears to bear some similarity with
the badly worn, highly gypsiferous material described by Swinton (1930) in which he states that ‘the teeth
remains are closely similar to those of Leiodon anceps but probably the crowns were more elongated and
slender’.

Postcranial elements. A straightforward description of the postcranial elements of M. nigeriensis will not be
repeated here as they have been described at length by Swinton (1930) and Azzaroli et al. (1975).

Caution must, however, be exercised when trying to establish mosasaur relationships on vertebral characters
alone as significant variations exist even within genera. Swinton (1930) points out that whereas M. nigeriensis
has no zygosphenes on the dorsals, a specimen of undoubted M. camperi ( M . hojfmannii) in the BMNH
certainly has zygosphenes. A similar variation exists in the chevron attachment of the caudals with the
chevrons being either fused or free. Taking such factors into account, I find that there is insufficient evidence
other than to state that the vertebrae of G. nigeriensis show definite nrosasauroid characters— for many of
the characters discussed are shared at least among the Mosasaurinae and Tylosaurinae.

Crushing and Distortion. Because of the critical importance of the jugal in Azzaroli et al.' s (1975) erection of
a new subfamily the Goronyosaurinae, I would like to further clarify certain points which are relevant to
their misinterpretation.

SOLI AR: CRETACEOUS MOSASAUR FROM NIGERIA

757

The individual bones which make up the jugal consist of what I believe to be at least three distinct
fragments of bone but appear to lack very definite seams or joins. This I believe is not unusual and may be
accounted for by the fossilization process during which fusion of the fragments occurred. Instances are seen
where during extreme crushing and poor fossilization, e.g. Platecarpus coryphaeus (BMNH R.2947) fusion
of the bones in certain areas show little or no trace of joins between identifiably separate bones. Specifically
in G. nigeriensis there are areas where similar fusion has taken place. For example, on the left side of the
skull below maxillary teeth 7/8 there is evidence of complete fusion between fragments (unidentified bones,
possibly parts of the lower jaw or palatine) and the medial side of the right maxilla. A further area of
complete fusion involves a part of the splenial with a fragment (possibly part of the lower jaw). There also
appears to be fusion between the lower part of the right premaxilla with the maxilla.

It seems most unlikely that the distinct grooves between the ‘lamina’ and postorbitofrontals mark suture
points as Azzaroli et al. (1975) state. These bones in life would have been in close contact and the severe
compression suffered by the skull would have in all probability caused fusion of the bones as witnessed in
other parts of the skull where the bones were even more remote.

CONCLUSIONS

Re-examination of the specimens of Goronyosaurus leads to the view that the interpretation of
some of the key characters by Azzaroli et al. (1972, 1975), especially with regard to the jugal, was
based on a misinterpretation of the crushed elements. The functional significance of the broad
ascending ramus of the jugal and its kinetic movement, as described by Azzaroli et al. (1975)
appears doubtful. However, an unusual feature of G. nigeriensis , the wide and robust internarial
bar, appears to be misrepresented in the reconstruction by Azzaroli et al (1975) as a considerably
narrower and more delicate bone.

text-fig. 7. Comparative reconstructions of skull of Goronyosaurus nigeriensis. a, lateral view, b, dorsal view
of generalized mosasaur skull (not to scale), c, d, Azzaroli et al. reconstruction, e, f, Soliar reconstruction.

758

PALAEONTOLOGY, VOLUME 31

G. nigeriensis presents a curious mixture of characters (text-fig. 6). The jugal is like that of
Leiodon anceps (Deperet and Russo, 1925, p. 340), the internarial bar is like that of a tylosaurine,
somewhat resembling that of T. proriger; the maxillaries including maxillary teeth extend
postorbitally unlike any other mosasaur. Table 2 indicates that G. nigeriensis had a uniquely narrow
skull among the larger mosasaurs comparing more closely with members of the plesiotylosaurs and
tylosaurs and closest to T. nepaeolicus. Such evidence does point to the possibility that G. nigeriensis
may be more closely related to the members of the subfamily Tylosaurinae.

A major character that distinguishes members of the subfamily Tylosaurinae, is the large
premaxillary rostrum although it may vary among members of the genus Hainosaurus. There is
not much material available on the hainosaurs although Dollo (1904) does remark on the variations
including that of the rostrum in a new species H. lonzeensis with that of H. bernardi — ‘par son
Rostre plus conique et a face superieure plus arrondie, et par ses Dimensions moindres,— indique,
egalement, une espece differente du Hainosaurus bernardi , Dollo, 1885, du Senonien superieur du
Hainaut’ (1904, p. 213).

A tentative suggestion is made that G. nigeriensis may even be a juvenile of a very much larger
tylosaur such as H. bernardi which may grow up to 17 m (56 ft.) (see Russell 1967, p. 210). As
very little is known of juvenile stages of mosasaurs it may prove to be that certain aberrant features
such as the postorbital maxillaries may simply be present only in the juvenile stages. Swinton
(1930) mentions that the presence or absence of zygosphenes and zygantra in the vertebral column
may in some instances be related to the age of the individuals (zygosphenes and zygantra may be
present only in older individuals in order to take up the extra load).

However, the presence of the small premaxillary rostrum together with a wide internarial bar,
a closed canal on the undersurface of the frontal for the reception of the olfactory lobes, the
unusual ectopterygoidal processes of the pterygoid, plus a tentative acceptance of the postorbital
maxillary teeth necessitate at least the erection of a new genus within the subfamily Tylosaurinae.

Reconstruction

In their reconstruction Azzaroli et a/, calculated the body length based on skull to body ratios
being ‘approximately the same as in other mosasaurs’ (1975, p. 28). This is, however, quite
confusing as the head to body proportions vary considerably among mosasaurs and can produce
considerable discrepancies in body lengths. Table 2 gives an idea of the head and body proportions
in a few selected mosasaurs.

Working on Azzaroli et al.'s (1975) derived body length of 7-80 m it would appear that the ratio
of head to body they used was 9- 1:100. Taking into account that it is a rough estimate, it still
nevertheless appears particularly low when compared with the figures in Table 2. For instance, we
find an appreciable difference in body length if we increase head length from 91 % to just 10%.
The length of G. nigeriensis goes down from 7-8 m (25 ft.) to 7-1 m (23 ft.). Using the tylosaur
figure of 13-8 % head, then the overall length of Goronyosaurus is dramatically reduced to 514 m
(16-85 ft.). This, coincidentally, would not be incongruous with the tentative suggestion that G.
nigeriensis might be a juvenile of a larger tylosaur.

The skull of Goronyosaurus gives us some clues as to its way of life. Goronyosaurus had a
uniquely elongated muzzle provided with straight, long, strong teeth. Added to this we can, from

table 2

Genera Head length Body length

Platecarpus (from von Huene, 1911)
Platecarpus (from Lambe, 1914)
Tylosaurus (from Williston, 1896)
Tylosaurus (from Osborn, 1899)

9-8 %

5-60 m (18i ft.)

10-4%

6-25 m (201 ft.)

13-0%

6-34 m (20| ft.)

13-8%

8-83 m (29 ft.)

SOLIAR: CRETACEOUS MOSASAUR FROM NIGERIA

759

the ruggedness of the postorbitofrontals, and associated bones, assume that the processes from
the parietal to the supratemporals were of equally large dimensions and would have provided large
areas for the points of origin of massive jaw muscles, the depressor mandibulae, the major muscles
for closing the jaws. Such powerful, strongly toothed jaws indicate that Goronyosaurus was certainly
capable of attacking large prey.

White’s (1934) study of fossil fishes of the Sokoto region indicates the presence of several families
of fishes in the Upper Cretaceous deposits, including members of the families Lanmidae,
Pycnodontidae, and Eotrigonodontidae as well as the remains of indeterminable bony fishes. This
in itself would have provided rich feeding grounds for Goronyosaurus. However, in addition to the
prolific fish life discoveries of marine turtle and crocodile remains in the Dukamaje Formation
(see Halstead 1979//, b) indicate that the waters were probably rich in such fauna. Goronyosaurus
could have been capable of seizing the smaller members of these families but in addition it is quite
probable that Goronyosaurus may have preyed upon the juveniles of, for example, the giant
pelomedusid turtles (Halstead 1979c/, b) and the young of dyrosaurid crocodiles such as Sokotosuchus
ianwilsoni (Halstead 1975; Buffetaut 1976, 1979). Dollo (1897, p. 520) refers to the discovery of
turtle bones in the body cavity of the Belgian Hainosaurus.

Although Goronyosaurus was in all probability more pelagic in behaviour than the crocodiles
and turtles, it is more than likely that the members of the genus Goronyosaurus converged in
numbers in the rich littoral waters in which the Dukamaje Formation was deposited (see Petters
1977; Buffetaut 1979) with their only real rivals being the giant mosasaurs such as Tylosaurus.

From the streamlined skull we can ascertain that Goronyosaurus probably possessed hydro-
dynamic qualities superior to other mosasaurs except perhaps the much smaller Clidastes.

Evolution of A frican Mosasaurs

It is possible that the evolution of Goronyosaurus continued through a period when the larger
mosasaurs such as Tylosaurus had disappeared elsewhere. Certainly the latest tylosaurs in other
parts of the world belong to no later than the Campanian period. Azzaroli et al. (1975) commented
on and figured three very large vertebrae found at Tunga which they believed belonged to some
representative of the subfamily Tylosaurinae, with which I am in agreement. If this is so it would
make it a very interesting find for tylosaurs as the geological age of Tunga is Maastrichtian.
Nevertheless because of the incompleteness of the material this identification can only be accepted
tentatively. Azzaroli et al. (1975, p. 30) also describe three vertebrae, one collected at Taloka and
two at Tunga which they refer to the genus Halisaurus on the basis of the characteristic flattened
centra. This is as they stated ’noteworthy insofar as it is the first record of this genus outside North
America’ (1975).

It would appear that large mosasaurs such as the tylosaurs or hainosaurs continued to flourish
in the Sokoto region after they had disappeared from other parts of the world by the Senonian
(see Russell 1967, charts 1-7 and Dollo 1904). It is, therefore, not unreasonable to surmise that
conditions too may have been conducive to the evolution of a more dynamic streamlined and
smaller tylosaur such as G. nigeriensis.

Systematic position of Goronyosaurus

Using cranial characters alone it is possible to draw up a cladogram of generic relationships among
the members of the subfamilies Plioplatecarpinae and Tylosaurinae. The subfamily Mosasaurinae
(not represented) was considered as the outgroup. The following list of apomorphic characters was
used:

1 . Premaxilla with large rostrum.

2. Dorsal mid-line of premaxilla smooth.

3. Internarial bar arises from rectangular base.

4. Internarial bar unconstricted between external narial opening.

5. Canal or deep groove on floor of basioccipital and basisphenoid for basilar artery.

6. Broad triangular alar from supraorbital process on prefrontal.

760

PALAEONTOLOGY, VOLUME 31

Plioplatecarpinae

Tylosaurinae

Mosasaurinae

I

£

/

/

I

£

|

e*

<c

I

I

/

#

/

n

/

/

«9

I

/

to

(8)(9)(I0)

33,34

32,36

1 14

37, 38

16, 22

35

31

1

1

6, 15,18,
20,23

25, 27

2,3

(8) (9) (10)

I, 21, 26
12

I

13,28,29,30

4, 24

17, 19
5

( ) = Convergent characters

Platecarpus - 1 i ke
ancestor

Clidastes- like
ancestor

Aegialosaurs

text-fig. 8. Cladogram of affinities among some of the more important genera of mosasaurs including
Goronyosaurus, in the subfamilies Plioplatecarpinae and Tylosaurinae. The subfamily Mosasaurinae is used

as the outgroup.

7. Ala on supraorbital process is a small nubbin.

8. First two processes of POF connected under frontal by thin sheet of bone, separated by a groove.

9. POF process to the jugal is large.

10. Face of POF beneath the posterolateral corner of the frontal is smooth.

1 1 . Parietal foramen is large.

12. Parietal invades the posterodorsal surfaces of the frontal medially by firm suture providing
little mesokinetic movement.

13. Penetration of the basisphenoid posteriorly by bilobate tunnel.

14. Small flattened quadrate with very reduced tympanic ala.

15. Pterygoid teeth increases anteriorly from very small diameter to equal mandibular teeth.

SOL I A R : CRETACEOUS MOSASAUR FROM NIGERIA

761

16. Ectopterygoidal process projects from main body of the pterygoid at an acute angle.

17. Suprastapedial process very large.

18. Supra and infrastapedial process fused.

19. Tympanic ala highly developed.

20. Very blunt termination of premaxilla.

21. Jaws long.

22. Keel on quadrate shaft for origination of the M. depressor mandibulae.

23. Pterygoid teeth: large.

24. Prefrontal excluded from external narial border.

25. Frontal does not emarginate with orbits.

26. Broad dentary projection.

27. Quadrate-medial surface flat. In front of this a heavy ridge descends vertically.

28. Wide robust internarial bar.

29. Small orbits.

30. Skull narrow and long.

31. Massive quadrate.

32. Dentary teeth: sixteen to seventeen.

33. Massive jaws.

34. Heavy dentition.

35. Elliptical teeth.

36. Long narrow jaws.

37. Maxillaries extend beyond the posterior margin of the orbits.

38. The inner surface of the frontal bears a fully closed canal housing the olfactory lobes.

Acknowledgements. I am grateful to Professor A. Azzaroli and Dr M. Mazzini of the Geological and
Palaeontological Museum, Unversity of Florence, for their kindness and permission to use the mosasaur
material; in addition, for their generous donation of casts of the skull of G. nigeriensis , which are now housed
in the Cole Museum, University of Reading. I thank other members of staff', especially Cristina Andreani,
Elisabetta Cioppi, and Dr V. Borselli for helping locate specimens and assistance in many ways. My thanks
go to Sandra Chapman of the British Museum (Natural History) for help, and access to the mosasaur
material, and to Mr Cyril Walker and Dr Angela Milner also of the British Museum. I am particularly
grateful to Dr Beverly Halstead of the Departments of Zoology and Geology, University of Reading, for
valuable and constructive discussions of the manuscript.

REFERENCES

antunes, m. t. 1964. O Neocretacico eo Cenozoico do litoral de Angola; 1 Estratigrafia; Repteis, 257 pp.
Junta Invest. Port Ultramar , Lisbon.

arambourg, c. 1952. Les vertebres fossiles des gisement de phosphates (Maroc-Algerie-Tunisie). Serv. Geol.
Maroc Notes , 92, I -372.

azzaroli, a., de guili, c. and torre, d. 1972. An aberrant mosasaur from the Upper Cretaceous of North-
Western Nigeria. Accad. naz. Lincei , Re. Classe di scienze fisiche matematiche e naturali, (8) 52 (3), 53 56.

— 1975. Late Cretaceous mosasaurs from the Sokoto District, Nigeria. Lincei-Memorie Sc.
Fisiche. ecc. 13, sess. 2, 21-34.

buffetaut, E. 1976. Une nouvelle definition de la famille des Dyrosauridae De Stefano, 1903 (Crocodylia,
Mesosuchia) et ses consequences: inclusion de genres Hyposaurus et Sokotosuchus dans le Dyrosauridae.
Geobios , 9 (3), 333-336.

1979. Sokotosuchus ianwilsoni and the evolution of the dyrosaurid crocodiles. Niger. Fid. Monogr. 1,31 41 .
broom, r. 1912. On a species of Tylosaurus from the upper Cretaceous of Pondoland. Ann. S. Afr. Mus. 7,
332-333.

callison, g. 1967. Intercranial mobility in Kansas Mosasaurs. Univ. Kans. Pubis. Paper , 26, 1 15.
camp, c. l. and allison, h. j. 1961. Bibliography of fossil vertebrates 1949 1953. Bull. geol. Soc. Am. 84,
1-532.

deperet, c. and russo, p. 1925. Les phosphates de Melgou (Maroc) et leur faune de mosasauriens et de
crocodiliens. Bull. Soc. geol. Fr. 4, ser. 25, 329 346.

762

PALAEONTOLOGY, VOLUME 31

dollo, l. 1889. Premiere note sur les mosasauriens de Mesvin. Mem. Soc. beige Geol. 3, 271 304.

1890. Premiere note sur les mosasauriens de Maestricht. Ibid. 4, 151-169.

- 1897. Le hainosaur et le nouveaux vertebres fossiles du Musee de Bruxelles. Revue. Quest sclent. I ser.
22, 70-112.

- 1904. Les mosasauriens de la Belgique (1). Mem. Soc. beige Geol. 18, 207-216.

1917. Les vertebres vivants et fossiles. Guide de Touring Club de Belgique , 2, 126-161.

de guili, c., ficcarelli G, and torre, d. 1970. Missione Palaeontologica Nella Provincia Di Sokoto (NW
Nigeria). Nota Preliminare. Boll Soc. geol. ital. 89, 547-556.

Halstead, L. B. 1973. Hunting prehistoric reptiles in Nigeria. Niger. Fid. 38, 4-14.

- 1975. Sokotosuchus ianwilsoni n.g. et sp.— a new teleosaur crocodile from the Upper Cretaceous of
Nigeria. J. Min. Geol. Ibadan , 11, 101 103.

1979u. The International Palaeontological Expedition to Sokoto State 1977-1978. Niger. Fid. Monogr.
1, 3-30.

— 1979ft. A taxonomic note on new fossil turtles. Ibid. 48-49.

- 1979c. Type sections of the Cretaceous-Tertiary transition of Sokoto. Ibid. 50-63.

1980. A note on the Cretaceous Tertiary transition in Sokoto State. Niger. Fid. 45, 104-105.

- and middleton, j. a. 1976. Fossil Vertebrates of Nigeria. Part III. Ibid. 41, 166-174.

1982. Fossil Vertebrates of Nigeria. Part IV. Ibid. 47, 39 44.

jones, b. 1948. The Sedimentary Rocks of the Sokoto Province. Bull. geol. Surv. Nigeria , 18, 1-75.
kogbe, c. a. 1973. Geology of the Upper Cretaceous and Tertiary sediments of the Nigerian sector of the
Iullemeden Basin (West Africa). Geol. Rdsch , 62, 197-211.
leonardi, p. and malaroda, r. 1946. Prima segnalazione di un mosasauro del genera Globidens nel Cretaceo
dell’Egitto. Acta pontif. Acad. Sci., 10, 183- 190.

nopcsa, f 1925. On some reptilian bones from the Eocene of Sokoto. Geol. Surv. Nigeria Occ. Pap. 2, 1-16.
Parker, d. h. and carter, j. d. 1965. Geological Map of Nigeria. Sheet 2 (Sokoto). Scale 1:250000.
Parkinson, J. 1822. Outlines of Oryctology: An introduction to the study of fossil organic remains , 298 pp.
London.

persson, p. o. 1959. Reptiles from the Senonian of Scania. Ark. Miner. Geol. Stockholm , 2 (5), 431-480.
petters, s. w. 1977. Ancient Seaway across the Sahara. Niger Fid. 42, 22-30.

1979a. Stratigraphic history of the south-central Saharan region. Bull. geol. Soc. Am. 90, 753-760.
1979ft. Maastrichtian arenaceous foraminifera from north and western Nigeria. Palaeontology , 22,
947-963.

quass, a. 1902. Beitrag zur Kennlnis der Fauna der obcrsten Kreidebildungen in der Libyschen Wuste.
Palaeontographica , 30 (2), 153-336.

raeburn, c. and tattam, c. m. 1930. A preliminary note on the sedimentary rocks of Sokoto Province. Bull,
geol. Surv. Nigeria , 13, 57-60.

reyment, r. a. 1965. Aspects of the geology of Nigeria , 145 pp. Ibadan University Press.
russell, d. a. 1967. Systematics of American Mosasaurs. Bull. Peabody Mas. nat. Hist. 23, 237 pp.
stromer, e. and weiler, w. 1930. Beschreibung von Wirbeltier— Resten aus dem nubischen Sandsteine
Oberagyptens und aus agyptischen Phosphaten. Abh. bayer Akad. Wiss, neuFolge. 7, 1 42.
swinton, w. e. 1930. On Fossil Reptilia from Sokoto Province. Bull. geol. Surv. Nigeria Bull. 13, 1-56.
walker, c. a. 1979. New turtles from the Cretaceous of Sokoto. Niger. Fid. Monogr. 1, 42 48.
white, E. i. 1934. Fossil Fishes of Sokoto Province. Bull geol. Surv. Nigeria 14, 1-78.
williston, s. w. 1897. Range and distribution of the mosasaurs. Kansas Univ. Quart. 6, 177-189.
zdansky, o. 1935. The occurrence of mosasaurs in Egypt and Africa in general. Bull. Inst. Egypte 17, 83-94.

T. SOLIAR

Department of Pure and Applied Zoology

Typescript received 7 February 1987 University of Reading

Revised typescript received 9 December 1987 Reading RG6 2AJ UK

Note added in proof. A new plioplatecarpine mosasaur, Selmasaurus russelli (Wright and Shannon, 1988)
from the Upper Cretaceous of Alabama shares a number of features with Goronyosaurus; the similarity of
the two genera, however, is increased in the light of the present reinterpretation of Goronyosaurus and the
elimination of its ‘bizarre’ characters.

wright, k. r. and shannon, s. w. 1988. Selmasaurus russelli a new plioplatecarpine mosasaur (Squamata,
Mosasauridae) from Alabama. J. Vert. Paleont. 8, 102-107.

A NEW AESHNID DRAGONFLY FROM THE
LOWER CRETACEOUS OF SOUTH-EAST

ENGLAND

by E. A. JARZEMBOWSKl

Abstract. A comparatively advanced ‘hawker’ dragonfly ( Valdaeshna surreyensis gen. et. sp. nov., Aeshnidae:
Gomphaeschninae) is described from the late Hauterivian of the Weald. The single male specimen shows
body ‘colour’ markings as well as full venational details of the fore and hindwings and is the most complete
early aeshnid found to date. Fossil preservation, association, and palaeoenvironment are briefly discussed.

The most intact example of an early aeshnid dragonfly was found by the author in April 1986 in
the Wealden Series of south Surrey whilst on field-work with the West Sussex Geological Society
and is described below. It has been popularly dubbed the Surrey Dragonfly (Jarzembowski 1 987c/)
and is formally named here Valdaeshna surreyensis gen. et sp. nov.

The earliest record of a true dragonfly (Odonata: Anisoptera) is from the Upper Lias (late Lower
Jurassic), but doubts have been expressed recently as to whether any extant families of dragonflies
occurred as early as the Mesozoic (Hennig 1981). A Palaeocene origin of the family Aeshnidae,
the living ‘hawker’ dragonflies of north-west Europe, has been suggested (Carle 1982). However,
subsequent finds of isolated wings in Europe and Asia suggested that Aeshnidae had appeared by
the Lower Cretaceous (Hong 1982; Jarzembowski 1984). The new find shows body details as well
as a complete venation and provides evidence that comparatively specialized Aeshnidae existed by
the early Cretaceous.

Locality and horizon. The specimen is preserved in a phosphatic concretion from the disused pit
of the former Auclaye Brickworks, Surrey (national grid reference TQ 170 388). This pit was
worked in the Lower Weald Clay above the Okehurst Sand (Worssam 1978).

Associated Fauna. Other insects found in concretions from the same locality include beetles
(Coleoptera), bugs (Hemiptera), crickets (Orthoptera), wasps (Hymenoptera), true flies (Diptera),
caddis flies (Trichoptera), scorpion flies (Mecoptera), and lacewings (Neuroptera) (Jarzembowski
19876). The insects occur with clam shrimps (Crustacea: Conchostraca), sea slaters (Crustacea:
Isopoda), fish scales, coprolites, and comminuted plant debris, the latter often fusainized.

Palaeoeco/ogy. The insects are represented mainly by dissociated skeletal elements of adults
(imagines), including detached wings and body parts and plates (sclerites). Intact specimens are
rare. The conchostracan Cyzicus subquadratus (J. de C. Sowerby) is locally abundant; the valves
are commonly paired and unworn. Isopoda are represented by undescribed bodies of sea slaters
which are incomplete but not dissociated like the insects.

The isopods suggest a salt-water depositional palaeoenvironment which is consistent with the
absence of immature freshwater insects such as aeshnid larvae. The adult insects could have been
blown or washed in; intact Coleoptera (Carabidae?) and Hemiptera (Homoptera) have their wings
folded suggesting the latter. The insect remains are poorly sorted and any fluvial influence was
probably weak. C. subquadratus commonly occurs with freshwater Mollusca in the late Jurassic-
early Cretaceous of southern England (Morter 1984) but there is no such association at the Auclaye
Brickworks. Some extant species of Cyzicus occur in both fresh and brackish water (Tasch 1969).

I Palaeontology, Vol. 31, Part 3, 1988, pp. 763-769.|

© The Palaeontological Association

764

PALAEONTOLOGY, VOLUME 31

Salinity tolerance is suggested in some Mesozoic species because Cyzicus (= Estheria ) murchisoniae
(Jones) from the Middle Jurassic of Scotland occurs in both fresh and brackish water assemblages
in shallow lagoonal deposits (Hudson 1963a, b). The pit faces at the Auclaye Brickworks are too
degraded for detailed palaeoecological work but the fossil arthropods from concretions provide
environmental pointers consistent with lagoon-bay models for the Weald Clay (Allen 1981).

Preservation. The dragonfly is not on a single bedding plane, the various parts occupying a 10 mm
thickness of sedimentary rock. The thorax and wing folds are preserved in relief although the
former is slightly compressed. The right forewing, which is preserved at an angle of 25° to the
plane of the body and the other three wings, is also slightly compressed. The abdomen is bent
(text-fig. 3a). The pterostigma (text-fig. 2, p) and veins are brown-tinted but membraneous areas
of the wings are the colour of the matrix. The body is distinctly patterned (text-fig. 3a) with brown
markings.

Alongside the dragonfly is a forewing of a mesoblattinid cockroach. The only other fossils in
the same parting are sinuous burrows on the hindwings of the dragonfly. They are mainly on the
upper side of the membrane and veins may be depressed along their courses. Some of the worm
(nematode?) traces are very small (text-fig. 3b).

Taphonomy. The dragonfly was evidently buried rapidly and there was some compaction of
sediment prior to cementation. The burrows were clearly formed when sediment and wings were
soft and pliable, but the lack of traces around the body suggests that they were not produced by
exiting parasites.

Dragonfly wings are permanently outstretched in life at right angles to the body and the abdomen
is a flexible cylinder. In the fossil the right forewing and abdomen are bent suggesting that soft
cuticle had started to decay prior to burial. I have observed that dead dragonflies floating in a
laboratory fish tank readily develop breaks in the abdomen. The thorax is lying on its left side
and the posterior part of the abdomen is bent to the right which suggest that the dragonfly floated
with its dorsal side uppermost.

‘Colour’ pattern is commonly preserved in Wealden fossil insects. The pattern is due to the
survival of heavily tanned (sclerotized) and darkly pigmented areas. The black or brown coloration
in Wealden insects resembles that of Recent relatives although no traces of non-melanic pigments
have yet been observed (Jarzembowski 1984). The brown coloration of the veins and pterostigma
of V. surreyensis is similar to that of many Recent dragonflies. The wing membrane is commonly
clear (hyaline) in Recent Odonata and could have been originally so in V. surreyensis. The bodies
of living aeshnids are usually brown, spotted with blue or green, sometimes with yellow stripes or
spots on the sides of the thorax (Walker 1958). Only the brown body patterning has survived in
V. surreyensis delimiting non-melanic areas.

SYSTEMATIC PALAEONTOLOGY

Class insecta Linnaeus, 1758
Order odonata Fabricius, 1792
Suborder anisoptera Selys, 1840
Superfamily aeshnoidea Leach, 1815
Family aeshnidae Leach, 1815
Subfamily gomphaeschninae sensu Lieftinck, 1968
Genus valdaeshna gen. nov.

Diagnosis. Gomphaeschnine with one row of cells separating the anterior median vein (MA) from
the median supplementary vein (Mspl) in the hindwing, supplementary anal loop absent, subcostal
vein (Sc) continuing beyond the nodus (N), and an incomplete basal antenodal (Ax) in fore and
hindwings.

Type species. Valdaeshna surreyensis sp. nov.

JARZEMBOWSKI: CRETACEOUS DRAGONFLY

765

text-fig. 1. Valdaeshna surreyensis gen. et sp. nov. In. 64632. Left forewing and hindwing showing venation.

Valdaeshna surreyensis sp. nov.

Text-figs. I -3a, 4a, b

Derivation of name. Named after the Weald, County of Surrey and extant odonatan genus Aeshna.
Diagnosis. As for genus.

Holotype. In. 64632a, b, British Museum (Natural History).

Locality. Auclaye Brickworks pit, near Capel, Surrey, England (national grid reference TQ 170 388).

Horizon. Lower Weald Clay, clay interval above British Geological Survey sandstone 3a (Gallois et at. 1972);
early Cretaceous: (?)late Hauterivian (Worssam 1978).

Description. The holotype is described as exposed naturally in a frost-cracked phosphatic concretion. For
abbreviations see text-fig. 2.

The thorax (text-fig. 3a) is exposed posteriorly from the mesothoracic spiracle (s) immediately above the
mesokatepisternum. The dorsal carina (dc) and pre-alar ridge (pr) are prominent. Brown markings are
preserved on the pterothorax (synthorax auctt.) The mesepisternum (humeral region auctt.) shows two stripes
one of which, the mid-dorsal, is truncated anteriorly. The humeral stripe is present overlying the mesopleural
suture (ms) and is partly divided by the latter. The interpleural and metapleural stripes lie immediately behind
their respective sutures.

The wings are hyaline, veins dark brown, pterostigma (p) pale brown. The anal triangle (text-fig. 2, at) is
developed as in Recent male Anisoptera. An incomplete basal antenodal (Ax) is present. The subcostal vein
(Sc) continues beyond the nodus (N).

Other venational characters in the Surrey Dragonfly are as follows, the numbers referring to characters
used in a recent phylogenetic study of gomphaeschnine dragonflies by Wighton and Wilson (1986) and
discussed below.

1 hind and forewing triangles (t) are approximately equal in length;

2 all triangles are relatively narrow;

766

PALAEONTOLOGY, VOLUME 31

i i i i i i

mm

text-fig. 2. Valdaeshna surreyensis gen. et sp. nov. In. 64632. Right hindwing with explanation of venation,
al, anal loop; Arc, arculus; at, anal triangle; Ax, antenodal; CuP, posterior cubitus; cux, cubito-anal crossvein;
df, discoidal field; int, middle fork; IR3, third intercalary; MA, anterior median; Mspl, median supplementary;
N, nodus; O, oblique; p, pterostigma; R3, third radial; R4, fourth radial; Rspl, radial supplementary; Sc,
subcostal; Sn, subnodus; spt, supratriangle; t, triangle; 1A, anal.

text-fig. 3. Valdaeshna surreyensis gen. et sp.
nov. In. 64632. a, dorsal aspect of body excluding
small rneso- and metathoracic sclerites. Wing bases
shown in solid lines, position of right forewing
indicated in dotted lines, dc, dorsal carina; is,
interpleural suture; MD, mid-dorsal spot; ms,
mesopleural suture; P, posterior spot; pr, pre-alar
ridge; s, spiracle; T5, fifth tergite; tc, transverse
carina; 1, mid-dorsal stripe; 5, metapleural stripe.
b, vermiform casts in anal area of left hindwing.

JARZEMBOWSKI: CRETACEOUS DRAGONFLY

767

text-fig. 4. Valdaeshna surreyensis gen. et sp. nov. holotype. a, dorsal view. In. 64632a. b, ventral view. In.

64632 b. x 1 -6.

3 the anal loop (al) has few (four or five) cells;

4 a supplementary anal loop is absent;

5 a median supplementary vein (Mspl) is present;

6 the radial supplementary vein (Rspl) is well developed;

7 the basal discoidal field (df) in the hindwing has three rows of cells;

8 the posterior cubitus (CuP) and anal vein (1A) in the hindwing are separated by several rows of cells
distally;

9 there is more than one cubito-anal crossvein (cux);

10 the oblique vein (o) is far from the subnodus (Sn);

1 1 the fourth radial (R4) and anterior median (MA) veins are parallel distally;

12 three crossveins are present in the hindwing supratriangle (spt);

13 seven intermedian crossveins are present in the area basad of the middle fork (int) in the hindwing;

14 MA is separated from Mspl by a single row of cells in the hindwing;

15 the third intercalary vein (IR3) is simple.

The abdomen is preserved up to segment 6, the last segment being weathered and truncated al the edge of
the concretion. Typical (elongate) segments possess transverse and dorsal carinae (tc, dc) and show colour
pattern. The mid-dorsal spots (MD) are united to form a light band which tapers away from the dorsal
carina, i.e. MD is ‘triangular’. The tergum immediately behind MD is dark. The posterior spots (P) are
united to form a light area larger than MD.

768

PALAEONTOLOGY, VOLUME 31

DISCUSSION

The presence of triangles (t) and supratriangles (spt) in the fore and hindwings identifies V.
surreyensis gen. et sp. nov. as an anisopteran or dragonfly in the strict sense (Hennig 1981). As in
Recent Anisoptera, the presence of an anal triangle (at) in the hindwing shows that the type is a
male. The triangles are similar in shape, possess crossveins, and are equidistant from the arculus
(Arc) in V. surreyensis as in the superfamily Aeshnoidea and family Aeshnidae (Davies and Tobin
1985).

V. surreyensis is placed in the extant subfamily Gomphaeschninae on a combination of venational
characters: the third intercalary vein (IR3) is simple and not forked as in other subfamilies
Aeshninae (Fraser 1957, fig. 50) and Brachytroninae (F. Nanninga in O’Farrell 1970, fig. 13.4);
an anal loop (al) is present; the third radial vein (R3) has an anterior convex curve behind the
pterostigma (p); the fourth radial (R4) and anterior median (MA) veins are parallel; and the radial
supplementary (Rspl) and median supplementary (Mspl) veins are well developed, unlike in
Neopetaliinae (Wighton and Wilson 1986). The only other evidence of Gomphaeschninae in north-
west Europe is from the Bembridge Marls (late Eocene/early Oligocene) of the Isle of Wight
(Cockerell and Andrews 1916).

V. surreyensis has some unusual venational characters useful for identification. Sc extending
beyond the nodus (N) and the presence of an incomplete basal antenodal (Ax) in the fore and
hindwings are unusual characters in Anisoptera and the combination of these characters in V.
surreyensis appears to be unique. Sc extends beyond N in some extant Aeshninae ( Neuraeschna
and Staurophlebia : Professor D. A. L. Davies, Mr S. J. Brooks, pers. comms.) and an incomplete
basal Ax may occur in Brachytroninae (Periaeschncr. Fraser 1936, fig. 23). Species of Cephalaeschna
(= Indophlebia , Brachytroninae) may have an incomplete basal Ax in the hindwing (Fraser 1936,
fig. 17) or Sc continuing beyond N in the forewing (Fraser 1936, fig. 26) but not in combination
as in Valdaeshna.

The ‘colour’ pattern of V. surreyensis shows some aeshnid and gomphid features. The mid-
dorsal spots (MD) resemble Boyeria (Walker 1958, pi. 17, figs. 1 and 2) but the thoracic colour
pattern resembles extant Gomphus (family Gomphidae: Walker 1958, pi. 40) except that the mid-
dorsal stripe is well separated from the dorsal carina and the interpleural and metapleural stripes
are slightly more posterior in Valdaeshna. An incomplete basal Ax may also be developed in the
forewing of Gomphus (Fraser 1934, fig. 63). However, gomphid-like characters in late Mesozoic
Anisoptera are probably symplesiomorphies (Hennig 1981) and therefore do not affect the
systematic placing of Valdaeshna.

In the above description, I have numbered 1-15 characters used in a recent phylogenetic study
of gomphaeschnine genera (Wighton and Wilson 1986). If the characters are scored 0 = primitive
and 1 = advanced, then the character states for Valdaeshna are:

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

0100111110 1 1 1 0 0

Valdaeshna shares the largest number of advanced characters (seven) with extant Boyeria and
Tertiary Oplonaeschna in the subfamily Gomphaeschninae, but also shares nine advanced characters
with genera of Aeshninae and Brachytroninae. Wighton and Wilson considered that Boyeria
was one of the most advanced genera of Gomphaeschninae and most closely related to
Brachytroninae/Aeshninae, although its similarity to the gomphaeschnine Oplonaeschna was
interpreted as parallel or convergent evolution in the latter genus. Gomphaeschninae are clearly
not a holophyletic (monophyletic) group, but Valdaeshna , like Boyeria and Oplonaeschna , may be
considered a comparatively specialised aeshnid.

Acknowledgements. I thank Professor D. A. L. Davies, Cambridge University and Mr S. J. Brooks, British
Museum (Natural History) for their helpful comments on extant dragonflies; Mr H. Taylor, BM(NH) and

JARZEM BOWSKI: CRETACEOUS DRAGONFLY

769

Mrs B. Jarzembowski for help with illustrations; and Drs P. E. S. Whalley, P. C. Barnard BM(NH), and
R. Goldring, Reading University for reading the manuscript.

REFERENCES

allen, p. 1981. Pursuit of Wealden models. Jl Geol. Soc. 138, 375 495.

carle, F. l. 1982. The wing vein homologies and phylogeny of the Odonata: a continuing debate. Soc. ini.
odonatol. rapid. Comm. 4, 1-66.

cockerell, t. D. a. and Andrews, h. 1916. Dragon-flies from the English Oligocene. Proc. biol. Soc. Wash.
29, 89-92, pi. 2.

davies, d. a. l. and tobin, p. 1985. The dragonflies of the world: a systematic list of the extant species of
Odonata. 2. Anisoptera. Soc. int. odonatol. rapid Comm. ( Sapp /.), 5, xi+151 pp.

fraser, F. c. 1934. Odonata II. Fauna Br. India , xxiii + 398 pp., 4 pis.

- 1936. Odonata III. Ibid, xi + 461 pp., 2 pis., 1 map.

1957. A reclassification of the order Odonata , 133 pp., 1 pi. Royal Zoological Society of New South
Wales, Sydney.

gallois, R. w., thurrell, r. o., worssam, b. c. and bristow, c. R. 1972. Sheet 302 (1:63360) Horsham.
Geological Survey of Great Britain (England and Wales). London.

hennig, w. 1981. Insect phytogeny , xxii + 514 pp. J. Wiley and Sons, Chichester.

hong, y. 1982. Mesozoic fossil insects of the Jiuquan Basin in Gansu Province , 187 pp., 39 pis., 2 tables.
Geological Publishing House, Peking. [In Chinese.]

Hudson, J. d. 1963a. The recognition of salinity-controlled mollusc assemblages in the Great Estuarine Series
(Middle Jurassic) of the Inner Hebrides. Palaeontology , 6, 318 326.

- 1963 b. The ecology and stratigraphical distribution of the invertebrate fauna of the Great Estuarine
Series. Ibid. 6, 327-348, pi. 53.

jarzembowski, e. a. 1984. Early Cretaceous insects from southern England. Mod. Geol. 9, 71 93, pis. I 4.

1987a. The Surrey Dragonfly. Antenna , II, 12 13.

19876. Early Cretaceous insects from southern England. Ph.D. thesis (unpublished). University of
Reading, England.

morter, a. a. 1984. Purbeck- Wealden Mollusca and their relationships to ostracod biostratigraphy,
stratigraphical correlation and palaeoecology in the Weald and adjacent areas. Proc. Geol. Ass. 95, 217
234.

o’farrell, a. f. 1970. Odonata. In C.S.I.R.O. The insects of Australia , 241 261. Melbourne University Press,
Carlton, Victoria.

tasch, p. 1969. Branchiopoda. In moore, r. c. (ed.). Treatise on Invertebrate Paleontology , Part R. Arthropoda ,
4 (1), R 128 R 1 9 1 . Geological Society of America, University of Kansas, Lawrence, Kansas.

walker, e. m. 1958. The Odonata of Canada and Alaska , 2, xi + 318 pp., 2 pis. University of Toronto Press.

wighton, d. c. and wilson, m. v. H. 1986. The Gomphaeschninae (Odonata: Aeshnidae): new fossil genus,
reconstructed phylogeny, and geographical history. Syst. Ent. 11, 505 522.

worssam, b. c. 1978. The stratigraphy of the Weald Clay. Rep. Inst. Geol. Sci. 78 (1 1 ), 1 19.

E. A. JARZEMBOWSKI

Booth Museum of Natural History
Dyke Road
Brighton BN1 5AA

Typescript received 30 April 1987

Revised typescript received 24 September 1987

ACANTHODIAN FISH REMAINS FROM THE
UPPER SILURIAN OR LOWER DEVONIAN OF
THE AMAZON BASIN, BRAZIL

by p. janvier and j. h. g. melo

Abstract. Acanthodian spines, scales, and tooth whorls are described from the Pitinga Member of the
Trombetas Formation (Middle Amazon Basin, northern Brazil). The spines and dermal scutes are of climatiid
type, but the associated tooth whorl is more like that of an ischnacanthid. This acanthodian assemblage is
quite similar to that from the top of the Catavi Formation of Bolivia, where an association of thelodont scales
suggests an Early Devonian age. The new acanthodian remains from Brazil are unlikely to be older than the
latest Silurian, thus refuting previous ideas that the richly fossiliferous Pitinga Member was Ordovician in age.

The fossils described here were collected in the north-western portion of the State of Para (northern
Brazil), in a region where the Trombetas River and its major tributary, the Mapuera River, cross
the belt of outcrops of Siluro-Devonian rocks that delineates the northern flank of the Middle
Amazon Basin (text-fig. 1). Most of the fish remains were recovered from a core sample of the well
SM-504, a shallow borehole drilled by Enge-Rio Engenharia e Consultoria S.A., near Cachoeira da
Porteira village, in the outcrop of the upper section of the Pitinga Member of the Trombetas
Formation. A bone-bed was encountered at a depth of 28-60 m within a thick layer of grey, fine-
grained sandstone. The upper and lower surfaces of the fossiliferous sample have been prepared by
removing the bone fragments and cleaning their natural moulds with dilute hydrochloric acid, after
which a silicone cast was made from both surfaces (text-fig. 2), showing acanthodian scales, spines,
scutes, and a tooth whorl. In addition, an isolated acanthodian spine (text-fig. 3b, c) was found
along with bony fragments in siltstones which crop out at a cascade of the Sucuriju Creek (text-fig.
1), some 26 km to the west-south-west of well SM-504. It may belong to the same form as the spines
in the core sample, thus suggesting that the same bone-bed may be followed over a relatively large
distance.

These fossils are interesting because of the scarcity of Middle Palaeozoic vertebrate remains in
Brazil. For a long time, the only record of Devonian vertebrates from this country included mentions,
but not illustrations, of possible Machaer acanthus spines and Pteraspis plates from the Maecuru
Formation of the Middle Amazon Basin (Katzer 1897a, 6), and spines from the Parnaiba Basin
(Pimenteira Formation) assigned to such genera as Ctenacanthus , Machaer acanthus, and 1 Devon -
canthus' (sic) (Kegel 1953, 1957; Santos 1961; Guimaraes 1964; Mendes 1971; Copper 1977).
Feonardi (1982, 1983) described a tetrapod footprint from the Upper Devonian of the Parana Basin,
and Janvier and Melo (1987) recorded some isolated actinopterygian scales from the Fate Devonian
shales of the Upper Amazon Basin.

All fossil material under considereration is housed in the collection of Petrobras (Museu de
Paleontologia do Cenpes, Rio de Janeiro, Brazil), under register numbers CENPES 002-V and
003-V.

DESCRIPTION

On one surface of the core sample from well SM-504, the cast shows several more or less complete spines and
a large tooth whorl (text-fig. 2a, b). The most complete spines are a pectoral spine (b in text-fig. 2b) and an
intermediate spine or scute (d). A fragmentary spine with a broad posterior surface (0 is probably a median

| Palaeontology, Vol. 31, Part 3, 1988, pp. 771-777-1

© The Palaeontological Association

772

PALAEONTOLOGY, VOLUME 31

text-fig. I. Locality map. Numbers refer to fossil sites mentioned in text: 1, Pitinga Member, basement
contact outcrop at the Viramundo Waterfall; 2, Madame Island; 3, Praia Lisa Island; 4, Bota Island; 5,
Caramujo locality. Fossils occurring in each locality are indicated by symbols shown in the figure.

tin spine, as is another fragment ornamented with very straight ridges (e). Finally, a broad-based spine (c),
ornamented with strongly divergent ridges which are visible on the counterpart (c in text-fig. 2d), is most
probably a prepectoral spine. The tooth whorl (a) bear three teeth and the base of a fourth one. The teeth are
strongly bent posteriorly, and ornamented with sharp ridges, which become sinuous near the base laterally.
There are very small lateral denticles on one side only (dtl in text-fig. 3a); on the opposite side of the main
teeth, these denticles are replaced by shallow sharp ridges which prolong the flange of each tooth posteriorly.
Such an asymmetry is also visible on the bony support of the whorl, which is embayed posteriorly by a deep
notch. Such morphology may indicate that this tooth whorl did not occupy a median, symphysial position,
but rather a parasymphysial position.

JANVIER AND MELO: ACANTHODI ANS FROM BRAZIL

773

g

text-fig. 2. Acanthodian remains from well SM-504, Pitinga Member of the Trombetas Formation, State of
Para, Brazil (specimen no. Cenpes 002-V). a, b, lower surface of the core sample; silicone cast (a) and
explanatory scheme (b). c, d, upper surface of the core sample; silicone cast (c) and explanatory scheme (d).
Scale: 10 mm for a and b, and 1 mm for h. Identifiable remains: a, tooth whorl; b, pectoral fin spine; c,
prepectoral spine; d, intermediate spine; e, f, median fin spines; g, possible plate of sclerotic ring or dermal

bone of the cheek; h, scale in external view.

Ornamentation of the spines consists of relatively thin, sharp ridges, except on the intermediate and
prepectoral spines, where the ridges are smoother and more irregular in shape. The pectoral spine displays
about thirteen ridges on one side at the base. In places, particularly near the insertion base of the spines, these
ridges are noded, the nodes often being set closely together.

774

PALAEONTOLOGY, VOLUME 3

text-fig. 3. a, reconstruction of the tooth whorl in text-fig. 2a, b, showing the lateral denticles (dtl). B, c,
prepectoral spine from Sucuriju Creek outcrop, Pitinga Member of the Trombetas Formation, State of Para,

Brazil (specimen no. Cenpes 003-V). Scale: 10 mm.

On the other surface of the core sample (text-fig. 2c), the skeletal elements are much smaller or broken into
small pieces. This may be due to the fact that the mode of deposition of these remains has varied during the
formation of the bone-bed. This surface shows a large number of isolated scales and small dermal plates.
Strangely enough, most scales are exposed in basal view, and show the classic gibbose base of acanthodian
scales. Only one scale displays a well-preserved crown (h in text-fig. 2d), ornamented with a median boat-
shaped ridge and lateral stepped zones.

Finally, a small dermal bone (g) with a vermiculate ornamentation may be either a plate of the sclerotic ring
or a dermal bone of the cheek.

The isolated spine from the Sucuriju Creek (text-fig. 3b, c) is a large, flat prepectoral spine, ornamented
with broad, somewhat sinuous noded ridges. The proximal half of these ridges bears a double row of nodes.

COMPARISON AND DISCUSSION

The presence of broad-based fin spines, prepectoral spines, and large intermediate ventral spines
indicates a climatiid acanthodian. The ornamentation of spines (particularly that of the pectoral fin
spine) and scales is very like Ptomacanthus Miles (Miles 1973, fig. 1a, pi. 4, 19) from the Lower
Devonian (Gedinnian) of Europe. In contrast, the tooth whorl is not of climatiid type, but resembles
that of the ischnacanthid Gomphonchus Gross (Gross 1967). The only differences concern the very
large size, much shorter lateral denticles, and sinuous ridges on the main teeth in the Brazilian form.

Finally, the prepectoral spine of the Sucuriju Creek is of climatiid type, and matches the ones
found in the borehole core, yet its double rows of tubercles on each ridge represent quite an unusual
type of ornamentation.

Although climatiid and ischnacanthid acanthodians are known as early as the Late Silurian, this
assemblage of large forms is rather suggestive of an Early Devonian acanthodian fauna. If the
climatiid remains are to be referred to Ptomacanthus , a Gedinnian (Lochkovian) age would be
preferable for this part of the Trombetas Formation.

JANVIER AND MELO: ACANTHODIANS FROM BRAZIL

775

It is noteworthy that these Brazilian acanthodian spines are strikingly similar to those recorded
from a bone-bed in the Catavi Formation at Seripona, Bolivia (Goujet et al. 1984; Janvier and
Suarez-Riglos 1986; Gagnier et al., 1988). The Catavi Formation is regarded as Late Silurian
in age, and the bone-bed at Seripona was originally referred to the Pfidolian on the basis of
lithostratigraphical correlations with Clarkeia antisiensis- bearing localities. However, its acantho-
dian and thelodont assemblage is rather suggestive of the Early Devonian (possibly Siegenian or
even Emsian; Turner, in Gagnier et al. , 1988). This discrepancy in the dating of the Seripona bone-
bed is hard to resolve, although one should note that thelodont scales have proved to be quite
reliable stratigraphical fossils. Those from Seripona are inferred to represent a large species of
Turinia, comparable to Early and Middle Devonian forms from Australia and Antarctica (Turner
in Gagnier et al ., 1988). No thelodont scales of this genus have been recorded from the Silurian, yet
thelodont scales are abundant as early as the Early Silurian.

STRATIGRAPHICAL COMMENTS

The discovery of these acanthodian remains may have a major bearing on the determination of the
upper age limit of the Trombetas Formation. In fact, this has become a matter of concern in the
last few years, since Quadros (1985a and b) assigned the upper section of that unit (the Pitinga and
Manacapuru Members) from the Ludlovian/Gedinnian, upwards to the Siegenian, on the basis of
acritarchs and chitinozoans; he revalidated the viewpoint of Ludwig (1964), who envisaged the
possibility of a gradational contact between the Trombetas Formation and the overlying Devonian
Maecuru Formation, based on sedimentological evidence. In contrast to Ludwig’s view, an inter-
vening unconformity (corresponding to the Wenlockian/Siegenian gap) has been conventionally
recognized by most authors, according to whom the Trombetas Formation could be no younger
than the Llandovery (Lange 1967, 1972; Daemon and Contreiras 1971; Caputo et al. 1972), or at
most Wenlockian (Caputo 1984; Caputo and Lima 1984). Late Ordovician/Early Llandoverian
(Medinian) age assignments for the Pitinga Member have long persisted in the literature, ever since
the pioneer palaeontological and stratigraphical investigations of early workers (e.g. Derby 1878;
Clarke 1899; Katzer 1903; Schuchert 1906; Maury 1929; Ruedemann 1929) had pointed out the
occurrence of characteristic species in the local fauna, such as Arthrophycus harlani Conrad, Climaco-
graptus innotatus Nicholson, and Orthis callactis Dalmann. However, further research has provided
new evidence that is consistent with Quadros’ interpretation. Climacograptus, for instance, is known
to have survived into the Lower Devonian of Europe (see Jaeger 1979, for original reference).
Current palynological investigations by Dr Jane Gray (University of Oregon, USA) strongly suggest
that the spore assemblages of the Pitinga Member are of post-Llandoverian age, and the alleged O.
callactis of Clarke (1899) turned out after closer inspection to be a generically indeterminable
dalmanellid (A. J. Boucot, pers. comm. 1984, 1985).

The acanthodians discussed herein are certainly no older than the Upper Silurian (see discussion
above); they occur only in highly bioturbated siltstones and sandstones that are regarded as belonging
to the uppermost Pitinga Member (the only member of the Trombetas Formation present in the
Trombetas valley, according to recent geological interpretations). These bioturbated, sometimes
massive beds differ markedly from the underlying shales and sandstones of the Pitinga Member,
and have been previously mapped by Enge-Rio geologists as the lower part of the Devonian Maecuru
Formation (Jatapu Member). However, more recent palynological determinations by L. P. Quadros
of the fish-bearing sediments from the cascade of the Sucuriju Creek revealed acritarch assemblages
that are typical of the Pitinga Member (L. P. Quadros, pers. comm. 1986). Thus, the Jatapu Member
is considered to be absent in this region, the Maecuru Formation being represented here only by its
upper division, the Lontra Member, of Emsian/Eifelian age.

Unfortunately, little is known of the vertical distribution of critical taxa through the Trombetas
section. The following remarks are based on preliminary field observations made in 1986 by
J.H.G.M.

776

PALAEONTOLOGY, VOLUME 31

Less than 75 m below the bone-bearing interval of well SM-504, at the Viramundo waterfall (text-
fig. 1), the lowermost Troinbetas beds, resting on the crystalline basement, are finely laminated
siltstones and shales charged with climacograptid remains. A few metres higher, downstream of the
Viramundo waterfall, fine-grained sandstones, locally displaying a hummocky cross stratification,
contains the invertebrate fauna described by Clarke (1899), which includes brachiopods ( Anabaia ,
Heterorthella , etc.), similar to those of the Upper Silurian of Argentina, Paraguay, and Bolivia.
These fossils may be collected at the northern end of the Boto and Praia Lisa Islands, the southern
portion of Madame Island, and adjacent outcrops on the right bank of the river. Still higher in the
column (that is, further downstream), abundant trace fossils referable to Arthrophycus harlani
Conrad crowd the bedding planes of sandstone slabs exposed in a bluff at Caramujo locality, on
the left bank of the Trombetas River.

None of the fossils mentioned above has been observed in association with the fish remains in
the bioturbated rocks of the upper Pitinga Member; rather, they seem to be stratigraphically confined
to lower beds of distinct lithology. In the Sucuriju Creek, scarce, poorly preserved Arthrophycus-
like traces are found in sandstones exposed a few hundred metres upstream of the cascade outcrop
(i.e. stratigraphically a few metres above it), where acanthodian remains occur, but these traces may
represent a form other than A. harlani (A. J. Boucot, pers. comm. 1986).

CONCLUSIONS

Acanthodian remains from the Pitinga Member of the Trombetas Formation of northern Brazil
suggest an Early Devonian (possibly Gedinnian) age. An Ordovician or Early Silurian age can be
definitely ruled out for this section, but a late Silurian age is still possible.

Acknowledgements . The authors are indebted to Dr A. J. Boucot (Oregon State University, Corvallis, USA),
Dr J. Gray (University of Oregon, Eugene, USA), and Dr L. P. Quadros (Petrobras Research Center— Cenpes,
Rio dc Janeiro, Brazil) for having provided pertinent informations quoted in the text, and to Petrobras—
Petroleo Brasileiro S.A. for permission to publish this paper. English was corrected by Dr M. Pickford (Paris).

REFERENCES

caputo, m. v. 1984. Stratigraphy, tectonics, palaeoclimatology and palaeogeography of northern basins of
Brazil, 583 pp. Ph.D. thesis (unpublished). University of California, Santa Barbara, California.

— and lima, e. c. 1984. Estratigrafia, idade e correla^o do grupo Serra Grande. An. 33e Congr. brasil. Geol.
2, 740 753.

— rodrigues, r. and vasconcelos, d. n. n. 1972. Nomenclature estratigrafica da bacia do Amazonas;
historico e atvaliza^ao. An. 26e Congr. Brasil. Geol. 3, 35-46. Sociedade Brasileira de Geologia, Belem.

clarke, J. m. 1899. A fauna siluriana superior do rio Trombetas. Arch. Mas. nacn. 10, 1 48.

copper, p. 1977. Paleolatitudes in the Devonian of Brazil and the Frasnian Famennian mass extinction.

Palaeogeogr ., Palaeoclimatol., Palaeoecol. 21, 165-207.
daemon, r. f. and contreiras, c. j. a. 1971. Zoneamento palinologico da bacia do Amazonas. An. 25e Congr.
brasil. Geol. 3, 79 88.

derby, o. a. 1878. Contribufoes para a geologia da regiao do Baixo Amazonas. Arch. Mas. nacn. 3, 77-104.
gagnier, p. y., turner, s., friman, l., suarez-riglos, m. and janvier, p. 1988. The Lower Devonian vertebrate
and mollusc fauna of Seripona, Chuquisaca, Bolivia. N. Jb. geol. paldont. Abh. 176, 269-297 .
goujet, d., janvier, p. and suarez-riglos, m. 1984. Devonian vertebrates of South America. Nature, Lond.

312, 311.

gross, w. 1967. Uber das Gebiss der Acanthodier und Placodermen. Jl Linn. Soc. (Zook), 47, 121 130
guimaraes, d. 1964. Geologia do Brasil, 264pp. Minist. Minas e Energia, Rio de Janeiro.
jaeger, h. 1979. Devonian graptolithina. Spec. Pap. Palaeont. 23, 335-339.

janvier, p. and melo, j. h. g. 1987. Late Devonian actinopterygian scales from the Upper Amazon Basin,
north-western Brazil. An. Acad, brasil. Cienc. 59, 3.

and suarez-riglos, m. 1986. The Silurian and Devonian vertebrates of Bolivia. Bull. Inst. Fr. Etud. and.
15, 73-114.

JANVIER AND MELO: ACANTHODIANS FROM BRAZIL

111

katzer, f. I897«. Das Amazonas-Devon und seine Beziehungen zu den anderen Devongebieten der Erde.
Sitzungsber. Kgl. bohm. Gesel/sch. IViss. , Math.-naturwiss. Kl. 46, I 50.

1897ft. A fauna devonica do rio Maecuru a as suas relagoes com a fauna de outros terrenos devonicos
do globo. Bol. Mus. Para. Hist. nat. ethnogr. 2, 204-246.

— 1903. Grundziige der Geologie des unteren Amazonas-gebietes (des Staates Para in Brasiliens ), 298 pp. Max
Weg, Leipzig.

kegel, w. 1953. Contribufao para o estudo do Dcvoniano da bacia do Parnaiba. Bol. Div. Geol. Mineral. 141,
I 48.

— 1957. Das Parnaiba-Becken. Geol. Rundsch. 45, 522 540.

lange, f. w. 1967. Subdivisao bioestratigrafica e revisao da coluna siluro-devoniana de bacia do Baixo
Amazonas. In At as Simp, sobre a Biota Amazonica , I, 215-326, annex 1-4. Conselho Nacional de Pesquisas,
Belem.

1972. Silurian of Brazil. Spec. pap. geol. Soe. Am. 133, 33-39.
leonardi, G. 1982. Descoberta da pegada de urn anfibio devoniano no Parana. Cienc. Terra , 5, 36-37.

1983. Notopus petri nov. gen., nov. sp.: line empreinte d’amphibien du Devonian du Parana (Bresil).
Geobios , 16, 233-239.

ludwig, G. 1964. Nova divisao estratigrafica e correla^ao faciologica por meio de pcquenas estruturas intcrnas
dos sedimenlos silurianos e devonianos da bacia do Medio Amazona s, 72 pp. Petrobras-Cenap (cienc. -Tec.-
Petrol ., Seg. expl. Petrol ., I), Rio de Janeiro.

maury, c. j. 1929. Uma zona de graptolitos do Llandovery Inferior no rio Trombetas, Estado do Para, Brasil.
Monogr. Serv. geol. mineral. 7, I -53.

mendes, J. c. 1971. Geologia do Brasil , 344 pp. Instituto Nacional do Livro, Rio de Janeiro.
miles, r. s. 1973. Articulated acanthodian fishes from the Old Red Sandstone of England, with a review of the
structure and evolution of the acanthodian shoulder-girdle. Bull. Br. Mus. nat. Hist. (Geol.), 24, 113-213.
quadros, l. p. 1985m Distribui^ao bioestratigrafica dos Chitinozoa e Acritarchae na bacia do Amazonas, 179
pp. Ph D. thesis (unpublished), Universidade Federal do Rio de Janeiro, Rio de Janeiro.

— 1985ft. Natureza do contato entre as formaqoes Trombetas e Maecuru (bacia do Amazonas). In campos,
d. a., ferreira, c. s., brito, i. m. and viana, c. f. (eds.). Coletanea de trabalhos paleontologicos; trabalhos
apresentados no VIII Congresso Brasileiro de Paleontologia, 1983, 435 441. DNPM-MME (Ser. geol., 27,
Se£. paleont. estratigr., 2), Brasilia.

ruedemann, R. 1929. Descripijao dos graptolitos do rio Trombetas. Monogr. Serv. geol. mineral. 7, 20 24.
santos, r. s. 1961. Peixes fosseis do Devoniano Inferior de Picos, Estado do Piaui. An. Acad, brasil. Cienc. 33,
32.

schuchert, c. 1906. Geology of the Lower Amazon region. J. Geol. 14, 722 746.

PHILIPPE JANVIER

U.A. 12 du C.N.R.S.
Institut de Paleontologie
8, rue Buffon
75005 Paris
France

Typescript received 25 May 1987
Revised typescript received 9 July 1987

JOSE IIENRIQUE G. MELO
Petrobras/Cenpes/Divex/Sebipe
Cidade IJniversitaria, Qd. 7
1 1 h a do Fundao
21910 Rio de Janeiro
Brazil

A MIDDLE CAMBRIAN CH ELICER ATE FROM
MOUNT STEPHEN, BRITISH COLUMBIA

by DEREK E. G. BRIGGS and DESMOND COLLINS

Abstract. A recently discovered arthropod, Sanctacaris uncata gen. et sp. nov., from the Glossoplewa Zone,
Stephen Formation of Mount Stephen, British Columbia belongs in the Chclicerata. The head shield is wider
than long, convex axially, and extends laterally into two flat triangular projections. It bears at least six pairs
of biramous appendages. The first five are similar, increasing in size posteriorly and arranged with their
inner rami in a raptorial array of inwardly facing, segmented, spinose limbs, accompanied by antenna-like,
presumably sensory, outer rami. The outer ramus of the sixth appendage is also antenna-like, but the inner
is short and terminates in a fringe of radiating spines. The eyes are at the front of the head shield. The trunk
has eleven tergites, each with a convex axis and projecting pleurae. The corresponding somites of the first ten
each bear a pair of biramous appendages with an inner segmented spinose ramus and an outer lamellate ramus,
fringed with long setae, which functioned in swimming and respiration. The wide flat telson is adapted for
stabilizing and steering.

Sanctacaris displays characters which are all derived for some member of the chelicerates. These include: 1,
at least six pairs of appendages (the first five raptorial) on the head shield; 2, a cardiac lobe; 3, the division of
the body into tagmata comparable to the prosoma and opisthosoma of merostomes; and 4, the anus at the
rear of the last trunk segment. Such a combination is unique to the chelicerates. The apparent lack of chelicerae,
an advanced character present in all other chelicerates, is consistent with the primitive biramous appendages
on both the head and trunk. It places Sanctacaris in a primitive sister group of all other chelicerates.

Sanctacaris demonstrates that chelicerates, although rare, were present in Middle Cambrian seas. Moreover,
even at this early stage of chelicerate evolution, Sanctacaris had the number and type of head appendages
that are found in modified form in the eurypterids and xiphosurids, the major Palaeozoic groups that
succeeded it.

C. D. Walcott’s extraordinary discovery, the Middle Cambrian Burgess Shale of Yoho National
Park in southern British Columbia, has become celebrated for perhaps the most important biota of
soft-bodied organisms known from the fossil record (Whittington 1985). Walcott’s material came
from a single section on the west side of the ridge between Mount Wapta and Mount Field and was
collected from the main quarry in the ‘Phyllopod bed’ and the smaller Raymond quarry some
23 m above (Whittington 1971). The Burgess Shale section occurs in the lower two-thirds of the
Stephen Formation where the basinal shales abut against the steep face of the adjacent dolomite
reef of the Cathedral Formation. The conditions necessary for the preservation of the soft parts of
the organisms appear to have been controlled by the proximity of this reef front. Away from the
reef front the exceptional preservation is less common.

The Burgess Shale was long considered to be a unique occurrence. In 1977 Mcllreath demonstrated
that the reef front, or Cathedral Escarpment as it is known, could be traced for about 20 km south-
east of Walcott’s quarry and that the contact between reef and basinal shales cropped out again on
Mount Field, Mount Stephen, Mount Odaray, Park Mountain, and Curtis Peak. One of us (D.C.)
speculated that more localities of soft-bodied fossils might be found in the basinal shales near these
contacts, and, indeed, a few indications were later reported by Aitken and Mcllreath (1981) along
the line of the Escarpment. In 1981 and 1982 field-work organized by D.C. and involving D.E.G.B.
and others led to the discovery of about a dozen new localities (Collins et al. 1983).

The most promising of the new localities (locality 9 of Collins et al. 1983, fig. 1) occurred in a
large in situ block of pale grey-blue siliceous shale about 1500 m south-west of the outcrop of the
Cathedral Escarpment on the north shoulder of Mount Stephen. This is about 5 km almost directly

| Palaeontology, Vol. 31, Part 3, 1988, pp. 779-798, pis. 71-73.|

© The Palaeontological Association

780

PALAEONTOLOGY, VOLUME 31

south of the Burgess Shale quarries. The site was excavated by a Royal Ontario Museum party in
the summer of 1983 when the arthropod described here was discovered (Collins 1986).

The stratigraphic level where the block occurred is characterized by the trilobite, Glossopleura,
which is the local zone fossil for the basal part of the basinal Stephen Formation (Fritz 1971). In
the Stephen Formation section about 1000 m to the north on Mount Stephen measured by Fritz
(1971, fig. 6), the top of the Glossopleura Zone is 40 m below the level equivalent to the main Burgess
Shale quarry. The block excavated was at least 40 m below the top of the Glossopleura Zone, so
was 80 m or more stratigraphically below the level of the Burgess Shale ‘Phyllopod bed’.

The faunal assemblage from the block is dominated by the arthropods, Alalcomenaeus and
Branchiocaris , which are very rare in the Burgess Shale. Many other Burgess Shale animals were
found (Collins et al. 1983) but not the most common one, Marrella. A number of new forms are
also present (Collins 1986). It is evident, therefore, that this fauna is distinct from those in the
Burgess Shale. It is also older. This is the first of a number of papers describing the animals from
the Glossopleura faunal assemblage in the Stephen Formation.

Terminology and methods. The morphological terms used in the description are those of Stormer (1955) as far
as possible. The orientation of specimens relative to bedding is given as parallel (i.e. dorsoventral) or oblique
(Whittington 1971); the restoration (text-fig. 6) is based on the approach described by Briggs and Williams
(1981). The explanatory diagrams which face the plates were made from tracings of large colour photographs
of the specimens and camera lucida drawings using a Wild M7S microscope. The specimens were photographed
either immersed in water or dry, with the light directed at about 30° to the horizontal. The direction of
illumination was varied where necessary to illustrate different features.

A small amount of preparation was carried out using a needle inserted in a percussion hammer with an
adjustable throw driven by a dental drill motor.

Repository. All material is held by the Department of Invertebrate Palaeontology of the Royal Ontario
Museum, Toronto (abbreviated ROM).

Preservation. All five specimens are complete and appear to be carcasses rather than moults. A number of lines
of evidence suggest that the mode of deposition at this locality on Mount Stephen (locality 9 of Collins et al.
1983) was essentially similar to that of the beds in the Walcott quarry (Whittington 1971, 1980), reflecting a
similar geological setting. The fossils are likewise preserved in a variety of orientations to bedding, the
compacted layers separated by a veneer of sediment, indicating deposition from a turbulent cloud of sediment.
The intervening layers of sediment allow the specimens to be prepared in the same manner as those from the
Walcott quarry. However, the sediment does not separate as readily from the layers of the specimen and the
potential for ‘palaeodissection’ is consequently more limited. The layers show evidence of fining upward from
an erosive base, in the manner of deposits from a density current. There is no evidence of scavenging or much
decay and this, together with a lack of bioturbation, suggests that deposition was rapid and that bottom
conditions may have been anoxic. Thus, like those excavated in the Walcott quarry, the organisms were
deposited in a ‘post-slide’ environment inimical to life, which was very different to the ‘pre-slide’ living
environment (Conway Morris 1979; Whittington 1980). The Cathedral Escarpment is, at most, 1500 m away
from the locality, but the distance or direction of transport is unknown. Allison (1986) has shown that live or
freshly killed arthropods can undergo transport over extensive distances (more than 10 km) without significant
damage.

SYSTEMATIC PALAEONTOLOGY
Phylum ARTHROPODA

Subphylum chelicerata

Taxa of lower rank above genus. Not assigned, plesion (sensu Patterson and Rosen 1977), primitive sister group
of all other chelicerates.

Genus sanctacaris gen. nov.

Derivation of name. Latin sanctus (saint; sacred, holy), referring to Santa in Santa Claws, the field name for
the holotype of this arthropod (Collins 1986), and caris (crab).

BRIGGS AND COLLINS: CAMBRIAN CHELICERATE

781

Type species. Sanctacaris uncata sp. nov.

Diagnosis. Head shield with pronounced axial convexity and triangular lateral projections; bearing
at least six pairs of biramous appendages, first five similar, increasing in size posteriorly and arranged
with inner rami in concentric array of inwardly facing, segmented, spinose limbs; sixth inner ramus
short, with radiating spines, outer rami antenna-like; eyes situated anterolaterally at front of head
shield.

Trunk not subdivided into tagmata, comprised of eleven tergites, first ten each bearing a pair of
similar biramous appendages, decreasing in size posteriorly; inner ramus segmented with short
spines, outer ramus broad and lamellate with long setae. Anus at posterior of eleventh trunk
segment, beneath telson; telson wide, flat, and paddle-shaped.

Geological horizon. Middle Cambrian, Stephen Formation, Glossopleura Zone, British Columbia.

Sanctacaris uncata sp. nov.

Plates 71-73; text-figs. 1-6

Derivation of species name. Latin uncata (bent inward, hooked, barbed), referring to claws in 'Santa Claws’.
Holotype. ROM 43502, part and incomplete counterpart, Plate 71.

Other material. ROM 43503-43506, part and counterpart.

Diagnosis. As for the genus.

Locality and stratigraphical horizon. Locality 9 of Collms et at. ( 1983, fig. 1), c. 7000 feet (2286 m) elevation,
1500 m south-west of the north shoulder of Mount Stephen, British Columbia; 40+ m below top of Glossopleura
Zone, Stephen Formation.

Associated fauna. Listed in Collins et al. (1983, table 1).

Description

Head shield. All five specimens are preserved in parallel or oblique orientation. ROM 43506 (PI. 73, fig. 5; text-
fig. 5) most nearly approaches a lateral aspect, but the outline of the head shield is obscured. Hence the three-
dimensional appearance is difficult to restore.

The outline of the head shield in dorsal view is shown by ROM 43505 (PI. 73, figs. I and 2; text-fig. 3). Both
the anterior and to a lesser degree the posterior margins are curved convexly. The lateral areas are subtriangular.
A pronounced convexity of the axial area is evident in relief, though it has been largely reduced by folding
during compaction. The curved compaction wrinkles indicate that the axial region was dome-shaped. The
head shield of the holotype, ROM 43502 (PI. 71, figs 2 and 3; text-fig. 1b), is similar, although the apices of
the lateral areas are more acutely angled, and the head shield foreshortened, due to posterior tilting. The
strongly convex projection of the front of the head shield can be seen more clearly in this specimen.

The original outline of the lateral areas is best revealed on the left side of the head of ROM 43504. This
specimen affords a dorsal view, but is tilted slightly obliquely, mainly by rotation around the longitudinal axis
(PI. 72, figs. 1 and 3; text-fig. 2b). The left lateral area has a more acute apex and a less convex anterior edge
than that in ROM 43505 (PI. 73, figs. I and 2), and it lacks compaction wrinkles. These features indicate that
the plane of the left lateral area of ROM 43504 (PI. 72, figs. 1 and 3) was near parallel to bedding when buried.
In contrast, the right lateral area of this specimen has a more rounded apex, a more convex anterior edge, and
is covered in compaction wrinkles indicating that it was at a higher angle to bedding. Thus, the evidence
suggests that the lateral areas were inclined venlrally in life (text-fig. 6). If they had been horizontal (i.e. in the
same plane), both right and left lateral areas in ROM 43504 would have had the same outline (angled to the
same extent above or below the bedding plane) and similar compaction wrinkles. This interpretation also
explains why the apices of the lateral areas of ROM 43505 (PI. 73, figs. 1 and 2; text-fig. 3), although symmetrical
about the axis, are more rounded and less acute than the left lateral area of ROM 43504. In ROM 43503 (PI.
73, figs. 3 and 4; text-fig. 4) the head shield has been tilted downwards at a high angle to the bedding.

A paired row of dark reflective traces is present on the axial area of three of the specimens. They are most
distinct in ROM 43505 (PI. 73, fig. 2; text-fig. 3) where they occur as paired black spots. Two pairs are distinct
and two indistinct. In ROM 43502 (PI. 71, fig. 3; text-fig. 1b) the traces occur as two irregular black streaks.

EXPLANATION OF PLATE 7 1

Figs. I -3. Sanctacaris uncata gen. et sp. nov., holotype, ROM 43502, dorsal view. 1, counterpart (text-fig. 1a),
x4, dry, appendages projecting beyond the head shield, illuminated from the north: 2 and 3, part (text-fig.
1b), x 1-5; 2, immersed in water, showing structures beneath the dorsal exoskeleton; 3, dry, showing relief,
illuminated from the north-west.

PLATE 71

BRIGGS and COLLINS, Sanctacaris

784

PALAEONTOLOGY, VOLUME 31

EXPLANATION OF PLATE 72

Figs. 1-5. Sanctacaris uncata gen. et sp. nov., ROM 43504, oblique dorsal view, part (text-fig. 2a, b). 1, x 3-5,
dry, appendages projecting on both sides beyond the head shield, illuminated from the north-east. 2, x 2-5,
dry, segmented ramus of left trunk appendage I, lamellate rami of I to 7, illuminated from the north. 3,
x 1 -25, dry, illuminated from the north-east. 4, x 2-25, dry, segmented rami of right trunk appendages 4
and 5, flanking lamellate ramus of 5, illuminated from the north-west. 5, x 3, immersed in water, telson and
dark stain beyond anus.

PLATE 72

BRIGGS and COLLINS, Sanctacaris

786

PALAEONTOLOGY, VOLUME 31

and they are even more indistinct on the head shield axis of ROM 43504 (PI. 72, figs. 1 and 3). The occurrence
of streaks rather than distinct spots may be the result of greater distortion of the head shields of ROM 43502
and 43504, compared to that of ROM 43505. Similar dark traces have been interpreted as muscle attachment
sites in the heads of synziphosurines (Eldredge 1974), where they are equivalent to the cardiac lobe, and in
phacopine trilobites (Eldredge 1971).

Head appendages. The head appendages are best preserved on the holotype, ROM 43502 (PI. 71, fig. 1; text-
fig. I a). Five pairs of spinose, raptorial limbs are evident curving forward from below the head shield, flanked
by two pairs of antenna-like structures. The outer pair of the two is more completely preserved; the structures
are elongate and slender and both right and left show large isolated proximal spines. Outside these antenna-
like structures two paired projections with radiating spines at the end extend a short distance beyond the
margin of the head shield. The five raptorial limbs are arranged in series, each pair increasing in length
and lying below and outside that preceding it. The number of segments in the limbs increases from at least
four in the first to eight or more in the fifth. The limbs are otherwise similar in structure, the terminal
segment bearing three inwardly curving spines, the more proximal segments bearing projecting bundles
of three or more inwardly-angled spines. A similar radiating spread of head appendages is evident in
dorsal view on ROM 43505, but they are poorly preserved and details are difficult to discern (PI. 73, fig. 1;
text-fig. 3).

A different view of the head appendages is provided by ROM 43504. Those of the right side are straight
and inclined anterolaterally (PI. 72, fig. 1; text-fig. 2a). Five raptorial limbs, presumably equivalent to those
in ROM 43502, are preserved. In addition, five antenna-like limbs, similar to those flanking the raptorial limbs
in ROM 43502, are evident, suggesting that a raptorial limb and an antenna-like limb together make up a
biramous appendage. No further limb elements can be seen on the right side of ROM 43504. The left head
appendages are strongly curved and overlapping. They show the antenna-like ramus lying outside the raptorial
one indicating its probable relative position in the biramous appendage (PI. 72, fig. 1; text-fig. 2a).

Poorly preserved head appendages incline anteroventrally from the head of ROM 43506 (PI. 73, fig. 5; text-
fig. 5) in a similar orientation to those on the right side of ROM 43504.

The structure and arrangement of the first five pairs of appendages in the head are clear. They are biramous,
the raptorial rami facing inwards and presumably bearing gnathobases proximally, the antenna-like rami
flanking them on the outside. They appear to have been attached parallel to the mid-line: an indication of a
narrow gap between the limb bases, in which the mouth was presumably situated, is present in ROM 43502,
particularly in the frontal projection (PI. 71, figs. 1 and 2).

The arrangement of the structures preserved outside the raptorial limbs in ROM 43502 is more problematic.
On the counterpart (PI. 71, fig. I; text-fig. 1a), the outer, antenna-like ramus is seen to converge with a short
projection fringed with radiating spines or setae, on both sides of the head. Together, the two structures seem
to comprise a biramous sixth head appendage. If this is so, then the pair of antenna-like structures just inside
the outer pair probably belong with the fifth raptorial limb, and, at least on the left side, appear to curve
beneath the head shield parallel to it. Lastly, a small array of spines lies beneath the short spiny projection on
both sides of the head. Unfortunately, the rest of this structure is concealed by the head shield, so it is not
clear whether it is a third ramus of the sixth appendage or belongs to a seventh. However, whatever the
interpretation of the structures outside the raptorial limbs, it is evident that Sanctacaris has at least six pairs
of biramous appendages in the head.

Eyes. ROM 43505 preserves a well-defined dark rounded structure on the left side of the head shield (PI. 73,
figs. I and 2; text-fig. 3), just abaxial of the appendages, which is probably an eye. A narrow marginal rim
recalls the eyes of Odaraia (Briggs 1981) and other Burgess Shale arthropods. A similar round structure with
a marginal rim occurs in the same position at the front of the head shield on the right side of ROM 43502,
best seen on the counterpart (PI. 71, fig. 1; text-fig. 1a). A matching, but less distinct, structure occurs on the
left side, also seen best on the counterpart. None of the other specimens preserves clear evidence of eyes, but
they could be concealed in the matrix.

Trunk. The trunk consists of eleven tergites (presumably corresponding to somites) and a telson (PI. 71, figs.
2 and 3; text-fig. 1b). The first five increase in length slightly; tergites 6 to 1 1 are very similar in length. The
trunk widens slightly to the fourth tergite, then tapers gradually to the eleventh. The dimensions of the tergites
provide no obvious basis for a subdivision of the trunk (into thorax and abdomen, or pre- and postabdomen,
for example). The axial area of the trunk, like that of the head, shows a pronounced convexity which has been
reduced in large measure by folding during flattening. Short longitudinal ridges, one pair per tergite, define an
axis along the centre of this raised area (PI. 71, fig. 3; text-fig. 1b; PI. 73, fig. 4; text-fig. 4). The pleurae are

BRIGGS AND COLLINS: CAMBRIAN CHELICER ATE

787

similar in width to the raised area. The anterior margin of each pleura curves posteriorly at its lateral extremity
to meet the posterior margin at a high angle. A well-defined narrow ridge runs parallel to the lateral and
anterior margin of each pleura delimiting a narrow, steeply sloping border (PI. 71, fig. 3; text-fig. 1b). The
trunk was evidently flexible to some degree as indicated by the curvature in ROM 43505 (PI. 73, fig. 2; text-
fig. 3).

Evidence of the alimentary canal is limited. A dark linear trace, slightly concave in section, is preserved
along the axis of the posterior somites in ROM 43504. The relief trace terminates in the posterior margin of
the last trunk somite, presumably at the anus, but the dark trace expands in the anterior part of the telson (PI.

72, figs. 3 and 5; text-fig. 2b), probably representing seepage of organic material from the anus. Similar dark
stains occur at the posterior of other arthropods from the Burgess Shale. A relief trace in the eleventh somite
of ROM 43506 also extends just into the telson (PI. 73, fig. 5; text-fig. 5). Thus the anus lies at the posterior
margin of the eleventh segment, presumably beneath the insertion of the telson. The dark reflective material
in the axial area of the head shield of ROM 43506 may be the remains of the contents of the stomach (PI. 73,
fig. 5; text-fig. 5). A similar dark trace occurs in the head shield and first trunk segment in ROM 43503 (PI.

73, fig. 3; text-fig. 4).

Trunk appendages. Traces of the trunk appendages are evident on all the specimens, but only ROM 43504 (PI.
72, figs. 2-4; text-fig. 2b) preserves much detail of their morphology. Each segment, with the probable exception
of the last, bears a pair. The most obvious preserved structure of these appendages is a flat lamellate ramus.
Its outer anterior border is gently convex, the rest of the margin more strongly so and fringed with long setae.
These setae are also evident in ROM 43503 (PI. 73, fig. 3; text-fig. 4) as lineations on the overlapping lamellate
rami of the ventrally exposed posterior segments. The flat lamellate rami in ROM 43504 form a graded series,
with the largest at the front (PI. 72, figs. 2 and 3; text-fig. 2b). Their arrangement beneath the front trunk
segments can be seen in ROM 43502 (PI. 71, fig. 2; text-fig. Ib). In this specimen, the lamellate ramus of the
first right trunk segment appears to be much larger than the succeeding ones, but this disparity in size is not
evident in the other specimens. A second ramus, spinose and segmented, is evident in association with some
of the lamellae in ROM 43504 (PI. 72, figs. 2 and 4) and ROM 43503 (PI. 73, fig. 4), although the outline in
the latter is very faint. Attempts to reveal details of this ramus by removing parts of the overlying lamellae
have been moderately successful. The segmented ramus between left lamellae 1 and 2 of ROM 43504 reveals
the most detail (PI. 72, fig. 2). The articulations between the three or four distal podomeres are evident, and
the limb bears elongate spines on the posterior preserved margin.

The lamellae are preserved extending beyond the margin of the pleurae in ROM 43504, and overlapping
anteriorly (PI. 72, figs. 2 and 3; text-fig. 2b). In ROM 43506 (PI. 73, fig. 5; text-fig. 5) they overlap posteriorly.
This contrast is a function of the configuration of the limbs and the attitude of the specimens to bedding. A
similar difference in overlap occurs in the outer rami of the trunk limbs of the Burgess Shale crustacean,
Canadaspis perfecta , for example. In specimens in parallel or parallel-oblique aspect they are preserved
overlapping anteriorly; in lateral aspect they overlap posteriorly (compare Briggs 1978, figs. 83 and 1 16 with
figs. 1 1 1 and 1 1 5).

The outlines of both left and right lamellae in the first five trunk limbs of ROM 43504 (PI. 72, figs. 2 and
3; text-fig. 2b) form a graded series. The first is tilted at a relatively high angle to the bedding and is folded
and foreshortened (particularly evident on the right side). The following lamellae are oriented progressively
more nearly parallel to bedding and consequently their apparent size increases as foreshortening decreases.
However, this variation cannot be explained simply as the result of curvature of the trunk, with successive
lamellae maintaining the same attitude to the rest of the body (the trunk would have to curve dorsally to
account for the variation in this way). Instead, the attitude of the appendages may reflect successive positions
in the backward swing of the limb (cf. Briggs 1978, p. 463). Beyond appendage 5 only the left limbs are exposed
and the degree to which they extend beyond the pleurae is much reduced (PI. 72, fig. 3; text-fig. 2b). Their
outline appears to progressively narrow and this may represent foreshortening (perhaps as they are swung
forward in a recovery stroke). The trunk appendages of ROM 43506 (PI. 73, fig. 5; text-fig. 5) may also reflect
different positions during a metachronal swimming wave. The first pair are swung backward as if in a propulsive
stroke; the more posterior limbs are swung forward as in a recovery stroke.

The segmented rami that are evident in ROM 43504 (PI. 72, figs. 2 and 4) appear to belong with the lamellate
outer ramus lying immediately in front of them. This is suggested by the relative levels and preserved overlap
of the appendages. If the segmented ramus belonged with the lamella behind it, this would imply that the
lamellae lay posterior and adaxial of the segmented rami and were interleaved between them, which seems
unlikely on functional considerations. In life the segmented ramus would have been adaxial to the lamellate
ramus.

788

PALAEONTOLOGY, VOLUME 31

EXPLANATION OF PLATE 73

Figs. 1 5. Sanctacaris uncata gen. et sp. nov. 1 and 2, ROM 43505, dorsal view (text-fig. 3); 1, counterpart,
x 3-2, dry, head shield, appendages projecting beyond it, and eye, illuminated from the south-west; 2, part,
x 2, dry, showing reflective spots on head shield, and body flexure, illuminated from the east. 3 and 4, ROM
43503, ventral view of dorsal exoskeleton, counterpart (text-fig. 4), x 2; 3, immersed in water, showing setae
fringing lamellate ramus of trunk limbs which are adhering to dorsal exoskeleton; 4, dry, showing relief,
illuminated from the north-west. 5, ROM 43506, oblique view of right side, part (text-fig. 5), x 2-6, immersed
in water.

PLATE 73

BRIGGS and COLLINS, Sanctacaris

790

PALAEONTOLOGY, VOLUME 31

Telson. The telson widens posteriorly, the lateral margins roughly straight and paralleled by a ridge demarcating
a narrow sloping border, as on the pleurae (PI. 71, fig. 3; text-fig. 1b; PI. 73, fig. 4; text-fig. 4). The posterior
margin is convex posteriorly and fringed by very short spines or setae. The lateral aspect is unknown, but the
telson was presumably dorsoventrally flat in life, i.e. paddle-like.

Size. Size is difficult to assess accurately due to the effects of compaction at different orientations to the bedding
(the length of ROM 43502 (PI. 71, figs. 2 and 3), for example, is clearly reduced by foreshortening). ROM
43504 (PI. 72, fig. 3) is probably the largest individual and is about 93 mm long (measured along the curved
axis from the anterior border of the head shield to the posterior of the telson). ROM 43506 (PI. 73,
fig. 5) is probably the smallest with a length of 46 mm.

DISCUSSION

Generic assignment

The two genera to which Sanctacaris shows closest similarity are Alalcomenaeus Simonetta, 1970
and Actaeus Simonetta, 1970. The sole definitely assigned species of Alalcomenaeus , A. cambricus,
occurs at the same locality as S. uncata (locality 9 of Collins et al. 1983) and therefore the possibility
that the two are conspecific requires particular consideration. A. cambricus is rare in the Burgess
Shale (Whittington 1981) but a large collection from locality 9 is presently under study. This shows
that it has eleven trunk tergites following the head shield (Briggs and Robison 1984, p. 156) not ten
as reconstructed by Simonetta (1970) or twelve as reconstructed by Whittington (1981), and a flat
paddle-like telson (‘terminal plate’ of Whittington 1981), all characters shared with S. uncata. Hou
(1987) tentatively referred a new arthropod from the Lower Cambrian of Chengjiang, eastern
Yunnan, which appears to have twelve trunk tergites, to Alalcomenaeus (as A.? illecebrosus). Actaeus
armatus, which is based on a single poorly preserved specimen from the Burgess Shale (Whittington
1981), is similar in many ways to Alalcomenaeus cambricus.

Sanctacaris differs from Alalcomenaeus and Actaeus in a number of important respects. It has at
least six pairs of head appendages, of which the first five are raptorial and similar to each other.
Alalcomenaeus and Actaeus , in contrast, have only four pairs of head appendages. The first of both
is very distinctive, with a broad base and elongate distal extension that may be twofold; the remainder
of the head appendages are essentially similar to those of the trunk, whereas those in Sanctacaris
differ greatly from the trunk appendages.

The outline of the head shield is poorly displayed in the specimens of Alalcomenaeus and Actaeus
known from the Walcott quarry (Whittington 1981), but they preserve no evidence of the subtriangu-
lar lateral projections characteristic of Sanctacaris. This is borne out by dorsoventrally compacted
specimens of Alalcomenaeus from locality 9 on Mount Stephen (Collins et al. 1983) which show the
head shield to have a trapezoidal outline. The telson of Sanctacaris is relatively larger than that in
Alalcomenaeus. The outline of the telson in Actaeus is unknown.

Leanchoilia superlata is similar to S. uncata in possessing a head shield, followed by eleven trunk
tergites and a spinose telson. The head of Leanchoilia , however, bears a pair of great appendages
followed by two pairs of biramous limbs similar to those of the trunk (Bruton and Whittington
1983). This contrasts with the head of S. uncata which bears a series of at least six pairs of appendages
that are very different from those of the trunk.

The differences between S. uncata and previously described taxa are thus clearly sufficient to
warrant the erection of a new genus and species.

Functional morphology

The formidable array of inwardly facing raptorial limbs at the front leaves no doubt that Sanctacaris
was a predator. The arrangement of the raptorial limbs in a graded series with the smallest on the
inside and succeeding larger ones around and below, indicates that they functioned as a unit,
grasping prey below and to the front (i.e. Sanctacaris probably fed on bottom dwellers). It seems
likely that the raptorial limbs were equipped with gnathobases that aided in comminuting food and

BRIGGS AND COLLINS: CAMBRIAN CHELICERATE

791

pushing it into the mouth. It is possible that they were also ambulatory, like the walking legs on the
prosoma of eurypterids. The antenna-like outer ramus on each head appendage was probably
sensory; a sensory function was attributed to a similar ramus in the head appendages of Burgessia
(Hughes 1975). The short, spine-fringed ramus on the sixth appendage (PI. 71, fig. 1; text-fig. 1a)
probably had a sensory function, too, although it may also have been used in grooming.

The large flap-like rami of the trunk appendages and the paddle-shaped telson both indicate that
Sanctacaris was an active swimmer. The broad rami would have provided propulsion, moving in
metachronal rhythm, whereas the telson would have provided lift and steering in the vertical plane.
Steering in the horizontal plane would have been achieved mainly by differential movement of the
lamellate trunk limbs on either side of the body. The long fringing setae would have increased the
effective area of the rami in the propulsive stroke, and been folded back during the recovery stroke.
Presumably, movement of the broad rami through the water also helped in respiration. It is unlikely
that Sanctacaris used its telson for forward propulsion. This kind of swimming is unusual in
arthropods (see Lochhead 1961, for example) and Sanctacaris would have been inhibited in up and
down flexing of the trunk by the large tergites and the overlap between them. Furthermore,
Sanctacaris lacks the narrow trunk near the tail that is characteristic of animals using caudal fin
propulsion (Webb 1975, 1984; Plotnick and Baumiller (1988) apply similar arguments to the function
of the telson of pterygotid eurypterids). Conceivably, Sanctacaris could have moved rapidly back-
wards to escape from predators by flexing its trunk and tail as some crustaceans such as shrimps
do today. ROM 43505 shows evidence of ventral flexure (PI. 73, fig. 2; text-fig. 3). The segmented
ramus on the inside of the trunk appendages may have been ambulatory; if so, its distal spines
would have provided better footing on the bottom, and may also sometimes have helped in capturing
prey.

In addition to steering, the broad telson would have helped to stabilize Sanctacaris. As Lochhead
(1961) pointed out for crustaceans ‘the most usual method of controlling rotations around the
transverse axis is by the action of a flattened structure at the end of the abdomen’. The two large
triangular lateral areas on the head would also have provided stability.

AFFINITIES AND CLASSIFICATION

The discovery of Sanctacaris adds a further major type to the arthropods with preserved appendages
known from the Middle Cambrian of British Columbia. Most of these, however, cannot be classified
in any of the four major groups. Indeed, of the twenty genera with well-preserved appendages
described from the Burgess Shale, none is assigned to the chelicerates or uniramians, three are
assigned to the trilobites, one to the crustaceans, and sixteen are ‘not placed in any phylum or class
of Arthropoda’ (Whittington 1985, p. 138).

Compared to their modern counterparts, most Burgess Shale arthropods can be seen from their
morphology to be at the primitive end of the spectrum: a high proportion have ‘short’ heads with
few appendages (seven have three or fewer); in many the posterior head appendages are the same as
those of the trunk; about half have undifferentiated biramous trunk limbs. These Middle Cambrian
arthropods fall into three categories. First, some are not sufficiently advanced morphologically to
be included in any of the major groups; Marrella , for example, is primitive enough to have given
rise to any of them. Secondly, others like Yohoia and Branchiocaris have a body plan or specialized
appendages which exclude them from the trilobites and major living groups— indeed they are arguably
sufficiently advanced to represent separate taxa of equivalent rank. Thirdly, a small number can be
assigned to the living groups, although some of the features which characterize their Recent descen-
dants have yet to evolve. Canadaspis , for example, can be assigned to the crustaceans, even though
it retains some primitive features: the inner rami have a very large number of podomeres (up to
fourteen) and the posterior head appendages are little differentiated from those of the trunk (Briggs
1978, 1983). Sanctacaris likewise, although it can be assigned to a living group, the chelicerates,
retains some strikingly primitive features. Thus the Middle Cambrian arthropods comprise a much
wider morphological spectrum than the three distinct major groups of arthropods living today.

792

PALAEONTOLOGY, VOLUME 31

in a metachronal wave.

BRIGGS AND COLLINS: CAMBRIAN CHELICER ATE

793

The following characters indicate that Sanctacaris has a chelicerate affinity:

1 . At least six pairs of head appendages. This is more than any other arthropod group apart from
the chelicerates and Emeraldella. The majority of chelicerates have six pairs of prosomal (head)
appendages: a chelicera and five others. An additional appendage pair is present at the posterior of
the prosoma in the Devonian synziphosurine Weinbergina (Stiirmer and Bergstrom 1981), and this
appendage may be represented by the chilaria in Limulus and the metastoma in eurypterids. The
first five pairs of head appendages of Sanctacaris form a graded but otherwise undifferentiated series
and so could be equivalent to five pairs of prosomal appendages behind the chelicerae in later
chelicerates.

2. The nature of the head appendages. Functionally the first five head appendages were raptorial
and also possibly ambulatory. They increase in size toward the posterior and each podomere of the
raptorial rami bears distal spines. These features are characteristic of the prosomal appendages of
eurypterids (Plotnick 1983).

3. The presence of a cardiac lobe. This is represented by the dark reflective areas in the head
shield. Eldredge ( 1974) pointed out that a cardiac lobe is common to the merostomes (occurring in
eurypterids, xiphosurids, chasmataspids) although it also occurs in Aglaspida. The lobe ‘need not
be defined by sharply emplaced cardiac furrows, but instead may be distinguished by . . . simply
color’ (Eldredge 1974, p. 36).

4. The nature of the tagmosis. There is a clear morphological and functional separation of the
body into a head region specialized for catching prey, and a trunk with lamellate rami that presum-
ably served for swimming and respiration. Such a division is characteristic of the merostomes,
although it also occurs in other arthropods including the crustaceans (e.g. the Remipedia).

5. The position of the anus. The anus lies at the posterior margin of the last trunk segment,
ventrally below the insertion of the telson. This position is characteristic of the merostomes.

6. The nature of the telson. The telson is undivided and lacks associated appendages. This
morphology is characteristic of the merostomes, although comparable arrangements occur in other
arthropods.

Sanctacaris differs from chelicerates in two important respects. First, its limbs are biramous. A
biramous limb, however, has long been considered primitive for the chelicerates, not least on the
basis of the abdominal limbs of Limulus which are interpreted as comprising a short telopodite and
an expanded outer lobe bearing the gill lamellae (e.g. Stormer 1944, p. 69; Tiegs and Manton 1958,
p. 500). The outer rami in the prosoma (head) and inner rami in the opisthosoma (trunk) have been
lost in later chelicerates, apart from Limulus , which may be specialized (Schram 1978, p. 86). An
equivalent loss in Sanctacaris would result in tagmata with limbs specialized in a manner similar to
those in merostomes. Secondly, and more importantly , Sanctacaris preserves no evidence of chelicerae,
the chelate anteriormost limbs characteristic of the chelicerates. The primitive morphology of
chelicerae is unknown. They are not preserved in early eurypterids or xiphosurids. Studies of the
embryology of the living limulid Tachypleus indicate that the chelicera originates from a post-
cephalic lobe separate from the other limbs and is differentiated at a very early larval stage (Anderson
1973). It is therefore unlikely, but not impossible, that the anteriormost preserved raptorial limb in
Sanctacaris is equivalent to the chelicara.

In summary, the morphological characters present in Sanctacaris — number and nature of head
appendages, cardiac lobe, body tagmosis, position of the anus, nature of the telson— are all derived
for some member of the chelicerates and demonstrate the affinity of Sanctacaris to this major group.
Chelicerae are apparently absent, but this is perhaps not surprising considering the primitive
biramous nature of the appendages on both the head and trunk. Equally they may simply not be
preserved. The chelicerae of Megalograptus ohioensis, the earliest well-preserved eurypterid, are
revealed in detail on only two specimens although ‘several hundred specimens, a few of which are
essentially complete, have been found’ (Caster and Kjellesvig-Waering 1964, p. 301). Similarly, the
posterior head appendages of Canadaspis perfecta from the Burgess Shale are only clearly evident
in two or three of over 4000 specimens (Briggs 1978).

794

PALAEONTOLOGY, VOLUME 31

Is Sanctacaris a chelicerate? This is not the first time a question of this kind has been raised. Smith
(1984(7), for example, noted similar difficulties in incorporating the features of Palaeozoic echinoids
into a diagnosis of the Class Echinoidea. Diagnosing the Class is easy when only the living representa-
tives are considered. However, when the fossils are included, the Class Echinoidea ‘can only be
recognized on the basis of a unique combination of features which individually can be found in
other echinoderm groups’ (Smith 1984o, p. 158). Thus the absence of some characters diagnostic of
the Recent members of a higher taxon need not preclude more ancient representatives, provided
that these fossil forms display a combination of characters unique to that group. In this solution,
Sanctacaris is a chelicerate because its combination of characters occurs only in the chelicerates and
in no other group.

Should Sanctacaris be included in the subphylum Chelicerata? Here we are dealing with a question
of taxonomic practice rather than biological relationship. Fortunately there are precedents, including
one involving Burgess Shale echinoderms. Because it had ‘well-developed uniserial erect arms
(perhaps bearing tube feet)’. Sprinkle (1973, p. 178) could ‘see no alternative but to regard Echmato-
crinus as a true crinoid’ even though it lacks regular plating of the calyx and a columnal-bearing
stem, important characters present in all other Palaeozoic crinoids. Likewise, Sanctacaris can be
assigned to the Chelicerata because of its basic chelicerate morphology, such as the six pairs of
appendages (five raptorial) on the head, even though it apparently lacks chelicerae. The alternative
would be to erect a new taxon (e.g. the Protochelicerata) to include chelicerate-like arthropods
without chelicerae. This would not change our understanding of the biological relationships of
Sanctacaris or of the chelicerates. Moreover, it would create taxonomic problems in the future if
further material of Sanctacaris revealed chelicerae. Including Sanctacaris within the Chelicerata, on
the other hand, emphasizes its biological affinity. The diagnosis of the Chelicerata should therefore
be broadened to include biramous appendages and the possibility of a lack of chelicerae.

Smith ( 1 984/7, text-fig. 15) incorporated Echmatocrinus into the classification of the Crinoidea as
a plesion with generic rank (in the sense of Patterson and Rosen 1977), a primitive sister group of
all other crinoids. The biramous appendages (and lack of chelicerae, if real) define the position of
Sanctacaris as a primitive sister group of all other chelicerates and we likewise designate it a plesion.

EVOLUTIONARY SIGNIFICANCE

None of the previously described arthropods from the Stephen Formation preserves characters that
indicate as close a chelicerate affinity as Sanctacaris (Briggs and Whittington 1981; Briggs 1983,
1985). Bruton (1981) pointed out the striking similarity between the morphology of the trunk limbs
of Sidneyia and those of the living limulid Tachypleus. Further resemblance between Sidneyia and
the chelicerates is slight, however; it has a single pair of antennae and no other appendages in the
head.

Emeraldella has six pairs of head appendages. The first, however, is an antenna, and the others
are very similar to the trunk appendages, which differ only in the possession of an additional lobe
(Bruton and Whittington 1983). The only Burgess Shale arthropod showing the same degree of
differentation between head and trunk appendages as Sanctacaris is Yohoia (Whittington 1974), but
it has only four pairs of head appendages, the first of which is highly specialized.

The aglaspids were long considered to be chelicerates (Raasch 1939; Stormer 1955) based on the
interpretation of one appendage-bearing specimen from the late Cambrian of Wisconsin. This
specimen, assigned by Raasch to Aglaspis spinifer , was restudied by Briggs et al. (1979) who
demonstrated that it had only four or perhaps five pairs of appendages on the head, of which the
first could not be shown to be chelate, and the rest were walking legs like those on the front of the
trunk. Thus, Aglaspis cannot be recognized as a chelicerate.

Wahlman and Caster, in a 1978 abstract, reported a new Upper Cambrian chelicerate with
preserved appendages from the Hickory Sandstone of central Texas and suggested that it warranted
a subclass ‘on a par with the Xiphosura and Eurypterida’, as did the chasmataspids from the

BRIGGS AND COLLINS: CAMBRIAN CHELICERATE

795

Lower Ordovician of Tennessee (Caster and Brooks 1956). Further discussion of the affinities of
the Hickory Sandstone arthropod awaits the publication of a full description.

Fragments of possible limbs occur in association with Kodymirus from the Middle Cambrian of
Bohemia, assigned by Chlupac and Havlicek (1965) to the merostomes, but the morphology of the
appendages is unknown. The remaining Cambrian fossils described as chelicerates do not preserve
evidence of the appendages (Bergstrom 1968, 1975) and their affinities are therefore uncertain.
Sanctacaris is thus the only Cambrian chelicerate recognized at present.

The earliest known chelicerates with chelicerae are the eurypterids, which appear in the Ordovi-
cian. Sanctacaris shares a number of characters with ‘the primitive eurypterid' (Plotnick 1983).
These include subdued carapace relief with a cardiac lobe, similarity and increase in size of the
prosomal appendages toward the posterior, distal spines on the podomeres of these appendages,
and a trilobed opisthosoma. A paddle-shaped telson also occurs in some eurypterids, but is probably
derived from a primitive styliform morphology (Plotnick 1983, p. 206). In general morphology and
life habit (both were swimming benthic predators), therefore, Sanctacaris makes a fine progenitor
to the eurypterids. Indeed the raptorial head limbs in Sanctacaris are similar to the generalized
eurypterid appendage of the Hughmilleria type which Stormer (1974, text-figs. 1-10) used to derive
the different limbs of the eurypterid prosoma, including the swimming paddle (appendage VI). The
opisthosoma of eurypterids, on the other hand, is much more derived than that of Sanctacaris.
Twelve somites are divided into a pre-abdomen and a post-abdomen, and the abdominal appendages
are modified to form chambers that enclose gills on the ventral body wall. Such appendages, although
very different, could have been derived from the generalized Sanctacaris trunk appendages once the
swimming function was taken up by the eurypterid prosoma. There is no evidence that Sanctacaris
had the metastoma or genital appendages which are characteristic of eurypterids, but the posterior
appendages of the head may be their forerunners.

The characters that Sanctacaris shares with the eurypterids are for the most part primitive for
the xiphosurids as well (see Eldredge 1974, for a discussion of xiphosurid relationships). However,
Sanctacaris appears to be separated from the xiphosurids to a greater extent than from the eurypter-
ids. Thus the head of Sanctacaris is much smaller than the trunk, whereas the two are of nearly
equal length in xiphosurids. Sanctacaris shows no sign of any differentiation of the trunk into pre-
and post-abdominal sections and the first trunk somite is not reduced. Sanctacaris lacks a defined
interopthalmic area with ridges and furrows.

Sanctacaris demonstrates that chelicerates had evolved by the Middle Cambrian. Moreover, its
combination of primitive and diagnostic morphological characters places it near the origin of the
Chelicerata. On one hand, the presence of biramous appendages throughout and the lack of
differentiation in the trunk limbs indicate its position at the primitive end of the arthropod spectrum,
along with most of the Burgess Shale arthropods; on the other, the degree of ‘cephalization’, with
at least six pairs of appendages in the head, and the similarity of the five raptorial limbs to prosomal
appendages in ‘the primitive eurypterid’, indicate that it is none the less a chelicerate. This assessment
of Sanctacaris is paralleled by that of Canadaspis , which occurs at the same locality (Collins et al.
1983, table 1) and which is an early member of the other major living aquatic arthropod group, the
Crustacea (Briggs 1983). Canadaspis also has primitive morphological characters (large number of
podomeres in the head appendages; similar posterior head and anterior trunk appendages) but the
presence of two pairs of antennae, a mandible and a maxilla, for example, indicates that it is a
crustacean (Briggs 1978). Is it likely that chelicerae had developed by the early stage of chelicerate
evolution represented by Sanctacaris ? Indications are equivocal. On one hand, the primitive bi-
ramous nature of the head appendages suggests that chelicerae may not have yet evolved; on the
other, in view of the presence of highly modified head appendages in some Burgess Shale arthropods
( Yohoia, Branchiocaris , and Leanchoilia , for example) it is reasonable to presume that chelicerae
could have evolved by this time. However, whether or not chelicerae had evolved by the Middle
Cambrian, it is evident that Sanctacaris is a chelicerate.

From the five specimens known in the Glossopleura Zone and none in the Burgess Shale, it is
evident that chelicerates are rare in the Middle Cambrian rocks of British Columbia. This rarity is

796

PALAEONTOLOGY, VOLUME 31

probably a true reflection of their scarcity in Middle Cambrian seas, at least in the off-shore,
relatively deep water environments, because the tens of thousands of diverse exceptionally preserved
fossils in the Burgess Shale and Glossopleura Zone should provide a relatively complete sample of
the communities they represent. Why were chelicerates so rare then? Two reasons come to mind.
First, Middle Cambrian arthropods are very diverse morphologically and most do not fall into well-
defined groups (Briggs and Whittington 1981; Briggs 1983). Thus, except for the trilobites, no
arthropod group has many different representatives. Flowever, even by this minimal standard,
chelicerates are rare. Secondly then, the rarity of chelicerates may reflect their predatory way of life.
Even the most numerous predator in the Burgess Shale, Sidneyia , has only 177 individuals compared
to 15 092 Marrella , 4179 Canadaspis, and 2158 Burgessia (Conway Morris 1986), although the
numerical comparison is misleading because an adult Sidneyia is several times larger than Canadas-
pis', and many times larger than Marrella and Burgessia.

Why have the chelicerates been so successful while most of their Middle Cambrian arthropod
contemporaries have died out? Chance probably played a role, particularly during mass extinctions,
by sparing the chelicerates and eliminating the other arthropods. Alternatively chelicerates may
have some unique morphological characters that contributed to their success. The most obvious are
the large number and morphological flexibility of the head appendages, shown by the raptorial
limbs in Sanctacaris and the grasping, walking, balancing, and swimming prosomal appendages of
the eurypterids and xiphosurids. However, whether or not these were significant factors in the
success of the chelicerates, the basic pattern evident in Sanctacaris has persisted to the present while
those in most of the other Cambrian arthropods have long since disappeared.

Acknowledgements. Funding for the 1983 excavation, which yielded the material described in the paper, was
granted to D.C. by the Geological Survey of Canada (EMR Research Agreement 171), the Natural Sciences
and Engineering Research Council (Grant A8427), Summer Canada (for two summer interns), and the Royal
Ontario Museum. Assistance in the field was provided by Chen Jun-yuan of the Nanjing Institute of Geology
and Palaeontology, David Rudkin and Peter Fenton of the Royal Ontario Museum, Sean McFarland, Alex
Nikolajevich, and Kate and Matthew Collins. Parks Canada (Western Region) kindly gave permission to
excavate, and A1 Fisk, Gordon Rutherford, and particularly Eric Langshaw and Randy Robertson helped in
Yoho National Park. We are grateful to Steve Hesselbo, Harry Whittington, and Ed Bousfield for comments
on the manuscript. Photographs of the specimens were especially useful in this study and our thanks for these
go to Bill Robertson, Brian Boyle, and Allan McColl of the Royal Ontario Museum, Tom Easter of Goldsmiths’
College, London, and Simon Powell of the University of Bristol. Sophie Poray-Swinarski and Marianne
Collins of the Royal Ontario Museum drew the interpretive figures and reconstruction, respectively. John
Burke of the Royal Ontario Museum typed the manuscript. D. E. G. Briggs’s research was supported by the
Royal Society of London and the Royal Ontario Museum.

REFERENCES

aitken, j. d. and mcilreath, i. a. 1981. Depositional environments of the Cathedral Escarpment, near Field,
British Columbia. In aitken, j. d. and TAYLOR, m. e. (eds.). The Cambrian System in the southern Canadian
Rocky Mountains , Alberta and British Columbia , Second Int. Symp. Cambrian System , Guidebook for Field
Trip 2, 35-44. US geol. Surv., Denver, Colorado.

allison, p. a. 1986. Soft-bodied animals in the fossil record: the role of decay in fragmentation during transport.
Geology , 14, 979-981.

anderson, D. T. 1973. Embryology and phytogeny in annelids and arthropods , 495 pp. Pergamon, Oxford.
Bergstrom, j. 1968. Eolimulus , a Lower Cambrian xiphosurid from Sweden. Geol. For. Stockh. Fork. 90, 489-
503.

1975. Functional morphology and evolution of xiphosurids. Fossils and Strata , 4, 291-305.
briggs, d. e. G. 1978. The morphology, mode of life, and affinities of Canadaspis perfecta (Crustacea:
Phyllocarida), Middle Cambrian, Burgess Shale, British Columbia. Phil. Trans. R. Soc. B281, 439 487.

— 1981. The arthropod Odaraia a/ata Walcott, Middle Cambrian, Burgess Shale, British Columbia. Ibid.
B291, 541 584.

BRIGGS AND COLLINS: CAMBRIAN CHELICERATE

797

— 1983. Affinities and early evolution of the Crustacea: the evidence of the Cambrian fossils. In schram,
f. r. (ed. ). Crustacean Phylogeny. Crust. Issues , 1, 1-22. A. A. Balkema, Rotterdam.

1985. Les premiers arthropodes. La Recherche , 16, 340-349.

— bruton, d. l. and Whittington, h. b. 1979. Appendages of the arthropod Aglaspis spinifer (Upper
Cambrian, Wisconsin) and their significance. Palaeontology, 22, 167-180.

and robison, r. a. 1984. Exceptionally preserved nontrilobite arthropods and Anomalocaris from the
Middle Cambrian of Utah. Kansas Univ. Contribs. Paleont. Ill, 23 pp.

and Whittington, h. b. 1981. Relationships of arthropods from the Burgess Shale and other Cambrian
sequences. Proc. 2nd International Symposium on the Cambrian System. US geol. Surv. Open-file rept. 81-
743, 38-41.

and williams, s. H. 1981. The restoration of flattened fossils. Lethaia 14, 157-164.
bruton, d. l. 1981. The arthropod Sidneyia inexpectans , Middle Cambrian, Burgess Shale, British Columbia.
Phil. Trans. R. Soc. B295, 619-656.

-and Whittington, h. b. 1983. Emeraldella and Leanchoilia , two arthropods from the Burgess Shale,
Middle Cambrian, British Columbia. Ibid. B300, 553-585.
caster, k. e. and brooks, h. K. 1956. New fossils from the Canadian-Chazyan (Ordovician) hiatus in Tennessee.
Bull. Am. Paleont. 36, 157 199.

— and kjellesvig-waering, e. n. 1964. Upper Ordovician eurypterids of Ohio. Palaeontogr. am., 32, 301
358.

chlupac, i. and havlicek, v. 1965. Kodymirus n. g., a new aglaspid merostome of the Cambrian of Bohemia.

Shorn. Geol. Ved. Praha, 6, 7-20.

Collins, d. 1986. Paradise revisited. Rotunda , 19, 30-39.

— briggs, d. and conway morris, s. 1983. New Burgess Shale fossil sites reveal middle Cambrian faunal
complex. Science, NY, 222, 163 167.

conway morris, s. 1979. Middle Cambrian polychaetes from the Burgess Shale of British Columbia. Phil.
Trans. R. Soc. B285, 227-274.

— 1986. The community structure of the Middle Cambrian Phyllopod Bed (Burgess Shale). Palaeontology,
29, 423- 467.

eldredge, n. 1971. Patterns of cephalic musculature in the Phacopina (Trilobita) and their phylogenetic
significance. J. Paleont. 45, 52-67.

— 1974. Revision of the Suborder Synziphosurina (Chelicerata, Merostomata), with remarks on merostome
phylogeny. Am. Mus. Novit. 2543, 1-41.

fritz, w. h. 1971 . Geological setting of the Burgess Shale. In Proceedings of the North American Paleontological
Convention, Chicago, 1969, part I, 1 155-1170. Allen Press, Lawrence, Kansas.
hou, xian-guang, 1987. Two new arthropods from Lower Cambrian, Chengjiang, eastern Yunnan. Acta
Palaeont. Sinica, 26, 236-256.

hughes, c. p. 1975. Redescription of Burgessia betla from the Middle Cambrian Burgess Shale, British
Columbia. Fossils and Strata, 4, 415-435.

lochhead, j. h. 1961. Locomotion. In waterman, t. h. (ed.). The physiology of Crustacea, vol. 2, 313-364.
Academic Press, New York.

mcilreath, i. a. 1977. Accumulation of a Middle Cambrian, deep-water limestone debris apron adjacent to a
vertical, submarine carbonate escarpment, southern Rocky Mountains, Canada. Spec. Pubis Soc. econ.
Paleont. Miner. Tulsa, 25, 113-124.

patterson, c. and rosen, d. e. 1977. Review of ichthyodectiform and other Mesozoic teleost fishes and the
theory and practice of classifying fossils. Bull. Am. Mus. nat. Hist. 158, 81 172.
plotnick, R. E. 1983. Patterns in the evolution of eurypterids. Ph.D. thesis (unpublished). University of Chicago.

-and baumiller, t. 1988. The pterygotid telson as a biological rudder. Lethaia , 21, 13-27.
raasch, G. o. 1939. Cambrian Merostomata. Sp. Pap. geol. Soc. Am. 19, ix+ 146 pp.
schram, f. r. 1978. Arthropods: a convergent phenomenon. Fieldiana, Geol. 39, 61 108.
simonetta, a. m. 1970. Studies on non trilobite arthropods of the Burgess Shale (Middle Cambrian).
Palaeontogr. ital. 66 (ns 36), 35- 45.

smith, a. b. I984u. Echinoid Palaeobiology, 190 pp. Allen (George) & Unwin, London.

1984A Classification of the Echinodermata. Palaeontology, 27, 431 459.

sprinkle, j. 1973. Morphology and evolution of blastozoan echinoderms. Spec. Pub. Mus. Comp. Zool.
284 pp.

stormer, l. 1944. On the relationships and phylogeny of fossil and Recent Arachnomorpha. Skr. norske
Vidensk-Akad., Mat.-naturv. Kl. 5, 1 158.

798

PALAEONTOLOGY, VOLUME 31

stormer, L. 1955. Merostomata. In moore, r. c. (ed. ). Treatise on Invertebrate Paleontology. Part P. Arthropoda
2, P4-P41 . Geological Society of America and University of Kansas Press, Boulder, Colorado and Lawrence,
Kansas.

— 1974. Arthropods from the Lower Devonian (Lower Emsian) of Aiken an der Mosel, Germany. Part 4:
Eurypterida, Drepanopteridae, and other groups. Senck. leth. 54, 359-451.

STURMER, w. and Bergstrom, J. 1981. Weinbergina , a xiphosuran arthropod from the Devonian Hunsruck
Slate. Palaont. Z. 55, 237-255.

tiegs, o. w. and manton, s. m. 1958. The evolution of the Arthropoda. Bio. Rev. 33, 255-337.
wahlman, G. p. and caster, k. e. 1978 Bearing of new Texas Upper Cambrian arthropods on merostome
classification. Geol. Soc. Am. North-central section at Ann Arbor, Michigan. Abstracts with Programs, 10 (6),
286.

webb, p. w. 1975. Elydrodynamics and energetics of fish propulsion. Bull. Fish. Res. Bd Can. 190, 159 pp.
1984. Form and function in fish swimming. Sci. Am. 251, 58-68.

Whittington, h. b. 1971. The Burgess Shale: History of research and preservation of fossils. In Proceedings
of the North American Paleontological Convention , Chicago, 1969, part I, 1 170-1201. Allen Press, Lawrence,
Kansas.

— 1974. Yohoia Walcott and Plenocaris n. gen., arthropods from the Burgess Shale, Middle Cambrian,
British Columbia. Geol. Surv. Can. Bull. 231, 63 pp.

— 1980. The significance of the fauna of the Burgess Shale, British Columbia. Proc. geol. Ass. 91, 127-148.
1981. Rare arthropods from the Burgess Shale, Middle Cambrian, British Columbia. Phil. Trans. R. Soc.

B292, 329-357.

— 1985. The Burgess Shale, xiv + 151 pp. Yale University Press, New Haven and London.

DEREK E. G. BRIGGS

Department of Geology
University of Bristol
Wills Memorial Building
Queen’s Road
Bristol BS8 1RJ

DESMOND COLLINS

Department of Invertebrate Palaeontology
Royal Ontario Museum
Toronto

Ontario M5S 2C6

Typescript received 14 June 1987

Revised typescript received 23 September 1987

PATTERNS OF DIVERSIFICATION AND
EXTINCTION IN

EARLY PALAEOZOIC ECHINODERMS

by ANDREW B. SMITH

Abstract. It has been claimed that Cambrian and early Ordovician echinoderms show two phases of
diversification as recognized by Sepkoski for marine taxa in general. However, a more critical assessment of
the record, using cladistic analyses of all determinable echinoderm genera to identify sister group relationships,
allows sampling deficits to be taken into consideration. It is shown that the Upper Cambrian dip in diversity
amongst echinoderms is likely to be the result of preservation failure and that diversification more closely
approximates to a continuous process of expansion during the Cambrian and early Ordovician. Taxonomic
diversity cannot be used as a measure of morphological diversity because rank has been applied for a number
of different and incompatible reasons. There is no evidence to support the claim that morphological evolution
was occurring significantly faster during this period compared with later periods.

The publication of the Treatise on Invertebrate Paleontology during the last three decades has
provided a relatively comprehensive and authoritative data base summarizing the stratigraphical
ranges for genera in most major groups of marine invertebrates. This in turn has made it relatively
simple to carry out analyses of taxonomic ranges, and has stimulated the current interest in patterns
of evolution and changes in taxonomic diversity during the Phanerozoic. The recent interest in
patterns of evolution, as revealed by the available taxonomic data, dates back to the work of
Simpson (1953). Further interest in large scale patterns was created by the publication of Valentine’s
( 1 969) analysis developing Simpson’s thesis that different taxonomic levels showed different diversity
patterns when plotted through geological time. Valentine found that, whereas generic diversity
appears to have increased towards the present day, higher taxonomic groups reach their peak
diversity further back in time; the higher the categorical rank of the taxon, the earlier it appears
to have reached its maximum diversity. Thus phyla and classes were most numerous in the Lower
Palaeozoic and have declined since then (see Raup 1972 for a clear analysis of this phenomenon).
Valentine believed this pattern demonstrated that there was a rapid initial morphological
diversification in the Cambro-Ordovician which was later followed by a protracted phase of
competition in which the less successful groups were weeded out.

This general approach of using taxonomic data to interpret patterns of evolution has since been
refined and expanded upon by several workers, notably by Sepkoski (1978, 1978, 1981u, b , 1986),
Sepkoski and Raup (1986), and Raup and Sepkoski (1982, 1984), using a compilation of families
of marine invertebrates that is as up-to-date as possible (Sepkoski 1982, plus supplements) and an
as yet unpublished compendium of marine genera (see Sepkoski 1986). Various stimulating
hypotheses have been generated in the last few years on the basis of these data, the most notable
being the identification of cyclicity in extinction events with a periodicity of 26-28 million years
(my) (Raup and Sepkoski 1984; Sepkoski and Raup 1986). Similar sorts of data have also been
used to test a number of other hypotheses, including multiphase evolutionary diversification
(Sepkoski 1979), the Red Queen’s hypothesis (Van Valen 1973; Raup 1975), onshore-offshore
community replacement (Sepkoski and Sheehan 1983; Jablonski and Bottjer 1986), changing levels
of diversity within taxa through time (Flessa and Jablonski 1985). The ecological and genetic
implications for these patterns have been explored by Valentine (1980, 1986).

| Palaeontology, Vol. 31, Part 3, 1988, pp. 799-828. |

© The Palaeontological Association

800

PALAEONTOLOGY, VOLUME 31

1 0
0)
V)

w

CO

o

o

<S)

oO

E

3

z

text-fig. 1. Plot of number of echinoderm classes in the early Palaeozoic
(taken from Sprinkle 19806). Questionable occurrences unshaded.

The record of echinoderms has also been examined in detail by specialists interested in
documenting the pattern of evolution within this group (Paul 1977, 1979; Sprinkle 1980a, 6 , 1981,
1983) and has been shown to conform with that for invertebrate marine groups in general. Indeed,
the record of echinoderms appears to show an initial diversification during the Cambrian and early
Ordovician at class level which is more pronounced than in any other group. Paul (1979) was able
to demonstrate that diversity, as measured in genera per million years, increased to a maximum
in the Carboniferous then declined to a low at the Permo-Triassic boundary and then increased
once more to the present-day levels. But analysis of diversity at class level demonstrated ‘a clear
early radiation’ (Paul 1979, p. 417) with fifteen classes known from the Cambrian and nineteen
from the Ordovician after which the numbers gradually decline to the present level of five extant
classes. Sprinkle (19806, 1983) confirmed this view of echinoderm evolution and claimed that ‘no
new classes [of echinoderm] appeared in the fossil record after the Middle Ordovician’ (Sprinkle
1983, p. 5). Paul (1979) described this pattern as one of ‘colonisation-radiation/competition-
retrenchment’ and discussed the biological causes that could produce this.

Sprinkle (1981, p. 221) argued that ‘four of the five echinoderm subphyla probably crossed the
pre-Cambrian-Cambrian boundary’ and noted that ‘new classes appeared suddenly in the record
without obvious ancestors’. His (19806) analysis of timing of appearance of echinoderm classes
(text-fig. 1) showed that echinoderms appear to have undergone two or three phases of taxonomic
diversification, precisely as described for invertebrate metazoans by Sepkoski (1979, 19816).
Campbell and Marshall ( 1 986) have also analysed the record of echinoderms using higher taxonomic
groupings and concluded that there were two phases of morphological diversification, one in the
Lower to Middle Cambrian, the other in the Early Ordovician. They also believe that these classes
were morphologically highly distinct from their inception. They speculate that the echinoderm
genome might have been significantly different at this time to allow such jumps.

Thus analysis of the traditional taxonomic data base shows that echinoderms are not an aberrant
group, in so far as they show a similar pattern of taxonomic diversification to other marine
invertebrate groups and to the marine invertebrate biota as a whole. The question that I wish to
address in this paper is— how real is this view of the early evolutionary history of echinoderms?
Or, put another way, how much does it reflect taxonomic artefact? This question was first posed
by Derstler (1981), in a short but interesting paper, where he briefly outlined evidence suggesting
that much evolution in echinoderms occurred during the early Phanerozoic rather than in the
Precambrian, and that accounts of the early Phanerozoic diversification were probably based on
‘inappropriate assumptions about morphological change during evolution’ (Derstler 1981, p. 74).
In this paper I shall try to show how the assumptions and practice of taxonomists have indeed
resulted in a very misleading view of the early evolution of echinoderms.

SMITH: PATTERNS OF DIVERSIFICATION AND EXTINCTION

801

PROBLEMS IN ESTABLISHING TAXONOMIC DIVERSITY PATTERNS

1. Poor knowledge of the Cambrian fauna

Despite some really excellent and detailed work on selected primitive echinoderms (see for example
Ubaghs 1953, 1963 a , b\ Ubaghs and Robison 1985; Sprinkle 1973; Bell and Sprinkle 1978; Jell et
al. 1985) our understanding of a large number of forms remains sketchy and incomplete. In the
past there has been a tendency to describe new taxa based on very incomplete or badly preserved
material simply because it comes from the Cambrian and therefore ‘must be important’. Certainly
the Cambrian fauna is of interest because it provides us with direct evidence on the earliest and
most primitive echinoderms that existed. Yet all too often taxa are established on very scrappy
material. This has caused workers to misinterpret basic anatomical organization, with the result
that these taxa appear to have a morphology that is strikingly different from known echinoderms.
Thus Cymbionites and Peridionites were described as new classes of echinoderm and united together
in the subphylum Haplozoa by Whitehouse (1941), whereas they are now interpreted as fragments
(basal circlets) of ‘eocrinoids’ (Smith 1982). Camptostroma , another echinoderm known from only
a small number of distorted and badly preserved specimens, was initially described as a plated
jellyfish (Ruedemann 1933), then as a new class of echinoderm with plated tube feet (Durham
1966) and more recently as either eocrinoid (Broadhead 1980) or edrioasteroid (Derstler 1981) or
intermediate between eocrinoids and edrioasteroids (Paul and Smith 1984).

Leaving aside specimens that have been misinterpreted, there are also a number of Cambrian
species (and the genera, families, and sometimes orders established for them) which are based on
just one or a very few specimens and which remain incompletely known. There are only a handful
of specimens of the genus Echmatocrinus , and much of its anatomy remains unknown (we do not
even know how many arms it had), yet this is elevated to the rank of subclass (Sprinkle and Moore
1978). Some genera are so poorly known that they are for all intents and purposes unclassifiable.
For example, the genus Volchovia , on which the record of ophiocistioids is extended back to the
base of the Ordovician, is so incompletely known and shows so few features that it could just as
easily be a mitrate, or a Rhipidocystis- like eocrinoid. The record of holothuroids is extended back
to the Cambrian on even flimsier evidence. The supposed Cambrian record of holothuroids is
based on one body fossil, Eldonia , from the Burgess Shale, which lacks even a single echinoderm
character, and on microscopic spicules which are undiagnostic and could be juvenile elements of
almost any echinoderm.

Thus, because of the ‘mystique’ of the Cambrian, there are a disproportionately large number
of incompletely known taxa some of which have been elevated to high rank on the basis of
misinterpretation.

2. Sampling , preservation , and Lagerstdtten

Although some echinoderms such as echinoids have rigid skeletons that are readily preserved intact
in the fossil record, most do not. Primitive echinoderms had a membrane-embedded skeleton that
rapidly disarticulated upon death into individual plates to be scattered and lost. Whereas echinoderm
debris is often an important element in bioclastic sands and limestones, whole fossils are rare. It
requires special sedimentological conditions to preserve primitive echinoderms more or less
complete and intact, and these conditions can produce the so-called ‘starfish beds’ (see Goldring
and Stephenson 1972; Paul 1977). Sprinkle (1976u) has noted how rare echinoderm localities are
in the Cambrian, despite the apparent abundance of echinoderm plates within the sediments at
certain horizons. Occasionally a locality will yield only one or a few specimens of a single species
by chance, but it is more usual for echinoderm localities to yield sometimes large numbers of well-
preserved specimens of more than one species. These are deposits where conditions have been
favourable for preserving the echinoderm fauna and which can be referred to as echinoderm
Lagerstatten.

The distribution of echinoderm Lagerstatten will affect the apparent pattern of taxonomic
origination and extinction, since periods when conditions were favourable for the formation of

802

PALAEONTOLOGY, VOLUME 31

table 1 . Localities at which two or more taxa of echinoderm or carpoid have been collected

as articulated specimens.

Arenig

Schistes de St Chinian, Herault, France
Schistes de Landeyran, Herault, France
Gres de la Manerie, Herault, France
Anti-Atlas Mountains, Morocco
Asaphus Marl, Oslo, Norway
Whitland, Wales, UK
Kunvald Formation, Estonia
Ramsey Island, Wales, UK
Llangynog, Wales, UK
Upper Tremadocian

Fillmore Limestone, Nevada, USA
Anti-Atlas Mountains, Morocco
Schistes de St Chinian, Herault, France
Lower Tremadocian

Ubaghs (1983)

Ubaghs (1983)

Ubaghs (1983)

Chauvel (1966)

Bockelie (1984)

Fortey and Owens (1987)
Bassler and Moodey (1943)
Spencer (1918), Bates (1968)
Cope (1988)

Lane (1970), Paul (1972)
Chauvel (1966)

Ubaghs (1983)

Trempealeauian

Whipple Cave Formation, Nevada, USA
Franconian

Chatsworth Limestone, Queensland, Australia
Dresbachian

Upper Middle Cambrian

Secret Canyon Formation, Nevada, USA
Marjum Formation, Utah, USA
Jince Formation, Czechoslovakia
Porth-y-Rhaw beds, Pembroke, Wales, UK
Beds E, F, Ferrals-les-Montagnes, France
Median Middle Cambrian
Chisholm Shale, Nevada, USA
Lead Bell Shale, Idaho, USA
Spence Shale, Idaho/Utah, USA
Burgess Shale, British Columbia, Canada
Oelandicus Shales, Norrtrop, Sweden
Cateena Group, Tasmania
Lower Middle Cambrian

Beetle Creek Formation, Queensland, Australia
Lower Cambrian, Bonnia-Olenellus Zone
Upper Olenellus Beds, Newfoundland, Canada
Kinzers Formation, Pennsylvania, USA
Lower Cambrian, Nevadella Zone
Poleta Formation, Nevada, USA
Poleta Formation, California, USA

Sprinkle ( 1 976<rz)
Jell et al. (1985)

Sprinkle (1976a)

Ubaghs and Robison (1985)
Pompeckj (1896)

Jefferies et al. (1987)

Courtessole (1973), Ubaghs (1987)

Sprinkle (1976a)

Sprinkle (1976a)

Sprinkle (1976a)

Sprinkle (1976a)

Berg-Madsen (1986)

Jell et al. (1985)

Jell et al. (1985)

Smith (1986)

Derstler (1981)

Sprinkle (1976a)

Durham (1967)

echinoderm Lagerstatten will tend to show falsely higher rates of origination and extinction of
taxa than periods when conditions were less favourable and the fauna only patchily preserved. So
rates of origination and extinction will be artificially depressed during times of low preservation
potential brought about by factors such as major marine regression. Sprinkle (1981) was aware of
this problem and pointed out that, although the Upper Cambrian appears to show a drop in total
diversity, this may be at least partially due to non-preservation. Sprinkle pointed out that the
Upper Cambrian was a period in which broad carbonate shelves formed, unfavourable for the
preservation of echinoderms. Whereas echinoderm debris is often an important constituent of these
Upper Cambrian limestones, whole specimens are particularly rare.

SMITH: PATTERNS OF DIVERSIFICATION AND EXTINCTION

803

The rapid appearance of a number of groups at the base of the Ordovician may, therefore, be
more apparent than real; the product of our poor knowledge of Upper Cambrian echinoderm
faunas. Campbell and Marshall (1986), however, believed that it was not so, arguing that
morphological innovation really was concentrated in two distinct phases. They based their argument
on the observation that other groups (namely trilobites, brachiopods, and molluscs) continue to
diversify through the Upper Cambrian, which for them proved that the observed pattern of
taxonomic origination for echinoderms must be genuine. Whether the observed low diversity in
echinoderms during the Upper Cambrian was a genuine phenomenon or is a result of sampling
deficiency can, however, be tested.

Table 1 lists all localities from the Lower Cambrian through to the Arenig that have yielded
articulated specimens of two or more taxa of echinoderm or ‘carpoid’. (As all previous analyses
of echinoderm diversity patterns have treated ‘carpoids’ as echinoderms, here and throughout the
paper carpoid and echinoderm data have been combined so that results are directly comparable.
This does not imply that carpoids and echinoderms necessarily form a monophyletic group.) When
plotted as number of Lagerstatten per time interval (text-fig. 2) there appears to be a very close

804

PALAEONTOLOGY, VOLUME 31

Number of Lagerstatten

text-fig. 3. Regression analysis of number of echinoderm
Lagerstatten for each time interval plotted against number of
apparent family-level taxonomic originations (data as in text-
fig. 2).

match between the number of Lagerstatten and the number of families originating during each
period (data taken from Sepkoski (1982, plus supplements)). This is confirmed by regression
analysis of the number of echinoderm Lagerstatten in each time interval plotted against apparent
number of first appearances of families during that period (text-fig. 3).

Thus a null hypothesis that the observed pattern for family level originations does no more than
reflect the quality of sampling of the fossil record cannot be rejected. However, a more positive
approach to deciding whether the drop in taxa observed in the Upper Cambrian is a real event or a
reflection of the number of fossil Lagerstatten can be adopted. To distinguish between these two
options it is necessary to analyse for missing taxa (see Paul 1982). Taxa may disappear at the end
of a particular time interval because they have gone extinct, because they have suffered a change of
name or because of inadequate sampling: the latter two cases produce pseudoextinctions. Taxa
which must have been present but which have not yet been found (named Lazarus taxa by Jablonski,
in Flessa and Jablonski 1983), have been calculated using the generic data base
compiled for this paper (Table 2). The method by which missing taxa was recognized is detailed
below and requires a cladistic analysis of the taxa. A plot of known taxa plus Lazarus taxa
against time is shown in text-fig. 4. From this it is clear that the fossil record of Upper Cambrian
echinoderms is very poor indeed, with up to 80 % of the taxa known to have been

table 2. Analysis of Cambrian and early Ordovician echinoderm diversity, based on text-fig. 9.
Figures in brackets = fauna excluding 'carpoids’.

No. of genera
recorded

No. of Lazarus
genera

No. of
originations

Minimum % genera
not yet recorded

Arenig

43 (27)

6 (6)

39 (23)

12

Upper Tremadocian

8 (7)

16 (14)

5 (5)

67

Lower Tremadocian

1 (1)

16 (13)

12 (11)

94

Trempealeauian

5 (2)

13 (13)

7 (4)

72

Franconian

3 (3)

12 (10)

3 (3)

80

Dresbachian

3 (2)

10 (8)

3 (3)

77

Upper Middle

22 (12)

6 (5)

13 (7)

21

Median Middle

15 (11)

6 (4)

12 (7)

29

Lower Middle

8 (8)

4 (3)

7 (7)

33

Upper Bonnia-Olenellus

3 (3)

4 (3)

4 (4)

57

Lower Bonnia-Olenellus

5 (4)

0 (0)

4 (4)

—

Nevadella

2 (2)

1 (1)

3 (2)

33

Fallotaspis

0 (0)

1

1

—

SMITH: PATTERNS OF DIVERSIFICATION AND EXTINCTION

805

text-fig. 4. Histogram of number of recorded
genera (black) and ‘Lazarus’ taxa, those that are
known to have been present but have not yet been
discovered (stippled), in each time interval.

50-1

present still undiscovered (this is a minimum estimate). During the entire Upper Cambrian and
Tremadoc, Lazarus taxa outnumber known taxa by more than two to one. The only other early
Palaeozoic geological period where Lazarus taxa form more than 50 % of the ‘total’ fauna is during
the Upper Bonnia-Olenellus Zone where they reach 57 %. (A comparable loss of taxa attributable
to preservation failure was noted for Lower Silurian cystoids by Paul 1982.) Thus the apparent
drop in taxonomic origination during this period is unlikely to be a real phenomenon, but a
reflection of sampling. Any calculation of echinoderm patterns of standing diversity, origination,
and extinction must take into account this artificial drop in diversity brought about through
inadequate sampling of the fauna.

3. Taxonomic artefact

Whereas the poor knowledge of Cambrian echinoderms can be put right by further finds and more
detailed revision of the taxa, and the effect of fluctuating levels of sampling produced by Lagerstiitten
distribution can be taken into account, a much more fundamental question can be raised about
the comparability of the taxa analysed.

Prior to the advent of cladistics, the practice of taxonomy lacked any clear or agreed methodology
and proceeded in a rather haphazard manner. In effect, taxonomic decisions about how to group
species were arbitrary and basically authoritarian, and there was no objective criterion by which

806

PALAEONTOLOGY, VOLUME 31

to judge rival schemes. Echinoderm taxonomy during the last three decades has fared particularly
badly with taxonomists tending to stress differences between taxa while ignoring shared derived
characters, which form the basis for identifying relationships. Paraphyletic taxa which have some,
but not all, characters of an established higher taxon have tended to be separated off as a new
higher taxon (for example the creation of the Coronoidea by Brett et al. 1983). This has led to the
creation of a plethora of taxa of high rank. I have criticized this approach elsewhere (Smith 1984)
for side-stepping problems of relationships between taxa and for destroying the hierarchical nature
of the classificatory system.

The lack of a clearly stated method that can be applied consistently in traditional systematics
has resulted in a taxonomic data base that is riddled with inconsistency. Thus there are major
problems in extrapolating evolutionary patterns from this sort of data, two of which will be
explored below:

(i) Extinctions and pseudoextinctions. In traditional systematics there has been no clear distinction
made between monophyletic and paraphyletic groupings, and paraphyletic groupings abound in
the non-cladistic taxonomic data base (see Patterson and Smith 1987). Paraphyletic groupings,
being arbitrary units whose upper boundary is created by a taxonomist’s decision, tend to obscure
evolutionary patterns and extinction patterns. Whereas the disappearance of a clade from the fossil
record is a real event and represents the extinction of an evolutionary lineage, the disappearance
of a paraphyletic group reflects a taxonomist’s decision to change the group’s name, usually
because he/she thinks sufficient morphological change has taken place to merit this. Paraphyletic
groups are created when a taxonomist removes a derived portion of a clade, leaving an ’ancestral
group’ characterized only by its ‘primitive’ features.

Where a number of more specialized members are removed from a clade the remaining primitive
taxa are lumped together in a paraphyletic catch-all group. This is precisely what has happened
in the creation of class Eocrinoidea (see Smith 1984). The traditional group Cystoidea was split
during the 1960s into a number of classes based on apparently well-defined novel structures
associated with gaseous exchange (Diploporita, Rhombifera, Paracrinoidea, Blastoidea, Parablast-
oidea, Coronoidea). The remaining primitive members were then grouped together within the class
Eocrinoidea whose only uniting feature is their lack of those defining characters on which the
more derived classes are recognized. The group has no reality in biological or evolutionary terms
and probably contains ancestors to all of the more derived classes. Its extinction, and the extinction
of the various lineages of which it is comprised, are taxonomic artefacts.

The arbitrariness of decisions about where a paraphyletic group should end is well demonstrated
by the arguments surrounding the glyptocystitid rhombiferans. An analysis of the Cambrian fauna
shows that typical glyptocystitid rhombiferans such as Cheirocrinus can be traced back through a
sequence of ‘eocrinoid’ genera that includes Macrocystella, Ridersia , and probably to Cambrocrinus
(see cladogram, text-fig. 6). Traditionally it has been the appearance of respiratory structures called
pectinirhombs in the theca that has been used to define this group of rhombiferans (viz. Kesling
1967). These are present in Cheirocrinus but absent in the ‘eocrinoid’ members. However, Paul
(19686) argued that Macrocystella had all the characters associated with glyptocystitid rhombiferans
(including the characteristic large lateral periproct, unique stem construction, and identical thecal
plating) save for the pectinirhombs and should be classified as a glyptocystitid. Sprinkle (19766)
objected to this, arguing that pectinirhombs were the all-important character for the group. More
recently, Jell et al. (1985) described the genus Ridersia which has the unique stem morphology of
glyptocystitid rhombiferans and similar thecal plating but lacks the large lateral anal area, one of
the circlets of thecal plates found in Macrocystella and Cheirocrinus and the pectinirhombs of
Cheirocrinus. Jell et al. concluded that it was ancestral to the glyptocystitids but should remain
classified as an eocrinoid because of its lack of pectinirhombs. Clearly then there is a clade defined
by the shared presence of pectinirhombs and a more inclusive clade defined by the shared presence
of a large lateral anus and a still more inclusive clade that is defined by the shared presence of the
unique stem morphology. Whether this branch of ‘eocrinoids’ is made to go extinct by taxonomists

SMITH: PATTERNS OF DIVERSIFICATION AND EXTINCTION

807

at the first appearance of a distinctive stem morphology (probably with Cambrocrinus), or the first
appearance of a definitive thecal plating ( Ridersia ), or the first appearance of a lateral anus
(Macrocy Stella), or the first appearance of pectinirhombs ( Cheirocystella ) is purely arbitrary.
Furthermore, there is no reason why one character acquisition should be taken as a ‘class-
level jump’, while another is treated as a ‘genus-level’ or ‘family-level’ jump as has been well illus-
trated by Runnegar (1987) for primitive molluscs. Designating one step within a sequence as
marking the start of a ‘class’ only serves to mislead non-specialists as to the significance of such
an event.

For any analysis that claims to be looking at rates and timing of taxonomic origination and
extinction, it is absolutely crucial to have data which incorporate only real groups (i.e. monophyletic
taxa) since paraphyletic groupings can, and usually do, introduce a large amount of taxonomic
artefact into the analysis. In effect, this requires a cladistic data base.

(ii) Taxonomic rank. Whereas clades are real entities which have meaning in the biological world,
the taxonomic rank which is assigned to them is arbitrary. The designation of taxonomic rank has
in the past proceeded in a very haphazard way. Even Agassiz, who had a set of criteria for
taxonomic rank, noted (1868, p. 110) that there was ‘. . . difficulty ... in determining the natural
limits of such groups . . . for individual investigators differ greatly as to the degree of resemblance
existing between the members of many Families, and there is no kind of group which presents
greater diversity of circumspection in the classification of animals’. Despite over 100 years of
taxonomic endeavours this remains as true today as it ever was and traditional taxonomy has
made little advance in defining how rank is to be assigned.

Yet despite this, there has continued a general and largely unspoken belief that taxonomic
categories such as class, order, family represent approximately equivalent chunks of evolutionary
trees that can be analysed meaningfully. So for example, Valentine (1980, 1986) and Campbell and
Marshall (1986) are able to equate morphological distance to categorical rank and derive
evolutionary models based on ‘phylum-level’ and ‘class-level’ jumps without considering how these
ranks are defined in practice.

Whereas it is generally agreed that a species is recognized on the basis of a morphologically
homogeneous sample population that is demonstrably discrete from other sample populations,
there exist no rules or method by which higher ranks are designated. Thus taxonomists rarely
agree about the precise composition of a taxon. Echinoderms have suffered particularly at the
hands of traditional taxonomists and Paul (1979, p. 417) noted that ‘echinoderm workers, unlike
those of some other groups, have unashamedly created classes for many fundamentally different
“designs” of echinoderms that appear early in the fossil record irrespective of their size (number
of genera or species) or longevity’. Like most evolutionary taxonomists, Paul (1979, p. 427) argued
for morphological distinctiveness as the guiding criterion on which higher taxonomic categories
should be defined, not taxonomic size or longevity. Sprinkle (1983, p. 8) took a similar approach
in trying to assess the ‘real’ number of classes into which echinoderms should be divided. He
identified three factors of importance in determining whether high taxonomic rank should be given,
morphological distinctiveness, success (measured by diversity and longevity), and survival to the
present day.

Most recently, Campbell and Marshall (1986) have attempted to justify the use of taxonomic
rank as some measure of morphological distinctiveness, claiming, amongst other things, that classes
of echinoderm did not converge in morphology towards their time of origin. The analysis of
specific and generic data presented here demonstrates that characters are hierarchically arranged
and lends no support lo their claim (see discussion on glyptocystitid rhombiferans, above).

Evolutionary systematists have assigned categorical rank arbitrarily and for reasons which are
often unclear. Although morphological distinctiveness (perceived phenetic distance) is currently
the most popular criterion on which this judgement is made, this certainly is not the only reason
why high rank has been applied. In echinoderms a high taxonomic rank has been designated for
at least five different reasons, outlined below, which means that taxa of the same rank are unlikely
to be commensurate entities.

808

PALAEONTOLOGY, VOLUME 31

1. A high categorical rank may be given to a species or small group of species which , because of
poor preservation or misinterpretation of morphological structure , are not well understood. This
is equivalent to stating that the species are problematic. A number of echinoderm classes
proposed in the past fall into this category, and are usually monotypic and based on scrappy
material (e.g. ‘Camptostromatoida’, ‘Haplozoa’).

2. A group is given a high categorical rank because it has achieved considerable diversity through
time. Such groups have to be given relatively high taxonomic rank because of the hierarchical
nature of the Linnean classificatory system. All extant classes of echinoderm fall into this
category. Crinoidea is a large group that includes a very large number of species and genera
which are grouped into various higher categories. However, it is important to realize that
high taxonomic rank here does not imply that there was significant morphological divergence
at the initiation of the group (the differences between the crinoid Echmatocrinus and the
cystoid Lepidocystis are slight), only that significant levels of diversity were achieved at some
time after the establishment of the clade.

3. A high categorical rank may be given to a group which is neither morphologically diverse, nor
particularly distinctive morphologically, but which has as its closest relative a group which has
achieved high internal diversity. Thus, the subclass Echmatocrinea was created for the
monospecific genus Echmatocrinus , not because it was so very different from other primitive
echinoderms, but because it is supposed to have given rise to the crinoids. Since crinoids are
given high rank, a similar rank was given to Echmatocrinus. Here Echmatocrinus, known
from a handful of specimens from one locality and horizon, is the plesiomorphic sister group
to all other crinoids.

4. A high categorical rank may be given to well-defined groups on the grounds of perceived phenetic
distinction. Sprinkle (1981, p. 220) claimed that ‘new classes appeared suddenly in the record
without obvious ancestors’. Although morphological distinctiveness from the outset may be
genuine, it is more commonly the result of an incomplete and patchy fossil record, or
taxonomy in which differences are stressed and shared derived features ignored. An example
of this is the newly erected class Concentricycloidea (Baker et al. 1986), where a single species,
sister group to the highly derived asteroid family Caymanostellidae (Smith, in press) has been
established on the basis of its unusual ambulacral arrangement. This represents the misuse
of classificatory schemes to express an opinion, not to group species.

5. A high categorical rank is given to what remains after abstracting a number of monophyletic
groups from a larger group. The Cystoidea ( Blastozoa) are a large group which has subsequently
been split into a number of groups, most of which are defined on derived character states
(rhombiferans, diploporites, etc.). This has left a number of genera with only primitive
morphological characters that have been grouped together in the Eocrinoidea, largely as a
by-product of abstracting better-defined clades.

Because taxonomic rank can be assigned for such a variety of disparate reasons it seems
untenable to believe that precise categorical rank can convey any meaningful biological significance.
Groups may be given the same rank because of the diversity that they subsequently achieve,
because of the morphological distance (apparent or real) separating them from other groups,
because their nearest relatives subsequently achieved high diversity, or because they are poorly
understood or misinterpreted.

Before leaving the problems of rank, it is worth pointing out that Raup (1983) demonstrated
that it was an unavoidable product of tree topology that major groups appear early in the history
of a clade. Consequently, as the Linnean system of nomenclature is also hierarchical in nature and
high rank is inevitably given to major groups that have achieved high subsequent diversity, then
it must also be true that most of the high ranking groups appear early in the history of a clade.
This, however, certainly does not imply that these groups of high rank were separated by
large morphological distances from their inception. Valentine’s (1969, 1980) and Raup’s (1983)
observation that groups with high rank appear early in the Phanerozoic is thus a topological

SMITH: PATTERNS OF DIVERSIFICATION AND EXTINCTION

809

feature of clade diversification and needs no biological explanation. To erect models in which
taxonomic rank is equated with morphological distance at inception is misleading and the biological
explanation spurious.

AN ALTERNATIVE APPROACH

If meaningful analyses are to be carried out on the pattern of specific diversity, origination, and
extinction through time then the first step must be to construct a relatively artefact-free data base
using cladistic methodology.

As an example, I present here an analysis of the early history of the echinoderms from the start
of the Cambrian through to the Arenig. For comparison with the results based on the Sepkoski
data base (1982, plus supplements) I have compiled a cladistic analysis and stratigraphic range
chart for all known species of echinoderm from this interval. Because all previous analyses have
treated carpoids as echinoderms I also include a stratigraphical range chart for this group, for
comparable results. The same time-scale as used by Sepkoski (1979, 1982) is adopted here except
that I have tried to use approximately equal-length geological periods of between 5 million years
(using the time-scale presented in Conway Morris 1987) and 7-10 my (using the time-scale of
Harland et al. 1982) each. Thus, the Tremadocian ( 1 7 my) is split into upper and lower Tremadocian,
and the Atdabanian and Lenian (30 my on the Harland et al. time-scale) are split into the
Fallotaspis, N evade Ha, and lower and upper Bonnia-Olenellus Zones.

Although this has proved a difficult task, hindered by our current poor understanding of many
of the taxa, a reasonable attempt can be made. Certain taxa posed major difficulties and no doubt
some of the relationships suggested here will be proved wrong as new data on primitive echinoderms
come to light. Indeed, my analysis is at variance in a small number of details with an independent
analysis carried out by Paul (1988), and these details clearly need further investigation. But my
aim here is not to provide a definitive phylogeny of Cambrian echinoderms, only to show that this
sort of approach provides a more accurate method of assessing their evolutionary history. Hopefully
though, the analysis provides a reasonable approximation to the evolutionary relationships of
echinoderms during the early Palaeozoic. Every known species has another that is its closest
relative and cladistic analysis is the best method for identifying these relationships. Combining the
stratigraphic and cladistic data produces the best corroborated phylogenetic tree and places very
specific constraints on the occurrence of Lazarus taxa. Comparison of the resultant patterns from
these data with those derived from previous, less critical analyses occupies the last part of this
paper.

Although this paper will deal primarily with radiate echinoderms, it has been necessary to
consider the record of carpoids since they have a similar skeletal structure and are equally prone
to the vagaries of preservation. The aims of this paper are to examine the diversification pattern
of a major clade during the early Phanerozoic, so that whether carpoids are treated as primitive
echinoderms, as by Ubaghs (1971/), 1975), Sprinkle (1983), and Philip (1979) or as primitive
chordates (Jefferies 1986) is largely irrelevant for present purposes. If carpoids are treated as
echinoderms then the pattern is one of ‘phylum-level’ diversification (in traditional taxonomic
terms), if treated as chordates then it plots the initial diversification of a larger segment of the
animal kingdom.

Cladistic analysis of Lower Cambrian to Lower Ordovician taxa

There are some eighty-five species of echinoderm and carpoid known from the Cambrian and
Tremadocian and a further forty or so species in the Arenig (Table 3). The great majority of
Cambrian and Tremadocian genera are monospecific, so that, with the exception of Gogia, analysis
at generic level is more or less the same as a specific level analysis. Where a genus has more than
one species, relationships of individual species have been examined to check for the possibility of
paraphyly. In such cases, species or species clusters are treated separately, but where the species

810

PALAEONTOLOGY, VOLUME 31

table 3. Echinoderm and ‘carpoid’ taxa from the Cambrian and Lower Ordovician.
Species are numbered for reference to text-fig. 9.

Lower Cambrian, Nevadella Zone
(Poleta Lormation)

, Helicoplacus gi/berti Durham and Caster
H. curtisi Durham and Caster
1 H. everndeni Durham

Durham and Caster (1963)
Durham and Caster (1963)
Durham (1967)

2

H. fir by i Durham
H. nelsoni (Durham)
Polyplacus kilmeri Durham

Durham (1967)
Durham (1967)
Durham (1967)

Lower Cambrian, Bonnia-Olenellus Zone
(Kinzer Lormation)

3 Unnamed solute

Paul and Smith (1984)

4

Camptostroma roddyi Ruedemann

Paul and Smith (1984)

5

Kinzercystis durhami Sprinkle

Sprinkle (1973)

6

Lepidocystis wanneri Loerste

Sprinkle (1973)

7

stromatocystitid

Derstler (1981)

8

(Upper Olenellus Beds)
Stromatocvstites walcotti Schuchert

Smith (1986)

9

S. pentangularis Pompeckj

Smith (1986)

10

(Bristolia subzone)
Gogia ojenai Durham

Durham (1978)

Lower Middle Cambrian

(Plagiura- Poliella Zone, North America)
1 1 G. prolifica Walcott

Sprinkle (1973)

12

Gogia sp. 1

Sprinkle (1973)

13

(Albertella Zone, North America)
G. hobbsi Sprinkle

Sprinkle (1973)

14

(Coonigan Lormation, Australia)
Cambaster sp. (isolated plates)

Jell et al. (1985)

15

(Beetle Creek Lormation, Australia)

Edriodiscus primotica (Henderson and Shergold)

Jell et al. (1985)

16

IStromatocystites sp.

Jell et al. (1985)

17

(basal Middle Cambrian, north-eastern Australia)
Cymbionites craticula Whitehouse

Smith (1982)

18

Peridionites navicula Whitehouse

Smith (1982)

Median Middle Cambrian

(Glossopleura Zone, North America)

19

unnamed cothurnocystid

Sprinkle (1976u)

20

Ctenocystis utahensis Robison and Sprinkle

Sprinkle and Robison (1978)

21

Gogia palmeri Sprinkle

Sprinkle (1973)

22

G. granulosa Robison

Sprinkle (1973)

23

G. guntheri Sprinkle

Sprinkle (1973)

24

G. longidactylus (Walcott)

Sprinkle (1973)

25

G. multibrachiatus (Kirk)

Sprinkle (1973)

26

Totiglobus nimius Bell and Sprinkle
(Bathyuriscus Elrathina Zone, North America)

Bell and Sprinkle (1978)

(21)

G. palmeri Sprinkle

Sprinkle (1973)

27

G. kitchnerensis Sprinkle

Sprinkle (1973)

(22)

G. granulosa Robison

Sprinkle (1973)

(23)

G. guntheri Sprinkle

Sprinkle (1973)

28

1G. radiata Sprinkle

Sprinkle (1973)

SMITH: PATTERNS OF DIVERSIFICATION AND EXTINCTION

29 Echmatocrinus brachiatus Sprinkle

30 ?trachelocrinid arms

31 W alcottidiscus typicalis (Walcott)

(Cateena Group, Australia)

32 Cambraster tastudorum Jell, Burrett and Banks

33 Ctenocystis jagoi Jell, Burrett and Banks
(E. oelandicus Zone, Sweden)

34 Ceratocystis sp.

35 Cigar a sp.

(16) Stromatocystites sp.

Upper Middle Cambrian

(Bolaspidella Zone, North America)

36 Caslericystis vali Ubaghs and Robison

37 Marjumicystis mettae Ubaghs and Robison

38 Gogia spiralis Robison

39 Eustypocystis minor Sprinkle

40 Walcoltidiscus lloydi (Sprinkle)

(Jince Formation, Czechoslovakia)

41 Etoctenocystis bohemica Fatka and Kordule

42 Trochocystoides parvus Jaekel

43 Trochocystites bohemicus Barrande

44 Ceratocystis perneri Jaekel

45 Cigara dusli Barrande

46 Acanthocystites briareus Barrande

47 A. jani (Prokop)

48 Acathocystites nuntius (Prokop)

49 Luhocrinus monicae Prokop and Fatka

50 Lichenoides priscus Barrande

(9) Stromatocystites pentangularis Barrande
(Beds E, F, Montagne Noire, France)

51 Ctenocystis smithi Ubaghs

52 Ceratocystis vizcainoi Ubaghs

53 Gogia gondi Ubaghs

54 " Eocystites" languedocianus Ubaghs

55 Trochocystites theronensis (Cabibel et al.)

56 Gyrocystis barrandei (Munier-Chalmas and Bergeron)

57 Decacystis hispanicus Gislen

(two other genera of cinctan from here are of dubious status)

58 Cambraster cannati (Miquel)

59 Undescribed genus resembling Cambraster but with annular

aboral plating like a cyclocystoid
(Beds G, H, Montagne Noire, France)

60 Gyrocystis pardailhanicus (Termier and Termier)

(upper Paradoxides paradoxissimus Zone, St Davids, Wales,
UK)

61 Protocystites meneviensis (Flicks)

62 ctenocystoid

Middle Cambrian (undifferentiated)

(Pirineo, Spain)

(57) Decacystis hispanicus Gislen
(Atlas Mountains, Morocco)

(43) Trochocystites bohemicus Barrande

Upper Middle Cambrian or Lower Upper Cambrian
(Siberia, USSR)

63 Pareocrinus ljubzovi Yakovlev

Sprinkle (1973)
Sprinkle (1973)
Smith (1986)

Jell et al. ( 1985)

Jell et al. (1985)

Franzen, in Berg-Madsen (1986)
Franzen, in Berg-Madsen (1986)
Franzen (pers. comm. Jan. 1987)

Ubaghs and Robison (1985)
Ubaghs and Robison (1985)
Sprinkle (1973)

Sprinkle (1973)

Sprinkle (1985)

Fatka and Kordule (1985)
Ubaghs (19676)

Ubaghs (19676)

Jefferies (1969)

Ubaghs (1967a)

Fatka and Kordule (1984)
Prokop (1962)

Prokop (1962)

Prokop and Fatka (1985)
Ubaghs (1953)

Smith (1986)

Ubaghs (1987)

Ubaghs (1987)

Ubaghs (1987)

Ubaghs (1987)

Cabibel et al. (1958)

Ubaghs (19676)

Ubaghs (19676)

Smith (1986)

Termier and Termier (1973)

Jefferies et al. (1987)
Jefferies et al. (1987)

Melendez (1954)

Chauvel (1971a)

Ubaghs (1967a)

812

PALAEONTOLOGY, VOLUME 31

TABLE 3 (cont.)

Dresbachian

64

(Cedaria Zone, North America)
unnamed solute

Bell and Sprinkle (1980)

65

Nolichuckia casteri Sprinkle

Sprinkle (1973)

66

(' Olenus Beds, Holy Cross Mountains, Poland)
Cambrocrinus regularis Orlowski

Orlowski (1968)

Franconian

(Conaspis- Prosciukia Zones, North America)

67 Trachelocrinus resseri Ulrich Sprinkle ( 1973)

(Chatsworth Limestone, Peichiashania secunda/ Prochuangia glabella Zone, Australia)

68 Ridersia watsonae Jell, Burrett and Banks Jell et al. (1985)

69 unnamed isorophid Jell et al. (1985)

Trempealeauian

(Whipple Cave Formation, USA)

70 Minervaecystis sp.

71 Nevadaecystis americana Ubaghs

72 cornute

73 possible rhombiferan
(Montana, USA)

74 hybocrinid-like crinoid

Ubaghs (19636)
Ubaghs (19636)
Ubaghs (19636)
Paul (1968a)

Derstler (1981)

Tremadocian

(Lower and Upper: Wales, UK)

75 Macrocystella mariae Callaway
(Lower: Herault, France)

(75) Macrocystella sp.

(Czechoslovakia)

(75) M. Ibavarica (Barrande)

(Australia)

(75) Macrocystella sp.

(Upper: Fillmore Limestone, USA)

76 Cheirocystella antiqua Paul

77 'Hybocrinus' sp.

78 Pogonipocrinus antiquus Kelly and Ausich
(Upper: Anti-Atlas Mountains, Morocco)

79 lAristocystites sp.

80 Palaeosphaeronites sp.

(75) Macrocystella bohemica Barrande
(75) M. tasseftensis Chauvel
(75) M. cf. mariae Callaway

8 1 Rhopalocystis destombesi Ubaghs
(Uppermost Tremadocian/basal Arenig: H

82 Aethocrinus moorei Ubaghs

83 Minervaecystis vidali Ubaghs

Paul (19686, 1984)

Ubaghs (1983)

Ubaghs (1983)

Jell et al. (1985)

Paul (1972)

Lane (1970)

Kelly and Ausich (1978)

Chauvel (1966)

Chauvel (1966)

Chauvel (1969)

Chauvel (1969)

Chauvel (1969)

Ubaghs (1963#), Chauvel (19716)

, France)

Ubaghs (1969a, 19726)

Ubaghs (19696)

Arenig

(Basal: Schistes de St Chinian, Herault, France)

(83) Macrocystella vidali Ubaghs

84 Phyllocystis blayaci Thoral

(84) P. crassimarginata Thoral

85 Cothurnocystis fellinensis Ubaghs

(85) C. courtessolei Ubaghs

86 Chauvelicystis spinosa Ubaghs

Ubaghs (19696)
Ubaghs (19696)
Ubaghs ( 19696)
Ubaghs ( 19696)
Ubaghs (19696)
Ubaghs (1983)

SMITH: PATTERNS OF DIVERSIFICATION AND EXTINCTION

8

87

Thoralicystis griffei (Ubaghs)

Ubaghs (19696)

88

Amygdalotheca griffei (Ubaghs)

Ubaghs (19696)

89

Galliaecystis lignieresi Ubaghs

Ubaghs (19696, 1983)

90

Chinianocarpos thorali Ubaghs

Ubaghs (19696)

91

Peltocystis cornuta Thoral

Ubaghs ( 19696)

92

Balantiocystis thorali Ubaghs

Ubaghs (1972a, 1983)

93

Chinianaster levyi Spencer

Spencer (1951 )

94

Villebrunaster thorali Spencer

Spencer ( 1951 )

95

Pradesura jacobi Thoral

Spencer ( 1951 )

(75)

Macrocystella azaisi (Thoral)

Paul (19686)

96

‘ Hemicystis' boehmi Thoral

(Lower: Schistes de la Maurerie, Herault, France)

Thoral (1935)

(85)

Cothurnocystis primaeva Thoral

Ubaghs (19696)

(93)

Chinianaster levyi Spencer

Spencer ( 1951 )

(95)

Pradesura jacobi Thoral

Spencer ( 1951 )

(75)

Macrocystella azaizi (Thoral)

Paul (19686)

(76)

Cheirocystella languedociana (Thoral)
(Lower: Gres du Foulon. Herault, France)

Paul (19686, 1972)

97

Lingulocystis elongate i Thoral

(Lower: Schistes du Landeyran, Herault, France)

Ubaghs (1960)

(92)

Balantiocystis sp.

Ubaghs (1983)

(85)

Cothurnocystis melchiori Ubaghs

Ubaghs ( 1983)

98

Ramseyocrinus vizcainoi Ubaghs
(Lower: Ramsey Island, Wales, UK)

Ubaghs (1983)

(98)

R. cambriensis (Hicks)

Bates (1968)

99

Petraster ramseyensis (Hicks)

(Lower: Anti-Atlas Mountains, Morocco)

Spencer (1918)

(86)

Chauvelicystis ubaghsi (Chauvel)

Chauvel ( 1 97 1 <7)

(84)

Phyllocystis sp.

Chauvel (197 In)

(92)

Balantiocystis regnelli Chauvel

Chauvel (197 In)

(87)

Thoralicystis zagoraensis
(Lower: Llangynog, Wales, UK)

Chauvel (1971a)

100

Blastoidocrinus antecedens Paul and Cope

Paul and Cope (1982)

(98)

Ramseyocrinus sp.
(Greenland)

Cope (1988)

101

Compagicrinus fenestratus Jobson and Paul
(Upper: Estonia, USSR)

Jobson and Paul (1979)

102

Glyptosphaerites leuchtenbergi (Volborth)

Jaekel (1899)

103

‘ Cheirocrinus ’ giganteus (Leuchtenberg)

Paul (1972)

104

Cheirocystis radiatus (Jaekel)

Paul (1972)

105

Blastocystis rossica Jaekel

Jaekel (1918)

106

Echinosphaerites aurantium (Gyllenhahl)

Bockelie (19816)

107

Echinoencrinites angulosus (Pander)

Bassler and Moodey (1943)

108

Cryptocrinites similis Bockelie

Bockelie (1981a)

109

Rhipidocystis sp.

Bockelie (1981a)

110

111

Asteroblastus sublaevis Jaekel
Bolboporites spp.

(Upper: Sweden)

Jaekel (1899)

112

Sphaeronites pomum Eichwald

Paul and Bockelie (1983)

(112)

S. minor Paul and Bockelie

(Upper Asaphus Marls, Oslo, Norway)

Paul and Bockelie (1983)

113

IHemicosmites sp.

Bockelie (1979a)

114

IBockia sp.

Bockelie (1981a)

115

Volchovia norvegica (Regnell)

Regnell (1945)

814

PALAEONTOLOGY, VOLUME 31

TABLE 3 ( cont .)

(Upper Arenig, Whitland, Wales, UK)
(85) Cothurnocystis sp.

1 1 6 Reticulocarpos sp.

117 Lagynocystis sp.

118 Balanocystites sp.

1 19 Guichenocarpos sp.

120 Anatifopsis sp.

121 Mitrocystites sp.

122 Mitrocystella sp.

(other localities)

123 Protocrinites sp.

124 Monocycloides oelandicus Berg-Madsen

125 Perritocrinus transitor (Beyrich)

Jefferies, in Fortey and Owens (1987)
Jefferies, in Fortey and Owens (1987)
Jefferies, in Fortey and Owens (1987)
Jefferies, in Fortey and Owens (1987)
Jefferies, in Fortey and Owens (1987)
Jefferies, in Fortey and Owens (1987)
Jefferies, in Fortey and Owens (1987)
Jefferies, in Fortey and Owens (1987)

Bockelie (1984)

Berg-Madsen (1987)

Ubaghs (197 In)

are differentiated on only minor variations and appear as a polychotomy when analysed cladisti-
cally, (i.e. they are, so far as our resolution allows us to determine, all equally related to their
sister group) then the genus is treated as a terminal taxon. This cladistic analysis has been carried
out only for radiate echinoderms.

The genus Gogia has been subdivided into a number of species or species clusters as follows: (i)
the G. spiralis group ( G . ojenai , G. granulosa , G. guntheri , and G. spiralis ), for species with spiral
brachioles. The spiralling of the brachioles is a shared derived feature and is found in no other
echinoderm; (ii) the G. prolifica group (G. prolifica and G. palmeri ), for species with extensively
developed epispires forming prominent external grooves on plate margins; (iii) the G. hobbsi group
(G. hobbsi and G. gondi ), species in which the holdfast is considerably reduced in size; (iv) G.
multibrachialis, a species in which there is no apparent holdfast differentiated; (v) the G. kitchnerensis
group ( Gogia sp. 1 of Sprinkle 1973, G. longidactylus, and G, kitchnerensis), for species in which
the epispires are greatly reduced and confined to the oral area of the theca. G. radiata Sprinkle
appears to represent yet another group, but it is so poorly known that its assignment to a genus
is impossible. However, from what little we do know of this species, it closely resembles Eocystites
languedocianus (Ubaghs 1987) and the two have been grouped together.

Previous cladistic analyses of Cambrian echinoderms or carpoids are few. Paul and Smith (1984)
produced a cladogram for Lower Cambrian taxa, Jefferies (1986) has published an analysis of
mitrate and cornute carpoids, and Smith (1986) has published an analysis of eleutherozoan taxa.
A cladistic analysis for blastozoan echinoderms is in press (Paul 1988), as is one for primitive
crinoids (Donovan 1988). But nowhere previously have all Cambrian taxa been drawn together
before in such an analysis. No cladistic analysis is attempted here for Cincta or Ctenocystoida;
much work on these groups remains to be done.

Text-figs. 5 to 8 present the character analyses on which the phylogenetic groupings have been
based. Individual character states are listed below and discussed where necessary. Table 3 provides
a list of all known taxa, with reference to the most informative description available.

Discussion of characters

1. Skeletal histology composed of stereom. (Secondarily lost in some more derived groups— synaptid
holothurians, ?chordates.)

2. Larval development asymmetrical. This is inferred in fossil forms from the presence of a single
asymmetrically positioned hydropore/gonopore (solid circles), as opposed to the paired openings in
henrichordates.

3. Radial symmetry. This is most prominently displayed in the arrangement of the radial water vessels
around the peristome. It is three-fold in hclicoplacoids and primitively five-fold in more derived echinoderms,
although there is a great deal of variation in later forms.

SMITH: PATTERNS OF DIVERSIFICATION AND EXTINCTION

815

text-fig. 5. Cladogram of early Cambrian echinodernrs. For a discussion of
characters 1 I 1 see text.

4. Ambulacra integrated into body wall. In helicoplacoids and in most crown group echinoderms the
ambulacra form an integrated pari of the main body of the animal. In crinoids and in some cystoid groups,
however, the distal portion of ambulacra extend free of the body as arms (see character 9). In solutes the
solitary ambulacrum is a free appendage and does nol form part of the body wall.

5. Differentiation of an aboral and ora! plated surface. In Helicoplacus the entire body is composed of
spirally arranged rows of plates. Polyplacus appears to show a zone of non-spiral plating which may coincide
with the oral area, although the anatomy of this genus is still largely unknown. In all other primitive
echinoderms there are well-defined oral and aboral plated surfaces that differ in their organization. Again
this feature is variably developed in some later cystoid groups.

6. Pentaradial symmetry. Present in all primitive crown group echinoderms (see Paul and Smith 1984) but
not uncommonly modified in more derived groups (see from example Bockelie 1982). This is expressed as a
2:1:2 pattern in the arrangement of ambulacra around the mouth.

7. Mouth and anus situated close together at the thecal summit. In most carpoids, and in larval echinoderms,
the mouth and anus are at opposite poles of the body. Helicoplacoids have a laterally positioned mouth and
may or may not have a terminally positioned anus. In other primitive echinoderms the mouth or anus has
rotated so that the two openings lie close together on the oral surface. This is true of primitive eleutherozoan
echinoderms but not for some derived groups, such as echinoids and some asteroids.

8. Epispires on oral surface. A derived character by outgroup comparison.

9. Brachioles. Paul and Smith (1984) have argued that brachiolcs are derived from ambulacral cover plate
series. As Sprinkle (1973) has previously pointed out, they cannot be considered as homologous with arms.

10. Ambulacra extend free of the theca as arms. Here the arms of fistuliporite cystoids, coronates,
aristocystitids, crinoids, and some primitive ‘eocrinoids' (e.g. Nolichuckia and Trachelocrinus) are treated as
homologous structures derived from extension of the ambulacra outside the theca as free appendages.

1 I. Loss of aboral holdfast. In primitive eleutherozoan echinoderms the aboral and oral surfaces are of
similar extent and there is no attachment holdfast. In Camptostroma the aboral surface is conical and shows
evidence of having had spirally arranged musculature (Paul and Smith 1984).

12. Food gathering appendages composed of brachioles only. Food is gathered via appendages which may
be composed of brachioles arising from ambulacra on the theca, or may incorporate brachiole-bearing
ambulacra that extend free of the theca.

13. Aboral surface extended into a stalk. The absence of a stalk in some cystoids is treated here as a
secondary loss, on the strength of other characters.

14. Cup plating is dearly differentiated from holdfast plating. Only in some of the most primitive members
is there little differentiation between the plating of the stalk and the cup.

15. Stalk supported by holomeric columnals. The recent discovery that Acanthocystites has holomeric
columnals draws into question the distinction between this genus and Akadocrinus. Furthermore, the newly
created genus Luhocrinus may also turn out to be a juvenile form of Acanthocystites.

816

PALAEONTOLOGY, VOLUME 31

V)

Q)

C

O

c

0)

o

c

■c

o

Uj

nj

v. >

O O

o o

<D d)

•C -C

o o

to

<D

£ nj

n .i:
5 cc

to

3

C

h.

o

o

•o

S

03

o

3 —

D) Q.

.s «

-J 0:

to

0) *-
*- C

■- o

a>

03

•c

a

to

o

Q.

3

a> <d

03

03 Q.

a. co

a

D

O

o>

03

0J O)
o o

o

co

a*

03 -c
O o

o 3

a

■D

O

(O 3

'(O 2

c cn

0)

to

CL) to

■C 03 ■'

£ .t: to

•t: o. ^

•3c to O

u.

03 03 <D

tt » C
O O —

O O *

o o o o o

o o o o o

0*000
o o o o o

A — B O O O

o o o ~ o o

O O O O o

22 • 0 O------- ----------

21 *-■■ »■-* OOOOOOOOOOOOOOOOO
20 • — • — * — • 00-- — - -- -- -- - - --
19 •— • — * — * — « 700000000000000

18 * — * — *—*—*—* 0000 — — o — — — — — —

17 *- * * * — • — ?— • O ■— H OOOO — — - — — -

16 * — * — • — * — * — * — * — •- O -• — * — * — « 0 0 0 0 3 0 0

15 • — t — « — * — * — » — • — * — « — •— 7 - -*000000

14 * » * * * * * * « * • - - * * * - * « o

13 * — • — ® * « * * 0 — * — • • - - 0 0 • - 0 • o

12 0 0—0 * O— ? -• *— * * • • #— * • • # 0 0—0

29

O

0

o

o

o

o

0

0

0

o o • — <§>

o

o

28

0

0

o

0

0

?

0

0

0

O • • <£>

0

0

27

0

0

0

0

0

?

0

0

0

o • — « — •

0

o

26

0

0

0

0

o

?

0

0

A — A — B = B=B

o

0

25

o

0

0

o

o

0

A — A— A—

a/b-b=b=b

0

0

24

0

0

0

0

0

0

% — •

0

OOOO

o

0

23

o

o

0

o

0

0

o

0

text-fig. 6. Cladogram of selected Cambrian
to Lower Ordovician cystoid groups. For a
discussion of characters 12-29 see text.

16. Loss of epispires. Camptostroma and primitive pelmatozoans such as Kinzercystis and Gogia all have
well-developed epispires over their oral surface. In the G. kitchnerensis group epispires are greatly reduced
in size and extent. In other groups the epispires are either lost or have been replaced by more sophisticated
respiratory structures.

17. Basal circlet fused ( solid squares) or composed of four basals (solid circles).

18. Xenomorphic stem. In Macrocystella , Ridersia, and glyptocystitids there is a very pronounced difference
between the proximal and distal parts of the stem. The same appears to be true of Cambrocrinus judging
from published photographs. Where known, the proximal portion of the stem has an extremely large lumen
and columnals are arranged alternately as an inner and outer series with synarthrial articulation.

19. Cup plating organized into discrete circlets with BB, ILL , LL recognizable.

20. RR circlet of plates developed: anus lateral , lying between ILL , LL , and one radial plate.

2 1 . Dichopores developed.

22. Dichopores disjunct.

23. Brachioliferous plates present. In some genera the brachioles are attached to a single thecal plate which
has the attachment facet, in others the brachioles are attached to two thecal plates and the attachment facet
lies across a plate suture. The single brachioliferous plate is treated here as the derived condition. Ambulacral
plating is not differentiated in either Palaeosphaeronites or Sphaeronites, the brachioles arising from facets
on thecal plates.

24. Theca flattened with well-developed marginal frame. Traditionally Lingulocystis and Rhipidocystis have
always been treated as closely related because of their similar body form, although their brachiole structure
differs somewhat (see Ubaghs 1960; Bockelie 1981a).

25. Stem reduced (A) or lost (B). In Protocrinites some species have a reduced stem, others have no stem
(Bockelie 1984).

SMITH: PATTERNS OF DIVERSIFICATION AND EXTINCTION

817

text-fig. 7. Cladogram of selected Cambrian to Lower Ordovician
pelmatozoan groups. For a discussion of characters 30-48 see text.

26. Respiratory pits formed: (A) internal pits or (B) diplopores which penetrate almost the entire plate
thickness. Sphaeronites, Palaeosphcieronites , and Gylptosphaerites all have diplopores that perforate the thecal
wall. Protocrinites has sealed perforations (Bockelie 1984) probably formed by resorption from the interior.
Rhopalocystis has sutural epispires but the interior of plates appears to be similarly covered in deep pits
comparable to those in Protocrinites.

27. Epithecal food grooves. In many diploporite cystoids the brachioles are connected to the mouth by
shallow epithecal grooves rather than discrete ambulacral grooves with recognizable ambulacra. This is
treated as a derived state.

28. Mouth covered by a palate of oral plates.

29. Attached directly to the substratum. In Sphaeronites and Palaeosphcieronites there is no stem and the
base of the theca is moulded to fit the substratum.

30. Aboral surface extended into a stalk (as character 13).

31. Arms extend free of the theca (as character 10). In most cases it is clear that it is the ambulacra that
extend extra-thecally to produce a filtration fan. It is not yet certain whether the subvective system in
Marjumicystis is ambulacral, brachiolar, or a mixture. Similarly, the fact that in Gogia kitchnerensis there is
a ?coelomic pore running through the biserial ‘brachioles’ (Sprinkle 1973) might suggest that these are
ambulacral extensions not brachioles. However, the same structure has now been observed in G. gondi
(Ubaghs 1987).

818

PALAEONTOLOGY, VOLUME 31

32. Arms uniserial.

33. Clip composed of organized circlets of plates.

34. Stem clearly differentiated from the cup.

35. Arms attaching to a single brachial-bearing plate (as character 23). In crinoids each arm is attached to
a radial plate. The arms in some other groups are also attached to a single plate, not shared between adjacent
flooring plates, and this is treated as a derived character.

36. Stem ossicles meric. The ossicles of the stem are unorganized in Gogia spp. but become organized into
vertical rows of stout ossicles in primitive crinoids and in fistuliporite cystoids. Nolichuckia has a stem that
appears to show semi-organized rows of stout, brick-like ossicles very similar to those of fistuliporite cystoids,
judging from photographs in Sprinkle (1973, pi. 29, fig. 4).

37. Free arms branch. Primitively the free ambulacra appear to be unbranched, but in some crinoids the
arms branch dichotomously at least once.

38. Ana! sac present. Hybocrinids lack an anal sac, as does Echmatocrinus, but other primitive crinoids all
have a well-developed anal sac.

39. Cup composed of three or more organized circlets of plates. Whether the monocyclic arrangement of
plating, as seen in hybocrinids, or the dicyclic arrangement, as seen in Cupulocrinus , is the more primitive
arrangement is unknown. Aethocrinus differs from Cupulocrinus and Compagicrinus in having a fourth circlet
of cup plates while Ramsayocrinus appears to have either one or two circlets in its cup. This character
separates Aethocrinus , Compagicrinus , and Cupulocrinus from Ramsayocrinus and Hybocrinus , but may turn
out to be symplesiomorphic.

40. Cup composed of infrahasals, basals , and radials. Aethocrinus differs from the very similar Compagicrinus
and Cupulocrinus in having a fourth circlet of plates incorporated into the cup. Jobson and Paul (1979) have
argued that the condition seen in Aethocrinus is the more primitive.

41. Epispires lost (as character 16). The open structure of the anal sac in the crinoid tegmen is interpreted
as homologous and derived from the condition of having sutural epispires scattered over the oral surface. If
Lane (1984) is correct in interpreting the anal sac as housing the gonads then its sutural pores serve a
comparable function.

42. Oral area produced into a spout-like structure. Here the adoralmost plates are modified into a spout-
like structure from which the free arms extend. Nolichuckia probably has such a spout but the only known
specimen does not show the structure of this area.

43. Stem supported by holomeric columnals. Unlike the holomeric columnals of Akadocrinus and glyptocystitid
rhombiferans, these columnals are disc-like with only a small central lumen.

44. Free arms bearing brachioles. These are the so-called pinnate arms. Eustypocystis and Balantiocystis are
so similar that I have treated them as synonymous. They have simple arms without brachioles. Bockia is
almost identical to Balantiocystis in body form but differs in having brachioles developed on the free arms.
Trachelocrinus also has ‘pinnate’ arms. The arms of Hemicosmites are unknown but Bockelie (1979a) assumed
that they are pinnate from the occurrence of pinnate arms in the very closely related Caryocrinites (see
Sprinkle 1975). The ambulacral structure in Blastoidocrinus is comparable to that of Bockia and more derived
members of this clade (eublastoids), and parablastoids are interpreted here as having secondarily recumbent
‘pinnate’ arms.

45. Anus positioned laterally , well outside the food gathering area. In primitive crinoids and cystoids such
as Gogia , the anus lies close to the mouth within the area of the subvective filtration fan. In some more
derived cystoids, however, the anus has shifted to a lateral position well outside the oral area. The position
of the periproct is unknown in Blastoidocrinus , but has been assumed to be near the apex of the test by
comparison with the better known Meristoschisma (Sprinkle 1973).

46. Theca with three basals. These are not of equal size, there are two large and one small basal plates.
Trachelocrinus , which is known from one specimen, shows three basals in profile and is thus likely to have
either four or five basals. The number of basals in Blastoidocrinus , or for that matter in any parablastoid is
unknown, but has been assumed to be five.

47. Dichopore-type respiratory structures with internal thecal folds. Hemicosmitids have traditionally been
placed with glyptocystitid rhombiferans into the larger group Rhombifera, because of the similarity of their
dichopore-type respiratory structures, which straddle plate sutures and form diamond-shaped regions of
thecal folding for gaseous exchange (Paul 1968c). Thin-walled zones of thecal folds also occur in blastoids
(where they also straddle plate sutures) and parablastoids (where they are confined to the deltoid plates: the
so-called cataspires). However, major differences distinguish hemicosmitids (with their three-fold oral plating
symmetry) and parablastoids (with their five-fold symmetry) and the presence of dichopore-type of respiratory
structures of uncertain homology is not a strong character. Both hemicosmitids and parablastoids were left

SMITH: PATTERNS OF DIVERSIFICATION AND EXTINCTION

819

unplaced in the analysis of Paul ( 1988) and are here tentatively placed as sister group to Bockia , Cryptocrinites ,
and their relatives, the eublastoids. These two taxa are the most difficult to place.

48. Theca attached directly to the substratum by a rosette-like attachment disc. Ubaghs and Robison (1985)
described the attachment rosette of Marjumicystis and a similar structure is seen on aristocystitids. (Possibly
the same as in sphaeronitid cystoids.)

49. Aborcd surface flat, composed of tesselate plating. All of these echinoderms differ from pelmatozoans in
lacking extensive development of the aboral surface into a holdfast. Camptostroma has a short aboral holdfast
with spiral contraction zones and represents an intermediate condition.

50. Stout ring of marginal ossicles between oral and aboral plated surfaces.

51 . Aboral surface much reduced in area compared with the oral surface.

text-fig. 8. Cladogram of primitive eleutherozoan
groups. For a discussion of characters 49-63 see text.

o

CO

CL)

3

£

3

*3

.

£

(0

o

O

a>

CO

3

CO

o

■0

<D

3

CO

•3

O

3

3

CO

O

3

■Q

o

£

CD

CO

^3

O

O

CO

>

o>

fO

c

3

3

■3

Q.

O

w

Q.

5

0)

■o

0

0

•Q

O

•2

•Q

6

o

s

E

O

c

a>

<3

o

3

■3

3

0

•3

o

CO

K

o

Lu

O

O

a

0

Q.

630

o

o

» m

O

O

0

0

0

O

62 0

o

o

o

o

0

?

0

-

0

0

61 O

o

o

o

o

0

0

0

O

O

O

60?

o

0

?

o

0

0

?

-

590

o

o

o

o

0

0

6

O

— •

58 0

o

o

o

o

?

0

9 -

-9-9 -

57 0

o

o

0

0

9

0

9

O

- 9 «==•

56 0

o

o

o

o

6

0

•

-

0

V

9

•

55 0

0

o

o

o

0

0

0

6

6

0

54 O

o

o

0

0

0

0

0

0

0

O

530

0

o

o

0

0

0

O

52?

o

•

?

•

?

0

O

-

0

0

0

0

51 6

o

• •

o

6

0

O

0

0

0

0

O

50 O

o

o

0

0

O

49 O

52. Single large interradial ossicle forming the mouth frame. These plates were interpreted by Smith (1986)
as composed of fused ambulacral plates.

53. Peripheral skirt of plates present outside margined ring.

54. Aboral plates with a central perforation.

55. Margined ossicles specialized with an inner crest and an outer cupule zone.

56. Arms extend free of the disc. In Cambraster and an undescribed species from the Middle Cambrian of
the Montaigne Noire, the arms extended slightly beyond the marginal ring (see Jell et ed. 1985).

57. Madreporite developed. The hydropore is developed into a discrete calcified body.

58. Loss of anus. The presence of an anus is difficult to detect in some fossils, but does genuinely appear
to be absent in primitive asteroids and ophiuroids.

59. Stellate body form: vagile , living mouth downwards. Precisely when an oral face downwards posture was
adopted is impossible to say but it is here taken to coincide with the loss of the oral anus.

60. Mouth angle plates articulated and no longer forming a fixed frame.

61. Virgalia developed. Adjacent to ambulacra in somasteroids there are series of aligned interambulacral
plates known as virgalia. These are only very feebly developed in Archegonaster.

62. Radial water vessel interned.

63. Tesselate oral plating without epispires.

820

PALAEONTOLOGY, VOLUME 31

DISCUSSION

1. Apparent and real diversity patterns

Taxonomic diversity is usually calculated by simply counting the number of taxa of equivalent
rank present at each time interval. Using this method the pattern observed from the generic data
compiled here (text-fig. 4), closely matches that obtained by using standard taxonomic data at
family level (see text-fig. 2) and class level (text-fig. 1). All three sets of data show a rise in diversity
which reaches a peak in the Middle Cambrian and a second, larger rise in the Lower Ordovician.
The two peaks are separated by a distinct trough in the Upper Cambrian. Clearly then the pattern
of taxonomic origination seen at family and class level provides a reasonable approximation to
sampled species diversity (since the great majority of genera in the Cambrian are monospecific).
However, this is not necessarily a real pattern, since we know that there is a very poor fossil record
of echinoderms and carpoids in the Upper Cambrian. Using the cladistic analysis, it is possible to
make some compensation for the vagaries of the fossil record. Missing taxa can be identified in
two ways:

(i) Where the primitive sister group predates and is separated by a stratigraphical gap from the
derived sister group , then at the very least there must have been one taxon that has not yet been
found which existed between the last record of the primitive sister group and the first record of
the derived sister group. This gap could be filled by extension of the range of the primitive sister
group upwards, by extension of the range of the most primitive member of the derived sister group
downwards, or by interpolation of one or more as yet unknown taxa that are intermediate in form.
Furthermore, if the primitive sister group is not directly ancestral to the derived sister group
(something that cannot be determined from the cladogram), then the range of the missing taxon
may extend below the last appearance of the primitive sister group. Thus extension of the primitive
sister group’s range gives the absolute minimum interpolation of missing taxa.

(ii) Where the earliest member of the derived sister group stratigraphically predates the earliest
record of the primitive sister group , then the range of the primitive sister group must extend down
to the level at which the derived sister group first appears. Again this represents only the absolute
minimum interpolation of taxa.

By using these two criteria, ranges of Cambrian to Arenig taxa known to have existed but which
have not yet been discovered (i.e. Lazarus taxa) can be interpolated into the data set to compensate
for the poor fossil record. In text-fig. 9 known occurrences of taxa are shown in solid lines, and
minimum inferred missing taxa as dashed lines. Clearly the proportion of missing taxa increases
greatly during the Upper Cambrian (text-fig. 4; Table 2) showing that this is indeed a period for
which sampling is exceedingly poor in comparison with either the Middle Cambrian or the Arenig.
A plot of estimated diversity (combining taxa both described and Lazarus taxa as yet undiscovered)
still shows a small dip in the Upper Cambrian, though nowhere near as large as one based only
on recorded diversity (text-fig. 4). Because only the absolute minimum number of taxa present can
be determined, rate of origination at intervals where Lazarus taxa are known to be more numerous
than sampled taxa is likely to be significantly underestimated. Generic diversity through the
Cambrian has therefore been plotted using only those time periods which appear reasonably well
sampled (text-fig. 10). This suggests that a more realistic interpretation of the data is of continuous
exponential growth during the Cambrian and Lower Ordovician.

The number of extinctions identified from non-cladistic taxonomic data differs significantly from
the number calculated from the data presented here. This is because a taxon may disappear from
the record because of: (i) biological extinction or (ii) pseudoextinction. Traditional (non-cladistic)
taxonomic data bases have not distinguished between these two very different events (extinction
and morphological divergence) whereas a cladistic data base can provide a minimum estimate of
genuine extinctions, as follows.

A branch of the cladogram with two or more species (i.e. united by an autopomorphy) that
disappears from the stratigraphical record can be assumed to be an extinction event. A branch

SMITH: PATTERNS OF DIVERSIFICATION AND EXTINCTION

821

Trempea-

lauian

Dresbach-

fan

Bonnia

Olenellu;

Zone

Fallotaspis

Zone

07 120 122 104

8587119 121 107 103

: I 1 | 848688899091 | § | | 8 S

3| I 1 I I I 8 I 8

■i>

el I” I51
42LJ43

P:

Uj

45|
34 35|

109 11012 102 106 113 105 108 101 128 124 99 95

IvUMI If LjLM ! | I!i««

1 1 1 i*

1

el, 1 2? Li25:

el ?|63

(I6..49 53U50 37|

u

L

1 10 1 1

? ?

8

i-

1 •:

H

*i

I

i" I

[I

text-fig. 9. Stratigraphical distribution of all published Cambrian to Arenig (Lower Ordovician) echinoderms.
Known ranges are shown as heavy black lines; interpolated ranges as dotted lines; phylogenetic relationships,
derived from the character analysis presented here are indicated by fine lines. Table 3 lists all occurrences
plotted here and provides the key to species, which are numbered 1 125 on this diagram. Broken vertical

line separates ‘carpoids’ from radiate echinoderms.

text-fig. 10. Plot of generic diversity for each time
interval. Only those intervals in which estimated number
of Lazarus taxa forms less than 50 % of the total data
(solid dots) are used to construct the diversity curve.
Open circles represent known diversity in time periods
where Lazarus taxa form more than 50 % of the calcu-
lated total diversity and which are likely to underestimate
real diversity considerably due to poor sampling.

s

822

PALAEONTOLOGY, VOLUME 31

with only a single species may be produced by having a taxon that is ancestral to its derived sister
group or a taxon that forms an evolutionary side branch but which shares a common ancestor
with its derived sister group. Thus single species branches cannot be assumed to have gone extinct
unless they are demonstrably derived themselves or post-date the derived sister group. In practice,
a genuine extinction event is accepted where it affects a multitaxon branch on the cladogram or
where it affects a single taxon with a unique autapomorphy or where a plesiomorphic sister species
is stratigraphically younger than its derived sister group. This will provide a minimum estimate of
genuine lineage terminations.

Whereas Sepkoski’s family-level compendium (1982, plus supplements) recognizes thirty extinc-
tion events during this period, mostly concentrated at the end of the Middle Cambrian, the generic-
level analysis here suggests that genuine extinctions are relatively rare events. For echinoderms,
there is one obvious extinction of the G. spiralis group at the end of the Middle Cambrian and a
possible second of the G. kitchnerensis group at about the same time. Helicoplacoids, which were
apparently quite diverse during the Lower Cambrian probably represent another extinction.
Without a cladistic analysis for all carpoids, it is impossible to state how many extinction events
there have been in the early history of this group. However, it seems likely that there were at least
two, one terminating the ctenocystoid clade and another terminating the cinctan clade (or a branch
within the group if it is paraphyletic).

Thus genuine extinctions are rather few during the Cambrian diversification of echinoderms.
Although the limited extinctions seem to be restricted to the median or upper Middle Cambrian,
it is not certain that this pattern is correct, because of the poor fossil record from the Upper
Cambrian. In summary, the evidence presented here shows that for echinoderms the Cambrian
was a period of exponential increase in taxonomic diversity and low extinction rate. The very
different picture that emerges from analysing taxonomic categories at family and/or class level is
largely artefact.

2. Multiphase models of taxonomic diversification

Sepkoski (1979, 1981fi), on the basis of his compilation of non-cladistic taxa at family level,
proposed that marine metazoan diversification during the Palaeozoic occurred in two phases. An
initial phase of diversification took place in the Cambrian to produce the ‘Cambrian fauna’,
followed by a second phase of diversification during the Ordovician to produce the ‘Palaeozoic
fauna’. A third phase was later postulated to produce the ‘modern fauna’.

As demonstrated above, the apparent peak in taxon origination in the Middle Cambrian, the
decline in the Upper Cambrian, and the second peak of origination in the Lower Ordovician are
purely artificial for echinoderms and reflect a poor Upper Cambrian record. Whether sampling is
also the cause of this pattern in other taxonomic groups remains to be tested.

Furthermore, a number of the groups Sepkoski included within his ‘Cambrian fauna’ are
paraphyletic. Sepkoski described the fauna as being dominated by trilobites, hyolithids, eocrinoids,
inarticulate brachiopods, and monoplacophoran molluscs. The last three of these are demonstrably
paraphyletic (though including a number of good clades) and, as shown here for eocrinoids, must
contain a number of lineages that are ancestral to later, more derived groups. The relationships
of Ordovician trilobite families, most of them true clades, to the Cambrian trilobite families is a
matter of contention. Many Ordovician families appear de novo above the Ordovician boundary
but their Cambrian sister taxa have not yet been identified (R. A. Fortey, pers. comm.). The
implication is that some of the Upper Cambrian families are paraphyletic— hence even trilobite
extinction at the Cambro-Ordovician boundary is partially a taxonomic artefact. This is borne out
by the recent analysis of trilobite family extinctions at this boundary (Briggs et al. 1987), where
more than 50 % of family disappearances are attributed to pseudoextinction.

It is hardly surprising that paraphyletic groups such as eocrinoids ‘go into decline’ after the
Cambrian, since taxonomists have pruned off all the successful post-Cambrian lineages originating
from these groups and placed them into other taxa. The decline of these elements of the ‘Cambrian
fauna’ is thus no more than taxonomic artefact.

SMITH: PATTERNS OF DIVERSIFICATION AND EXTINCTION

823

Sepkoski (1979) searched for a biological reason for this apparent two-phase pattern of
diversification, and suggested that the Cambrian radiation favoured the appearance of generalist
forms, whereas the Ordovician radiation produced more specialized forms that outcompeted the
Cambrian fauna. However, an alternative explanation is that it is the product of taxonomists
creating paraphyletic Cambrian groupings which are terminated (at arbitrary points) by the
abstraction of monophyletic groups.

3. Rank , morphological distance, and macroevolution

Several workers (Paul 1979; Valentine 1980; Sprinkle 1983; Campbell and Marshall 1986) have
put forward the idea that there was something rather different going on in evolutionary terms
during the Cambrian with the appearance of so many high level taxa (phyla, classes). Valentine
(1980) explained this in terms of vacant ecological space availability, suggesting that it is easier to
make ‘phylum-level or class-level jumps' during the early radiation of metazoans while ecological
space was relatively empty. Clearly, these workers believe that morphological innovation was
proceeding much faster and by many fewer steps than later in the Phanerozoic. Hence Paul (1979,
p. 417) was able to claim that ‘virtually all echinoderm evolution was over by the end of the
Ordovician’.

The evidence for rates of evolution in the production of marine invertebrate phyla is difficult to
study because much of the morphological diversification must have gone on prior to the evolution
of skeletal systems and we thus have no fossil record. In echinoderms, however, we do have a
fossil record with which to assess morphological distance in the origination of ‘classes’. Evidence
provided in this paper for the first 60-90 million years of echinoderm diversification does not
support claims of major macroevolutionary jumps in the creation of ‘classes’. Were echinoderms
to have gone extinct at the end of the Arenig, it is doubtful whether many of the classes recognized
today by traditional taxonomists would have been created. The same conclusion has been reached
by Runnegar (1986) for early Palaeozoic molluscs. Furthermore, the nested pattern of character
distribution identified in cladograms such as text-fig. 6 suggests that diversification was more
gradual and stepwise than has previously been recognized.

The arbitrariness with which ‘classes’ have been recognized in the past has been discussed above.
Echinoderm taxonomists have been inconsistent when it comes to designating rank. To mention
just a couple of examples, the evolution of uniserial arms is seen as the primary character that
separates Echmatocrinus from blastozoans and places it in the subphylum Crinozoa (Sprinkle 1973,
19766). Yet some gomphocystitid cystoids have also evolved uniserial ambulacra (Bockelie 19796),
and Rhipidocystis, which is not even separated at family level (Sprinkle 1973), has uniserial
‘pinnules’. The class Coronoidea was erected for a group of cystoids (Blastozoa) with erect, pinnate
arms (Brett et al. 1983) yet both Bockia and Trachelocrinus with almost identical pinnate arms are
left as genera within the Eocrinoidea. Taxonomic rank in non-cladistic data has been applied for
such non-commensurate reasons that it seems unlikely that any biologically meaningful results can
come from analysis that uses such data purporting to measure morphological distance.

Acknowledgements. 1 should like to thank Dr C. R. C. Paul for letting me read a draft typescript of his
phylogenetic analysis of cystoid groups and for assistance and encouragement during the development of
this work. Dr C. Patterson, Dr C. R. C. Paul, and Dr R. P. S. Jefferies provided helpful criticism of an
earlier draft of this paper.

Note added in proof. The Upper Arenig Al Rose Formation of California has been omitted from the list of
echinoderm Lagerstatten in error. Ausich (1986) has described two crinoids from there, Proexenocrinus
inyoensis Strimple and McGinnis and Inyocrinus strimplei Ausich.

824

PALAEONTOLOGY, VOLUME 31

REFERENCES

agassiz, l. 1868. Methods of study in natural history, 319 pp. Ticknor and Fields, Boston.
ausich, w. i. 1986. The crinoids of the A1 Rose Formation (Early Ordovician, Inyo County, California,
USA) Alcheringa , 10, 217 224.

baker, a. N., rowe, f. w. e. and Clark, h. e. s. 1986. A newclass of Echinodermata from New Caledonia.
Nature, Lond. 321, 862-864.

bassler, r. s. and moodey, m. w. 1943. Bibliographic and faunal index of Paleozoic pelmatozoan echinoderms.
Spec. Pap. geol. Soc. Am. 45, 734 pp.

bates, d. e. b. 1968. On ‘ Dendrocrinus' cambriensis Flicks, the earliest known crinoid. Palaeontology , 11, 406-
409.

bell, b. l. and sprinkle, j. 1978. Totiglobus, an unusual new edrioasteroid from the Middle Cambrian of
Nevada. J. Paleont. 52, 243-266.

1980. New Homoiostelean echinoderms from the Late Cambrian of Alabama. Geol. Soc. Ain.
Abstr. Progr. 12 (7), 385.

berg-madsen, v. 1986. Middle Cambrian cystoid ( sensu lato) stem columnals from Bornholm, Denmark.
Lethaia, 19, 67-80.

1987. A new cyclocystoid from the Lower Ordovician of Oland, Sweden. Palaeontology, 30,
105-116.

bockelie, j. 1979a. Taxonomy, functional morphology and palaeoecology of the Ordovician cystoid family
Hemicosmitidae. Palaeontology, 22, 363 406.

\919b. Celticystis n. gen., a gomphocystitid cystoid from the Silurian of Sweden. Geol. For. Stockh.
Forh. 101, 157 166.

1981a. The Middle Ordovician of the Oslo region, Norway, 30. The eocrinoid genera Cryptocrinites,
Rhipidocystis and Bockia. Norsk geol. Tidsskr. 61, 123-147.

19816. Functional morphology and evolution of the cystoid Echinosphaerites. Lethaia, 14, 189-
202.

1982. Symmetry and ambulacral pattern of the rhombiferan superfamily Caryocystitida and the
relationship to other Blastozoa. Geol. For. Stockh. Forh. 103, 491 498.

1984. The Diploporita of the Oslo region, Norway. Palaeontology , 27, 1-68.
brett, c. E., frest, t. J., sprinkle, j. and clement, c. r. 1983. Coronoidea: a new class of blastozoan
echinoderms based on taxonomic reevaluation of Stephanocrinus. J. Paleont. 57, 627 -651.
briggs, d. e. g., fortey, r. a. and Clarkson, e. n. k. 1987. Extinction and the fossil record of the arthropods.
In larwood, G. p. (ed.). Extinction and survival in the fossil record. Systs Assoc. Spec. Vol. 34. Clarendon
Press, Oxford.

broadhead, t. w. 1980. Blastozoa. In broadhead, t. w. and waters, j. a. (eds.). Echinoderms: notes for a
short course. Univ. Tenn. Dept. geol. Sci., Stud. Geol. 3, 118-127.
cabibel, J., termier, h. and termier, G. 1958. Les Echinodermes mesocambriens de la Montaigne Noire.
Amis Paleont. 44, 281-294.

Campbell, k. s. w. and marshall, c. R. 1986. Rates of evolution among Palaeozoic echinoderms. In Campbell,

k. s. w. and day, m. f. (eds.). Rates of evolution, 61 100. Allen & Unwin, London.
chauvel, j. 1966. Echinodermes de l'Ordovician du Maroc. Cah. Paleont. (1966), 120 pp.

1969. Les echinodermes macrocystellides de l'Anti-Atlas Marocain. Bulk Soc. Geol. miner. Bretagne, (C)

l, 21 32.

1971a. Les echinodermes carpoides de Paleozoique inferieur Marocain. Notes Mem. Serv. geol. Maroc,
31, 49-60.

19716. Rhopalocystis Ubaghs: un echinoderme eocrinoide du Tremadocien de l'Anti-Atlas Marocain.
Mem. Bur. Rech. geol. min. 73, 43-49.

conway-morris, s. 1987. Cambrian enigmas. Geology Today (May-June 1987), 88-92.
cope, j. c. w. 1988. A reinterpretation of the Arenig crinoid Ramsey ocrinus. Palaeontology, 31, 229-235.
courtessole, R. 1973. Le Cambrien moyen de la Montague Noire: biostratigraphie. Laboratoire de Geologie,
CEARN, Toulouse.

derstler, K. L. 1981. Morphological diversity of early Cambrian echinoderms. In taylor, m. e. (ed.). Short
papers for the second International Symposium on the Cambrian System. US geol. Surv. Open File. Rep.
81-743, pp. 71-75.

donovan, s. k. The early evolution of the Crinoidea. In paul, c. r. c. and smith, a. b. (eds.). Echinoderm
phytogeny and evolutionary biology , pp. 235-244. Oxford University Press.

SMITH: PATTERNS OF DIVERSIFICATION AND EXTINCTION

825

Durham, j. w. 1966. Camptostroma , an early Cambrian supposed scyphozoan, referable to Ecliinodermata.
J. Paleont. 40, 1216 1220.

1967. Notes on the Helicoplacoidea and early echinodernrs. Ibid. 41, 97 102.

1978. A Lower Cambrian eocrinoid. Ibid. 52, 195-199.

and caster, k. e 1963. Helicoplacoidea: a new class of echinoderms. Science, NY. 140, 820-822.
fatka, o. and kordule, v. 1984. Acantliocystites Barrande, 1887 (Eocrinoidea) from the Jince Formation
(Middle Cambrian) of the Barrandian area. Vest, ustred. Ust. geol. 59, 299-302.

1985. Etoctenocystis bohemica gen. et sp. nov., new ctenocystoid from Czechoslovakia (Echino-
dermata. Middle Cambrian). Ibid. 60, 225-230.
flessa, k. w. and jablonski, d. 1983. Extinction is here to stay. Paleobiology, 9, 315-321.

1985. Declining Phanerozoic background extinction rates: effect of taxonomic structure? Nature,
Lond. 313, 216-218.

fortey, r. a. and Owens, R. M. 1987. The Arenig series in South Wales. Bull. Brit. Mas. Nat. Hist. (Geol.),
41, 61-307.

goldring, R. and Stephenson, d. G. 1972. The depositional environment of three starfish beds. Neues Jb.
Geol. Pa I don t. Mb. (1972), 611 624.

HARLAND, W. B., COX, a. V., LLEWELLYN, P. G., PICKTON, C. A. G., SMITH, A. G. and WALTERS, R. 1982. A
geological time scale, 131 pp. Cambridge University Press.
jablonski, d. and bottjer, d. j. 1986. Onshore offshore evolutionary patterns: discordance across hierarchical
levels. Geol. Soc. Am. Abstr. Progr. 18, 644.

jaekel, o. 1899. Stammesgeschichte der Pelmatozoen. 1, Thecoidea and Cystoidea, 442 pp. Springer, Berlin.
1918. Phylogenie und System der Pelmatozoen. Paldont. Z. 3, 1 128.

Jefferies, R. p. s. 1969. Ceratocystis perneri — a Middle Cambrian chordate with echinoderm affinities.
Palaeontology, 12, 494-535.

— 1986. The ancestry of the vertebrates, 376 pp. British Museum (Natural History), London.

— lewis, m. and donovan, s. k. 1987. Protocystites menevensis— a stem group chordate (Cornuta) from
the Middle Cambrian of South Wales. Palaeontology, 30, 429 484.

jell, p. a., burrett, c. f. and banks, m. r. 1985. Cambrian and Ordovician echinoderms from eastern
Australia. Alcheringa , 9, 183-208.

jobson, l. and Paul, c. r. c. 1979. Compagicrinus fenestratus, a new Lower Ordovician inadunate crinoid
from North Greenland. Rapp. Gronlands geol. Unders. 91, 7181.
kelly, s. m. and ausich, w. i. 1978. A new Lower Ordovician (Middle Canadian) disparid crinoid from
Utah. J. Paleont. 52, 916-920.

kesling, R. v. 1967. Cystoids. In MOORE, r. c. (ed.). Treatise on invertebrate paleontology, part S: Ecliinodermata
I, S85-267. Geological Society of America and University of Kansas Press, Lawrence, Kansas.
lane, n. g. 1970. Lower and Middle Ordovician crinoids from west-central Utah. Brigham Young Univ. Geol.
Stud. 17, 3-17.

1984. Predation and survival among inadunate crinoids. Paleobiology. 10, 453 458.
melendez, b. 1954. Los Trochocystites del Pirineo. Boln R. Soc. esp. Hist. nat. 51 [for 1953], 97-105.
orlowski, s. 1968. Upper Cambrian fauna of the Holy Cross Mts. Acta geol. pol. 18, 257-291.
patterson, c. and smith, a. b. 1987. Is the periodicity of extinctions a taxonomic artefact? Nature, Lond.
330, 248-252.

paul, c. R. c. 1 968a. Notes on cystoids. Geol. Mag. 105, 413 420.

— 1968A Macrocystella Callaway, the earliest glyptocystitid cystoid. Palaeontology , 11, 580 600.

— 1968c. Morphology and function of dichoporite pore-structures in cystoids. Ibid. 11, 697 730.

— 1972. Cheirocy Stella antiqua gen. et sp. nov. from the Lower Ordovician of Western Utah, and its
bearing on the evolution of the Cheirocrinidae (Rhombifera: Glyptocystitida). Brigham Young Univ. Geol.
Stud. 19, 15-63.

— 1977. Evolution of primitive echinoderms. In hallam, a. (ed.). Patterns of evolution, 123 158. Elsevier,
Amsterdam.

— 1979. Early echinoderm radiation. In house, m. r. (ed.). The origin of major invertebrate groups. Syst.
Ass. Spec. Publ. 12, 415-434. Academic Press, London.

— 1982. The adequacy of the fossil record. In joysey, k. a. and Friday, a. e. (eds.). Problems of phylogenetic
reconstruction. Syst. Ass. Spec. Publ. 21, 75-1 17. Academic Press, London.

1984. British Ordovician cystoids, Palaeontogr. Soc. [Monogr.], part 2, 65-152.

— 1988. Phylogeny and evolution of cystoids. In paul, c. r. c. and smith, a. b. (eds.). Echinoderm phytogeny
and evolutionary biology, 199-213. Oxford University Press.

826

PALAEONTOLOGY, VOLUME 31

faul, c. R. C. and bockelie, I. F. 1983. Evolution and functional morphology of the cystoid Sphaeronites in
Britain and Scandinavia. Palaeontology , 26, 687-734.

— and cope, j. c. w. 1982. A parablastoid from the Arenig of South Wales. Ibid. 25, 499-507.

— and smith, a. b. 1984. The early radiation and phylogeny of echinoderms. Biol. Rev. 59, 443-481.
philip, g. m. 1979. Carpoids — echinoderms or chordates? Ibid. 54, 439 471.

pompeckj, J. f. 1896. Die Fauna des Cambrium von Tejrovic und Skrej in Bohmen. Jb geol. Bundesanst.,
Wien , 45, 495 614.

prokop, R. J. 1962. Akadocrinus nov. gen., a new crinoid from the Cambrian of the Jince area (Eocrinoidea).
Sb. ustr. Ust. geol. 27, 31 39.

— and fatka, o. 1985. Luhocrinus monicae gen. et sp. n. (Eocrinoidea) from the Middle Cambrian of
Bohemia. Vest, ustred. Ust. geol. 60, 231-234.

raup, c. m. 1972. Taxonomic diversity during the Phanerozoic. Science , NY , 117, 1065-1071.

— 1975. Taxonomic survivorship curves and Van Valen’s law. Paleobiology , 1, 82-96.

— 1983. On the early origins of major biological groups. Ibid. 9, 107-115.

and sepkoski, j. j. 1982. Mass extinction in the marine fossil record. Science , NY, 215, 1501-
1503.

— 1984. Periodicity of extinctions in the geological past. Proc. natn. Acad. Sci. USA , 81, 80 1

805.

regnell, g. 1945. Non-crinoid Pelmatozoa from the Palaeozoic of Sweden, a taxonomic study. Meddn Lunds
geol. -miner. Instn , 108, 255 pp.

ruedemann, r 1933. Camptostroma, a Lower Cambrian floating hydrozoan. Proc. US natn. Mus. 82 (no.
13), 8 pp.

runnegar, b. 1986. Rates and modes of evolution in the Mollusca. In Campbell, k. s. w. and day, m. f.

(eds.). Rates of evolution , 39 60. Allen & Unwin, London.
sepkoski, j. j. 1978. A kinetic model of Phanerozoic taxonomic diversity. I, Analysis of marine orders.
Paleobiology , 4, 223-251.

179. A kinetic model of Phanerozoic taxonomic diversity. II, Early Phanerozoic families and multiple
equilibria. Ibid. 5, 222-251.

— 1981fl. The uniqueness of the Cambrian fauna. In taylor, m. e. (ed.). Short papers for the second
International Symposium on the Cambrian System. US geol. Surv. Open File Rep. 81-743, 203-207.

19816. A factor analytic description of the Phanerozoic marine fossil record. Paleobiology , 7, 36-53.
1982. A compendium of fossil marine families. Contr. Biol. Geol. Milwaukee publ. Mus. 51, 125 pp.
1986. Phanerozoic overview of mass extinction. In raup, d. m. and jablonski, d. (ed.). Patterns and
processes in the history of life , 277 -295. Springer Verlag, Berlin.

— and raup, d. m. 1986. Periodicity in marine extinction events. In elliott, d. k. (ed.). Dynamics of
extinction , 3-36. John Wiley & Sons, New York.

— and sheehan, p. m. 1983. Diversification, faunal change and community replacement during the
Ordovician radiations. In tevesz, m. j. s. and mccall, p. l. m. (eds.). Biotic interactions in Recent and fossil
benthic communities , 673-717. Plenum Press, New York.

simpson, g. g. 1953. The major features of evolution, 434 pp. Columbia University Press, New York.
smith, a. b. 1982. The affinities of the Middle Cambrian Haplozoa (Echinodermata). Alcheringa, 6,
93 100.

1984. Classification of the Echinodermata. Palaeontology, 27, 431 459.

1986. Cambrian eleutherozoan echinoderms and the early diversification of edrioasteroids. Ibid. 28,
715-756.

— (in press). To group or not to group: the taxonomic position of Xyloplax. Proc. 6th internat., Echinoderms
Conf.

spencer, w. k. 1918. A monograph of the British Palaeozoic Asterozoa. Palaeontogr. Soc. [Monogr.], part
3, 109 196.

1951. Early Palaeozoic starfish. Phil. Trans. R. Soc. Lond. (B), 235, 87-129.

sprinkle, j. 1973. Morphology and evolution of blastozoan echinoderms. Spec. Pub. Mus. comp. Zool. Harv.
284 pp.

1975. The ‘arms’ of Caryocrinites, a rhombiferan cystoid convergent on crinoids. J. Paleont. 49, 1062-
1073.

I976u. Biostratigraphy and paleoecology of Cambrian echinoderms from the Rocky Mountains. Brigham
Young Univ. Geol. Stud. 23, 61 73.

— 19766. Classification and phylogeny of ‘pclmatozoan’ echinoderms. Syst. Zool. 25, 83-89.

SMITH: PATTERNS OF DIVERSIFICATION AND EXTINCTION

827

1980m An overview of the fossil record. In broadhead, r. w. and waters, j. a. (eds.). Echinoderms:
notes for a short course. Univ. Tenn. Dept. geol. Sci., Stud. Geol. 3, 15-26.

19806. Early diversification. Ibid. 86-93.

- 1981. Diversity and evolutionary patterns of Cambrian echinoderms. In taylor, m. e. (ed.). Short papers
for the second International Symposium on the Cambrian System. US geol. Surv. Open File Rep. 81-743,
219-221.

1983. Patterns and problems in echinoderm evolution. Echinoderm Studies , 1, 118.

- 1985. New edrioasteroid from the Middle Cambrian of western Utah. Paleont. Contr. Univ. Kansas ,
Pap. 1 16, 4 pp.

— and moore, r. c. 1978. Echmatocrinea. In moore, r. c. and teichert, c. (eds.). Treatise on invertebrate
paleontology. Part T: Echinodermata 2, T405-407. Geological Society of America and University of Kansas
Press, Lawrence, Kansas.

— and robison, r. a. 1978. Ctenocystoids. In moore, r. c. and teichert, c. (eds.). Treatise on invertebrate
paleontology. Part T: Echinodermata 2, T998 1002. Geological Society of America and University of
Kansas Press, Lawrence, Kansas.

termier, h. and termier, g. 1973. Les Echinodermes Cincta du Cambrien de la Montaigne Noire (France).
Geobios , 2, 131-156.

thoral, m. 1935. Contribution a T etude paleontologique de T Ordovicien inferieur de la Montague Noire et
revision sommaire de la fauna cambrienne de la Montague Noire , 362 pp. Montpellier.
ubaghs, G. 1953. Notes sur Lichenoides priscus Barrande, eocrinoi'de du Cambrien moyen de la Tchecoslo-
vaquie. Bull Inst. r. Sci nat. Belg. 29, no. 34, 24 pp.

— 1960. Le genre Lingulocystis Thoral. Annls Paleont. 46, 81 115.

— 1963m Rhopalocystis destombesi n. g., n. sp., eocrinoi'de de l’Ordovicien inferieur (Tremadocien superieur)
du Sud marocain. Notes Serv. geol. Maroc, 23, 25-44.

— 19636. Cothurnocystis Bather, Phyllocystis Thoral and an undetermined member of the order Soluta
(Echinodermata, Carpoidea) in the uppermost Cambrian of Nevada. J! Paleont. 37, 1133 1 142.

— 1967m Eocrinoidea. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part S: Echinodermata
1, S455 495. Geological Society of America and University of Kansas Press, Lawrence, Kansas.

19676. Flomostelea. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part S: Echinodermata
1, S565-581. Geological Society of America and University of Kansas Press, Lawrence, Kansas.

— 1969m Aethocrinus moorei Ubaghs, n. gen., n. sp., le plus ancien crinoide dicyclique connu. Paleont.
Contr. Univ. Kansas , Pap. 38, 25 pp.

19696. Les echinodermes carpoi'des de l’Ordovicien inferieur de la Montagne Noire (France). Cali.
Paleont. (Paris), 112 pp.

— 1971m Un crinoide enigmatique ordovicien: Perittocrinus Jaekel. Neues Jb. Geol. Palaont. Abh. 137,
305-306.

— 19716. Diversite et specialisation des plus anciens echinodermes que Fon connaise. Biol. Rev. 46,
157-200.

— 1 972<3. Le genre Balantiocystis Chauvel (Echinodermata, Eocrinoidea) dans l’Ordovicien inferieur de la
Montaigne Noire (France). Annls Paleont. 58, 3-27.

— 19626. More about Aethocrinus moorei Ubaghs, the oldest known dicyclic crinoid. J. Paleont. 46,
773-775.

1975. Early Palaeozoic echinoderms. Rev. Earth planet. Sci. 3, 79 98.

1983. Echinodermata. Notes sur les echinodermes de FOrdovicien inferieur de la Montagne Noire
(France). In courtessole, r., market, l., pillet, j., ubaghs, g. and vizcaino, d. (eds.). Calymenina ,
Echinodermata et Hyolitha del' Ordovicien inferieur de la Montagne Noire (France meridonale). Mem. Soc.
Etud. sci. Aude, (1983), 62 pp.

1987. Echinodermes nouveaux du Cambrien moyen de la Montagne Noire (France). Annls Paleont.
73, 1 -27.

— and robison, r. a. 1985. A new homoiostelean and a new eocrinoid from the Middle Cambrian of
Utah. Paleont. Contr. Univ. Kansas, Pap. 115, 24 pp.

valentine, j. w. 1969. Patterns of taxonomic and ecological structure of the shelf benthos during Phanerozoic
times. Palaeontology, 12, 684 709.

— 1980. Determinants of diversity in higher taxonomic categories. Paleobiology, 6, 444 -450.

— 1986. Fossil record of the origin of Bauplane and its implications. In raup, d. m. and jablonski, d.
(eds.). Patterns and processes in the history of life, 209-222. Springer Verlag, Berlin.

van valen, l. 1973. A new evolutionary law. Evol. Theory, 1, I 30.

828

PALAEONTOLOGY, VOLUME 31

whitehouse, f. w. 1941 . The Cambrian faunas of north-eastern Australia, part 4: early Cambrian echinoderms
similar to the larval stages of Recent forms. Mem. Qd Mus. 12, 1-28.

Typescript received 9 August 1987

Revised typescript received 16 November 1987

A. B. SMITH

Department of Palaeontology
British Museum (Natural History)
London SW7

THE STRATIGRAPHICAL DISTRIBUTION AND
TAXONOMY OF THE TRILOBITE ONN1A IN
THE TYPE ONNIAN STAGE OF THE
UPPERMOST CARADOC

by alan w. owen and j. keith ingham

Abstract. The litho- and biostratigraphy of the type section of the Onnian Stage in the Onny River, south
Shropshire, is reassessed on the basis of detailed sampling over an extended period, including years when
the river level was unusually low. The base of the Onny Formation is redefined at a level within the upper
part of the Onnian and thus the base of the stage lies within the Acton Scott Formation. Four biozones are
defined on the basis of closely spaced samples of the trinucleid trilobite Onnia , a peri-Gondwanan immigrant.
In ascending order these are: the O. superba cobboldi Local Range Zone, the O. s. creta Local Range Zone,
the O. gracilis Acme Zone, and the O. s. superba Local Range Zone. The second of these is based on a new
subspecies, the others on a reassessment of previously named taxa. Within the O. superba subspp. zones,
fringe pit distribution of successive samples of Onnia shows considerable stasis, although early and late
populations of O. s. superba can be recognized. The changes between the subspecies can be viewed as reflecting
either an evolutionary lineage or subtle fluctuations in environmental controls on a cline or set of
ecophenotypes within a variable species.

The richly fossiliferous type Caradoc succession of south Shropshire has been the subject of
considerable interest since the publication of Murchison’s Silurian System in 1839 (see Hurst 1979<r/,
pp. 185-189 for historical review). Most importantly, as befits an international standard section,
detailed work by Bancroft (1929-1949), Dean (1958-1964), and Hurst (1979//, b ) has contributed
significantly to the stratigraphy, trilobite and brachiopod systematics, sedimentology, and palaeo-
ecology of the Shropshire Caradoc. However, there are still major problems with the definition of
the uppermost (Onnian) stage and the correlation potential of its zone fossils, species of the
trinucleid trilobite Onnia.

The present study addresses these problems and is based on mapping and annual sampling of
the type Onnian section in the Onny River [Grid Ref. SO 425 854] between 1974 and 1983. This
included some years when the water level was anomalously low and enabled bulk samples to be
taken from horizons which were inaccessible to previous workers; in all, twenty-three levels have
been extensively sampled. Bancroft also collected from this section and his sample points were
well localized on a sketch map (Bancroft 1949, fig. 39). These samples, along with his original map
and detailed field notes, are housed in the British Museum (Natural History), as is Hurst’s
collection. Comparison of these earlier samples with our own has proved most illuminating. For
reasons of site conservation, our map is not reproduced in this paper but copies are available for
consultation by bona fide researchers at the Hunterian Museum, Glasgow, British Museum
(Natural History), and National Museum of Wales, Cardiff.

The present work demonstrates that the base of the Onnian stage does not correspond to a clear
lithological change in the Onny River section (cf. Hurst 1979//, b) and we draw a much clearer
distinction between litho-, bio-, and chronostratigraphy than did some earlier workers. Moreover,
assessment of the successive changes in samples of Onnia enables a revised set of biozones for
the type Onnian to be established. These samples also provide a case study demonstrating the

| Palaeontology, Vol. 31, Part 3, 1988, pp. 829-855, pis. 74-77. |

© The Palaeontological Association

830

PALAEONTOLOGY, VOLUME 31

text-fig. 1 . The revised stratigraphy of the type Onnian Stage showing the horizons sampled in the present study along with those of
Bancroft. The range, mean, and one standard deviation on each side of the mean of the radius number of the posterior E! pit in samples
of Onnia are illustrated. This is a measure of the number of pits in the half fringe of this arc but 'rounds down' halt pit values. Note that
only the sample from the cliff section includes specimens collected by Bancroft. The rest were collected in the present study. Possible
relationships between the Onnia taxa are suggested but it is unclear whether these are evolutionary or ecological in origin.

OWEN AND INGHAM: CARADOC TRILOBITE

831

difficulties of distinguishing between evolutionary and ecophenotypic/clinal variation in a series of
populations at a single site.

THE TYPE ONNIAN STAGE

The term Onnian was introduced by Bancroft for the uppermost stage of the Caradoc Series in
his privately published correlation tables (1933). These show that Bancroft placed the base of the
stage below the ‘Fossil-bed with 0[nnia\ cobboldi' in the Onny River section. This latter horizon
was undoubtedly his locality Px (see Bancroft 1949, fig. 39; also text-fig. 1 herein). Bancroft
recognized three zones within the Onnian characterized by, in ascending order: O. cobboldi ,
O. gracilis , and O. superba. All three species were named by him in 1929 (Bancroft 1929 b) and
referred to his new genus Onnia in 1933.

O. gracilis has also been described from Welshpool in the Welsh Borderlands (Cave 1965) and
the Cross Fell Inlier in northern England ( Dean 1 962). The occurrence at Welshpool is in association
with a similar binodicope ostracod fauna to that in the Onny River (Jones 1987, p. 108). At Cross
Fell O. gracilis is succeeded by ‘O. s. pusgillensis' Dean, 1961 which also occurs in the uppermost
Onnian of nearby Cautley (Ingham 1966, 1974). Thus, Bancroft’s O. gracilis Zone has been at
least tentatively recognized outside the type area and "O. s. pusgillensis ' has been considered to
indicate the O. superba Zone in northern England (but see below p. 853).

The base of the type Onnian is defined primarily on the first appearance of "O. cobboldi ’ (see
below for revised taxonomy) at a level about 6 m below Bancroft’s locality Px (text-fig. 1). There
are also changes in several other elements of the shelly fauna as detailed by Dean (1963, pp. 8,
13-14) and Hurst (19796, p. 212). Of particular interest for international correlation (e.g. Owen
1980, 1987) is the occurrence of another trinucleid trilobite, Tretaspis ceriodes Angelin favus Dean
at levels immediately above and below this boundary (Dean 1963, p. 8). Another subspecies, T. c.
alyta Ingham, occurs in the uppermost Onnian of northern England (Ingham 1970, pp. 50, 52).

The Onnian strata in Shropshire are overstepped eastwards, with small angular unconformity
and overlap, either by the upper Llandovery Hughley Shale Formation (as in the Onny River
section) or by the underlying Pentamerus Beds to the north-east. However, a continuous succession
is present at Cautley where the base of the Pusgillian Stage (and therefore the base of the Ashgill
Series) is defined (Ingham 1966; Ingham and Wright 1970; Wright in Whittington et al. 1984). This
boundary is marked (inter alia ) by the disappearance of Onnia from Britain and the first appearance
of members of the T. seticornis species group.

The precise correlation of the Onnian Stage with the standard graptolite and conodont zones
remains unclear. Present evidence suggests that the stage may equate with the uppermost D.
clingani and most of the P. linearis graptolite zones (Ingham and Wright 1970; Wright in
Whittington et al. 1984). The base of the A. ordovicicus conodont zone may also lie within the
stage (Savage and Bassett 1985, p. 683) in spite of statements to the contrary (Orchard 1980;
Bergstrom and Orchard 1985).

REVISED BIOSTRATIGRAPHY

Bancroft’s work on the upper Caradoc faunas of the Onny River section was based on bulk
samples taken from exposures on the river bank. In the Onnian Stage, these samples were widely
spaced (Bancroft 1949, fig. 39) and give only a broad picture of the biostratigraphy. The present
study involves more closely spaced samples, many from the river bed (text-fig. 1). Nevertheless,
Bancroft’s material was so well localized that it can be easily accommodated in our analysis.

The zone fossils of the Onnian Stage, species of Onnia , belong to a group of trilobites with a
cephalic fringe which is pitted in a regular and quantifiable pattern (see Hughes et al. 1975 and
text-fig. 2 herein). This pitting is widely used taxonomically and even poorly preserved fringe
fragments can yield useful data. Where the trilobites are sufficiently abundant (as in the type
Onnian), statistical analysis of fringe pit distribution can be used to assess sequential changes

832

PALAEONTOLOGY, VOLUME 31

in populations. They also allow a semi-quantitative rather than purely typological definition
of taxa.

The successive changes in samples of Onnia from the Onny River are discussed in detail below.
It is clear that whilst O. gracilis is distinctive there is considerable overlap in pit distribution and
gross morphology in the ranges of variation of O. superba and O. cobboldi from their type horizons.
These latter taxa are redefined as subspecies of O. superba', their definitions encompass samples
from other horizons near the top and base of the type Onnian (text-figs. 1, 3, 4). A third subspecies,
O. s. creta subsp. nov. is here established both in terms of pit distribution and fringe shape for
the samples from horizons immediately below those containing O. gracilis. Lacking any evidence
on the occurrence of these three subspecies of O. superba outside the Onny River, we take a
cautious view of their correlative potential. Nevertheless, the type Onnian is here redefined as
comprising four biozones (text-fig. 1) in ascending order: the O. s. cobboldi Local Range Zone, the
O. s. creta Local Range Zone, the O. gracilis Acme Zone, and the O. s. superba Local Range
Zone.

The bases of all four zones are defined on the first occurrence of the eponymous species or
subspecies. O. gracilis persists as a rare element of the earliest O. s. superba Zone and hence the
underlying strata, where O. gracilis is abundant, are defined as an acme zone. The occurrence of
O. gracilis at Welshpool and Cross Fell may be at a broadly similar level to its presence in the
type Onnian but this is poorly constrained. O. gracilis also occurs at Cardington, Shropshire along
with specimens of T. ceriodes which are morphologically closer to T. c. alyta and a morph of T.
c. angelini Stormer than to T. c.favus. Owen (1980, p. 722) suggested that the strata here may be
Onnian in age but the presence of Flexicalymene salteri Bancroft and the brachiopods Onniella
depressa Bancroft ( sensu Hurst 19796) and Chonetoidea cf. radiatula (Barrande) (D. A. T. Harper,
pers. comm. 1987) now confirm the Actonian age given by Dean (1963, pp. 8-9). Hurst (19796,
p. 204) noted that O. depressa appears high in the Actonian Stage in the Onny Valley and thus
the strata at Cardington may be equivalent to this level. None the less, it appears that Onnia
gracilis ranges both above and below its Acme Zone.

The taxonomic affinities of O. s. pusgillensis Dean in northern England are unclear and may
even be closer to O. gracilis than to O. superba. It is therefore regarded as a distinct species, O.
pusgillensis (see text-fig. 2).

LITHOSTRATIGRAPHY

Hurst (19796, figs. 2, 3, 11) summarized the historical development of the terms applied to the
type upper Caradoc and established a modern lithostratigraphical terminology. He assigned all
the strata of Onnian age to the ‘Onny Shale Formation’ — a usage which broadly followed that of
{inter alia) La Touche (1884, ‘Onny Shales’) and Dean (1958-1963, ‘Onnia Beds’). In contrast,
Bancroft (19296, 1933) restricted the terms ‘Trinucleus Shales’ or ‘Onny Shales’ to the uppermost
Onnian Stage, assigning the lower Onnian to the underlying ‘Acton Scott Beds’. He placed the
base of the latter at about the Marshbrookian- Actonian boundary (cf. Hurst 19796, fig. 2). Hurst
(19796, p. 178) defined the Onny Shale Formation in the Onny River as comprising 20 m of
bioturbated, very fossiliferous blue-black mudstones overlain by 5 m of laminated blue-grey
mudstone succeeded by yellow-weathering blocky mudstones of the river cliff section (perhaps
another 18 m). He defined the base of the Onny Shale Formation as lying in a 3 m gap in exposure
below which are the poorly fossiliferous calcareous mudstones and siltstones of the Wistanstow
Member of the Acton Scott Formation. Hurst termed the Onnian fauna the ‘ Onniella broeggeri-
Sericoidea homolensis Association’ and considered it to have lived in a distal shelf setting (1979a,
pp. 223-228, 238-239).

Hurst sampled twenty-six horizons in his Onny Shale Formation (1979a) but our analysis of his
samples in the British Museum (Natural History) shows that only the lowest two (thought by
Hurst to be from the lowest 2-5 m of the unit) contain Onnia superba cobboldi. The succeeding
five samples contain O. s. creta and the next five O. gracilis. The highest of these (no. 32), from

OWEN AND INGHAM: CARADOC TRILOBITE

833

text-fig. 2. The fringe in Onnia pusgillensis Dean (as redefined herein) shown diagrammatically as
corresponding upper and lower lamellae and labelled to show the location of pit arcs and other fringe features
discussed herein. Original illustration of Ingham (1974, text-fig. 20), published by kind permission of the

Palaeontographical Society.

c. 15 m above the base of Hurst’s section, also contains O. s. superha which persists through the
rest of the sequence. There is no doubt, therefore, that the lowest 10 m of the Onnian was not
sampled by Hurst who underestimated the stratigraphical thickness between his lowest sample and
those yielding Actonian faunas. This unsampled part of the sequence constitutes most of the O. s.
cohholdi Zone, including the type horizon of the eponymous subspecies (Px of Bancroft, sample
B herein).

The rubbly, calcareous lower Onnian strata show a greater lithological similarity to the underlying
Wistanstow Member of the Acton Scott Formation than to the overlying blocky mudstones and
are here included in the lower unit (text-fig. 1). The base of the laminated blue-grey mudstone
noted by Hurst is here taken as the base of the Onny Formation and lies within the O. s. superha
Local Range Zone. As only this lowest e. 5 m is at all shaly, we recommend that the term ‘shale’
be omitted from the formation name.

At sample locality O in the lower part of the O. s. superba Zone (uppermost Acton Scott
Formation as herein understood) the rubbly mudstones contain largely comminuted shelly debris
and more complete specimens are rare. Nevertheless, this 5 cm horizon is particularly interesting
in that it contains abundant, hard irregularly shaped phosphatic nodules in which the fine shelly
debris is well preserved. The episode of slow deposition represented by this horizon may be broadly
coeval with similar events which also produced bands of phosphate nodules in, for example, the
Nod Glas and Blaen y Cwm formations in mid Wales.

ONNIA IN THE TYPE ONNIAN

Onnia was a late Caradoc immigrant into the British area and stayed but a short time. Its origins
were in higher latitudes around Gondwanaland where it has a much greater stratigraphical range

834

PALAEONTOLOGY, VOLUME 31

(early Caradoc to Ashgill; Hughes et al. 1975, p. 575). It was derived from another middle
Ordovician marrolithine, Deanaspis, which is not known from the British Isles (see Hughes et al.
1975). The appearance of Onnia in Shropshire may reflect the circulation of cold, fairly deep waters
of the outer neritic regime which was the climax of the Caradoc transgression in the area. This
correlates with the widespread Nod Glas deepening in mid Wales and with the probable circulation
of colder waters at even greater depths which brought a peri-Gondwanaland cyclopygid biofacies
to the margins of Laurentia at Girvan— the Upper Whitehouse Group (Ingham 1978). Specimens
of Onnia far outnumber the relatively few other trilobites at most levels in the type Onnian and
thus provide an effectively continuous record of the genus in Shropshire over a period in excess
of a million years.

text-fig. 3. Histograms showing the radius number of the posterior
E, pit in pooled samples of the successive subspecies of Onnia superba
highlighting the overlap (shaded) between O. s. cobboldi and samples
of O. s. superba from the highest part of the Onny River section.
Both of these histograms include data from the Bancroft Collection.

The revised zonation of the type Onnian is founded on the successive appearance, without
overlap, of three subspecies of O. superba ; the second and third of these being separated by the
only distantly related O. gracilis. The fringe pits enable a semi-quantitative, graphical assessment
of successive changes between population samples of the same taxon and between taxa. These
changes are not size dependent; pit distribution in an individual becomes fixed at an early stage
in ontogeny, as was also demonstrated by Hughes ( 1 970). The gross changes between the subspecies
of O. superba in all the pit arcs are shown on text-figs. 3 and 5, whilst the text-figs. 1 and 4 also
illustrate the range, mean, and one standard deviation on each side of the mean for arcs Ej, Il5
and In in each sample. These three arcs extend around the whole fringe and are the outermost arc
(Ej), innermost arc (In), and first arc inside the girder (I,) (see Ingham 1974, pp. 59-60 for fringe
pit terminology in Onnia ; Hughes et al. 1975). All values refer to half-fringe counts and only the
range is given in very small samples. Table 1 shows changes from sample to sample in arcs I2, I3,
and the F pit series in the subspecies of O. superba.

OWEN AND INGHAM: CARADOC TRILOBITE

835

Successive changes in the type Onnian

The earliest subspecies, O. s. cohholdi , shows stasis in all its fringe characters except for the
stratigraphically highest sample which shows a reduction in pit number. This heralds O. s. creta ,
which is also characterized by a strongly inflated area of the posterior fringe along the Ij arc— a
feature seen only in a subdued form in some specimens of O. s. cohholdi. Chi-squared tests show
that the reduction in pit number in O. s. creta compared with the earlier subspecies is significant
at less than the 01 % probability level for arcs El5 Il5 In, and I2. Most arcs show little or no
change from sample to sample in O. s. creta but the mean number of pits in arcs E! and I2 shows
a progressive decrease, effectively continuing the ‘trend’ from O. s. cohholdi.

The sequence of O. superha populations is interrupted by the appearance of O. gracilis , a very

Radius number of posterior It pit

Sample

Radius number of ln cut-off

n 57

- V

38

40m -

12

late

I 1 n 2

early

O. s. superba

Zone 30m

n 46

n 1 8

I 1 n 2

n 5

O. gracilis

Zone n 18

-| 20m

n 24

n 24

n 26

O. s. creta

Zone

n 6

n 5

I 1 n 4

10m -

O. s cobboldi

Zone

I 1 n 2

late

s. superba

Zone

early

- M

O. gracilis

L Zone
K

n 9 |-

H n 28

n 7 h

^ n 20
I n 20

O. s. creta

Zone

H n 5

I 1 n 3

1 n 4

O. s. cobboldi

Zone

l 1 1 1 1 1 1 1 1 1 1 1 1 r 0 J i r 1 1 1 1 1 1 1 1 1 1 1 1 1

14 16 18 20 22 24 26 8 10 12 14 16 18 20 22

text-fig. 4. Changes in the radius numbers of the posterior 1, pit and the In cut-off in successive samples of
Onnia from the type Onnian Stage. Range, mean, and one standard deviation on each side of the mean
shown for the larger samples. Note that each taxon is restricted to its own zone except that a few specimens
of O. gracilis occur with the lowest sample of O. superba superba and these provided the two high pit counts
in arc I,. Sample V incorporates specimens from the Bancroft Collection.

836

PALAEONTOLOGY, VOLUME 31

1 " 73

H4> 1 n 84

I — N> 1 n 79

I — I n 150

i — i — i — i — i — i — i — i

4 6 8 10

Radius number
of first l3 pit

I —

I — I " 127

i — i — i — i — i — i — i — i — i — i — i

8 10 12 14 16 18

Pits in l3

[ n 38

" 81 1

n 67 1

n 128 | ■ ■ 1

I I 1 1 1 1 1 1 1 1 1

6 8 10 12 14 16

Radius number of first F pit

n

n 79

n 80

162 I—

\ — — O-H
n 83 l-HIN — I
— I

— I

O. s. superba late
O. s. superba early
O. s. creta
O. s. cobboldi

1 1 1 1 1 1 1 1 1

14 16 18 20 22

Radius number of posterior
I, pit

" 85 | |

n 89 MO" 1

I — “4^™ — I n
f 4> — I

77

n 166

1 1 1 1 1 1 1 1 1 1

8 10 12 14 16

Radius number of ln cut-off

H>H n 67

HD n 89

n 78 h-4>H

" iso

n 61 1— CN

n 77 f ■ U 1

I — H> — H n 78

26 f

14

—i 1 1 1 1 1 1 1 1 1 1 1 1 1 1

16 18 20 22 24 26 28 30

Radius number of posterior El pit

i 1 1 — i 1

1 2 3 4 5

Radius number
of first l3 pit

i p 1 i 1 1 1 1 1 1 1

12 14 16 18 20 22

Pits in l2

text-fig. 5. Summary of the fringe pit distribution in the subspecies of Onnia superba in the type Onnian
Stage showing range, mean, and one standard deviation on each side of the mean. The samples of O. s.
cobboldi and late O. s. superba include specimens from the Bancroft Collection.

different species in several respects; of the features depicted on text-tigs. 1 and 8, the most substantial
difference is shown by arc Ej. The very large number of pits in this arc exceeds even the upper
end of the range in O. superba. Two morphs can be recognized within the samples of O. gracilis
based on the presence or absence of arc I4. The percentage of individuals with this arc decreases
upwards through the O. gracilis Acme Zone from 94% (sample K, n = 18) through 91 % (L, n =
43) to 42 % (M, n = 119), although both specimens of O. gracilis in the lowest O. s. superba
sample have this arc.

O. s. superba is closer to O. s. cobboldi than to O. s. creta not only in lacking the strongly
inflated posterior part of I, but also in having a greater mean number of pits in every arc (text-
fig. 5). There is therefore a reversal of the pit reduction that marks replacement of O. s. cobboldi
by O. s. creta. This is further emphasized when successive samples of O. s. superba are analysed
(e.g. text-figs. 1 and 4; Table 1). These fall readily into two groups. ‘Early' populations (samples
N-S) have a fairly planar fringe surface and a range and mean values for each pit arc equal to or
slightly greater than those in O. s. cobboldi. In contrast, ‘late’ populations (T-V) have a more
com ex fringe surface and, in the case of arcs E, and In, an increased mean pit count. Chi-squared
tests show that the numbers of pits in E, and I2 in O. s. cobboldi and early O. s. superba are
significantly different at the OT % level. The same applies to E, and In when the early and late

table 1. The range, mean (x), and sample size (n) of selected fringe features of the successive samples of subspecies of Onnia superba from the
Onny River section. Such changes in arcs E1? Il5 and In are illustrated graphically on text-figs. 1 and 4. Sample V incorporates specimens from

the Bancroft Collection.

OWEN AND INGHAM: CARADOC TRILOBITE

X

GO

Cd

X

X

Cd

o

oo

os

GO

Cd

CO

X

CO

o

G— i

o

a

CO

GO

f"H

Cd

X

co

J-(

<D

X>

1— (

1

|

1

1

1

Cd

i

Cd

o

o

os

1

1

1

1

1

— 1

Cd

£

• X

1— <

1

1

1—1

1—1

1—1

1

1

1

1

1— «

c 'a

C/5

<L>

OXJ

.2 ^

CO

CO

or

or

Cd

oT

GO

or

CO

GO

1—1

Cd

Cd

CO

or

or

’3 ^

c

1

1

i

1

i

1

|

i

1

1

1

1

1

1

1

1

cd £

OC cc

cd

X

O

OO

Os

os

Os

oo

oo

oo

o

oo

—

OS

OS

o

O

Cl

c~

or

CO

Cd

GO

os

r-

X

Cd

or

cd

co

X

GO

a

GO

or

r'-1

Cd

X

CO

or

or

1

or

or

,

1

1

1

1

Cd

Cd

IX

—

'

1

'

'

'

1—1

’ — 1

'

1

1

1

1

1—1

l“'

C

0)

CJ)

X

GO

GO

GO

c

oo

CO

CO

CO

GO

CO

X

or

or

(75

c

1

’ — 1

” 1

|

|

i

1

1

cd

C*

—

or

o

r-

Cd

Cd

or

o

oo

o

os

—

CO

CO

—

Cd

os

o

o

CJ

X

Cd

X

cd

CO

os

Cd

Cd

GO

Cd

CO

c

os

C— <
O

C

X

GO

Cd

CO

X

or

i-H

a>

X

1

1

I

i

1

1

1

1

1

1

g

ix

c

c-

1

1

X

1

X

c*

r->

X

1

1

1

1

1

oo

oo

C ‘Bh

C/5 ^

<L>

W)

c

cd

.2 ^
o3 ^

cd £

os

oo

r-

os

c-

oo

oo

oo

o

OS

00

oo

oo

os

os

o

* X

&

or

X

GO

X

X

oj-

or

c-

or

GO

GO

or

c-

or

X

oo

GO

GO

oo

Cd

r-

Cd

c~

X

Cd

CO

OS

X

X

Cd

GO

Cd

Cd

X

GO

oo

c

in

,—H

or

Cd

X

Cd

oo

1

|

1

|

oo

1

c-

or

GO

GO

1

1

1

1

1

X

c*

IX

1

1

1

1

1

1

1

1

1

1

<N

1—1

a

(D

bD

Cd

oo

OS

os

oo

os

o

Os

c

X

oo

GO

c-

os

oo

os

os

C/5

C

cd

1

1

i

1

1

1

Cd

1

1

1

1

—

1

1

1

1

X

cd

GO

c-

c

'd-

X

GO

X

c-

GO

Cd

co

Cd

CO

or

oo

c

or

or

CO

Qh

X

os

Cd

c

Cd

Os

r-

Cd

Cd

or

c~

CO

GO

CO

Cd

r-

GO

GO

o

C

GO

GO

Cd

CO

c-

co

J-l

<u

X

l

l

1

l

1

1

1

1

1

1

1

g

IX

Cd

1

1

1

1

1

CO

1

CO

CO

CO

CO

1

1

1

1

1

co

CO

c a

C/5

<L>

bli

2

' o -*—*
as £

C

cd

or

CO

CO

CO

CO

co

or

GO

or

CO

CO

CO

or

or

at «

’ — 1

Cd

Cd

CO

CO

Cd

Cd

Cd

Cd

Cd

Cd

Cd

Cd

CO

cd

1—1

Cd

<u

a

S

00

>

£

H

cn

&

Q

p

o

£

>“1

HH

X 0

Dh

w

a

u

CQ

<

i§

<>

2

-O

<o

i>T

o

Th

cd

cd

>5

W

»5

05

6

d

d

838

PALAEONTOLOGY, VOLUME 31

samples of O. s. superba are compared. The change from early to late O. s. superba occurs at about
the base of the Onny Formation as redefined herein.

In contrast to the general 'trends’ in O. superba noted above, some fringe pit data show a more
complex pattern of change. The position of the first F pit shows a marked zigzag series of changes
(text-fig. 5). In contrast, the mean position of the first I3 pit shows a significant adaxial shift from
O. s. cobboldi through O. s. creta to early O. s. superba , but late O. s. superba shows a reversal of
this ‘trend’.

INTERPRETATION

The changes in pit number and distribution seen in the successive samples of O. superba include
several which are (albeit significant) shifts in mean values largely within the considerable overlap
in the range of values shared by the different subspecies. However, two features in particular
indicate that the changes in the type Onnian are not simply random fluctuations in pit number
within an essentially conservative species. First, successive samples of the same subspecies (or the
early and late forms of O. s. superba) show a considerable degree of stasis in the range and mean
values of most pit counts. Secondly, the range of values in the most variable fringe pit feature, the
number of pits in Ej, shows marked differences between subspecies (text-fig. 3). This is especially
true in the case of O. s. creta in which only 19-5 % (n = 82) of specimens have enough pits in Ej
to fall within the overlap in range between O. s. cobboldi and the late O. s. superba. Moreover, the
admittedly small sample (F) from the top of the O. s. cobboldi Zone shows a downward shift in
the number of pits in Els and is therefore transitional towards the range seen in the slightly younger
O. s. creta. Other pit counts (e.g. Il5 In— see text-fig. 4) also show this transitional condition but
the swelling along the lateral part of It is much weaker than in O. s. creta.

If the changes seen in O. superba are not random fluctuations, they must reflect either an
evolutionary lineage (or lineages) or fluctuations in environmental conditions affecting one very
variable species whose morphology is ecologically controlled. As O. superba is unknown outside
the type area there is insufficient evidence to confirm either hypothesis but, in view of the possible
biostratigraphical importance of Onnia, some discussion and speculation is merited.

Evolution

The presence of O. gracilis and consequent gap in the record of O. superba in the middle of the
type Onnian complicates any evolutionary interpretation of the O. superba subspecies. The change
from the relative stasis of O. s. cobboldi to that of O. s. creta could be viewed as a punctuational
event with only the youngest sample of O. s. cobboldi being intermediate in pit number if not fringe
swelling. O. s. superba appears above the O. gracilis Zone and has a fringe shape and pit number
in each arc that are closer to those of O. s. cobboldi than O. s. creta. This reversion to a higher
pit count is continued in O. s. superba with the change from ‘early’ to ‘late’ populations. The rate
of this change in O. s. superba cannot be assessed as it takes place in a poorly fossiliferous part
of the sequence.

None of these changes is considered to be of sufficient magnitude to indicate the formation of
a new species but they can be described in an analogous way. The O. gracilis interval masks the
critical evidence which would indicate whether a single lineage or a branching event is represented
in the evolution of O. superba (see text-fig. 1). In the former case, O. s. superba would have been
derived from O. s. creta by a reversal of the earlier trend ( = ‘detour trend’ of Henningsmoen 1964).
Alternatively, O. s. creta may represent a side branch of an otherwise fairly conservative lineage
from O. s. cobboldi to O. s. superba , a substantial part of which is not represented (for ecological
reasons) in the Onny section. In either model, the appearance of O. s. creta (and possibly late O.
s. superba) might best be viewed as an example of punctuated equilibria (Gould 1985 and references
therein). In the single lineage hypothesis it would also conform to the ‘punctuated gradualism’
documented by Malmgren et al. (1983, 1984) in planktonic foraminifera. This was reinterpreted
by Gould (1985, p. 10) as ‘punctuated anagenesis’ and reflects changes of short duration (but with

OWEN AND INGHAM: CARADOC TRILOBITE

839

intermediates) separating periods of stasis but without lineage splitting. Maynard Smith (1983) has
discussed the possible genetic controls on stasis and punctuation.

Ecological control

Both suggested evolutionary models for the changes in O. superba involve at least some ecological
control on the presence or absence of particular subspecies, or even O. superba itself, in the type
Onnian. An extreme development of this would be to regard the various subspecies as entirely
ecologically controlled morphologies. This could be as portions of an intergradational cline
distributed along an environmental gradient (e.g. Cisne et al. 1982) or as ecophenotypes developed
in response to particular sets of environmental conditions (e.g. Mayr 1963; Johnson 1981; Hurst
1978, 1982 and references therein).

The only major lithological changes in the type Onnian are at the base of the Onny Formation
where the sparsely fossiliferous laminated mudstone is developed and overlain by blocky mudstone.
More subtle environmental controls (or selection pressures) must have operated earlier, yet it is in
these lower three zones that a coherent (if simple) positioning of subspecies in a morphoseries can
be postulated. Taking the two most variable features— the number of pits per arc (especially Ej)
and the shape of the fringe— the series extends from O. s. creta with a low pit count and strongly
swollen posterior fringe, through O. s. cobboldi with an increased pit count and gentle posterior

text-fig. 6. Reconstructions in dorsal view of cephala of the three subspecies of Onnia superba recognized
herein, showing typical morphological differences between them, c. x 3. A, O. s. cobboldi (Bancroft), b, O. s.
creta subsp. nov. c, O. s. superba (Bancroft), early form, d, O. s. superba (Bancroft), late form (which includes

the type material of O. s. superba).

840

PALAEONTOLOGY, VOLUME 31

swelling, to early O. s. superba with a similar or even larger number of pits and a flatter fringe
profile (text-fig. 6). The Onny River O. superba faunas began, therefore, in the middle of this
morphoseries and after a period of stability were replaced, with slight gradation in terms of pit
number, by the O. s. creta ‘end member’. After another period of stability a much more profound
environmental shift brought a different species, O. gracilis, into the area. This may reflect a
deepening event as the broadly contemporaneous appearance of O. gracilis at Welshpool is thought
to have been in response to the ‘Nod Glas transgression’ (Dean 1963; Cave 1965). Whatever the
change was, it was sufficient for the ‘early’ O. s. superba morphology to be ‘missed out’. The
subspecies only appeared later with, and eventually completely replacing, O. gracilis— perhaps
indicating a slight regression. The base of the Onny Formation and the broadly coeval appearance
of late O. s. superba is associated with a depleted fauna that was interpreted by Hurst (1979a,
pp. 23 1 -232) as reflecting poorly oxygenated conditions caused by upwelling of oxygen-poor waters
from deeper levels in the basin. Late O. s. superba shows an increased pit count and in this respect
can be placed at the ‘high’ end of the postulated morphoseries. Its fringe profile, however, is closer
to that of O. s. cobboldi than early O. s. superba, and thus does not fit this simple picture.

The subdivision of O. superba into subspecies adopted in this paper implies either a punctuated
evolutionary explanation or at least discrete ecologically controlled, entities rather than arbitrary
points along completely intergradational chronoclines, topoclines, or ecophenotypic series. The
subdivision is, however, partly a pragmatic solution to the available data. Any of these hypotheses
could be correct but they can only be tested if O. superba is found outside its type locality.

SYSTEMATIC PALAEONTOLOGY

The terminology used herein is that advocated by Ingham (1974; see also text-fig. 2 herein) and Hughes et
al. (1975), and pit counts refer to half-fringe values. Although we cite ranges in variation in fringe pit
distribution in diagnoses, we do not intend the values from our samples to be completely prescriptive. Thus
the terms ‘approximately’ and ‘about’ are used in order to avoid (say) a specimen with one more pit in an
arc being excluded from the taxon or a new diagnosis being required. Specimens are housed in the Hunterian
Museum, Glasgow University (HM) and the British Museum (Natural History) (BM).

EXPLANATION OF PLATE 74

Figs. 1-13. Onnia superba superba (Bancroft) from the O. s. superba Local Range Zone, Onnian Stage, Onny
River section, south Shropshire. Note that figs. 1-6, 8, 9 are from early populations and figs. 7, 10 13
from late populations. These are also from the uppermost Acton Scott and Onny formations respectively.
All specimens testate or largely so unless otherwise stated. 1, BM In520 11/1, oblique anterolateral view of
cephalon, Bancroft Collection loc. Pc (equivalent to sample N herein), x3. 2, HM A 1 5 1 45, frontal view
of cephalon, sample N, x 3. 3, BM In49028, dorsal view of almost complete individual, Bancroft loc. Pc
(= N herein), x 3, figured by Dean (1960, pi. 19, fig. 1) as 'OP cobboldi' in the mistaken belief that it came
from the type locality of that form (Bancroft’s Px, our B); the specimen bears Bancroft’s original loc. Pc
label. 4 and 5, HM A21759, oblique anterolateral and dorsal views of cranidium showing healed severe
damage to right side of fringe, sample N, x 3 and x4 respectively. 6, HM A21758, oblique anterolateral
view of cephalon, sample N, x 3. 7, HM A21751, partially exfoliated cephalon with parts of three thoracic
segments, sample U, x 2. 8, HM A15148, dorsal view of partially exfoliated cephalon showing long

occipital spine, sample P, x 3. 9, HM A21741, oblique anterolateral view of complete individual sample
N, x 3. 10, BM In49029, dorsal view of exfoliated almost complete specimen with ventral mould of lower
lamella of fringe; cliff section, x 1-5, figured by Dean (1960, pi. 19, figs. 13 and 14). 11, HM A21757,

internal mould of lower lamella of fringe, sample U, x 3. 12, HM A21767a, dorsal view of rather flattened
cranidium, cliff section, x 3. 13, HM A217536 and HM A217546, latex peel of external moulds of small
cranidium and cephalon respectively, both showing broad reticulated band on mesial part of glabella,
loc. U, x 6.

PLATE 74

*

?*ai*

OWEN and INGEIAM, Onnia

842

PALAEONTOLOGY, VOLUME 31

Family trinucleidae Hawle and Corda, 1847
Subfamily marrolithinae Hughes, 1971
Genus onnia Bancroft, 1933

Type species. Cryptolithus superbus Bancroft, 1929 6, p. 95, pi. 2, fig. 10, from the Onny Formation (as
redefined herein), Onny River section, south Shropshire, England; by original designation.

Discussion. The recognition of the In cut-off on the fringe of Onnia , together with the identification
of the position of the true girder, undoubtedly places Onnia in the Subfamily Marrolithinae (see
Ingham 1974, p. 59; Hughes et al. 1975, p. 570). It is common for marrolithines to exhibit lateral
fringe swelling and pit enlargement (seen in Marrolithus, Marrolithoides , Costonia, and some
Deanaspis ), although the tendency is by no means confined to this subfamily, having been
independently developed in the Trinucleinae ( Telaeomarrolithus ) and Hanchungolithinae ( Ningkian -
olithus). Some Onnia taxa also exhibit this feature to a degree, none more so than O. s. creta subsp.
nov. (described below).

Exfoliated specimens of Onnia in all our samples show areas of distinctive, closely spaced pitting
(in reality they are spiculate areas on the underside of the test). One is a roughly rectangular area,
situated immediately anterior to the anterior fossula, i.e. between the fossula and the innermost
arc on the fringe. The other area is longer and crescentic in form and occupies a similar position
with respect to the fringe but at the lateral periphery of the genal lobes (text-fig. 7g). These features
may be areas of muscle attachment.

Onnia superba (Bancroft, 1929 b)

Plates 74-76; text-figs. 1, 3-7; Table 1

Emended diagnosis. Profile of upper lamella of fringe almost planar or variably convex, moderately
declined. Arcs E! and I, complete, containing approximately 14-29 and 14-22 pits respectively.
Arc In complete frontally and truncated posteriorly by I3 which extends to the posterior margin
but lacks about 3-10 pits mesially. Posterolaterally pits of I3, In and the anterior F pits may share
sulci. I2 complete posteriorly but with up to about 4 pits absent mesially.

Discussion. Our analysis of population samples of Onnia from the Onny River indicates that O.
superba and O. cobboldi should not be maintained as separate species and that they are best viewed as
subspecies. Both taxa were established by Bancroft in 1929 but although ‘ cobboldi ' was described
earlier in his paper (1929 b, pp. 92-94 cf. 95-96), as First Revisers under ICZN article 24(b) (1985),
we here choose superba as the senior specific name. Cryptolithus superbus was designated the type

explanation of plate 75

Figs. 1-1 F Onnia superba cobboldi (Bancroft). Acton Scott Formation, O. s. cobboldi Local Range Zone,
Onnian Stage, Onny River section, south Shropshire. All specimens testate unless otherwise stated. 1 and
2, HM A21761, oblique anterolateral and frontal views of cephalon, sample B, both x 3. 3, HM A21732,
oblique anterolateral view of cranidium, sample B, x 3. 4, HM A 151 58/1, 2, oblique views of two cranidia
the smaller with reticulation on the mesial glabella and genal lobe, the larger smooth, sample D, x4.
5, HM A 1 5 1 84, oblique anterolateral view of partly exfoliated cephalon, sample E, x 3. 6, HM A 1 5 1 83/1 ,
oblique anterolateral view of portion of damaged cephalon showing subdued I, swelling, sample F, x 3.
7, HM Al 5159/1, dorsal view of small, partly compressed cranidium showing deeply pitted genal lobes
and fine reticulation in narrow mesial band on glabella, sample D, x9. 8, HM A 151 78/1 oblique

anterolateral view of part of cranidium showing subdued R swelling, sample F, x 4. 9, HM A2 1 742/1 ,
oblique anterolateral view of cranidium with very subdued L swelling, sample B, x4. 10, HM A21734,

oblique anterolateral view of incomplete cranidium showing subdued R swelling, sample A, x 4. 11, HM
A 1 5 1 59/2, oblique anterolateral view of partly exfoliated cranidium showing slight L swelling, sample D,
x 6.

PLATE 75

OWEN and INGHAM, Onnia

844

PALAEONTOLOGY, VOLUME 31

species of Onnia by Bancroft in 1933 and it would be unduly disruptive to synonymize this well-
established name with the hitherto less well-understood O. cobboldi. Moreover, topotype material
of O. superba is widely dispersed through British and other museum collections. In the interests
of stability therefore, we designate O. superba as the preferred species name.

Three subspecies of O. superba are recognized here. Dean (1960) gave full descriptions and
synonymies of two of these, O. s. superba (as O. superba) and O. s. cobboldi (as 0.1 cobboldi ), and
thus only emended diagnoses are given herein. However, specimens of both subspecies are illustrated
along with summary statistics of the fringe pit distribution. More detailed histograms of fringe
data have been deposited with the British Library, Boston Spa, Yorkshire, UK, as Supplementary
Publication No. SUP 14034 (5 pages). Discussion of all three subspecies is given after the description
of O. s. creta subsp. nov. It should be stressed, however, that Dean (1960) misinterpreted the first
internal pseudogirder for the true girder; hence his descriptions refer to two E arcs, whereas only
Ej is actually present (see Hughes et a/. 1975, p. 575).

Onnia superba superba (Bancroft, 19296)

Plate 74; text-figs. 1, 3-5, 6c, d; Table 1
19296 Cryptolithus superbus Bancroft, p. 95, pi. 2, fig. 10.

1933 Onnia superba ; Bancroft, table 1 (non Dufton Shales = O. pusgillensis Dean, 1961).
non 1948 Onnia superba (Bancroft); Bancroft in Lament, p. 416 ( = O. pusgillensis Dean, 1961).
1960 Onnial cobboldi (Bancroft); Dean, pi. 19, fig. 1.

1960 Onnia superba (Bancroft); Dean, pp. 133 136, pi. 19, figs. 4-6, 8, 9, 11, 13, 14.

1960 Onnia aff. superba (Bancroft); Dean, pp. 136 137, pi. 19, fig. 10.

1975 Onnia superba (Bancroft); Hughes et al ., pi. 9, fig. 107.

19796 Onnia superba (Bancroft); Hurst, p. 210, fig. 36.

For complete synonymy see also Dean (1960, p. 133).

Holotype. An internal mould of a cephalon (BM In42070) from the upper part of the Onny Formation (level
of sample V herein) (upper Onnian), cliff section, Onny River, south Shropshire.

Occurrence. Some complete specimens are known and disarticulated sclerites are abundant in the Onny cliff
section and at some horizons in the river bed (when not covered by river gravels), in the upper 24 m of the
type Onnian Stage. This distribution constitutes the O. s. superba Local Range Zone and extends across the
boundary between the Acton Scott and Onny formations as recognized herein (text-fig. 1).

Emended diagnosis. External surface of glabella and genal lobe smooth except in small specimens.
Fringe moderately declined, upper lamella only gently convex in early forms, more so in later
populations. Arcs E! and Ix complete, containing approximately 20-29^ and 1 54-224 pits
respectively. Arc In contains about 12-17 pits, cut off posteriorly by I3 which lacks approximately
3-9 pits mesially. Up to about 3 I2 pits missing frontally.

Onnia superba cobboldi (Bancroft, 19296)

Plate 75; text-figs. 1, 3-5, 6a; Table 1
19296 Cryptolithus cobboldi Bancroft, p. 92, pi. 2, figs. 6 and 7.

1960 Onnial cobboldi (Bancroft); Dean, pp. 128-132, pi. 19, figs. 3 and 12 (non fig. I = O. superba
superba).

1975 Onnia cobboldi (Bancroft); Hughes et a! ., pi. 9, figs. 104-106.

1979a Onnia cobboldi (Bancroft); Hurst (pars), pp. 204, 227 (samples 97, 98 only non 35, 99-102 = O.

superba creta subsp. nov), fig. 16.11
19796 Onnia cobboldi (Bancroft); Hurst, p. 210 (pars), fig. 37.

1983 Onnia cobboldi (Bancroft); Owen, pi. 34, figs. 1 and 5.

For complete synonymy see also Dean (1960, p. 128).

OWEN AND INGHAM: CARADOC TRILOBITE

845

Lectotype. Selected by Dean (1960, p. 132), an incomplete cephalon (BM In42074) from the upper part of
the Wistanstow Member of the Acton Scott Formation (Bancroft loc. Px = loc. B herein) (lower Onnian),
Onny River section, south Shropshire.

Occurrence. Disarticulated sclerites are abundant at the type horizon and levels immediately above and below
it. They are less common in the upper part of the O. s. cobboldi Local Range Zone (text-fig. 1). Complete
specimens are extremely rare.

Emended diagnosis. External surface of glabella and genal lobes smooth in mature specimens,
reticulated in small individuals. Upper lamella fairly steeply declined; fringe convex upwards,
with some specimens also gently swollen along the lateral part of arc Ij. Arcs Et and Ij complete,
comprising approximately 18-26} and 14-22-} pits respectively. Arc In contains about 9-17 pits,
cut off posteriorly by I3 which lacks approximately 3-10 pits mesially. Up to about 3 I2 pits
missing frontally.

Onnia superba creta subsp. nov.

Plate 76; text-figs. 1, 3-5, 6b, 7; Table 1

1979a Onnia cobboldi (Bancroft); Hurst (pars), pp. 204, 227 (samples 35, 99 102).

19796 Onnia cobboldi (Bancroft); Hurst (pars), p. 210 (pars).

Holotype. A testate cephalon (HM A 15087) from 14-8 m above the base of the Onnian Stage (sample H,
text-fig. 1), upper Acton Scott Formation (O. s. creta Local Range Zone), Onny River section, south
Shropshire.

Paratypes. Two cephala (HM A15083, A15086/2), four cranidia (HM A 15067/1, A15073/1, A15075, A15076),
and a lower lamella (HM A 15067/2). Other skeletal parts are not included here as the best specimens are
from other sample horizons within the local range zone.

Occurrence. Disarticulated sclerites are common at four horizons within the 5 m of the O. s. creta Local
Range Zone in the Onny River section. Complete specimens are known.

Derivation of name. From the Latin cretus, arisen; sprung/descended from; born of— referring to the possible
derivation of this subspecies from the stratigraphically lower subspecies in the Onny River section.

Diagnosis. External surface of glabella and genal lobe variably reticulate, pitted, or smooth. Upper
lamella of fringe markedly convex along very strong ridge-like swelling over lateral part of I, arc,
beginning between about R5 and R9 beyond which the pits of Ij are also enlarged. Arcs Ej and
I, complete, containing approximately 14-23 and 15-20} pits respectively. Arc In contains about
8}- 16} pits, cut off posteriorly by I3 which lacks approximately 3-10 pits mesially. Up to about
4 12 pits missing frontally.

Description. Cephalon almost semicircular in outline (excluding spines) but with sagittal length slightly more
than half the posterior width. Strongly swollen (tr. ), clavate, glabella achieves maximum width a short
distance behind anterior fossula. Outer part of occipital ring ridge-like, directed abaxially downwards and
forwards at about 45° to the sagittal line and defined anteriorly by deep, slot-like apodemal pit. Mesially,
occipital ring differentiated from rest of glabella by only a slight break in slope and extended rearwards and
slightly upwards as a stout spine whose sagittal length is equal to almost half that of preoccipital part of
glabella. The rearward tapering of this spine is continuous with the general narrowing of rest of glabella. LI
developed as diminutive swelling marked anteriorly by small pit-like SI. Axial furrow broad and shallow
bearing small but distinct fossula near its anterior end. Genal lobe strongly convex (tr., exsag.), quadrant
shaped to reniform in outline. Posterior border narrow, convex (exsag.) directed transversely for a short
distance before being moderately deflected rearwards and downwards to form posterior margin of fringe;
inner part defined anteriorly by shallow furrow bearing posterior fossula distally. Long genal spines diverging
gently at first but gradually becoming subparallel distally.

Many mature specimens and some smaller individuals have totally smooth glabella and genal lobes.
Nevertheless, some mature specimens show surface sculpture. Pseudofrontal lobe of glabella in some specimens
bears an ill-defined, broad, mesial strip of sculpture which is manifested either as a fine, occasionally coarser

846

PALAEONTOLOGY, VOLUME 31

reticulation or sometimes as a fine pitting in which pits may be clustered together in irregular groups of two
to four, particularly towards front of glabella (PI. 76, figs. 4-6; text-fig. 7a, b). This kind of pattern is
occasionally also found on genal lobes, albeit in very subdued form. More commonly, sculpted specimens
show fairly evenly spaced, shallow pits on genal lobes, except for their peripheral regions which are always
smooth. Very small specimens have both glabella and genal lobes reticulated. An ill-defined glabellar node
is situated at about the midlength of preoccipital part of glabella and at its highest point (sometimes difficult
to detect on external surface of sculpted specimens, but invariably visible on internal moulds). Shape and
position of spiculate areas on inner surface of test (see discussion of genus) in O. s. creta corresponds with
those peripheral parts of genal lobes which are invariably smooth on outer surface.

Fringe moderately steeply declined mesially, upper lamella increasingly more convex upwards abaxially.
This is caused by development of an almost ridge-like swelling along course of arc Il5 beginning between Rs
and R9 (mean and mode 7, sample standard deviation 1, n = 82) such that inner part of fringe is gently
declined, almost horizontal, or even concave upwards, and outer part, along E^ is so steeply declined that
a substantial part is not visible in dorsal view. Ij pits are noticeably enlarged along this inflated sector of
fringe. Details of fringe pit number given on text-figs. I, 3-5, table 1, and in the supplementary material in
deposition. Arcs and IL complete, containing 14-23 and 15-20| pits respectively in samples studied. In
contains 8^-16| pits and is cut off posteriorly by I3 which contains 8^-1 4| pits and lacks 3-10 pits mesially.
12- 18 pits present in I2 which lacks up to 4 pits mesially. F pit series begins between R8 and R15. Lower
lamella fairly steeply declined, lacking any swelling equivalent to that along Ij on upper lamella. Figured
specimens (PI. 76, fig. 2; text-fig. 7c, f) all show clearly the distinction between the true girder and the first
internal pseudogirder.

Thorax typically trinucleid in plan, comprising six segments of which third and fourth occupy greatest
width. Axis moderately convex but ill defined, occupying little more than one-fifth width of thorax throughout.
Each axial ring is gently convex (sag., exsag.) and narrowest mesially, posterior margin arched forwards
somewhat. Laterally, a shallow furrow originating in axial furrow at posterolateral extremity extends across
each ring and shallows before becoming confluent with its counterpart. Articulating furrows sharply incised,
defining simple articulating half-rings. Pleurae transverse for most of their length but deflected sharply
posteroventrally towards their tips at a distinct fulcrum. Terminations blunt on all but first segment, which
is shorter and more tapered to a rounded point. Pleural furrows broad and deeply impressed, confluent with
axial articulating furrows, directed gently rearwards, deepest where they traverse the fulcrum but end abruptly
just inside pleural termination. Convex posterior band thus tapers abaxially and ridge-like anterior band
expands to fulcrum.

Pygidium broadly triangular in outline, larger specimens have sagittal length about 35 % of maximum
anterior width, although smaller specimens proportionally longer. Posterolateral margins slightly sinuous in
outline, with shallow concavities to either side of posterior, obtusely rounded termination. A posteriorly
widening convex marginal band (sag., exsag.) extends around lateral and posterior margin. It is steeply
declined and sharply recurved ventrally into a narrow doublure. Dorsally, the angulation between marginal
band and pleural lobes is elevated as a narrow ridge. Axis only gently convex (tr.), occupying a little over
one-fifth of maximum pygidial width anteriorly, relatively ill-defined by shallow axial furrows which converge
gradually rearwards and become effaced before they reach marginal band. First axial ring well-defined both
anteriorly and posteriorly by sharp furrows which bear apodemal pits abaxially; it is gently convex (sag.,

EXPLANATION OF PLATE 76

Figs. I -9. Onnia superba (Bancroft) creta subsp. nov., Acton Scott Formation, O. s. creta Local Range Zone,
Onnian Stage, Onny River section, south Shropshire. All specimens testate unless otherwise stated. 1 and
3, HM A15087, oblique anterolateral and dorsal views of holotype cephalon, sample H, both x 3.
2, HM A 15067/1, 2, oblique anterolateral view of cranidium and oblique ventral view of lower lamella,
both paratypes, sample H, x 4. 4, HM A21745, oblique anterolateral view of cephalon with reticulate
glabella and pitted genal lobe, sample I, x 6. 5, HM A21746, dorsal view of large cephalon with finely
reticulate glabella and sparsely pitted genal lobes, sample J, x 2. 6, HM A21747, oblique anterolateral
view of portion of cranidium showing reticulate glabella and pitted genal lobes, sample J, x4. 7, HM
A 15075, oblique anterolateral view of paratype cranidium and incomplete thorax (pygidium present but
not seen in this view), sample H, x 3. 8, HM A 15083, dorsal view of partly exfoliated paratype cephalon,
sample H, x 3. 9, HM A 15086/2, oblique anterolateral view of paratype cephalon, sample H, x4.

PLATE 76

OWEN and INGHAM, Onnia

848

PALAEONTOLOGY, VOLUME 31

exsag.) and gently arched forwards. Second ring a little more arched anteriorly but less well-defined posteriorly,
the shallower furrow there still bears traces of apodemal pits abaxially. Successive rings progressively less
well defined, particularly laterally but mesially they are a little clearer and may be impressed there as short,
straight, transverse furrows with shallow depressions at their outer extremities. Seven or eight rings are
present in all. Pleural lobes relatively depressed but gently convex adjacent to axis and gently concave
abaxially. They are traversed by four distinct and slightly divergent pleural ribs, the anterior one or two
following a slightly sigmoidal path towards submerged rim, which they almost reach. Ribs confluent with
first four axial rings. Fifth pair of ribs barely discernible. Surface of pygidium largely smooth but marginal
band and submarginal rim bear many fine anastomosing thread-like ridges.

Discussion of O. superba subspecies. The changes in pit number of successive populations of the
subspecies of O. superba are shown in text-figs. 1 and 4 and table 1, whilst text-figs. 3 and 5
summarize the differences in pit distribution between the separate subspecies as a whole. These
changes and differences are discussed in the section on ‘ Onnia in the type Onnian’ (above). Suffice
it to note here that the fringe pitting of O. s. creta subsp. nov. differs from that of the other two
subspecies in its lower mean number of pits in each arc. This is especially true in arc E3 where the
lower part of its range extends well below the values of the other subspecies. O. s. superba , however,
has a significantly higher mean value for arcs Ej and I3 than even O. s. cobboldi , with the former
arc showing a marked overall increase in pits from early to late samples of the nominate subspecies.
In addition to pit numbers, O. s. superba can usually be distinguished by the clearer separation of
arcs I2 and I3 laterally. Moreover, the profile of the upper lamella ranges from near planar in early
O. s. superba , through gently convex upwards in late O. s. superba and strongly convex in O. s.
cobboldi , to the extreme convexity caused by the highly inflated lateral and posterior parts of arc
Ij in O. s. creta. Some specimens of O. s. cobboldi have a gentle swelling here but never as strongly
developed as in O. s. creta.

Outside the Anglo-Welsh area, species of Onnia have been described from Caradoc and Ashgill
rocks in north-west France, Iberia, Czechoslovakia, and Morocco (Hughes et al. 1975, pp. 574-
575). Whilst it is clear that some of these peri-Gondwanan species are similar in many respects to
O. superba subspp., most are in need of modern documentation and description. None has the
markedly swollen posterior part of I3 shown by O. s. creta. O. [or Deanaspisl ] vysocanensis Pribyl
and Vanek, 1980 (pp. 268-269, pi. 3, figs. 1-3; text-fig. 1 a, 6), from the middle Caradoc Zahorany
Group in Bohemia, has a very much broader glabella than is seen in the British species and there
is a marked prolongation of the mesial part of the pygidial border. Details of the fringe are not
clear from Pribyl and Vanek’s photographs, except that I3 is absent at least anteriorly and
anterolaterally. This arc is also missing in some illustrated specimens of O. abducta Pribyl and
Vanek, 1969, from the upper Caradoc Bohdalec Formation in Bohemia (see Pribyl and Vanek
1980, pi. 6, fig. 6; Cech 1975, pi. 4, fig. 1). Examination of topotype specimens of O. abducta in
the British Museum (Natural History) has confirmed this and has also shown that the pit
distribution for most arcs lies well within the overlap in range shown by the three subspecies of
O. superba , although the number of pits in E, is at or slightly beyond the upper part of the range
in O. s. cobboldi. Like O. [Dl] vysocanensis , the posterior margin of the pygidium of O. abducta
has a sinuous outline.

The material described by Hammann (1976, p. 40, pi. 1, figs. 1-10; pi. 2, figs. 11-14; text-fig. 3;
table 2) as OP. n. sp. aff. grenieri (Bergeron), from probable Ashgill strata (W. Hammann, pers.
comm. 1984) in the eastern Sierra Morena, Spain, belongs in Deanaspis, a genus more typical of
somewhat older strata. The girder and first internal pseudogirder are equally well developed
anteriorly and anterolaterally, with the girder the more strongly developed beyond this. ‘0. grenieri ’,
redescribed by Coates (1966, pp. 84-87, text-fig. 5 a-e) on the basis of type and other material
from the early Caradoc ‘Vauville Formation’ (now La Sangsuriere Formation, Hammann et al.
1982, p. 8), also appears to have a moderately well-developed true girder frontally. This species
also may be better placed in Deanaspis. It is at least broadly similar to O. s. superba and O. s.
cobboldi in its cephalic and pygidial characters but better material needs to be described before a
detailed comparison can be made.

OWEN AND INGHAM: CARADOC TRILOBITE

849

text-fig. 7. Oiuiia superba (Bancroft) creta subsp. nov., Acton Scott Formation, O. s. creta Local Range
Zone, Onnian Stage, Onny River section, south Shropshire. All specimens testate unless otherwise stated.
a, c, HM A21738, dorsal and ventral views of cephalon in enrolled individual, sample J, both x 3. b. HM
A 15073/1, frontal view of paratype cranidium with fine glabellar reticulation, sample H, x6. d, HM A21748,
dorsal view of pygidium, sample I, x 6. e, HM A15076, dorsal view of paratype small cranidium with
reticulate glabella and genal lobes; note Ij swelling subdued, sample H, x9. F, HM A21763, ventral view of
lower lamella, sample I, x3. G, HM A21766, anterolateral view of part of damaged cephalon in which the
right genal lobe has been stripped of test revealing, on internal mould, impressions of spiculate areas adjacent
to anterior fossula and lateral margin of genal lobe, sample I, x 4. h, HM A21743, dorsal view of pygidium,
sample H, x6. i, HM A21740, dorsal view of partly exfoliated small cranidium showing reticulate genal lobe
and smooth internal mould of glabella, sample I, x 9.

850

PALAEONTOLOGY, VOLUME 31

Onnia gracilis (Bancroft, 192%)

Plate 77; text-figs. 1, 4, 8

19296 Cryptolithus gracilis Bancroft, p. 94, pi. 2, figs. 8 and 9.

1960 Onnia gracilis (Bancroft); Dean, pp. 130-132, pi. 19, figs. 2 and 7.

1962 Onnia gracilis (Bancroft); Dean, p. 84, pi. 8, figs. 12 and 13.

1965 Onnia gracilis (Bancroft); Cave, pp. 282, 286, 287, pi. 12, figs, a, b, m, q.

1975 O. gracilis (Bancroft); Hughes et al ., p. 574.

1979a Onnia gracilis (Bancroft), Hurst, p. 204 (samples 32-34, 36, 37).

19796 Onnia gracilis (Bancroft); Hurst, p. 210.

1983 Onnia gracilis (Bancroft); Owen, pi. 34, fig. 2.

For a complete synonymy see also Dean (1960, p. 130).

Lectotype. Selected by Dean (1960, p. 132), an incomplete cephalon (BM In42074) from the upper part of
the Wistanstow Member of the Acton Scott Formation (= samples M and N herein) (middle Onnian), Onny
River, south Shropshire.

Occurrence. Rare complete specimens and abundant disarticulated sclerites occur in the 4-2 nr of the O.
gracilis Acme Zone in the Onny River section, and a few sclerites are known from the lowest part (sample
N) of the overlying O. s. superba Local Range Zone. Bancroft's locality Pc was largely in the O. gracilis Zone
but the presence of a few specimens of O. s. superba indicate that the lowest part of the overlying zone was
also sampled. Our two samples M and N more precisely delimit the zonal boundary and demonstrate the
nature of the co-occurrence of the two taxa. Disarticulated sclerites are also known from possible equivalents
of the Onny River O. gracilis Zone at Welshpool (Cave 1965) and Cross Fell (Dean 1962). The species is
also a rare component of strata of probable late Actonian age at Heath Brook near Cardington, south
Shropshire.

Emended diagnosis. External surface of glabella and genal lobe smooth. Fringe moderately declined,
surface of upper lamella essentially planar. Arcs E,, Ils and I2 complete, containing about 30-41^,
19-J-27, and 19-25y pits respectively. Arc In of about 124-22 pits cut off posteriorly by either I3
or (when present) I4 which anteriorly lack 1-6 and (when present) 4-10 pits respectively.

Description. Dean (1960) gave an extensive description of O. gracilis which need not be repeated here save
to enlarge upon and update his assessment of the fringe pitting. Number of pits in arcs E1; I1; and In in
successive samples of O. gracilis are summarized in text-figs. 1 and 4, whilst text-fig. 8 shows total range of
pits in these arcs together with arcs I2_4 and radius number of first pits in arcs I3, I4 and F pit series. Two
distinct morphs can be recognized based on presence or absence of arc I4. Moreover, when this arc is
developed, it comprises at least ten pits. Like arc I3, it is always incomplete frontally. Range, mean, and one
standard deviation on each scale of mean is shown for various fringe variables of the two morphs on text-
fig. 8. For most features, there is little difference other than a slight increase in pit number when I4 is absent.
In the case of arc In, however, this increase is substantial. The three samples of O. gracilis from the O. gracilis
Zone in the Onny River show a progressive decrease in percentage of specimens lacking arc I4, from 94 %

EXPLANATION OF PLATE 77

Figs. 1-12. Onnia gracilis (Bancroft), Acton Scott Formation, O. gracilis Acme Zone, Onnian Stage, Onny
River section, south Shropshire. All except fig. 4 from Bancroft Collection loc. Pc (equivalent to our
samples M and N, probably the former). All specimens testate or largely so unless otherwise stated. 1, BM
In49032, dorsal view of paralectotype individual showing repaired damage to left side of fringe, arc I4
present adjacent to In anterolaterally, x 3, figured by Dean (1960, pi. 19, fig. 7). 2 and 5, BM In52014/2,
oblique anterolateral and lateral views of cephalon lacking arc I4, x 3. 3, BM In52010, anterolateral view
of cephalon lacking I4, x3. 4, HM A15018, dorsal view of partly exfoliated crushed cranidium showing
short occipital spine and I4 present, sample L, x 2. 6-8, BM In52014/3, oblique anterolateral, dorsal, and
lateral views of cranidium, x3. 9 and 12, BM In52014/1, dorsal and oblique anterolateral views of
somewhat flattened cranidium lacking I4, x 3. 10, BM In52017/2, oblique anterolateral view of cranidium
lacking I4, x 3. 11, BM In520 17/3, oblique anterolateral view of cranidium lacking I4, x 3.

PLATE 77

* »trSw''
* * **«■*§'

OWEN and INGHAM, Onnia

852

PALAEONTOLOGY, VOLUME 31

I4 absent n 1 0

1 4 present

n 33

17 19 21

Pits in I 3

23

I4 absent n 16
I4 present |-

-NZN— I

12 14 16 18

R. No. first F pit

20

21 22 23

Pits in 1 0

I4 absent
I4 present

Pits in E1

text-fig. 8. Summary of the fringe pit distribution in Onnia gracilis in the type Onnian Stage based on our
own and Bancroft Collection specimens. Note that two morphs are present: one with and one lacking arc
I4. Differences in other fringe features between these morphs are indicated by the illustration of the range,
mean, and one standard deviation on each side of the mean.

(sample K, n = 18), through 91 % (L, n = 43), to 42 % (M, n = 19). Both specimens from lowest O. s. superba
Zone have this arc as do all eight suitably preserved BM specimens from Cross Fell Inlier. Similarly, the
four specimens from Welshpool have pits in I4.

Discussion. The broader fringe with more numerous E3 pits (text-fig. 1), arc I3 complete frontally,
and (in some specimens) I4 developed all serve to distinguish O. gracilis from the subspecies of O.
superba. In addition, the mean number of pits in arcs IL_3 is greater than the numbers of pits seen
in these arcs in O. superba but there is some overlap in total range (text-figs. 3, 4, 7). Only in the
case of specimens lacking I4, however, is this marked difference seen in the mean value ot pits

in !,r

Arc I4 is invariably present in O. s. pusgillensis Dean, from the Dufton Shales ot Cross Fell and
equivalent Onnian strata in the Cautley Mudstones near Cautley (Dean 1961, 1962; Ingham 1974,
pp. 60-63, pi. 10, figs. 1-18, text-figs. 20 and 21; see also text-fig. 2 herein). The complete development
of I2 frontally and, commonly, the greater anterior extension of I3 (only about 3-5 pits missing
frontally) also place the North of England form closer to O. gracilis than to O. superba. However,
the number of pits in arcs E! and I3 and the R number of the In cut off lie almost entirely within
the range of the latter species, being 224-31 (n = 14), 16-21 (n = 15), and 11-16 (n = 14) in the
more abundant, better preserved material from Cautley (Ingham 1974, text-fig. 21). The gently
convex profile of the upper lamella is also like that of late O. s. superba and some O. s. cobboldi.
A numerical taxonomic analysis of British trinucleids by Temple (1981, text-fig. 9) showed that
the species of Onnia plot close to each other in terms of the y3 and y2 axes of ordination, but

OWEN AND INGHAM: CARADOC TRILOBITE

853

whereas ‘ gracilis' has a low positive score on y3, ‘ superba ’, ‘ cobboldi' , and ‘pusgillensis’’ have low
negative scores. It must be stressed, however, that Temple’s approach differs markedly from that
used herein as it is based on a different set of attributes measured on small topotype samples of
each taxon.

O. \s-.’ pusgillensis differs from both O. superba and O. gracilis in its much more subdued first
internal pseudogirder which approaches the condition seen in Deanaspis where the girder and first
internal pseudogirder are developed to about the same extent (Hughes et al. 1975, p. 573). Thus
the North of England form shows a distinctive set of characters and is here given separate specific
status. O. pusgillensis may have been derived either from O. superba or O. gracilis but its affinities
are unclear.

Acknowledgements. We thank Dr L. R. M. Cocks and Mr S. F. Morris for access to specimens in their care,
two anonymous referees for their helpful comments, and Mrs Jenny Orr for typing the manuscript.

REFERENCES

Bancroft, b. b. 1929a. Some new genera and species of Strophomenacea from the Upper Ordovician of
Shropshire. Mem. Proc. Manchr lit. phil. Soc. 73, 33 65.

19296. Some new species of Cryptolithus ( s.l .) from the upper Ordovician. Ibid. 67-98.

1933. Correlation tables of the Stages Costonian-Onnian in England and Wales , 4 pp. Blakeney,
Gloucestershire (privately printed).

1945. The brachiopod zonal indices of the Stages Costonian to Onnian in Britain. J / Paleont. 19, 181
252.

1949. Upper Ordovician trilobites of zonal value in South-east Shropshire (edited by a. lamont). Proc.
R. Soc. B 1 36, 291-315.

bergstrom, s. M. and orchard, m. j. 1985. Conodonts of the Cambrian and Ordovician Systems from the
British Isles. Pp. 32-67, table I. In higgins, a. c. and Austin, r. l. (eds.). A stratigraphical index of
conodonts. Ellis Horwood, Chichester.

cave, r. 1965. The Nod Glas sediments of Caradoc Age in North Wales. Geol. Jb 4, 270 298.
cech, s. 1975. Cranidial reticulation and functional morphology of the cephalic fringe in Tnnucleidae
(Trilobita). Vest. ustr. Ust. geol. 50, 173-177.

cisne, J. l., chandlee, G. o., rabe, B. d. and cohen, J. a. 1982. Clinal variation, episodic evolution, and
possible parapatric speciation: the trilobite Flexicalymene senaria along an Ordovician depth gradient.
Lethaia, 15,325 341.

coates, a. 1966. Stratigraphie et paleontologie des synclinaux de Siouville et de Jobourg dans le Cap de la
Hague (Cotentin, Normandie). Bull. Soc. linn. Normandie , (10), 7, 77 103.
dean, w. t. 1958. The faunal succession of the Caradoc Series of south Shropshire. Bull. Br. Mas. nat. Hist.
(Geol.), 3, 191-231.

1960. The Ordovician trilobite faunas of south Shropshire, I. Ibid 4, 73 143.

1961. Trinucleid trilobites from the higher Dufton Shales of the Caradoc Series. Proc. Yorks, geol. Soc.
33, 119 134.

1962. The trilobites of the Caradoc Series in the Cross Fell Inlier of northern England. Bull. Br. Mus.
nat. Hist. (Geol.), 7, 65-134.

— 1963. The Ordovician trilobite faunas of south Shropshire, IV. Ibid. 9, 1 18.

- 1964. The geology of the Ordovician and adjacent strata in the southern Caradoc district of Shropshire.
Ibid. 251-296.

gould, s. J. 1985. The paradox of the first tier: an agenda for paleobiology. Paleobiology , 11, 2-12.
hammann, w. 1976. Trilobiten aus den oberen Caradoc der ostlichen Sierra Morena (Spanien). Senckenberg.
leth. 57, 35-85.

robardet, m. and romano, M. 1982. The Ordovician System in South-western Europe (France, Spain,
and Portugal). Pub/. Int. Union geol. Sci. II, 1-47.

HAWLE, I. and corda, a. j. c. 1847. Prodrom einer Monographic der bohmischen Trilobiten , 176 pp. Prague.
henningsmoen, g. 1964. Zig zag evolution. Norsk geol. Tidsskr. 44, 341-352.

854

PALAEONTOLOGY, VOLUME 31

hughes, c. p. 1970. Statistical analysis and presentation of trinucleid (Trilobita) fringe data. Palaeontology ,
13, I 9.

- 1971. The Ordovician trilobite faunas of the Builth Llandrindod Inlier, Central Wales, Part II. Bull.
Br. Mus. nat. Hist. (Geol.), 20, 117-182.

Ingham, j. k. and addison, r. 1975. The morphology, classification and evolution of the Trinucleidae
(Trilobita). Phil. Trans. R. Soc. B272, 537-607.

hurst, J. m. 1978. A phenetic strategy model for dalmanellid brachiopods. Palaeontology , 21, 535-554.

1979a. Evolution, succession and replacement in the type upper Caradoc (Ordovician) benthic faunas
of England. Palaeogeogr. Palaeoclimat. Palaeoecol. 27, 189 246.

19796. The stratigraphy and brachiopods of the upper part of the type Caradoc of south Salop. Bull.
Br. Mus. nat Hist. (Geol.), 32, 183-304.

1982. Dalmanellid brachiopod phenetics. Lethaia , 15, 342.
ingham, J. K. 1966. The Ordovician rocks in the Cautley and Dent districts of Westmorland and Yorkshire.
Proc. Yorks, geol. Soc. 35, 455-505.

1970 1974. A monograph of the upper Ordovician trilobites from the Cautley and Dent districts
of Westmorland and Yorkshire. Palaeontogr. Soc. [Monogr.], Part I (1970), 1 58; Part 2 (1974),
59 87.

- 1978. Geology of a continental margin, 2: middle and late Ordovician transgression, Girvan. In bowes,
d. r. and leake, B. E. (eds.). Crustal evolution in northwestern Britain and adjacent regions. Geol. Jl Spec.
Iss. 10, 163-176.

and wright, a. d. 1970. A revised classification of the Ashgill Series. Lethaia , 3, 233-242.
international commission on zoological nomenclature. 1985. International Code of Zoological Nomencla-
ture (3rd edn.), 338 pp. University of California Press, London, Berkeley and Los Angeles.

Johnson, a. l. a. 1981. Detection of ecophenotypic variation in fossils and its application to a Jurassic
scallop. Lethaia , 14, 277-285.

jones, c. R. 1987. Ordovician (Llandeilo and Caradoc) beyrichiocope Ostracoda from England and Wales.
Palaeontogr. Soc. [Monogr. ], Part 2, 77-114.

lamont, a. 1948. B. B. Bancroft’s geological work in the Cross Fell Inlier. Quarry Mgrs. Jl , 31,
416-418.

la touche, j. d. 1884. A Handbook of the geology of Shropshire, 91 pp. London and Shrewsbury.
malmgren, b. a., berggren, w. a. and lohmann, G. p. 1983. Evidence for punctuated gradualism in the Late
Neogene Globorotalia tumida lineage of planktonic foraminifera. Paleobiology, 9, 377 389.

1984. Species formation through punctuated gradualism in planktonic Foraminifera. Science,

225, 317-319.

maynard smith, j. 1983. The genetics of stasis and punctuation. Ann. Rev. Genet. 17, 1 1-25.
mayr, e. 1963. Animal species and evolution, 797 pp. Cambridge, Massachusetts.

Murchison, R. i. 1839. The Silurian System, founded on geological researches in the counties of Salop , Hereford,
Radnor , Montgomery, Caermarthen, Brecon , Pembroke, Monmouth, Gloucester , Worcester and Stafford;
with descriptions of the coalfields and overlying formations, xxxii + 768 pp. London.
orchard, m. j. 1980. Upper Ordovician conodonts from England and Wales. Geologica Palaeont. 14,
9-44.

Owen, a. w. 1980. The trilobite Tretaspis from the upper Ordovician of the Oslo region, Norway.
Palaeontology , 23, 715-747.

— 1983. Abnormal cephalic fringes in the Trinucleidae and Harpetidae (Trilobita). In briggs, d. e. g. and
lane, p. d. (eds.). Trilobites and other early arthropods: papers in honour of Professor H. B. Whittington,
F.R.S. Spec. Pap. Palaeont. 30, 241 247.

1985. Trilobite abnormalities. Trans. R. Soc. Edinb. (Earth Sci . ), 76, 255-272.

- 1987. The trilobite Tretaspis at the Middle -Upper Ordovician boundary in Vastergotland. Geol. For.
Stockh. Fbrh. 109, 259-266.

pribyl, a. and vanek, j. 1969. Trilobites of the family Trinucleidae Hawle et Corda, 1847 from the Ordovician
of Bohemia. Sb. geol. Ved. Praha (Paleont.), 1 1, 85-137.

1980. Neue Erkenntnisse uber einige Trilobiten aus dem bohmischen Ordovizium. Cas. Miner.
Geol. 25, 263 274.

savage, n. m. and bassett, m. g. 1985. Caradoc- Ashgill conodonl faunas from Wales and the Welsh
Borderland. Palaeontology, 28, 679-713.

temple, j. t. 1981. A numerical taxonomic study of species of Trinucleidae (Trilobita) from the British Isles.
Trans. R. Soc. Edinb. (Earth Sci.), 71, 213-233.

OWEN AND INGHAM: CARADOC TRILOBITE

855

WHITTINGTON, H. B., DEAN, W. T., FORTEY, R. A., RICKARDS, R. B., RUSHTON, A. W. A. and WRIGHT, A. D. 1984.
Definition of the Tremadoc Series and the series of the Ordovician System in Britain. Geol. Mag. 121,

17-33.

Typescript received 10 September 1987
Revised typescript received 11 November 1987

ALAN W. OWEN

Department of Geology
The University
Dundee DD1 4HN

J. KEITH INGHAM
Hunterian Museum
The University
Glasgow G12 8QQ

Note added in proof. In his recent review of British trilobites, Morris (1988, p. 155) has drawn attention to
an abstract written by us for the Palaeontological Association’s Evolutionary Case Histories Symposium in
1983. In that preliminary report about our work on Onnia we (as Ingham and Owen) suggested the name
O. cobboldi creta for the taxon here described as O. superba creta. The collection of abstracts for the meeting
was not paginated. It was intended purely for the information of likely delegates to the conference and was
distributed as an annexe to the Palaeontological Association Circular. Since the subsequent publication of the
1985 ICZN Code, the Circular includes a taxonomic disclaimer confirming that it is not valid for taxonomic
purposes. Thus the abstracts were not ‘for the purpose of providing a permanent scientific record’ (see Article
8(tf)(i) of the 1985 International Code of Zoological Nomenclature). Indeed, amongst the papers in the formal
publication arising from the meeting (Cope and Skelton 1985) there is, for example, a formal abstract (op.
cit. p. 185) by another author which was intended to be a permanent record of his work. No type specimen
(or even museum collection, cf. Morris 1988, p. 155) was indicated by us. Moreover, owing to a typographical
error in our abstract, the only phrase which could be construed as a taxonomic ‘description or definition’
(see Article I3(a)(i)) is nonsensical and reads ‘later specimens have the outer parts of arc I situated on a
distinct ridge’ [there are four or five I arcs present in all the Shropshire specimens of Onnia], Thus we consider
O. c. creta to be a nomen nudum and creta therefore is an available name. The formal establishment of Onnia
s. creta is in the present work.

cope, j. c. w. and skelton, p. w. (eds.). 1985. Evolutionary case hisories from the fossil record. Spec. Pap.

Palaeont. 33, 202 pp.

morris, s. F. 1988. A review of British trilobites, including a synoptic revision of Salter’s Monograph.

Palaeontogr. Soc. [Monogr.], 316 pp.

A NEW CAPITOSAURID AMPHIBIAN FROM THE
EARLY TRIASSIC OF QUEENSLAND, AND
THE ONTOGENY OF THE CAPITOSAUR SKULL

by A. A. WARREN and M. N. HUTCHINSON

Abstract. Capitosaurid temnospondyls are the most widespread and among the most abundant of the Triassic
amphibians, but their phylogenetic relationships are not well understood. The superfamily Capitosauroidea
(Capitosauridae, Benthosuchidae, and Mastodonsauridae) appears to be well characterized by several synapo-
morphies, but taxa within the superfamily are often less firmly established. A new capitosaurid species,
Parotosuchus ciliciae , is described from the earliest Triassic (Scythian A1 ) of Queensland. The hypodigm of the
new species consists of immature animals, including three identified as barely metamorphosed, which provide
the first information on the earliest post-larval growth stages of capitosaurids. Many character states present
only in juvenile capitosaurids are known to be retained in the adults of several Triassic temnospondyl families,
providing strong evidence that paedomorphosis was a dominant mode of evolutionary change in these groups.
P. aliciae is in some respects one of the most primitive capitosaurids, but it has several unique features which
do not indicate a sister-species relationship with any of the known Parotosuchus species.

Relationships among capitosaurs have until recently been assessed primarily on the basis of the
skull proportions, culminating in the system of indices developed by Welles and Cosgriflf (1965). We
have commented unfavourably on this approach (Warren and Hutchinson 1983) and have attempted
to establish relationships among temnospondyls by searching for shared derived character states.
Cladistic theories of relationships among capitosaurid genera have been suggested by Ingavat and
Janvier (1981) and Morales and Kamphausen (Kamphausen and Morales 1981; Morales and
Kamphausen 1984; Morales 1987). The difficulty of using the cladistic approach with capitosaurs
arises in part from the uncertain familial boundaries and lack of knowledge concerning interfamilial
relationships in the Superfamily Capitosauroidea. We have used as our starting point the scheme
of family-level phylogenetic relationships suggested by Warren and Black (1985), where the Family
Capitosauridae is regarded as belonging to a ‘capitosaurian’ lineage. This lineage also includes
the Rhinesuchidae, Benthosuchidae, Mastodonsauridae, Almasauridae, and Metoposauridae, and
possibly the Luzocephalidae (not recognized by Warren and Black 1985) and Lydekkerinidae
(tentatively assigned by them to the ‘trematosaurian’ lineage).

Within the capitosaurian lineage the Superfamily Capitosauroidea is usually considered to com-
prise three families (Capitosauridae, Benthosuchidae, and Mastodonsauridae; Morales 1987). At
present none of these three families has been adequately defined by means of derived character
states since the few potential apomorphies are all found in parallel elsewhere.

For example, the Benthosuchidae and Mastodonsauridae may be separated from the Capitosauri-
dae by the shared presence of paired, or butterfly-shaped, anterior palatal vacuities (Morales and
Kamphausen 1984). This presumed apomorphy is present also in other temnospondyl families, e.g.
the Trematosauridae. In adopting it. Morales and Kamphausen have chosen to accept parallel
development of the semi-closed and closed otic notch (in the ‘capitosaurids’ Parotosuchus and
Cyclotosaurus and the ‘benthosuchids’ Odenwaldia and Eocyclotosaurus) as more likely than parallel
development of paired anterior palatal vacuities, but no case has been presented for preferring the
former scenario.

Ingavat and Janvier ( 1981 ) defined a select group of genera as ‘Capitosauridea s.str .’ on the basis
of their having a well-defined suture between the exoccipital and pterygoid. This excludes P. gunganj ,

IPalaeontology, Vol. 31, Part 3, 1988, pp. 857-876.)

© The Palaeontological Association

858

PALAEONTOLOGY, VOLUME 31

P. helgolandicus, and the new parotosuchian described below, in all of which the pterygoid is
prevented from suturing with the exoccipital by a foramen (or notch), and also P. rewanensis in
which the two bones suture on the occiput. While we agree that these three species and some others
may form a plesiomorphic group of capitosauroids, we nevertheless include them in the Family
Capitosauridae. The fact that this character is also present in some (but not all) rhytidosteids and
some (but not all) brachyopids and that those genera lacking the character are the more plesio-
morphic members of their families indicates that it is a ‘grade' character perhaps associated with
increase in size of later genera.

In our opinion, the arguments used by Morales and Kamphausen (1984) for establishing the
boundaries of the Benthosuchidae and Capitosauridae, and of Inga vat and Janvier (1981) for
grouping the Capitosauridae s.str. are unconvincing. The Mastodonsauridae likewise may not be
distinct at the family level. However, there is good evidence that the genera included in these families
are close relatives, united by several apparently unique apomorphies, and can be discussed together
as capitosauroids. These genera were most recently divided by Morales (1987) into Capitosauridae
s.s. ( Parotosuchus , Eryosuchus , Paracyclotosaurus, Stenotosaurus, Cyclotosaurus), Benthosuchidae
s.l. (Benthosuchus, Benthosphenus, Kestrosaurus , ' Par otosanr us' lapparenti, Thoosuchus, Trematoteg-
men, Odenwaldia , Eocyclotosaurus), and Mastodonsauridae ( Heptasaurus , Mastodonsaurus). Chief
among the genera considered problematic by Morales is Wetlugasaurus which, although usually
associated with the primitive open-notched capitosaurids with a single anterior palatal vacuity, does
not have the frontals entering the orbital margins. He also noted that Parotosuchus is almost
certainly paraphyletic, since it includes most of the open-otic-notch capitosaurids.

We became particularly aware of these taxonomic problems when confronted with specimens of
a new species of capitosaurid recently collected from the Early Triassic Arcadia Formation of
Queensland. In determining that the very small juveniles described here were indeed capitosaurids,
we identified several other characters which are apomorphic either for the superfamily or for the
family. Without using these characters, we could not have determined the smallest specimens as
capitosaurs, as their proportions were in no way capitosaurian.

In the following discussion the adjective ‘capitosauroid’ pertains to the genera included (Morales
1987) in the Families Capitosauridae, Benthosuchidae, and Mastodonsauridae, while ‘capitosaurian’
refers to the broader assemblage of families regarded as a monophyletic lineage by Warren and
Black (1985). ‘Capitosaurid’ refers to members of the Family Capitosauridae (Morales 1987).

CHARACTER STATES USED IN THIS PAPER

Capitosaurians

The presence of an oblique ridge on the quadrate ramus of the pterygoid. This character was used by
Warren and Black (1985) as derived for capitosaurians.

Capitosauroids

Crista falciformis of the squamosal. This crest is a flattened flange of bone on the otic-occipital
margin of the squamosal, which projects towards the tabular horn. In later capitosauroids the crista
becomes progressively broader and more horizontal in orientation and contributes to the restriction
and eventual closure of the otic notch. In other Triassic temnospondyls the margin of the squamosal
does not project or projects only as a low ridge which is rounded in section rather than flattened.
The Late Permian rhinesuchoids appear to show a modest development of the squamosal which
approaches the state seen in capitosauroids, providing further evidence for the relationship of these
two groups of genera.

The arrangement of muscular crests on the parasphenoid. The posteroventral face of the parasphenoid
bears an area for the attachment of some of the neck musculature, the transverse ridge (Cosgriff
1974; crista muscularis of several authors). The attachment area is a depression, set off anteriorly
by a ridge which starts at the level of the trailing edge of the pterygoid. The ridge usually curves

NEW TRIASSIC CAPITOSAUR FROM QUEENSLAND

859

posteromedially and the ridges on each side generally meet, forming a V-shaped outline. In some
early forms (e.g. P. orientalis ) the two ridges do not meet, while in some (especially) later forms the
posterior curvature disappears and the ridge becomes a straight transverse line. The rhinesuchoids
(including Uranocentrodon and Lydekkerina ) possess a pair of semicircular depressions on the
parasphenoid, usually enhanced by a flange of bone which projects around the anterior and lateral
margins of the depressions. These were dubbed ‘pockets’ by Watson (1962) who regarded the ridges
of capitosaurs as derived homologues of the rhinesuchoid pockets. Most other families in the
capitosaurian lineage (Warren and Black 1985) have lost all trace of pockets or ridges, the only
exception apparently being the metoposaur genus Eupelor which shows a capitosauroid V-shaped
transverse ridge (Colbert and Imbrie 1956).

Dentary teeth. Benthosuchids and capitosaurids uniquely share a very large number, fifty or more,
of small marginal dentary teeth, and this also seems to be an apomorphy within the capitosaurians.

Capitosaurids

Hamate process. Jupp and Warren (1986) described a number of distinguishing features of the
capitosaurid lower jaw. A unique, clearly apomorphic character state is the prearticular or hamate
process, defined as a dorsal projection of the prearticular on the anterior margin of the glenoid
fossa, which rises above the level of the articular and surangular. Jupp and Warren considered that
only the capitosaurids possessed a well-developed hamate process. In this respect, capitosaurids
(e.g. Parotosuchus) are derived with respect to benthosuchids (e.g. Benthosuchus sushkini ) in which
the prearticular does not rise above the level of the articular.

Raised orbits. A further characteristic of Parotosuchus , as well as genera such as Wetlugasaurus and
Cyclotosaurus , is the elevation of the orbital rims above the level of the surrounding skull surface.
This is especially pronounced anteriorly where the prefrontal slopes down sharply from the leading
edge of the orbit. A result of this is that, whatever the degree of flattening or other changes in skull
proportions, the orbits always face almost directly upwards.

Lateral line system. A last point which seems useful to note is that most capitosaurids, including
Parotosuchus in particular, have poorly incised lateral line systems. Lateral line grooves are usually
only continuous, if at all, on the anterior parts of the supraorbital and infraorbital canals. Grooves
on the cheeks, skull table, and interorbital area are ofter reduced to chains of pits or are absent.

Parotosuchus

Frontal bones enter orbital margins. Parotosuchus species are characterized by frontals entering the
orbital borders, a derived state also found in most other capitosaurs, but absent from the two species
placed in Wetlugasaurus ( W. angustifrons and W. samarensis). The latter two species are in all other
respects similar, not simply to Parotosuchus , but to its Early Triassic species, with tapering horns
and relatively narrow pterygoid-parasphenoid contact. It is possible that Wetlugasaurus was derived
from these primitive Parotosuchus species via secondary contact of the prefrontal and postfrontal,
rather than retaining a primitive contact of these bones.

Otic notch. The most easily observed evolutionary change which occurred within Parotosuchus
( sensu Kamphausen and Morales 1981) was the development of a partly closed otic notch. The
plesiomorphic state of the otic area is shown by those species having tapering, pointed, posteriorly
directed tabular horns and only a moderate development of the crista falciformis. More derived
taxa show the development of a rounded lappet on the end of the tabular horn, a more lateral
orientation of the horn, and expansion of the crista towards the tip of the tabular horn. Many of
the later capitosaurs show a marked broadening and flattening of the skull compared with earlier,
apparently more primitive species. The broad-skulled forms also show a reduction in the relative
size of the orbits, which is accompanied by a reduction in the extent of the jugal bordering the orbit.
These evolutionary changes are often found as a mosaic in different species.

860

PALAEONTOLOGY, VOLUME 31

SYSTEMATIC PALAEONTOLOGY

Superfamily capitosauroidea Save-Soderbergh, 1935
Family capitosauridae Watson, 1919
Genus parotosuchus Otschev and Shishkin, 1968

Type species. Capitosaurus nasutus Meyer 1858, by subsequent designation.

Diagnosis of genus. Capitosaurid temnospondyls with open otic notches, a single anterior palatal
vacuity, and with both the frontals and jugals taking part in the orbital border (Kamphausen and
Morales 1981; Morales and Kamphausen 1984). Full discussions of intrageneric variation are
provided by Welles and CosgrifT (1965), and Cosgrifif and de Fauw (1987).

Parotosuchus aliciae n. sp.

Text-figs. 110

Derivation of name. The species is named in honour of Alice Crosland Hammerly who found the small juvenile
specimens referred to this species.

Type specimens. Holotype. QM FI 2281 (text-figs. 1, 2, 4-6a, 7), a partial skeleton consisting of most of the
skull and attached lower jaws, parts of the anteriormost vertebrae and ribs, most of the right hind limb, the
right ilium, and other fragmentary postcranial remains. The nearly complete dermal pectoral girdle was
destroyed in order to expose the palate, but is preserved as a polyester resin cast.

Paratype. QM F12282 (text-figs. 3-5, 6b), a skull and lower jaws minus the snout, with a partial shoulder
girdle.

Referred specimens. QM FI 2286, a weathered specimen consisting of the skull posterior to the level of the
orbits, with the rear portions of both lower jaws and most of the dermal pectoral girdle still in place. QM
FI 2287, a weathered right hand rear quadrant of the skull (plus associated dermal girdle) of a smaller
individual. QM FI 2290- 12292 (text-figs. 8-10), three small juvenile skulls with associated mandibles and
skeletal fragments.

Type locality. Collected by R. Jupp, A. C. Hammerly, A. A. Warren, R. Lane, and D. Harrison at AAW field
locality Q6, on Duckworth Creek south-west of the town of Bluff, Queensland. Bluff lies on the Tropic of
Capricorn, approximately 195 km west of the coastal city of Rockhampton.

Horizon. Lower Upper Arcadia Formation, Rewan Group, Early Triassic (Scythian). Jensen (1975) and
Warren (1980) discuss the stratigraphic position of these deposits and their Lystrosaurus zone fauna. Based
largely on these studies Cosgriff (1984) assigned the Arcadia Formation fauna to his earliest, Al, division of
the Scythian. Q6 is also the type locality for three other temnospondyl amphibians: Xenobrachyops alios
(Howie, 1972) (Brachyopidae), Keratobr achy ops australis Warren, 1981 (Chigutisauridae), and Arcadia myri-
adens Warren and Black, 1985 (Rhytidosteidae).

Diagnosis. Distinguished from all other species of Parotosuchus by the following combination of
character states: oblique ridge of pterygoid greatly expanded, forming a fan-shaped, dorsomedially
directed plate flooring the otic area; crista falciformis of the squamosal well developed and oriented
nearly vertically, forming a high wall bordering the otic notch laterally and terminating abruptly at
the squamosal-quadratojugal suture; posteroventral margins of tabular and postparietal with an
unusually well-developed crista muscularis which partially occludes the post-temporal fossa; crista
tahularis externa absent; ectopterygoid tusks present; posterior meckelian foramen of lower jaw
exceptionally small, not bordered by the postsplenial. The species is currently known only from
small specimens (skull length less than 40 mm) which show such juvenile features as a relatively
short, rounded snout, very large orbits, weakly sutured cranial roofing bones, and a parietal foramen
centred at the level of the posterior margins of the orbits.

Description

The description of the new species is based on the holotype and paratype specimens.

NEW TRIASSIC CAPITOSAUR FROM QUEENSLAND

861

Preservation. The types were collected in small nodules. The matrix surrounding the fossil bone is basically
the same red mudstone which predominates in the Arcadia Formation (Jensen 1975), but is distinctive in being
heavily impregnated with small gypsum crystals. These crystalline inclusions have apparently given the matrix
sufficient solidity to weather as nodules or cobbles rather than breaking down into the silty mud typical of the
formation. The holotype was extracted from two nodules, the break between them having occurred at the level
of the anterior margins of the interpterygoid vacuities. Weathering had destroyed the dorsal surface of the
anterior nodule and abraded the right cheek and lower jaw. The paratype was also in two sections, the posterior
right-hand corner of the skull, jaw, and girdle being separated from the rest. Both skulls apparently suffered
some damage prior to preservation. The holotype has been subjected to compression which has depressed the
right cheek region and laterally compressed the interorbilal region, producing a depressed fracture along the
mid-line suture. The paratype shows some damage to the skull roof anterior to the orbits and has had the
right tabular area pushed down crushing the right paroccipital process.

Skull roof (Text-figs. 1a, c, 2a, c, 3a, c, d, f, 4a.) The skull in dorsal view is bluntly triangular with a relatively
broad, rounded snout. We estimate the mid-line length of the holotype as 39 mm and the maximum width as
36 mm. The orbits are elongate ovals, 27 % of the length of the skull, and they are centred just behind the mid-
point of the skull length and are separated by a shallow trough running along the skull mid-line. They are
raised above the level of the adjacent skull roof, the elevation being most pronounced anteriorly where the
prefrontal slopes markedly downwards. A circular parietal foramen is centred on a line level with the posterior
extremities of the orbits. Each deeply incised otic notch is bounded laterally by a well-developed flange on the
occipital edge of the squamosal (the crista falciformis, Bystrow and Efremov 1940), which appears as a
pronounced fin-like projection when the skull is seen in lateral view. The tabular horns of the paratype and
holotype differ in shape, those of the paratype being disproportionately smaller and more slender that those
of the holotype.

The dermal roofing bones of the holotype are covered with a fine ornament which is well preserved on the
newly exposed surfaces. On the skull table and between the orbits the ornament consists primarily of pits, but
becomes a ridge-groove on the cheeks. In the paratype the ridge-groove pattern extends more on to the dorsal
surface of the skull; presumably the more pitted pattern seen in the holotype is the result of cross-bridge
development during ontogeny, as discussed by Bystrow (1935). Sensory canals are not obvious on the skull
bones preserved. Taken together, the type skulls provide complete outlines of the postparietals, tabulars,
squamosals, supratemporals, quadratojugals, jugals, postorbitals, postfrontals, and parietals, while the maxil-
lae, prefrontals, and frontals lack only their anterior extremities. Parts of the lachrymals and probably the rear
portions of the nasals are also present, but their outlines are difficult to determine. The outer margins of the
premaxillae, with several of the teeth remaining, give the shape of the lip of the snout. The pattern of the skull
bones is typically parotosuchian (Welles and CosgrifT 1965). The frontals enter the orbital margins, as do the
jugals. There is only a small intrusion into the jugal of the lateral process of the postorbital. The supratemporal
is excluded from the otic margin by the contact of the squamosal and tabular.

Occiput. (Text-figs. Id, 2d, 4c.) In occipital view the skull is moderately deep, with the cheeks descending more
abruptly than in many Parotosuchus species. A prominent feature of the rear of the skull table is a descending
flange of bone, borne by both the tabular and postparietal on each side. This crista muscularis (Bystrow and
Efremov 1940; Cosgriff and de Fauw 1987) partly overgrows the post-temporal fossa which, as a result, takes
the form of an obliquely oriented slot. The paroccipital processes are made up of the tabulars and exoccipitals
with no exposure of the opisthotic between them. The ventral surface of the tabular portion of the paroccipital
process lacks any trace of the crista tabularis externa (Bystrow and Efremov 1940) which is normally present
in capitosaurs (CosgrifT and de Fauw 1987). A relatively well-developed crista tabularis interna is present on
the otic margin of the tabular. The exoccipital bone bears a large foramen (cranial nerve X) at the base of the
paroccipital process while the inner margin of its ascending ramus gives rise to a well-developed processus
lamellosus. The occipital condyle is set off laterally from the body of the exoccipital by a strongly incurved
neck of bone. Its articular surface is elliptical and is directed posteromedially. The occipital portions of the
squamosal and quadratojugal provide a convex surface for the origin of the depressor mandibulae muscle. On
its occipital surface the squamosal is bordered laterally by the crista falciformis which is not continued along
the quadratojugal but ends abruptly at the squamosal quadratojugal suture. A feature unusual in capitosaurs,
but possibly present in all juveniles, is the failure of the ascending ramus of the pterygoid to meet the skull roof,
so that a gap is present. This represents the more dorsal part of a palatoquadrate fissure, the ventral part being
obliterated by a suture between the squamosal and the ascending ramus. The oblique ridge of the pterygoid
is exceptionally well developed, its trailing edge being oriented diagonally upwards so that its dorsal limit is

862

PALAEONTOLOGY, VOLUME 31

text-fig. I . Parotosuchus aliciae sp. nov., holotype skull and mandible, QM F12281. a, dorsal view; B, ventral
view; c, left lateral view; d, posterior view, x 2 natural size.

obscured by the paroccipital process. Between the ascending ramus of the pterygoid and the edge of the oblique
ridge is a smoothly curved trough underlying the tympanic area. The quadrate is poorly ossified and has a
pointed dorsal process wedged between the squamosal and the ascending ramus of the pterygoid.

Palate. (Text-figs. 1b, 2b, 3b, e, 4b.) The palate shows the usual suite of vacuities seen in temnospondyls. Parts
of the dorsal surface of both vomers are exposed on the holotype snout and contain much of the shallow
convex posterior margin of the anterior palatal vacuity. The marginal teeth of the premaxilla and maxilla are
small, lanceolate, even in height (IT to T3 mm), and number between fifty and fifty-five. In ventral view the
snout fragment shows the dentition of the right vomer. Following a pair of vomerine tusks (each about T5 times the
size of a maxillary tooth) is a series of six slender teeth running almost directly posteriorly and apparently
delimiting the inner margin of the choana. Two teeth situated posterolaterally to these vomerine teeth probably
represent palatine teeth, although no vomer-palatine suture was preserved. The snout has broken along a line

NEW TRIASSIC CAPITOSAUR FROM QUEENSLAND

863

text-fig. 2. Parotosuchus aliciae sp. nov., holotype, skull and mandible, QM FI 2281 . a, dorsal view; B, ventral
view; c, left lateral view; d, posterior view, x 2 natural size. Broken bone surface hatched. Matrix stippled.
Abbreviations; A, angular; a.p.v., anterior palatal vacuity; AR, articular; ch, choana; cr.fal., crista falciformis;
cr.mus., crista muscularis; c.t.f., chorda tympanic foramen; cul.pr., cultriform process; D, dentary; ecpt.t.,
ectopterygoid tusk; EOC, exoccipital; fi.pq.. palatoquadrate fissure; J, jugal; m.s., mandibular sulcus; MT,
metatarsal; MX, maxilla; ob.r., oblique ridge; P, parietal; pal.t., palatine tusk; PAR. prearticular; PF, postfron-
tal; p.m.f., posterior meckclian foramen; PO, postorbital; PP, postparietal; PRF, prefrontal; PSP, para-
sphenoid; PSPL, postsplenial; pt.fen., post-temporal fenestra; PTG, pterygoid; Q, quadrate; QJ, quadratojugal;
SA, surangular; SPL, splenial; SQ, squamosal; ST, supratemporal; STA, stapes; sym.t., symphysial tusk; T,
tabular; tr.r., transverse ridge; V, vomer; X, foramen for tenth cranial nerve.

864

PALAEONTOLOGY, VOLUME 31

text-fig. 3. Parotosuchus aliciae sp. nov., paratype, skull and mandible, QM F12282. a, d, dorsal view; b, e,
ventral view, c, f, left lateral view, x 2 natural size. Broken bone surface hatched. Matrix stippled. Abbrevia-
tions: F, frontal; remainder as in text-fig. 2.

NEW TRIASSIC CAPITOSAUR FROM QUEENSLAND

865

text-fig. 4. Parotosuchus aliciae sp. nov., restoration of skull based mainly on holotype, QM F12281, with
details of interorbital and parasphenoid regions added from paratype, QM F12282. a, dorsal view; b, ventral

view; c, posterior view, x 2 natural size.

running through the palatine tusks. The palatal dentition is distinctive in including an ectopterygoid tusk at
the front of each ectopterygoid tooth row. An expansion of the ectopterygoid in the region of the tusk contacts
the margin of the interpterygoid vacuity, separating the pterygoid and palatine bones. The teeth on the palatines
and ectopterygoids are smaller than the marginal teeth, and their crowns are angled lingually.

The palatal ramus of the pterygoid has a sharply downturned flange on its trailing edge which borders the
subtemporal fossa. A mid-line strip of the palatal ramus bears a coarse ornament which becomes more
elaborate on the body of the pterygoid. The greatly expanded oblique ridge of the pterygoid has already been
described. In ventral view it can be seen to merge with the quadrate ramus at a low ridge which is confluent
with the posterior edge of the body of the pterygoid. The pterygoid forms a sinuous medial suture with the
parasphenoid but does not make ventral contact with the exoccipital. The parasphenoid bears a pair of shallow,
transversely aligned grooves (‘transverse ridge’ of Cosgriff 1974; ‘pockets’ of Watson 1962; crista muscularis
of e.g. Otschev 1972) posteriorly. The cultriform process has a flattened median crest posteriorly which reduces
to a narrow ridge anteriorly. The stapes is preserved on both sides of the holotype and appears to have been
of the usual, rather massive capitosaurian type. Both left and right stapes are slightly displaced.

Hyoid Element. (Text-fig. 6a.) A small dumb-bell-shaped bone was found in the oral cavity of both type
specimens and the largest referred specimen (QM FI 2286). The bone in each case lay just in front of the
anterior extremity of the interclavicle. We have tentatively identified this as a median hyoid element, probably
the copula. No other remains referable to the hyoid apparatus were found, and there was no trace of any
branchial bars.

866

PALAEONTOLOGY, VOLUME 31

text-fig. 5. Parotosuchus aliciae sp. nov., restoration of mandible (oriented parasagittally) based
on holotype and paratype specimens, a, labial view; B, lingual view, x 4 natural size.

Lower Jaw. (Text-figs, lc, 2c, 3c, f, 5.) No complete mandibular ramus was recovered, but enough partial
jaws are available for many of the mandibular features to be described. The mandible resembles that of other
capitosaurs (see Jupp and Warren 1986) in many respects, including overall shape, relative tooth size, presence
of a well-developed hamate (prearticular) process preceeding the glenoid area, a Type I postglenoid area (PGA,
Jupp and Warren 1986), and a single row of teeth on the posterior coronoid. There are several features which
differentiate P. aliciae from some or all other capitosaurs. The labial surface of the rear of the mandible shows
no extension of the angular on to the PGA, the surangular meeting the angular along a vertical suture at the
level of the glenoid area. In lingual view, some unique features are visible. The chorda tympanic foramen is
large and situated on the articular prearticular suture about midway between the glenoid cavity and the ventral
margin of the jaw. The posterior meckelian foramen is exceptionally small and fails to contact the postsplenial,
so that it is bordered solely by the prearticular and the angular. Damage to the jaw prevents us from determining
the presence or absence of an anterior meckelian foramen, although if present it must have been small. The
posterior coronoid bore a series of at least three small teeth. The surfaces of the lower jaw covered by the
middle and anterior coronoids were difficult to prepare and much of this area remains covered by a thin layer
of matrix. However, at least two small teeth are present in this region and have been tentatively restored in
text-fig. 5 as lying on the middle coronoid. At the point where the left ramus of the holotype mandible is
broken, there is a thickened bump of bone which we interpret as the origin of an enlarged tusk-like tooth.

Pectoral Girdle. (Text-fig. 6.) The pectoral girdle is represented in the holotype by a nearly complete dermal
girdle and a partial scapulocoracoid, while fragments of the dermal girdle and scapulocoracoid are also present
in the paratype. In both specimens the girdles were preserved in almost their natural positions, and had to be
removed in order to expose the posterior palate and basicranium. The description of the dermal elements is
based principally on a polyester resin cast made of the holotype girdle prior to its destruction.

The ventral plate of the clavicle is roughly triangular with a relatively narrow, concave posterior margin
and shallow convex, elongate anterior and medial margins. Ridge-groove ornamentation radiates from a
pitted centre of ossification situated at the posterolateral corner of the clavicle. The dorsal process of the
clavicle is preserved in external view in the paratype and in posterior view in the holotype cast. The dorsal
process is slender and tapering, and lacks any sigmoid flexure or cleidomastoideus scar. In posterior view the
process can be seen to consist of a columnar shaft bordered laterally by a flange of bone which merges with
the shaft about half-way up. The interclavicle is rhomboidal with an extended anterior arm. The ornament of
its ventral surface is similar to that of the clavicles. No specimen retains an intact posterior edge to the
interclavicle, although the missing portion does not appear to have been large.

All capitosaurs for which the clavicle has been described have a dorsal process which is markedly different
from that of P. aliciae. Clavicles of Paracyclotosaurus davidi (Watson 1958), Parotosuchus peabodyi (Welles
and Cosgriff 1965), P. pronus (Howie 1970), P. orenhurgensis , P. tverdochlebovi, and P. garjainovi (Otschev
1966, 1972) all have a dorsal process, the base of which runs forward along the anterolateral edge of the

text-fig. 6. (left). Parotosuchus aliciae sp. nov., pectoral girdle and hyoid elements as preserved, x 2 natural
size, a, clavicles, interclavicle, and copula of holotype, QM FI 2281, based on polyester resin cast; b, lateral
view of right partial clavicle and scapulocoracoid of para type, QM FI 2282. Abbreviations: CL, clavicle; COP,
copula; dor.proc., dorsal process of clavicle; ICL, interclavicle; SCAP, scapula; remainder as in text-fig. 2.

text-fig. 7. (right). Parotosuchus aliciae sp. nov., right hindlimb elements of holotype, QM FI 2281. a d,
femur; a, dorsal view; b, anterior view; c, ventral view; d, posterior view, e-h, tibia; e, posterior view; f, lateral
view; G, anterior view; h, medial view, i l, fibula; i, anterior view; J, medial view; k, posterior view; l, lateral
view, m-o, three metatarsals in dorsal view, p-r, three proximal phalanges in dorsal view, x 2 natural size.

clavicle, so that the process in lateral view has a squat, triangular shape, terminating in a short slender
projection. The leading edge of the base of the process bears a well-developed scar or depression for the
cleidomastoideus muscle. In posterolateral view, the dorsal process shows a marked sigmoid flexure, curving
outwards at the base, then inwards, and outwards again towards the apex. The dorsal process of P. aliciae is
a simpler structure, in which the cleidomastoideus area is not developed and the sigmoid flexure of the tall
slender dorsal process is scarcely apparent.

Warren and Hutchinson (1983) attempted to define the ‘typical’ structure of clavicles for many of the Triassic
temnospondyls, an attempt which now appears to have been unsuccessful. When compared with fig. 27 in
Warren and Hutchinson (1983, p. 42), the dorsal process of the clavicle of P. aliciae is most similar to the
brachyopoid Siderops. Recently Snell (1986) has described the clavicle of an Arcadia Formation capitosaur
(QM FI 2278) which can be referred, on the basis of an associated skull, to P. rewanensis. This specimen (skull
length approx. 150 mm) includes a partial right clavicle with an almost complete dorsal process which is
slender, tapering, lacks a sigmoid curvature and, in short, resembles that of P. aliciae very closely.

In the series of juvenile to adult clavicles of Benthosuchus sushkini (Bystrow and Efremov 1940, fig. 77) the

868

PALAEONTOLOGY, VOLUME 31

dorsal process is hardly visible. However, in a second illustration (fig. 78) which shows individual variation in
the dorsal processes of eight specimens, it is apparent that those of the smaller individuals are more slender
and have a less-developed muscle scar. This indicates that the slender unscarred dorsal process of P. aliciae
and QM FI 2278 may be related to their small size and possible immaturity.

The scapulocoracoid of the paratype is incomplete dorsally and ventrally, but the holotype fragment shows
the posterior margin and the ventral limits of the coracoid and the supraglenoid buttress. The latter two regions
were unfinished ventrally, so that the supraglenoid foramen was open. Such unfinished scapulocoracoids are
the rule in the Australian Early Triassic (Warren and Hutchinson 1983).

text-fig. 8. Parotosuchus aliciae sp. nov., referred small juvenile skull, QM FI 2290. a, dorsal view showing
impressions of the ventral surface of the cranial roof; B, ventral view; c, sketch of specimen shown in a,
indicating bone outlines; d, sketch of specimen shown in b, indicating bone outlines. Abbreviations: mand.,
mandible; N, nasal; ot.n., otic notch; remainder as text-fig. 2. x 5 natural size.

NEW TRIASSIC CAPITOSAUR FROM QUEENSLAND

869

text-fig. 9. Parotosuchus aliciae sp. nov., referred small juvenile skull, QM FI 2291 . a, skull seen in dorsolateral
view, b, interpretive drawing of specimen shown in a. x 5 natural size.

Pelvic Girdle. The right ilium and both ischia were preserved as counterparts in the nodule containing the main
part of the holotype skull. The ilium is 12 mm long and is notable for its gracile proportions, with a narrow
subcylindrical shaft becoming flattened and slightly swept back at its dorsal extremity and with a thickened
basal area behind the acetabulum. In posterior view the ilium is bowed outwards. The ischia appear to have
been poorly ossified and are visible only as indistinct but bony impressions, each roughly trapezoidal in shape
with the narrow end facing posteriorly. No traces of the pubes are visible.

The few capitosaurid ilia which have been described (Watson 1958; Howie 1970) are from larger animals
and are considerably more robust in shape than that of P. aliciae. The principal difference seen in the larger
species is that the expanded dorsal blade of the ilium extends much further ventrally so that the shaft is reduced
to a ‘waist’ separating expanded dorsal and ventral regions.

Limbs. (Text-fig. 7.) An almost complete right hind limb was found with the holotype skull. The femur, tibia,
fibula, three or four metatarsals, and several phalangeal bones were preserved draped across the left cheek and
orbit of the skull, some of the metatarsals having fallen into the matrix which filled the orbit. Apart from their
more slender build, the limb bones are very similar to those described by Howie (1970) for P. promts. No trace
of ossified tarsals was detected, and in view of the good preservation of the rest of the limb, it seems likely
that the tarsal region was not ossified.

Vertebrae and Ribs. Several neural arches and associated proximal rib fragments were attached to the holotype
skull. Their preservation was not good and the very thin bone proved difficult to separate from the matrix. As
far as can be determined the neural arches are similar in shape to those of other capitosaurs such as P. promts
(Howie 1970) or Paracyclotosaurus davidi (Watson 1958). The ribs are broad-based without any ossified
bicipitate structure, but, like the vertebrae, they are preserved in a very fragile state which makes detailed study
difficult. No determinable remains of intercentra were recovered, apart from several impressions associated
with the ischial remains. No traces of pleurocentra were identified.

Juvenile Skulls

Collection and preparation. Three very small temnospondyl skulls (text-figs. 8-10) were recovered
at the same site, within a few metres of the Parotosuchus aliciae types. As discussed below, we believe
that these are referable to P. aliciae. With skull lengths of just over 10 mm, these are by far the
smallest specimens to have been identified as capitosaurids.

870

PALAEONTOLOGY, VOLUME 31

text-fig. 10. Parotosuchus aliciae sp. nov., restoration of referred small juvenile skull, based on QM F12290
and F 1 229 1 , in dorsal (a) and ventral (b) views, x 5 natural size.

The three skulls were found embedded in small nodules. Two (QM F12290 and F 1 229 1 ) were
found one on top of the other within the same nodule and are better preserved that the third
specimen (QM F12292) which has suffered more weathering and distortion. QM F12290 (text-
fig. 8) became detached from its nodule, leaving the skull roof attached to the matrix. After a
sketch was made of the bony sutures, the exposed underside of the skull roof was filled with poly-
ester resin to provide support while the nodule was mechanically prepared to expose the dorsal side.
During this preparation the second skull (QM FI 2291; text-fig. 9) was discovered lying on its left
side on top of the skull of QM FI 2290. It was also noted during this preparation that numerous
postcranial bones were present, including girdles, limbs, neural arches, and possible ribs. These were
little more than fragile films of bone and could not be saved but were sketched before being destroyed
as preparation of the skulls proceeded.

In making the reconstructions of the skull shown in text-fig. 10, information from QM F12290
and FI 2291 was combined. QM FI 2290 retained a detailed impression of most of the internal
surface of the skull roof and provided the most complete palatal surface, as well as the overall
proportions of the skull and orbits. The dorsal surface of its skull roof, which could only be partially
prepared, gave additional information on the sutures and ornament of the interorbital area. QM
FI 2291 preserved the rear of the skull table including the external surfaces of the tabular horns and
otic notches, and provided the surfaces of the lateral skull bones and a complete labial view of the
right mandibular ramus. This specimen also provided extra detail of the palate, including the
parasphenoid ridges and the ectopterygoid tusk.

Description. The mid-line length of QM F12290 is 12-5 mm, and the less complete QM F1229 1 and QM F12292
are of similar size. The skull is broadly rounded, with a short, blunt snout and very large orbits (length of
orbit 36 % of skull length). The orbital borders are raised above the level of the adjacent skull bones, especially
anteriorly. The mid-line region of the skull is shallowly concave. The parietal foramen is large and centred on
a line level with the posterior margins of the orbits. The otic notches are deeply incised but broadly open
posteriorly. The bones of the skull roof bear a pitted ornament which is absent from the sutural margins of
the bones, especially on the skull table. No impressions of sensory canals are evident. The arrangement of the
skull bones is typical of many Triassic temnospondyls, with the following exceptions: the frontals enter the
orbital margins and the jugals broadly border the orbits laterally; there is a broad jugal -prefrontal suture
running to the ventrolateral rim of the orbit; the otic margin of the squamosal bears a distinct crista falciformis;
the tabular horn projects only slightly beyond the body of the tabular, and it is well buttressed ventrally by

NEW TRI ASSIC CAPITOSAUR FROM QUEENSLAND 871

the tabular portion of the paroccipital process; the tabular and squamosal contact to exclude the supratemporal
from the otic margin.

In palatal view, the choanal openings are large and their posteromedial borders bulge into the interpterygoid
vacuities. The anterior palatal vacuity is not preserved, but it is restored here as single. Tusks are present on
the vomers, palatines, and ectopterygoids; it is not possible to determine if smaller palatal teeth were also
present. The body of the pterygoid is flat and in moderately broad contact with the parasphenoid. The
cultriform process of the parasphenoid is relatively broad. The body of the parasphenoid bears a pair of
transverse ridges which start just posterior to the pterygoid parasphenoid suture and run anteromedially. The
exoccipitals were apparently poorly ossified and have not been adequately preserved, as is true also of the
quadrates.

The lower jaw is known primarily from its external surface as preserved in QM F 1 229 1 . The pattern of
sutures completely matches that seen in the lower jaw of P. aliciae QM F12281 (text-fig. 5), and the pattern
of ornamentation, with a pitted surangular and ridged angular, is also very similar. The postglenoid area is
Type I (Warren and Black 1985). The internal surfaces of the jaws show few sutural details, but it is clear from
QM FI 2290 that the prearticular gave rise to a pronounced hamate process.

Allocation to P. aliciae. The following character states collectively indicate that QM FI 2290- 12292
are small capitosaurids, and should probably be allocated to P. aliciae.

1. Otic notches distinct and semicircular. A primitive character state which characterizes the
capitosaurian lineage but is lost by trematosaurians, the other major Triassic temnospondyl assem-
blage (Warren and Black 1985).

2. Tabular horn well buttressed ventrally by the paroccipital process. Again a primitive character
state, but one which is typical of capitosaurians.

3. Frontals enter orbital borders. A derived state found in most capitosaurids although also
occurring in other families (e.g. Dissorophidae).

4. Parasphenoid with transverse ridges. The form of the ridges in the small specimens is somewhat
aberrant in that the ridges are directed anteriorly as well as medially but in this respect they resemble
the smaller P. aliciae (paratype) specimen (QM F 1 2282).

5. Parasphenoid-pterygoid suture. The referred specimens resemble early capitosaurs, including
P. aliciae , in possessing an intermediate stage of this character, in which the corpus of the pterygoid
has a flattened ventral surface and the suture with the parasphenoid is sinuous but not greatly
extended posteriorly. This is derived with respect to more archaic groups such as eryopoids and
dissorophoids, in which the pterygoid corpus is narrow and curved ventrally and in virtual point
contact with the parasphenoid. Later capitosaurids, as well as most other Triassic families, show a
more derived state in which there is a posterior lengthening of the pterygoid-parasphenoid suture.

6. Orbital borders raised above the level of the adjacent skull surface.

7. Squamosal with flattened, fin-like crista falciformis. Such a crista is diagnostic for capitosaurids.
The form of the crista , well preserved in QM FI 2291, is very similar to that seen in the P. aliciae
types, and it differs only in its relatively smaller size.

8. Ectopterygoid tusks present. The only capitosaurids known to retain ectopterygoid tusks are
P. (= Benthosuchus ) madagascariensis (Warren and Hutchinson, in press) and P. aliciae.

9. Mandibular features. The Type I PGA and hamate process, both indicate a capitosaurid.

The characters discussed above all suggest that QM F 1 2290 1 2292 are capitosaurids, and in
particular a species of Parotosuchus. Characters 4, 7, and 8 indicate a special resemblance to P.
aliciae, and in view of the fact that the small specimens were apparently preserved at the same time
and place as the P. aliciae types, we are confident that QM FI 2290- 12292 should be regarded as
very young specimens of P. aliciae.

CAPITOSAURID ONTOGENY AND TEMNOSPONDYL PHYLOGENY

Boy’s (1974) analysis of temnospondyl ontogeny, based on the Permian eryopid Sclerocephalus,
summarized the changes occurring during larval development to early postmetamorphic stages. Our

872

PALAEONTOLOGY, VOLUME 31

capitosaurid specimens appear to complement Boy's material, and extend his staging of temno-
spondyl ontogeny through to the adult.

Boy’s criteria for determining the point of metamorphic climax were loss of gills (including gill
rakers), development of a sclerotic ring, ossification of the exoccipital, and definite presence of
vertebral centra. Both latest larva and earliest adult showed a complete dermatocranium, labyrin-
thine teeth, ossified copula, lateral line grooves, and ossified limbs and girdles (except coracoid and
pubis). Our very small specimens (QM FI 2290- 12292) appear to be at this stage and show, where
it is possible to ascertain, a combination of late larval and early adult character states. Adult features
include apparent loss of gills, no traces of which (or their more durable branchial teeth) were
found, and partial ossification of the exoccipitals. However, a sclerotic ring was not preserved,
nor were any remains of centra, although neural arches were preserved in partial articulation.
Thus, these small specimens represent the starting point for postmetamorphic ontogenetic changes
in capitosaurids.

table I . Characteristics of the skull of capitosaurids at metamorphosis.

1. Bones weakly sutured.

2. Ornament of coarse pits not extending to sutural boundaries.

3. Short, broadly rounded snout.

4. Palatine and ectopterygoid relatively short and ‘crowded’ towards the front of the interpterygoid vacuities.

5. Very large orbits (> 30% of skull length) centred in anterior half of skull.

6. Pineal foramen centred level with the rear margins of the orbits.

7. Tabular horns not strongly projecting; otic notches not deeply incised and widely separated.

8. Cultriform process relatively broad.

9. Transverse ridges of parasphenoid directed anteromedially but not contacting medially.

10. Poorly ossified exoccipitals and quadrates.

1 1 Palatoquadrate fissure present.

12. Occiput deep.

At the end of metamorphosis the capitosaur skull was evidently very different from that of a
mature individual (Table 1). In many respects, such as the broad, parabolic skull outline, anteriorly
centred orbits, weakly projecting tabular horns, broad cultriform process, and presence of a palato-
quadrate fissure, the skull of a juvenile capitosaur resembled that of a mature brachyopoid or
trematosauroid (Warren and Black 1985). However, all of these features were lost during subsequent
growth, via positive allometry of the antorbital and cheek regions and increased ossification.

The next stages in growth are shown by the P. aliciae types (QM F12281-12282), as well as
in the smaller individuals of several growth series of B. sushkini (Bystrow and Efremov 1940),
Archegosaurus decheni and Actinodon latirostris (Romer 1939), and Zatrachys serratus (Steen 1937,
as Acanthostoma vorax ). In these, the proportions are intermediate between the metamorphling
and adult, although the skull is well ossified and family characteristics more obvious. Based on
the growth series reported for P. peabodyi (Welles and Cosgriff 1965), and on the proportions
of the small Australian species P. wadei (Cosgriff 1972), both described from later in the Triassic
than P. aliciae , capitosaurids seem to have attained almost adult proportions by a skull length
of 60 to 70 mm, although growth proceeded to a much larger adult size (in excess of 250 mm for
P. peabodyi). Thus by this stage, ossification of the skull was complete, and allometric growth
had become much less important. A similar pattern of growth also seems to have been the case for
the well-documented B. sushkini series described by Bystrow and Efremov (1940). A more recent
study of benthosuchid ontogeny (Getmanov 1981) was based on skulls said to belong to two
species, B. korobkovi and Thoosuchus jakovlevi. Getmanov identified two phases of positive
allometric growth in these skulls which were of much larger (older?) individuals than the biggest

NEW TRIASSIC CAPITOSAUR FROM QUEENSLAND

873

text-fig. 11. Diagrammatic drawings of the changes in skull roof proportions seen in the post-metamorphic
growth of capitosaurids. a, at metamorphosis; b, juvenile; c, immature; d, adult. Scale bar in all drawings

equals 10 mm.

specimen of P. aliciae. As Getmanov does not describe the character states by which he identified
his specimens as capitosauroids, benthosuchids, or members of their respective genera and species,
we are unable to determine whether he was indeed studying a growth series or just a collection of
different-sized temnospondyls.

In summary, temnospondyls appear to have gone through four post-metamorphic stages, starting
with (1) recently metamorphosed individuals, retaining larval cranial proportions; (2) juveniles,
during which allometric growth is pronounced; (3) immatures, in which adult proportions are
essentially achieved, grading into (4) adults, in which maximum size is reached. Allometry continues
in these last two stages but to a much reduced extent.

Text-fig. 1 1 shows a series of four generalized early capitosaurid skulls, based on P. aliciae , P.
wadei , and P. rewanensis, showing the ontogenetic changes occurring during the post-metamorphic
growth of a capitosaur.

The juvenile characteristics of young capitosaurs include several which have been regarded as
significant for phylogenetic investigations. Among these are skull outline, anteriorly centred orbits,
and palatoquadrate fissure, and these now appear to be the result of paedomorphosis. A change in
the timing of the development of a character is a relatively ‘simple’ evolutionary step (Hecht and
Edwards 1977), and therefore more prone to parallel evolution. In addition, character reversal, with
the re-establishment of a more developed ‘adult’ condition, would be expected to be an easily
acquired source of confusion. The fact that most Triassic temnospondyl families appear to possess
unique mosaics of juvenile and adult character states supports the idea that similar juvenile character
states have been independently retained by unrelated lineages. It should, therefore, be clear that
such retained juvenile character states are not likely to be sufficient to diagnose monophyletic taxa;
rather, they must correlate with a number of other, ideally non-paedomorphic, derived character
states before they can contribute to the recognition of natural groups.

RELATIONSHIPS OF P. ALICIAE

The capitosaurs with the greatest phenetic similarity to P. aliciae are the other Early Triassic
species. Those which we consider determinable are P. madagascariensis (Lehman 1961; Warren and
Hutchinson, in press) from Madagascar; the Australian forms P. wadei (Cosgriff 1972) and the

874

PALAEONTOLOGY, VOLUME 31

two other parotosaurs from the Arcadia Formation, P. gunganj Warren 1980 and P. rewanensis
Warren 1980; the type capitosaurid, P. nasutus (Meyer 1858) and the other European species P.
helgolandicus (Schroder 1913), P. orientalis (Otschev 1966), and P. orenburgensis (Konzhukova
1965); and the southern African P. haughtoni ( Broili and Schroder 1 937) and Wetlugasaurus magnus
Watson 1962. These are the more primitive, often deeper skulled capitosaurs with tapering
posteriorly directed tabular horns and quadrate condyles aligned behind the level of the occipital
condyles. They differ from the similar species usually included in Wetlugasaurus (e.g. W. angustifrons
(Riabinin 1930), W. samarensis Sennikov 1981) in having the frontal included in the orbital margin.

P. aliciae can be distinguished from all of the early species of Parotosuchus using the diagnostic
features given earlier, especially the greatly fanned hypertrophied oblique ridge of the pterygoid and
the absence of a crista tabularis externa. Its skull topography is closest to that of P. madagascariensis
with which it shares the otherwise unique (for capitosauroids) presence of ectopterygoid tusks. This
overall similarity to P. madagascariensis may reflect the immature nature of the holotypes of P.
madagascariensis and P. aliciae, but the presence of ectopterygoid tusks is probably not a juvenile
feature of capitosauroids as they are absent in the smallest specimens of B. sushkini (Bystrow and
Efremov 1940).

The relatively broad skull of these two species is found in another small capitosaur, P. wadei,
from the Early Triassic of the Sydney Basin (Cosgriff 1972). Although P. wadei is small, it contrasts
with P. aliciae and P. madagascariensis in that its proportions are essentially adult with small orbits
centred well posterior to the mid-point of the skull and closely spaced otic notches. It also differs
from P. aliciae in having a less abruptly defined crista falciformis. Cosgriff (1972) noted that the
frontal entered the right orbital margin of the holotype of P. wadei but was excluded from the left
orbit. We consider this asymmetry unproven as the specimen is not well preserved.

Of the two larger Australian forms, P. rewanensis has a markedly heart-shaped anterior palatal
vacuity bordered by a V-shaped transvomerine tooth row; in this respect it is similar to P. madagas-
cariensis but not to P. aliciae which resembles the other Queensland capitosaur, P. gunganj, in
having a kidney-shaped vacuity and a straight transvomerine tooth row. P. aliciae and P. madagas-
cariensis share with P. gunganj a ‘notch’ on each side of the parasphenoid lateral to the transverse
ridges (‘ventral notch’ of Warren 1980, figs. 3, 4, 6, 7). This has been illustrated in two other Early
Triassic species, P. (= Eryosuchus) tverdochlebovi ( foramen ventrale of Otschev 1972, fig. 18) and
P. helgolandicus (Welles and Cosgriff 1965, fig. 24).

P. aliciae differs from P. gunganj in the shape of the transverse ridges on the parasphenoid. In
P. aliciae each ridge turns sharply posteriorly leaving a raised median area separating them.
In P. gunganj and most other capitosaurids these two posterior deflections meet, forming a V. In
some capitosaurids, especially the African species, the posterior deflection is lacking so that a single
straight ridge runs across the parasphenoid. The only large capitosaurid to have transverse ridges
shaped like those of P. aliciae is P. orientalis which has a skull approximately 470 mm long, indicating
that the feature is not a juvenile one.

Among the Australian Early Triassic forms, P. aliciae is closest to P. gunganj. As the former is
small (skull length 39 mm) and the latter much larger (skull length 227 mm) and as both specimens
come from the Arcadia Formation, it seems possible that P. aliciae is a partly grown P. gunganj
and that the features which separate them are in fact juvenile characters of P. aliciae. However, if
we consider these characters as shown by the juvenile to adult series in B. sushkini (Bystrow and
Efremov 1940), it is apparent that the two Queensland forms are not conspecific. In fact those
characters which are larger in P. aliciae than in P. gunganj (the oblique ridge on the pterygoid and
the crista muscularis above the occiput) are smaller in the juvenile B. sushkini than in the adult. The
transparasphenoid ridges of P. aliciae do not meet in the mid-line whereas in P. gunganj they meet
to form a V. In B. sushkini they are more widely separated medially in the adult than in the juvenile.
Ectopterygoid tusks, present in P. aliciae but not P. gunganj, are absent from all specimens of B.
sushkini. It is, therefore, apparent that those characteristics used by us to distinguish P. aliciae from
P. gunganj are not those of juvenile capitosauroids nor are they related to allometric growth.

NEW TRIASSIC CAPITOSAUR FROM QUEENSLAND

875

Acknowledgements. We thank our field assistants, in particular Alice Hammerly, Rob Jupp, and Ruth Lane
who discovered the site where the P. aliciae types and the small juveniles were found. Reg Goodwin of
‘Colorado’, Bluff, kindly allowed access to his property. Most of the drawings in this paper were painstakingly
executed by David Keen. David Walsh (La Trobe University Department of Zoology) took the photographs.
We thank Dr Philippe Janvier (Museum Nationale D'Histoire Naturclle, Paris) for casts of the type material
of P. madagascariensis , and Dr Sam Welles (University of California, Berkeley) for a cast of P. peabodyi. Dr
Alex Ritchie (Australian Museum, Sydney) loaned us the P. gunganj paratype and Dr Max Banks (University
of Tasmania, Department of Geology) loaned us some small specimens from Tasmania. Field-work and
support for M. N. H. were funded by an Australian Research Grants award to A. A. W. We are grateful to Dr
Andrew Milner and an anonymous reviewer for comments on an earlier draft of this paper. Their advice has
greatly improved it.

REFERENCES

boy, j. a. 1974. Die Larven der rhachitomen Amphibien (Amphibia: Temnospondy(i); Karbon-Trias). Paldont.
Z. 48, 236-268.

broili, F. and schroder, h. c. 1937. Beobachtungen an Wirbeltieren der Karooformation. XXVII. Uber einen
Capitosauriden aus der Cynognathus zone. Sber. bayer. Akad. Wiss 1937, 97-117.
bystrow, a. p. 1935. Morphologische Untersuchungen der Decknochen des Schiidels der Wirbeltiere. Acta
Zool. , Stockh. 16, 1 141.

— and efremov, I. a. 1940. Benthosuchus sushkini Efremov— a labyrinthodont from the Eotriassic of the
Sharzhenga River. Trudy paleozool. Inst 10, 1 152. [In Russian.]

colbert, e. h. and imbrie, j. 1956. Triassic metoposaurid amphibians. Bull. Am. Mus. nat. Hist. 1 10, 399-452.
cosgriff, j. w. 1972. Parotosaurus wadei , a new capitosaurid from New South Wales. J. Paleont. 46, 545-555.
1974. Lower Triassic Temnospondyli of Tasmania. Spec. pap. geol. Soc. Am. 149, I 134.

1984. The tenmospondyl labyrinthodonts of the earliest Triassic. J. vertebr. Paleont. 4, 30 46.

— and defauw, s. l. 1987. A capitosaurid labyrinthodont from the Lower Scythian of Tasmania. Alcheringa,
11,21 41.

Getmanov, s. n. 1981. On some regularities of skull growth in the benthosuchids. Paleont. Zh. 1981, 110 116.
[In Russian.]

hecht, m. k. and edwards, J. L. 1977. The methodology of phylogenetic inference above the species level. In
hecht, M. K., goody, p. c. and hecht, b. M. (eds.). Major patterns of vertebrate evolution , 3-51. Plenum,
New York.

howie, A. A. 1970. A new capitosaurid labyrinthodont from East Africa. Palaeontology , 13, 210 253.

1972. A brachyopid labyrinthodont from the Lower Trias of Queensland. Proc. Linn. Soc. NSW. 96,
268-277.

ingavat, r. and janvier, p. 1981. Cyclotosaurus cf .postumus Fraas (Capitosauridae, Stereospondyli) from the
Huai Hin Lat Formation (UpperTriassic), northeastern Thailand, with a note on capitosaurid biogeography.
Geobios , 14, 711-725.

jensen, a. r. 1975. Permo-Triassic stratigraphy and sedimentation in the Bowen Basin, Queensland. Bull. Bur.
miner. Resour. Geol. Geophys. Aust. 154, 187 pp.

jupp, r. and warren, a. A. 1986. The mandibles of the Triassic labyrinthodont amphibians. Alcheringa , 10,
99 124.

kamphausen, d. and morales, m. 1981. Eocyclotosaurus lehmani , a new combination for Stenotosaurus lehmani
Heyler, 1969 (Amphibia). Neues Jb. Geol. Paldont. Mh. 1981, 651-656.
konzhukova, e. d. 1965. A new capitosaur from the Triassic of Cisuralia. Paleont. Zh. 1965, 97 104. [In
Russian.]

Lehman, j. p. 1961. Les Stegocephales du Trias de Madagascar. Ann. Paleont. 47, 109-154.
meyer, h. von. 1858. Labyrinthodonten aus dem bunten Sandstein von Bernberg. Palaeontographica , 6,
221 245.

morales, M. 1987. A cladistic analysis ofcapitosauroid labyrinthodonts: preliminary results. J. vertebr. Paleont.
7 (Suppl.), 21A.

— and kamphausen, d. 1984. Odenwaldia heidelbergensis, a new benthosuchid stegocephalian from the
Middle Bunlsandstein of the Odcnwald, Germany. Neues Jb. Geol. Paldont. Mh. 1984, 673 683.

otschev, v. G. 1 966. Systematics and phytogeny of capitosaurid labyrinthodonts, 1 84 pp. Saratov State University
Press, Saratov. [In Russian ]

876

PALAEONTOLOGY, VOLUME 31

otschev, v. G. 1972. Capitosaurid labyrinthodonts from the Southeastern European part of the USSR , 269 pp.
Saratov State University Press, Saratov. [In Russian.]

— and shishkin, m. a. 1 986. In kalandadze, n. n. et al. (ed. ). Katalog permskikh i triasovykh tetrapod SSSR.
Dokl. Akad. Nauk SSSR 179, 72-91. [In Russian.]

riabinin, a. n. 1930. A labyrinthodont stegocephalian Wetlugasaurus angustifrons nov. gen., nov. sp. from the
lower Triassic of Vetluga Land in northern Russia. Ezheg. russk. paleont. Obshlch , 8, 49 -76. [In Russian.]
romer, a. s. 1939. Notes on branchiosaurs. Am. J. Sci. 237, 748-761.

save-soderbergh, G. 1935. On the dermal bones of the head in labyrinthodont stegocephalians and primitive
Reptilia with special reference to Eotriassic stegocephalians from East Greenland. Meddr. Gronland. 98,
1-211.

schroder, h. c. 1913. Ein Stegocephalen-Schadel von Elelgoland. Jb. preuss. geol. Landesanst. BergAkad. 33,
232-264.

sennikov, a. G. 1981. A new wetlugasaur from the Samara River Basin. Paleont. Zh. 1981, 143-148. [In
Russian]

snell, N. 1986. The postcranial skeleton of Mesozoic temnospondyl amphibians, xii+ 125 pp. B.Sc. (Honours)
thesis (unpublished). Department of Zoology, La Trobe University, Bundoora.
steen, M. c. 1937. On Acanthostoma vorax Credner. Proc. zool. Soc. Lond. 3, 491-500.
warren, a. a. 1980. Parotosuchus from the Early Triassic of Queensland and Western Australia. Alcheringa,
4, 25-36.

— 1981. A horned member of the labyrinthodont superfamily Brachyopoidea from the Early Triassic of
Queensland. Ibid. 5, 273-288.

— and black, T 1985. A new rhytidosteid (Amphibia, Labyrinthodontia) from the Early Triassic Arcadia
Formation of Queensland, Australia, and a consideration of the relationships of Triassic temnospondyls. J.
vertebr. Paleont. 5, 303-327.

— and hutchinson, m. n. 1983. The last labyrinthodont? A new brachyopoid (Amphibia, Temnospondyli)
from the Early Jurassic Evergreen Formation of Queensland, Australia. Phil. Trans. R. Soc. B303, 1-62.

— In press. The Madagascan Capitosaurs. Bull. Mus. natn Hist, nat., Paris.

watson, d. m. s. 1919. The structure, evolution and origin of the Amphibia, the 'orders’ Rhachitomi and
Stereospondyli. Phil. Trans. R. Soc. B209, 1-73.

— 1958. A new labyrinthodont (Paracyclotosaurus) from the Upper Trias of New South Wales. Bull. Br.
Mus. nat. Hist. (Geol.), 3, 233-264.

1962. The evolution of the labyrinthodonts. Phil. Trans. R. Soc. B245, 219-265.

welles, s. p. and cosgriff, j. w 1 965. A revision of the labyrinthodont family Capitosauridae and a description
of Parotosaurus peabodyi n. sp. from the Wupatki Member of the Moenkopi Formation of Northern
Arizona. Univ. Calif. Pubis geol. Sci. 27, 241-289.

A. A. WARREN
Department of Zoology
and

M. N. HUTCHINSON

Typescript received 5 May 1987
Revised typescript received 9 December 1987

School of Biological Sciences
La Trobe University
Bundoora

Victoria 3083, Australia

QUATERNARY DINOFLAGELLATE CYST
BIOSTRATIGRAPHY OF THE NORTH SEA

by REX HARLAND

Abstract. The dinoflagellate cyst biostratigraphy of Quaternary sediments in the North Sea is described. The
data accumulated demonstrate the recognition of glacial, interstadial, and interglacial periods but do not
necessarily date the relevant sediments. Certain major events such as the distinctive change from the Early
Pleistocene to Middle and Late Pleistocene conditions are particularly noted, as is the onset of the modern
oceanographic situation, all of which have distinctive signals in the dinoflagellate cyst record. The potential
for using dinoflagellate cysts in correlating shelf, slope, and ocean sediments is stressed.

The Quaternary is characterized by climatic fluctuations that have served long as the basis for its
subdivision. Indeed climatic fluctuation is accepted as the guiding principle for defining its various
stages (Shotton 1973). Imbrie (1985) admits that even after 150 years of study no fully satisfactory
theory exists to explain all climatic variations. Nonetheless it is now increasingly accepted that on
a 10 000 to 400 000 year time-scale, variation in the Earth’s orbit including eccentricity, obliquity,
and precession is the fundamental cause of climatic fluctuations. At the smaller scale there is evidence
that changes in solar activity, or episodes of volcanism may exert some influence.

Although the effects of climatic change may be quite differently recorded depending upon the
geographic position of the recipient site the driving force is almost certainly planetary. Therefore
the initiation of the various effects must be essentially isochronous even though the rate of response
of the physical and biological system will be different, not least because of the many complex feed-
back systems that operate.

Historically the possibility of significant climatic change was first realized in the terrestrial environ-
ment from the recognition of glaciogenic sediments in areas not currently affected by glacial activity.
More recently the marine record has come under increasingly closer scrutiny with the availability
of ocean sediment cores and the techniques of oxygen isotope and palaeomagnetic analysis. Chemical
(Arrhenius 1952), micropalaeontological (Ericson et al. 1956), and oxygen isotope analyses (Shackle-
ton 1969) have given way to such integrated studies as the work of the CLIMAP (Cline and Hays
1976) and OSKAP (Stabell and Thiede 1985) projects. This approach has vastly improved the
understanding of the nature, frequency, and effect of major climatic events in the marine Quaternary
record (West 1985).

Despite these major advances based upon deep-ocean marine sediments relatively little is known
of the contemporaneous continental shelves, which promise much in linking the deep ocean and
terrestrial records. Some significant progress has been made in the North Atlantic area around the
British Isles (Binns, Harland and Hughes 1974; Binns, McQuillin and Kenolty 1974; Caston 1977;
Holmes 1977; Thomson and Eden 1977; Pantin 1978; Skinner and Gregory 1983; Stoker et at. 1983,
1985«, b; Davies et al. 1984), the Netherlands (Jansen 1976, 1980; Jansen et al. 1979), Norway
(Feyling-Hanssen 1981, 1982; Knudsen 1985; Mangerud et al. 1984; Stabell and Thiede 1985), and
Canada (Mudie and Aksu 1984; Scott et al. 1984; Aksu and Mudie 1985).

A major problem in Quaternary shelf sediment studies is the provision of a reliable biostratigraphy.
Shotton (1973) points out that a biozonation based upon the appearance and extinction of species
is impractical because the duration of the Quaternary is insufficient to encompass more than one or
two biozones at most. In all groups the extant species are often dominant in Quaternary assemblages
and so any biostratigraphical divisions are necessarily the result of interpreted environmental change

IPalaeontology, Vol. 31, Part 3, 1988, pp. 877-903, pis. 78-82.|

© The Palaeontological Association

878 PALAEONTOLOGY, VOLUME 31

as exemplified by the changing sequential assemblages of pollen and Coleoptera (Moore and Webb
1978; Coope 1977).

Fossil groups used in the recognition of environment/climatic change within the marine realm
include planktonic and benthonic Foraminifera, ostracodes, molluscs, coccoliths, and diatoms.
However, of late, one group, the dinoflagellates, is becoming increasingly utilized. These marine
planktonic algae (Division Pyrrhophyta), contain genera and species that produce hypnozygotic
cysts resistant to bacterial decay and hence with fossilization potential. Dinoflagellate cysts in marine
Quaternary sediments can be used to decipher environmental and climatic history (Dale 1983, 1985).
Recent studies of such cysts have underlined their usefulness in the interpretation of the marine
Quaternary record on land (Wall and Dale 1 968c/), on the continental shelf (Harland 1977), and in
the deep ocean (Turon 1980). The potential for correlating the shelf with the deep ocean, a potential
not shared by either planktonic Foraminifera or coccolithophores, has not yet been fully realized
(Harland 1984c; Bakken and Dale 1986), nor has their use in charting climatic of oceanographic
change throughout the marine realm.

The present paper attempts to document the Quaternary dinoflagellate cyst biostratigraphy for
the North Sea area and to relate it where possible to oceanographic fluctuations in the North
Atlantic Ocean and to climate change in the Northern Hemisphere. It is based on studies at
the British Geological Survey (BGS) for the Marine Earth Sciences Research Programme and
the East Anglian Regional Mapping Programme and centres around the central North Sea.
Other work includes recent analyses of sediment cores from the outer continental shelf and the
continental slope of the north-west of the British Isles, and to work published on DSDP Legs 80
and 81 (Harland 1984<7, b).

MATERIAL AND METHODS

The study samples, collated from vibrocores and boreholes drilled as part of the BGS exploration of eastern
England and the continental shelf, are of clay, silt, or fine sand; finer grade material being preferred over
coarse because dinoflagellate cysts tend to act as sedimentary particles of fine silt size (Dale 1976).
Details of the vibrocores and boreholes may be found in BGS registers at Keyworth and Edinburgh
and many are described in the Institute of Geological Sciences (1974 et seq .) and British Geological Survey
(1984 et seq.).

All the samples are cleaned and only those portions thought free of outside contamination were processed.
Normal palynological processing was used throughout but the samples were subjected to the sintered glass
funnel technique of Neves and Dale (1963) for washing, concentrating, and staining. No oxidizing method
was used, if at all possible, in an attempt to reduce the loss of the more susceptible peridiniacean cysts
(Dale 1976). Strew slides were made by dispersing the microfossils on coverslips and then mounting
in Elvacite.

As a general rule a single slide per sample was counted for its dinoflagellate cyst content and to give the
proportions of the various species. Although the technique was standardized as far as possible to yield
consistent results, at this reconnaissance level the results can only be semi-quantitative at best. Rich and diverse
samples were counted to give a minimum of some 100 specimens for any one particular species. This method
has proved sufficient, in samples that contain widely different numbers of cysts, to recognize patterns of
fluctuation. Such counts for the majority of samples where less than twenty species are present give cyst
proportions with errors between 3 % and 9 % of the estimated percentages at two standard deviations,
depending upon numbers of specimens counted (Van der Plas and Tobi 1965).

The dinoflagellate cyst spectra illustrate the proportions of the various genera and/or species together with
the numbers counted. The number of counted cysts per slide is also a useful, if limited, ‘rule of thumb’ guide
to the richness of the samples.

Although the methodology outlined above is not statistically rigorous the patterns of dinoflagellate cyst
fluctuations and climatic change are thought to be real. They have largely been confirmed by the study of
other fossil groups, e.g. benthonic Foraminifera, and by other geological techniques.

All the slides, records, and illustrated specimens are housed in the palynological collections of the BGS at
Keyworth.

HARLAND: QUATERNARY DINOFLAGELLATE CYSTS

879

INTERPRETATION OF THE DINOFLAGELLATE CYST RECORD

The dinoflagellate cyst analysis of continental shelf sediments has used various interpretative
methods. Some of the earlier work relied heavily upon the recognition of sedimentary units favour-
able or unfavourable for dinoflagellate cysts which were largely equated with climatic ameliorations
(interglacials or interstadials) and deteriorations (glacials) respectively.

This led to the documentation of various climatic sequences and attempts at correlation (Harland
1973, Harland 1974, Binns, Harland and Hughes 1974 and Hughes et al. 1977) but suffered from
difficulties in the recognition of changes in dinoflagellate richness, because of lithological variations,
and lacked precision in the use of syn- and autecological data from the study of modern dinoflagel-
lates and their cysts.

The method was later supplemented by limited ecological data as it became available. Nonetheless
sequences were described in terms of patterns of favourability and unfavourability, plus the growing
recognition that certain dinoflagellate cyst species were important in imparting specific ecological
information, especially in respect of changes in water mass and hence the influence of the North
Atlantic Current (Harland 1977; Harland et al. 1978; Gregory and Harland 1978).

More recently work on sequences recovered from the Deep Sea Drilling Project (Harland 1979,
1984a, b), on dinoflagellate cyst thanatocoenoses (Reid 1975; Reid and Harland 1977; Wall et al.
1977; Turon 1980; Harland 1983; Bradford and Wall 1984; Mudie and Short 1985; Matsuoka
1985 6), and from living dinoflagellate cysts (Dale 1976, 1983, 1985; Balch et al. 1983; Lewis et al.
1984) has produced a growing volume of relevant data greatly assisting the understanding of the
ecological requirements of many dinoflagellates

It has also become possible to examine the contained dinoflagellate cyst assemblage, whether rich
or poor, in terms of species presence alone and interpreted from a knowledge of dinoflagellate
ecology. The literature now contains sequences for which diagrams have been drawn showing the
changing relative frequencies of cysts present (Turon 1980; Harland 1982; Cameron et al. 1984;
Harland 1984 <7, b, c; Scott et al. 1984; Dale 1985; Long et al. 1986). Thus dinoflagellate cyst spectra
have begun to be constructed for marine Quaternary sequences in the same way as pollen spectra
have been drawn for continental sequences.

Here sequences are categorized by their dinoflagellate cyst content, and diagrams are drawn to
illustrate the cyst assemblages for the various seismostratigraphic units. Units of rich dinoflagellate
cyst occurrence are easily recognizable and interpreted using autecological data.

Especially important, in the context of changing climatic environments, is the recognition of the
influence of Atlantic water, i.e. north-temperate and normally saline waters with rich and diverse
associations of Operculodinium centrocarpum (Deflandre and Cookson) Wall, Nematosphaeropsis
labyrinthea (Ostenfeld) Reid, Spiniferites membranaceus (Rossignol) Sarjeant, S. mirabilis (Rossig-
nol) Sarjeant, and S. ramosus (Ehrenberg) Loeblich and Loeblich, together with some Protoperidin-
ium species such as P. conicum (Gran) Balech, P. leonis (Pavillard) Balech, and P. pentagonum
(Gran) Balech; and more Arctic water with poorer and less diverse associations of Bitectatodinium
tepikiense Wilson, elongate Spiniferites spp., and such round brown Protoperidinium spp. as P.
conicoides (Paulsen) Balech. Transitional situations also exist and often the assemblage sequences
are complex with proportions of cysts not well known from modern environments.

The dinoflagellate associations reflect the same kind of changing environment as those that have
been documented by the CLIMAP project (Cline and Hays 1976; Ruddimann and McIntyre 1981)
using other fossil groups. Although they have not been applied in sufficient detail to test the
precision and sensitivity of the group, patterns of climatic change are most definitely reflected in the
dinoflagellate cyst assemblages.

Finally the interpretation of the cyst record has been somewhat complicated by problems in
rationalizing two systems of taxonomy, originating because of the separate study of living motile
dinoflagellates by phycologists, and the study of cysts by palaeo-palynologists. The use of incubation
experiments (Wall and Dale 19686) and more recently by Matsuoka (1984, 1985a), Matsuoka et al.
(1982), and Lewis et al. (1984) has allowed some integration of systems (Harland 1982), but not

880

PALAEONTOLOGY, VOLUME 31

without controversy (Dale 1983). At present several procedures are used which include the use of
the fossil nomenclature, modern biological nomenclature, and an amalgamation of the two systems.
This reflects our present knowledge but more particularly is an honest attempt to use the maximum
amount of information inferred by the use of any particular name.

STRATIGRAPHY

Introduction

Stoker et al. (1985a, b) presented a stratigraphic framework for Quaternary sediments in the central part of
the North Sea following the earlier works of Holmes (1977) and Thomson and Eden (1977). Their synthesis

7qo

text-fig. 1 . Sketch map of the north-east Atlantic Ocean and Norwegian Sea showing the general bathymetry
and location of the various boreholes, cores and vibrocores.

HARLAND: QUATERNARY D I NOFL AGELL ATE CYSTS

is based upon a seismostratigraphic approach but includes lithological, geotechnical, palaeomagnetic, and
micropalaeontological data including the analysis of dinoflagellate cysts.

Ten major Quaternary formations were formally recognized by Stoker et al. (1985Z>) that individually can
reach some 200 m in thickness. The base of the oldest Quaternary formation was not observed or sampled
and indeed the actual base of the Quaternary itself cannot be identified with any certainty.

In Britain the base of the Quaternary has been taken at the base of the Waltonian Red Crag (Shotton 1973)
or perhaps better termed the Pre-Ludhamian (Beck et al. 1972). However, Funnell in Curry et al. (1978) argues
that the Waltonian should be regarded as a part of the Pliocene such that the Plio/Pleistocene boundary must
lay somewhere within or above the Red Crag. Berggren et al. (1985) report a resolution to the IUGS
recommending the boundary be taken at the top of marker bed e at about 3-6 m above the Olduvai normal
polarity event within the Matuyama reversed epoch at the Le Castella Section. This is at I -6 My and commonly
used as the boundary in the central part of the North Sea.

Although details of the North Sea stratigraphy are presented in Stoker et al. (1985/r) the analysis of the
dinoflagellate cyst floras was not given there, and hence will be documented herein. The dinoflagellate cyst
floras are described formation by formation with relevant data from the North Sea and north-eastern North
Atlantic included where pertinent. The formations are discussed from oldest to youngest. The dinoflagellate
cysts mentioned are illustrated by stereoscan photomicrographs where possible but reference should be made

SW NE

text-fig. 2. Correlation and lateral variation of the North Sea stratigraphical succession along a south-west
north-east transect in relation to the north-west European and British Quaternary stages (based largely on

Stoker et al. 1 985Z?).

882

PALAEONTOLOGY, VOLUME 31

text-fig. 3. Dinoflagellate cyst biostratigraphy of the Aberdeen Ground Formation in Borehole 81/34, lat.
56 7-68' N., long. 1 35-21' E. a, Operculodinium centrocarpwn (Deflandre and Cookson) Wall with O.
israelianum (Rossignol) indicated in black, b, Bitectatodinium tepikiense Wilson with Tectatodinium pellitum
Wall indicated in black, c, Spiniferites cysts with Achomosphaera andalousiensis Jan du Chene in black, d,
Protoperidinium cysts. Small ticks in first column indicate sample levels.

to Harland (1977, 1983) for the taxonomy and to Dale (1983) and Harland (1983) for the ecology and cyst
distributions respectively.

Aberdeen Ground Formation

The type sequence occurs in Borehole 81/34 between 142.0 and 229.1 m, but unfortunately the borehole did
not prove the base. The location of Borehole 81/34 and all subsequent cores are shown in text-fig. 1. The
interpreted sequence and lateral variations are illustrated in text-fig. 2. The Aberdeen Ground Formation
consists of dark-grey to brown, very stiff to hard silty muds with some shelly and pebbly sands. Stoker et cd.
(1983) have identified the Brunhes/Matuyama palaeomagnetic boundary within the formation, and have
indicated a Tiglian to 'Cromerian Complex’ (late Antian to Cromerian) age (Early to Middle Pleistocene).

The dinoflagellate cyst spectrum for the Aberdeen Ground Formation in Borehole 81/34 is given in text-fig.
3. It is immediately apparent that samples between 200 m and 214 m yielded rich dinoflagellate cyst assemblages
in contrast to the remainder of the section. This suggests that only during this time were conditions favourable
enough to allow a relative rise in the recruitment (no. of cysts per gram of sediment being incorporated at any
particular time) of the dinoflagellate cysts. During this interval two distinct episodes are recognized. There is
an older period dominated by B. tepikiense (c. 50%) (text-fig. 4) and a younger dominated by Spiniferites spp.
(e. 60 %) (PI. 79, figs. I 6). The cyst Achomosphaera andalousiensis Jan du Chene (PI. 81, figs. 1 - 4) is consistently
present throughout the sequence. Although environmental conditions may be favourable, the presence of high
proportions of B. tepikiense and the persistence of A. andalousiensis suggest rather cold, north-temperate to
arctic-like environments. B. tepikiense is well known as a north-temperate cyst (Harland 1983; Dale 1983) and
although A. andalousiensis has rarely been recovered from modern sediments (Harland 1983; Balch et al. 1983),
it has been associated with cold north-temperate to arctic environments (Long et al. 1986). The upper part of
this section with higher proportions of Spiniferites spp., but not A. andalousiensis , and with lower proportions
of B. tepikiense may indicate the maximum occurrence of the amelioration.

HARLAND: QUATERNARY DINOFLAGELLATE CYSTS

883

text-fig. 4. Stereoscan photomicrographs of Bitectatodinium tepikiense Wilson, x 1200. A, specimen MPK
5276, Norwegian Sea, dorsal view with archeopyle and camerate V' apical margin and planate 4" margin, b,
specimen MPK 5277, Norwegian Sea, oblique dorsal view with planate 4" apical margin.

The presence of O. israelianum (Rossignol) Wall (PI. 82, fig. 1 1) and Tectatodinium pellitum Wall (PI. 82, fig.
10) in the younger assemblage probably does not suggest warmer-water conditions, as intimated in Stoker et
al. (19856), but may indicate reworking from Early Pleistocene sediments. Similarly Palaeogene reworking is
prevalent throughout the Aberdeen Ground Formation.

The remaining part of the dinoflagellate spectrum can also be interpreted as indicative of north-temperate
to arctic conditions but perhaps with some increasing uphole influence from the North Atlantic. The presence
of A. andalousiensis and B. tepikiense supports the north-temperate environment and the low proportions of
Protoperidinium spp. (round, brown cysts) (PI. 82, fig. 9) preclude the possibility of much sea-ice. Protoperidin-
ium dinoflagellates are heterotrophs, and therefore do not require the presence of light to survive (Bujak 1984;
Dale 1985). This is reflected in their distribution patterns along the Norwegian coast (Dale 1983) but less so
in the maps of Harland (1983). However, there is a noticeable rise in the proportions of Protoperidinium spp.
between 180 and 190 m in the sequence, possibly indicating a cooling of the environment and the introduction
of seasonal ice-cover.

In addition to the type borehole, dinoflagellate cyst analyses were completed upon other sequences of the
Aberdeen Ground Formation proved in additional North Sea boreholes. For instance Borehole 81/27 (see
text-fig. 5) yielded rich dinoflagellate floras dominated by T. pellitum with subsidiary Spiniferites spp. and
relatively low proportions of O. centrocarpum and O. israelianum. This kind of dinoflagellate cyst assemblage

text-fig. 5. Dinoflagellate cyst biostratigraphy of the Aberdeen Ground Formation in Borehole 81/27, lat.
56 32-7 T N., long. 0 23-10' W. Columns as in text-fig. 3.

884

PALAEONTOLOGY, VOLUME 31

is now regarded as indicating south-temperate to almost sub-tropical conditions in a neritic environment
(Harland 1983; Cameron el al. 1984) and not cool temperate environments as incorrectly interpreted by Wall
and Dale (1968a), following the pollen work of West (1961). The presence of both O. israelianum and T.
pellitum are indicative of quite different environmental conditions from the succeeding cyst floras in which
they are absent. Their presence is usually associated with Early Pleistocene sediments and the Matuyama
palaeomagnetic reversal. Indeed Stoker et al. (1983) have recorded reversed palaeomagnetism from sediments
of the Aberdeen Ground Formation in Borehole 81/27.

In summary, evidence indicates that the Aberdeen Ground Formation contains dinoflagellate cyst assem-
blages of wide-ranging environments including south-temperate to sub-tropical, north-temperate, and north-
temperate to arctic. The dinoflagellate cysts taken with the palaeomagnetic results indicates an older Early
Pleistocene part of the sequence and a younger ?Middle Pleistocene part, and indeed is part of the evidence
used by Stoker et al. (19856) to suggest a Tiglian to 'Cromerian Complex’ age range. The Aberdeen Ground
Formation is obviously a complex unit. It needs further study to circumscribe its age and environments of
deposition more closely.

The climatic ameliorations described from benthonic Foraminifera and dinoflagellate cyst evidence for the
sediments below the prominent seismic reflector in Borehole 75/33 (Harland et al. 1978; Gregory and Harland
1978) are now assignable to the Aberdeen Ground Formation. The dinoflagellate cyst and foraminiferal work
would appear to suggest a Middle Pleistocene and not a Fate Pleistocene age as originally suggested (Harland
1977; Harland et al. 1978; Gregory and Harland 1978). The radiocarbon dates quoted originally by Holmes
(1977), which led to an underestimation of age, are thought to be invalid (Stoker et at. 19856).

The recognition of a distinct change in the upper part of the Aberdeen Ground Formation between sediments
containing such dinoflagellate cysts as O. israelianum , and T. pellitum as common components, and to a lesser
extent by the presence of Amiculosphaera umbracula Harland (PI. 81, figs. 5 and 6) and Impagidinium multiple-
xum (Wall and Dale) Lentin and Williams (not illustrated) together with various undescribed Spiniferites spp.
and Protoperidinium spp., from sediments containing forms that commonly occur around the British Isles
today has been used as a practical guide to delineate an Early/Middle Pleistocene boundary. It also marks the
change between the fairly stable equable climate of the Early Pleistocene from the widely fluctuating situation
of the Middle and Fate Pleistocene. Studies by Wall and Dale (1968a), Reid and Downie (1973), and
Harland (unpubl. data) suggest that the marked dinoflagellate change falls within the presently defined Middle
Pleistocene possible as high as the Cromerian/Anglian boundary.

The early Pleistocene record is, nevertheless, characterized by sequences in which the proportions of various
cysts fluctuate markedly (Wall and Dale 1968a; Cameron et al. 1984) and it is likely that these fluctuations
together with some stratigraphical last and first appearances will lead to a dinoflagellate cyst biostratigraphy
for the ?Early Pleistocene. The work of Harland (1984a, 6) adds evidence from the oceanic record to these
suggestions and points to the possibility that this boundary may correlate with the NN 19/20 boundary.

EXPLANATION OF PLATE 78

All the stereoscan photomicrographs are illustrated at a magnification of x c. 1200 unless otherwise noted.
Full details of locality and horizon are to be found in the MPK registers of the BGS, Keyworth.

Figs. 1 and 2. Operculodinium centrocarpum (Deflandre and Cookson) Wall. 1 , Specimen MPK 5280, Norwegian
Sea, orientation unknown showing the nature of the cyst wall and processes. 2, specimen, MPK 5281, Bay
of Biscay, oblique dorsal view with the 1 P archeopyle formed by the loss of paraplate 3" and illustrating
the nature of the process morphology with the infundibular and multifurcate distal tips.

Fig. 3. Protoperidinium ( Protoperidinium sect. Selenopemphix) conicum (Gran) Balech. Specimen MPK 2958,
Firth of Forth, apical view showing the apical tuft of acicular processes, x c. 1000.

Fig. 4. P. {P. sect. Trinovantedinium) pentagonum (Gran) Balech. Specimen MPK 2956, Firth of Forth, dorsal
view illustrating overall cyst morphology and the broad hexa I archeopyle formed by loss of paraplate 2a,
xc. 1000.

Fig. 5. P. (P. sect. Quinquecuspis) leonis (Pavillard) Balech. Specimen MPK 2954, Firth of Forth, dorsal view
showing cyst morphology particularly the hexa I archeopyle formed by the loss of paraplate 2a and the
continuous paracingulum, xc. 1000.

Fig. 6. Nematosphaeropsis labyrinthea (Ostenfeld) Reid. Specimen MPK 5282, Bay of Biscay, orientation
unknown, overall cyst morphology and ribbon trabeculae.

PLATE 78

HARLAND, Operculodinium , Protoperidinium , Nematosphaeropsis

886

PALAEONTOLOGY, VOLUME 31

text-fig. 6. Dinoflagellate cyst biostratigraphy of the Ling Bank Formation in Borehole 81/34. Columns as

in text-fig. 3.

Ling Bank Formation

The type sequence for the Ling Bank Formation is to be found in Borehole 81/34 from 55 0 m to 142 0 m.
The formation consists of dense silts and silty sands with interbedded sands and clays especially in the upper
part. The sediments are normally magnetized and probably part of the Brunhes Normal Epoch. Dating of this
formation is difficult but Stoker et al. (1985a and b) have suggested a Flolsteinian to Saalian (Floxnian to early
Wolstonian) age.

The dinoflagellate cyst spectrum for the type sequence (text-fig. 6) reveals a series of favourable assemblage
that can be subdivided at about 83 0 m depth. The older is dominated by O. centrocarpum with Spiniferites
spp. and lower proportions of B. tepikiense and Protoperidinium spp. This phase indicates a marked influence
of the North Atlantic Current (Harland 1983), and without doubt can be attributed to an interglacial stage.
Also present in this interval is Achomosphaera andalousiensis , particularly towards the base and top with a
maximum proportion of 22-5 % at a level of 135-9 m (text-fig. 6), N. labyrinthea (PI. 79, fig. 6) which like A.
andalousiensis occurs towards the top and bottom, P. conicum (PI. 78, fig. 3; PI. 82, fig. 6) towards the middle
and base, P. pentagonum (PI. 78, fig. 4; PI. 82, figs. 3 and 4) in the middle part of the sequence only and various
Spiniferites spp. The Spiniferites spp. include S. elongatus Reid (PI. 80, fig. 6) which occurs throughout, and

EXPLANATION OF PLATE 79

All the stereoscan photomicrographs are illustrated at a magnification of x c. 1200 unless otherwise noted.
Full details of locality and horizon are to be found in the MPK registers of the BGS, Key worth.

Figs. 1 and 2. Spiniferites ramosus (Ehrenberg) Loeblich and Loeblich. 1, specimen MPK 5283, Bay of Biscay,
dorsal view to show the IP reduced archeople formed by the loss of paraplate 3", the paratabulation and
the trifurcate processes with bifid distal tipes. 2, specimen MPK 5284, Bay of Biscay, slightly oblique dorsal
view.

Figs. 3 6. S. lazus Reid. 3, specimen MPK 5285, north-eastern Atlantic Ocean, oblique dorsal view to illustrate
the IP reduced archopyle formed by the loss of paraplate 3" and the fenestrate nature of the process bases.
4, detail of fenestrate process base, x c. 1 2 000. 5, specimen MPK 5287, Bay of Biscay, dorsal view to show
archeopyle, paratabulation, and deeply trifurcate nature of the processes. 6, detail of trifurcation of process
together with the fenestrate process bases, x c. 2400.

PLATE 79

HARLAND, Spiniferites

PALAEONTOLOGY. VOLUME 31

text-fig. 7. Dinoflagellate cyst biostratigraphy of the Fisher Formation in Borehole 81/34. Columns as in

text-fig. 3.

S. mirabilis and S. ramosus (PI. 79, figs. I and 2) that occur more frequently in the middle part of the sequence.
This pattern can be interpreted as indicative of changing conditions within the interglacial with a cool initiation,
a warm middle period, and a cool final phase.

Above 82 0 m the character of the assemblages changes with samples showing a reduction in specimen
numbers. There is a marked decrease in the proportion of O. centrocarpum with a reciprocal increase in the
proportions of B. tepikiense and Spiniferites spp. especially A. andalousiensis. S. elongatus is consistently
present, with S. membranaceus (PI. 82, figs. 7 and 8) and S. ramosus occurring occasionally. This part of the
sequence can be interpreted as the onset of poorer environmental conditions at the end of the interglacial
possibly due to a more north-temperate to arctic influence. The penetration of the North Atlantic Current
may not be as great but it is unlikely that the area was much affected by ice-cover because there is a lack of
heterotrophic Protoperidinium species (Dale 1983, 1985).

Dating of the Ling Bank Formation is difficult from the dinoflagellate cysts alone but undoubtedly it
contains the record of an interglacial. Stoker et al. (1985u and b) favour a Holsteinian to Saalian age on general
stratigraphic relationships.

Fisher Formation

The Fisher Formation type sequence occurs in Borehole 81/34 between 15-3 and 55 0 m depth. The formation
consists of over-consolidated clays and silty sands with occasional shell fragments and pebbles. The sediments
are normally magnetized and are likely to be part of the Brunhes Normal Epoch, although a single reversed
horizon has been noted (Stoker et al. 1985b). A Saalian (Wolstonian) age has been suggested for the formation
(Stoker et al. 1985b) although Holmes (1977) recorded a late Devensian radiocarbon age of 23 170 years b.p.
from partially lignitized wood from a commercial borehole.

The dinoflagellate cyst spectrum (text-fig. 7) is poor with many of the recovered assemblages being rep-
resented by 100 specimens or less. This lack of recovery is in itself an indication of poor, unfavourable

EXPLANATION OF PLATE 80

All the stereoscan photomicrographs are illustrated at a magnification of x c. 850 unless otherwise noted. Full
details of locality and horizon are to be found in the MPK registers of the BGS, Keyworth.

Figs. 15. Spiniferites mirabilis (Rossignol) Sarjeant. 1, specimen MPK 5289, Bay of Biscay, dorsal view
showing the 1 P reduced archeopyle formed by the loss of paraplate 3" and the extensive antapical membrane.
2, specimen MPK 5291, Bay of Biscay, dorsal view showing a rather less extensive antapical membrane but
many trifurcate parasutural processes with bifid distal tips. 4, specimen MPK 5292, Bay of Biscay, oblique
dorsal view showing a cluster of processes surmounting the apex. 5, specimen MPK 5293, Bay of Biscay,
right lateral view showing the many parasutural processes.

Fig. 6. Spiniferites elongatus Reid. Specimen MPK 3990, Barents Sea, dorsal view showing the IP reduced
archeopyle, the elongate morphology and development of parasutural membranes, x c. 1200.

PLATE 80

HARLAND, Spiniferites

890

PALAEONTOLOGY, VOLUME 31

conditions. The assemblages observed are for the most part dominated by O. centrocarpum and Spiniferites
spp. A. andalousiensis is persistently present alongside S. elongatus which occurs as the ?ecophenotypic form
of 5. frigidus Harland and Reid (Harland and Sharp 1986). Protoperidinium cysts, as the round, brown
morphotypes, occur throughout the formation albeit in small proportions. More importantly is the presence
of Multispinula minuta Harland and Reid (PI. 82, fig. 12), a form commonly associated with arctic environments
in the Canadian offshore area (Harland et al. 1980; Mudie and Aksu 1984; Scott et al. 1984).

The dinoflagellate cyst spectrum is indicative of a somewhat intermediate situation between normal north-
temperate conditions and severe arctic environments. The poor assemblages, the presence of A. andalousiensis ,
B. tepikiense, M. minuta , and S. elongatus all point to cold environments whereas the presence of richer
assemblages may indicate rather open marine conditions with some influence from the North Atlantic. The
presence of sea-ice, for instance, may have been seasonal but undoubtedly the environment is difficult to
categorize. Variations in the cyst spectrum appear to suggest some short-lived climatic or environmental
changes at 24-0 and 32-0 m with the latter yielding assemblages overwhelmingly dominated by A. andalousiensis.
The factors causing such changes are unknown and indeed autecology data for A. andalousiensis is lacking
(Harland 1983), although it may be more typical of cooler north-temperate to arctic waters (Long et al. 1986).
The slight uphole increase in the proportions of O. centrocarpum and the decrease in B. tepikiense are probably
in anticipation of more favourable environments.

The dinoflagellate cyst evidence from the Fisher Formation cannot in itself give a definitive age but the
environmental interpretation suggests an assignment to a glacial and not an interglacial stage.

Coed Pit Formation

The type section for the Coal Pit Formation occurs from 32-0 to 107-5 m in Borehole 81/37. The formation
consists of dark-grey to brownish-grey, muddy pebbly sands and hard dark-grey, silty pebbly muds to silty
muds, sandy silts, and fine to very fine sands. The sediments are mostly normally magnetized and can be
assigned to the Brunhes Normal Epoch but some reversed polarity episodes have been identified possibly
corresponding to the Blake Event and Laschamp Excursion (Stoker et al. 19856). The Coal Pit Formation is
thought to be of Saalian to Weichselian (Wolstonian to Devensian) age and includes the Eernian (Ipswichian)
interglacial.

The dinoflagellate cyst spectrum for the Coal Pit Formation is divisible into three (text-fig. 8). This subdivision
results from the recognition of a sequence of sediments between 72 0 to 102 0 m depth that contains particularly
rich dinoflagellate cyst assemblages. These assemblages are all dominated by O. centrocarpum (up to 75 %)
with minor proportions of B. tepikiense , Spiniferites spp., and Protoperidinium spp. Included within the
Spiniferites column (text-fig. 8) is A. andalousiensis together with S. elongatus; occasionally present are the
species S. mirabilis and 51. ramosus. This assemblage is consistent with a more ameliorative environment of
deposition than the remaining sediments of the Coal Pit Formation despite the presence of the more northerly
cold water indicators A. andalousiensis and .S', elongatus. The presence of N. labyrinthea and S. mirabilis also
suggest an eastern Atlantic component (Harland 1983).

The remaining parts of the dinoflagellate cyst spectrum show a less productive aspect with lower proportions
of O. centrocarpum and greater proportions of B. tepikiense, Spiniferites spp., and Protoperidinium spp.
Although less favourable environmental conditions are envisaged, the lack of change in the dinoflagellate cyst
proportions suggest some input from the North Atlantic. However, it is possible that some of these changes
in productivity may result from lithological change.

EXPLANATION OF PLATE 81

All the stereoscan photomicrographs are illustrated at a magnification of x c. 1200 unless otherwise noted.
Full details of locality and horizon are to be found in the MPK registers of the BGS, Keyworth.

Figs. 1, 2, 4, 6. Achomosphaera andalousiensis Jan du Chene. 1, specimen MPK 5294, Bay of Biscay, oblique
dorsal view showing the IP reduced archeopyle, the subdued parasutural ridges and the reticulate process
tips. 2, specimen MPK 5295, Bay of Biscay, oblique dorsal view showing the reticulate process tips especially
in the region of the paracingulum. 4, detail of process tips, x c. 2400. 6, specimen MPK 5296, Bay of
Biscay, oblique dorsal view.

Figs. 3 and 5. Amiculosphaera umbracula Harland. 3, specimen, MPK 4339, Bay of Biscay, dorsal view showing
the periphragmal archeopyle, x c. 1000. 5, specimen MPK 5298, Bay of Biscay, dorsal view, x c. 1000.

PLATE 81

HARLAND, Achomosphaera , Amiculosphaera

892

PALAEONTOLOGY, VOLUME 31

50 50 50 50 100 200

A B C D Cysts/Slide

text-fig. 8. Dinoflagellate cyst biostratigraphy of the Coal Pit Formation in Borehole 81/37, lat. 56° 4743'

N., long. 1° 3147' E. Columns as in text-fig. 3.

EXPLANATION OF PLATE 82

All the photomicrographs are with Nomarski interference contrast and are illustrated at a magnification of
x 500. Full details of locality and horizon are to be found in the MPK registers of the BGS, Keyworth.

Figs. 1 and 2. Protoperidinium ( Protoperidinium sect. Quinquecuspis) leonis (Pavillard) Balech. Specimen MPK
2781, Firth of Forth. 1, ventral epicystal view illustrating the deeply inset parasulcus. 2, dorsal hypocystal
view by transparency with the continuous paracingulum and single intercalary operculum formed by the
loss of paraplate 2a.

Figs. 3 and 4. P. {P. sect. Trinovantedinium) pentagonum (Gran) Balech. Specimen MPK 1240, Irish Sea. 3,
dorsal view by transparency with the broad hexa single intercalary archeopyle formed by the loss of paraplate
2a. 4, ventral view illustrating the continuous paracingulum and nature of the processes.

Fig. 5. Nematosphaeropsis labyrinthea (Ostenfeld) Reid. Specimen MPK 2963, Barents sea, ?oblique ventral
view showing the overall morphology.

Fig. 6. P. (P. sect. Selenopemphix) conicum (Gran) Balech. Specimen MPK 2772, Firth of Forth, apical view
of cyst showing the morphology and the offset standard hexa single intercalary archeopyle formed by the
loss of paraplate 2a.

Figs. 7 and 8. Spiniferites membranaceus (Rossignol) Sarjeant. Specimen MPK 5299, North Sea. 7, dorsal view
illustrating the single reduced precingular archeopyle formed by the loss of paraplate 3" . 8, optical section
with the prominent and characteristic antapical membranous process.

Fig. 9. P. (P. sect. Brigantedinium ) conicoides (Paulsen) Balech. Specimen MPK 1232, Firth of Clyde, dorsal
view illustrating the single standard hexa intercalary archeopyle formed by the loss of paraplate 2a.

Fig. 10. Tectatodinium pellitum Wall. Specimen MPK 5595, BGS Ormesby Borehole, dorsal view showing the
nature of the single precingular archeopyle formed by the loss of paraplate 3" .

Fig. II. Operculodinium israelianum (Rossignol) Wall, specimen MPK 3117, Chillesford Clay, dorsal view
illustrating the broad, single precingular archeopyle formed by the loss of paraplate 3" .

Fig. 12. IMultispinuIa minuta Harland and Reid. Specimen MPK 1306, Beaufort Sea, Canadian Arctic,
orientation unknown.

PLATE 82

HARLAND, Quaternary dinoflagellate cysts

894

PALAEONTOLOGY, VOLUME 31

The Coal Pit Formation either in its entirety or in part appears to have been deposited in a more ameliorative
and favourable climatic environment than the underlying Fisher Formation. This suggests full interglacial
conditions and the establishment of the North Atlantic Current in a course not unlike that of today. Other
micropalaeontological evidence supports the recognition of an ameliorative episode between 72-0 to 102 0 m
depth but otherwise is indicative of a cold, harsh climate (Stoker et al. 1985b).

Wee Bankie Formation

This formation was first described by Thomson and Eden (1977) as the Wee Bankie Beds but has now been
formally adopted as a formation by Stoker et al. (1985b). The type sequence occurs in Borehole 72/20 from
sea-bed to about 33 0 m. Lithologically it consists of stiff, poorly-sorted polymictic till containing some
interbedded sands, pebbly sands, and silty clay.

Unfortunately and perhaps not unexpectedly there is no indigenous dinoflagellate cyst flora (Gregory et al.
1978). The deposit is interpreted by Stoker et al. (1985b) as being a basal till with the coarser sediment deposited
from sub-glacial streams. A late Weichselian (Devensian) age is most likely and its eastern geographical
boundary may mark the maximum offshore extent of the late Weichselian ice sheet (Stoker et al. 1985b).

Man- Bank Formation

The type section of the Marr Bank Formation occurs in Borehole 74/77 between 2-0 and 21 0 m depth. The
formation consists of very fine to coarse olive-grey to grey sands with occasional silty and gravelly horizons.

Dinoflagellate cyst recovery was poor. This recovery is consistent with other micropalaeontological evidence
(Gregory et al. 1978) suggesting deposition in a shallow, glacio-marine environment. A radiocarbon date of
17 734 + 480 years b.p. (Holmes 1977) has been recorded for this formation. Although the date confirms a late
Weichselian (Devensian) age Stoker et al. (1985b) regard the date as a minimum age.

text-fig. 9. Dinoflagellate cyst biostratigraphy of the Swatchway Formation in Borehole 75/33, lat. 58
4-30' N., long. 0° 33-83' E. Columns as in text-fig. 3

Swatchway Formation

The Swatchway Formation’s type section occurs in Borehole 75/33 between 17-3 and 26-3 m depth. The
formation comprises mainly muddy sands that pass northwards into clayey-silts and silty-clays with some
thin sands. Sediments examined from this unit have all been normally magnetized (Stoker et al. 1985b).

The dinoflagellate cyst spectrum (text-fig. 9) drawn for this formation is based upon limited evidence.
The moderately productive samples yielded assemblages co-dominated by O. centrocarpum and B. tepikiense
with only minor proportions of Spiniferites spp. and Protoperidinium spp. The high proportions of B.
tepikiense together with the presence of A. andalousiensis and S. elongatus indicate some northerly influence
but with a North Atlantic component. The low proportion of the heterotrophic Protoperidinium spp. may
suggest limited or non-existant sea-ice cover.

The evidence from the dinoflagellate cyst assemblages gives no direct indication of age, but taken with
the evidence from the Marr Bank Formation to which the Swatchway Formation may be, in part, laterally
correlated (Stoker et al. 1985b) indicates a period of slight warming. This may mark the beginning of a
North Atlantic Current influence as full glacial conditions began to give way to more amenable climates.

St Abbs Formation

The sequence recovered from 10.0 to 1 6-0 m in Borehole 73/1 1 was taken as the type section (Stoker et al.
1985b) following the earlier work of Thomson and Eden (1977). The formation consists of soft to stiff,
weakly laminated, grey to brown and pinkish muds and silty muds containing sporadic pebbles.

Dinoflagellate cyst analysis of the sequence proved to be unsatisfactory with most samples barren of
indigenous cysts. A single productive sample containing specimens of B. tepikiense , is consistent with other
micropalaeontological evidence (Gregory et al. 1978) in suggesting arctic marine environments.

HARLAND: QUATERNARY DINOFLAGELL ATE CYSTS

895

It has been suggested that the St Abbs Formation is equivalent to the Errol Beds of the Forth and Tay
estuaries (Thomson and Eden 1977) and, therefore, was deposited between 18 000 and 13 500 years b.p.
(Peacock 1981). If this is correct then the St Abbs Formation is of late Devensian age.

Witch Ground Formation

The Witch Ground Formation is divided into three members, which in ascending order are the Fladen, the
Witch, and the Glenn. The type section of the formation occurs in Borehole 75/33 from sea-bed to 17-3 m
depth. However, because the upper part of the sequence, the Glenn Member, is not well represented in Borehole
75/33, a vibrocore, 57/ + 00/9 at lat. : 57° 59.95' N., long.: 0 40- 15' E., was chosen as a supplementary type
sequence and is here represented by sediments from the sea-bed to T3 m. The formation consists of soft,
greenish-grey to greyish-brown clays and silts with the occasional sandy horizon. Sediments from the Witch
Ground Formation recovered from Borehole 75/33 all have normal polarity and, therefore, are thought to be
of a late Weichselian (Devensian) to Flandrian age (Stoker et at. 19856).

Witch Member
Fladen Member

A ^ u Cysts/Slide

text-fig. 10. Dinoflagellate cyst biostratigraphy of the Witch Ground Formation in Borehole 75/33. Columns

as in text fig. 3.

Dinoflagellate cyst analysis of the Witch Ground Formation from Borehole 75/33 (text-fig. 10) illustrates
three significant factors. First there is an increased dinoflagellate productivity uphole, second a change from
assemblages dominated by B. tepikiense to those dominated by O. centrocarpum , and thirdly the uphole
disappearance of A. andalousiensis. The significance of these factors lies in the fact that all these changes occur
at about the 10 0 m level in the borehole. This level may be interpreted as the change from a cold late Devensian
environment to the warm Flandrian as modern oceanographic conditions became established, or alternatively,
the onset of the Allerod Interstadial

This dinoflagellate cyst event coincides with the lithological boundary between the Fladen and Witch
Members (Stoker et al. 19856). Cyst evidence from the Fladen Member is consistent with deposition in a cold
climate, but without the undue effects of sea-ice, giving way to an environment increasingly influenced by the
North Atlantic.

The dinoflagellate cyst evidence suggests a late Devensian to early Flandrian age for the Fladen and Witch
Members of the Witch Ground Formation but the boundary between them may not coincide with the
Devensian/Flandrian boundary but with the onset of the Allerod Interstadial. The Fladen Member can be
correlated (Stoker et al. 19856) to the Fladen Deposits of Jansen et al. (1979) who suggested a 15 000 to
18 000 years b.p. age.

The Witch Member contains a dinoflagellate cyst flora similar to modern assemblages (Reid 1975; Harland
1983) with influence of the North Atlantic Current. The assemblages contain high proportions of O. centrocar-
pum and lower proportions of B. tepikiense with S. mirabilis , P. pentagonum. and P. conicum. The occurrence
of the last three species suggests conditions like those of today although the presence of A. andalousiensis , P.
conicoides, and S. elongatus indicate some influence from the north. Jansen et al. (1979) suggested a 8700 to
15 000 years b.p. age for the Lower Witch Deposits, a correlative of the Witch Member (Stoker et al. 19856).
The recently discovered Vedde Ash equivalent in vibrocore 58/ + 00/111 by Long et al. (1986), which is
dated at 10 600 years b.p. (Mangerud et al. 1984) would suggest a much older age for the top of the Witch
Member.

The uppermost Glenn Member was not analysed for dinoflagellate cysts in its type sequence but it has been
examined in vibrocore 58/ + 00/1 1 1 (Long et al. 1986). It is equivalent to Facies D of that sequence and includes
the Vedde Ash. The dinoflagellate cyst record illustrates a lower colder period (Facies C) thought to be
attributed to the Younger Dryas cooling between 10 000 and 1 1 000 years b.p., and an upper warm period
(Facies D) with the establishment of the present day oceanography.

896

PALAEONTOLOGY, VOLUME 31

Forth Formation

The type section of the Forth Formation occurs from sea-bed to 29 0 m depth in Borehole 71/33. The
formation is divided into four members; the Fitzroy, the Largo Bay, the Whitethorn, and the St Andrew’s Bay.
Lithologically the Forth Formation consists of muds overlain by pebbly muddy sands and soft silty muds,
and is considered (Stoker et al. 19856) to be laterally equivalent to the Witch Ground Formation and therefore
to late Weichselian (Devensian) to Flandrian age. This age is supported by a radiocarbon date of 7109 + 60
years b.p. (Holmes 1977). The various members of the Forth Formation are described more fully than those
of the Witch Ground Formation since the Largo Bay and St Andrew’s Bay Members occur to the west of the
central North Sea and the Fitzroy and Whitethorn Members to the east, in the Devil’s Hole area (Stoke et al.
1985b) (see text-fig. 2)

Borehole 74/1 provides the type section for the Largo Bay member which occurs between 5-0 nr to 25-0 m.
Dinoflagellate cyst recovery was poor but the cyst B. tepikiense was noted as the commonest species together
with a few Protoperidinium cysts. This evidence is consistent with cold environments but perhaps with little or
no sea-ice since so few Protoperidinium spp. were observed. Gregory et al. (1978) detail further micropalaeonto-
logical data suggestive of less than present day temperatures but not as cold as those suggested for the St Abbs
Formation. Judging the evidence of climate and the stratigraphical relationship of this member with its
associated strata Stoker et al. (19856) believe deposition occurred during the late Weichselian (Devensian)
between 10 000 and 13 500 years b.p.

The type section for the Fitzroy Member occurs in Borehole 81/39 between 110 and 60 0 m depth.
Palaeomagnetic analysis indicates normal polarity. The dinoflagellate cysts are generally sparse or absent in
the sediments. Assemblages recovered are dominated by O. centrocarpum together with B. tepikiense and
fewer Spiniferites spp. and Protoperidinium cysts. Evidence points to cool climatic conditions with some
influence from the North Atlantic Current and little sea-ice. The Fitzroy Member is thought to be in part
laterally equivalent to the Largo Bay Member and the Fladen Member of the Witch Ground Formation (text-
fig. 2), and between 10 000 and 13 500 years b.p. in age (Stoker et al. 19856).

The St Andrew’s Bay Member has its type section in Borehole 71/33 between the sea-bed and 23 0 m. The
dinoflagellate cyst recovery from the St Andrew’s Bay Member was extremely poor. Protoperidinium cysts, B.
tepikiense and O. centrocarpum were in evidence but no consistent picture emerged. Deposition in a cold,
unfavourable environment is suggested possibly in relation to sea-ice, and considerable reworking was noted
throughout. Further micropalaeontologial data, consistent with a cool environment, are given in Gregory et
al. (1978). The dinoflagellate cyst and micropalaeonlological evidence would, therefore, tend to disprove an
early Flandrian (7000 to 10 000 years b.p.) age as suggested by Stoker et al. (19856) and would appear to be
more consistent with a late Devensian age.

The type section for the Whitethorn Member occurs in BGS Borehole 81/39 between the sea-bed and
1 1 0 m. Palaeomagnetic work (Stoker et al. 19856) indicates normal polarity. Unfortunately only a single sample
has been analysed for its dinoflagellate cysts and this yielded an assemblage dominated by O. centrocarpum ( c .
85 %). This is comparable to modern assemblages from the area (Reid 1975; Harland 1983) and is indicative
of similar oceanographic conditions. At present there is insufficient data to compare with the analyses of the
Witch Ground Formation (Long et al. 1986) of similar age. Stoker et al. (19856) indicate a Holocene age for
this member.

Summary

Interpretation of the Quaternary dinoflagellate cyst record of the North Sea sequences allows a
subdivision into favourable and unfavourable units. It is apparent that apart from the thick and
extensive Aberdeen Ground Formation most of these units fall within defined seismostratigraphic
formations (Stoker et al. 19856). This reflects differences in the character of the units due to changes
in the environment of deposition and hence the engineering properties of the material.

Although the sequence is fully discussed in Stoker et al. ( 1985a, 6) it is worth stressing, that apart
from the complex of environments in the Aberdeen Ground Formation, a number of ameliorative
or interglacial episodes are noted. These interglacials occur in both the Ling Bank and Coal Pit
Formations.

Dating of the interglacial units is difficult as no definitive radiometric data are available. However,
the seismostratigraphic relationships, and the occurrence of the interpreted Blake and Laschamp
Excursions in sediments attributed to the Coal Pit Formation, indicate an Ipswichian age for the
Coal Pit Formation and therefore a Hoxnian age is inferred for the Ling Bank Formation. The

HARLAND: QUATERNARY DINOFLAGELLATE CYSTS

897

dinoflagellate cysts are not of any assistance in dating these units as similar ameliorative assemblages
are recorded from both.

However, of interest is the assemblage recovered from the ameliorative episode seen in Borehole
78/9 at lat.: 61° 30-65' N, long.: 0° 49-78' E. (Skinner and Gregory 1983), which because of its
association to the Blake palaeomagnetic event is attributed to the Ipswichian. The dinoflagellate
cyst assemblages are rich and dominated by O. centrocarpum below and B. tepikiense above.
Foraminiferal evidence suggests a strong amelioration and water depth exceeding 70 m. The dinoflag-
ellate cyst assemblage is most like that recorded from the ?Hoxnian Ling Bank Formation in proving
a transition from O. centrocarpum to B. tepikiense dominated floras.

COMPARISONS

Various comparisons can be made according to the stratigraphic level under discussion. Unfortu-
nately none are wholly satisfactory and allow only for a rather piecemeal approach.

Early Pleistocene

The Early Pleistocene of the North Sea is represented by the Aberdeen Ground Formation and is
recognized on the occurrence of reversed palaeomagnetic sediments and the presence of certain
species of dinoflagellate cysts and Foraminifera seemingly restricted to the Early Pleistocene around
the British Isles. Notable among the dinoflagellate cysts are Amiculosphaera umbracula , O. israeli-
anum, and T. pellitum with Impagidinium multiplexum and various undescribed Spiniferites and
Protoperidinium cysts.

Dinoflagellate cysts were first described from the Early Pleistocene of the British Isles by Wall
and Dale ( 1968c/) in their study of Ludhamian to Baventian strata from The Royal Society’s Borehole
at Ludham, Norfolk. Further work by Reid and Downie (1973) on the Bridlington Crag and mine
on the Pastonian Chillesford Beds (unpubl.) have indicated a last appearance of these assemblages
within the earliest part of the Middle Pleistocene. The cyst assemblages always indicate climatic
environments considerably warmer than those of today.

Recently, dinoflagellate work described in Cameron et al. (1984) documents assemblages of cysts
through a series of formations in the southern part of the North Sea. These assemblages substantiate
the general character of the Pliocene/Early Pleistocene as noted above. Fluctuations in the pro-
portions of some species appear to be controlled by changes in water mass characteristics, i.e.
influence of more oceanic water and changes in sea-level, rather than by changes in climate.

It has proved difficult to correlate the palaeomagnetics and dinoflagellate cyst analyses. Some
attempt was made in Cameron et al. ( 1984) but it is thought that the succession is badly affected by
breaks in sedimentation. Further attempts are in progress on the Early Pleistocene succession from
the Ormesby Borehole (text-fig. 1).

In the eastern Atlantic the Pliocene to Early Pleistocene dinoflagellate cyst record at Rockall
(Harland 19846) is poor whereas in the Goban Spur (Harland 1984//) the approximate Early to
Middle Pleistocene boundary can be recognized, although a Plio/Pleistocene boundary cannot. The
change in cyst assemblages across the Early/Middle Pleistocene boundary appears to coincide with
the nannofossil NN 19/20 boundary.

The Early Pleistocene is characterized by particular groups of dinoflagellate cysts and although
an abrupt change is noted between the Early Pleistocene and somewhere in the Middle Pleistocene
no such change occurs between the Pliocene and Early Pleistocene. No doubt some of the reasons
behind the latter are the result of few studies but nevertheless both Pliocene and Lower Pleistocene
sediments contain similar cyst floras indicative of equable and stable climatic conditions. There is
an obvious need to study more closely Pliocene and Early Pleistocene cyst floras not only to
document the many new species that are undoubtedly present, but also to understand better the
nature of the environmental changes.

PALAEONTOLOGY, VOLUME 31

Middle-Late Pleistocene

The Middle to Late Pleistocene record in the North Sea occurs within a number of seismostrati-
graphic units. It is characterized by normally magnetized sediments and severely fluctuating climatic
conditions. The dinoflagellate cyst record reveals, for the most part, severe cold, arctic-like environ-
ments with evidence of three climatic ameliorations. Cyst floras are generally of low diversity, low
cyst recruitment, and dominated by north-temperate to arctic forms. Ameliorations have been noted
toward the top of the Aberdeen Ground Formation, the Ling Bank and Coal Pit Formations and
can be recognized by the rise in cyst recruitment and the greater diversity of the cyst floras. Species
that signify North Atlantic water or an oceanic influence together with north-temperate species
become common.

Studies from the eastern Atlantic have recognized obvious interglacials both in the Rockall area
(Harland 1984/?) and the Goban Spur (Harland 1984u). However, first order correlations between
the dinoflagellate cyst floras and oxygen isotope records have not been achieved but it is thought
possible Ipswichian and Hoxnian sequences are present.

Unfortunately the Middle to Late Pleistocene of the southern part of the North Sea has not
yielded good sequences of dinoflagellate cysts and indeed Cameron et al. (1987) indicate major
hiatuses in southern North Sea sequences. Comparisons, therefore, cannot be made to the more
complete northern North Sea successions.

Latest Pleistocene-Holocene

Dinoflagellate cyst studies reveal that the changes from full glacial conditions through to the modern
oceanographic situation are clearly recorded even to the recognition of the Younger Dryas cooling
(Long et al. 1986). The chronostratigraphy is assisted by the recognition of the Vedde Ash equivalent
dated at 10 600 years b.p.

Particularly interesting is the clear signal, expressed by a change in the dinoflagellate cyst assem-
blages from those dominated by B. tepikiense to those dominated by O. centrocarpum , occurring at
the Late Devensian-Allerod/Bolling boundary. This signal may indicate the passage of the Polar
Front across the north-eastern part of the Atlantic and the retreat of ice-dominated waters from
the Atlantic Ocean and North Sea.

This singular event possibly dated from between 13 000 and 11 000 years b.p. is seen in the
dinoflagellate cyst signal from the North Sea (Long et al. 1986, herein), the Goban Spur (Harland
1 984c/; DSDP Holes 548 and 549A), the Rockall Plateau (Harland 19846; DSDP Hole 552A),
Norwegian Sea (Harland 1984c; Verna Core 23-76) and has been recognized also by Turon (1981)
from the Rockall Channel and by Harland (1987) and Stoker et al. (1987) in the Northern Rockall
Trough and Faeroe-Shetland Channel.

Unfortunately there is insufficient evidence from these studies to comment on the exact timing of
this event, which is known otherwise to be linked to the deglacial history of the North Atlantic
(Duplessy et cd. 1981 ; Ruddimann and McIntyre 1981 ) but it is worth stressing that the dinoflagellate
cyst signal is clear, characteristic, and can be traced over a large area of the North Atlantic and
North Sea. Work from the north-west continental shelf margin of the British Isles (Harland 1987;
Stoker et al. 1987) and in the Minch (unpubl.) on similar sequences of thick late Devensian and
Flandrian sediments indicate the possibility of further precision.

CONCLUSIONS

An attempt has been made to synthesize the contribution dinoflagellate cyst research has made to
the understanding of offshore Quaternary stratigraphy over the past decade or so. The use of
dinoflagellate cyst biostratigraphy in offshore marine Quaternary sequences has assisted in the
elucidation of events occurring within the Quaternary on the continental shelf. Dinoflagellate cyst
studies are capable of recognizing glacial/interglacial cycles as well as interstadial events and have
brought new information to the understanding of Early Pleistocene palaeoenvironments. High
resolution stratigraphy is also possible within the latest Pleistocene and Holocene and events

HARLAND: QUATERNARY DINOFLAGELLATE CYSTS

899

text-fig. I I The North Sea Quaternary succession in
context of the established chronostratigraphy, palaeo-
magnetics, and oxygen isotope stratigraphy (based
largely upon Jenkins et al. (1985)).

-HOLOCENE

UJ CJ

I- o

< I-

-> CO

LU

z

LU

o

o

I —

CO

Q

o

o

o

I —

CO

Jarami

Event

X

CJ

O

0.

Ilo

_l

DC

<

CO

cr

>

UJ

cr

Olduvai'

Event

Reunion

Event

20

25

50

70

120

130

190

247

276

336

352

453

480

500

551

619

649

662

712

10
1 1
12

13

14
1 5
16
1 7
18
19

730 20
21
22
23

900.

970

1.67

201
2 04

IPSWICHIAN

WOLSTONIAN

HOXNIAN

ANGLIAN

CROMERIAN

BEESTONIAN

PASTONIAN

PRE-PASTON

BRAMMERT.

BAVENTIAN

LUDHAMIAN

Witch

Ground

Formation

HSwatcEway-

Formation

Coal Pit
Formation

Fisher Fm.

Ling Bank
Formation

Aberdeen

Ground

Formation

900

PALAEONTOLOGY, VOLUME 31

occurring over two or three thousand years can be recognized. Unlike other organisms, dinoflagellate
cysts are unique in their correlation potential from deep-ocean sediments, continental slope, shelf,
and nearshore marine sediments. Dinoflagellate cyst spectra will undoubtedly prove as useful
offshore as pollen diagrams have proved onshore.

Finally there is potential in gaining palaeoceanographic information from dinoflagellate cyst
analysis. Research points to the recognition of water masses, the documenting of surface and deep-
water currents, and the recognition of the polar front and ice margins.

Text-fig. 1 1 attempts to place the North Sea sequence in terms of the currently recognized
chronostratigraphy, and oxygen isotope stages. I look forward to even greater precision and to a
better understanding of marine sequences from the use of dinoflagellate cyst analysis.

Acknowledgements. None of this work would have been possible without the multidisciplinary approach of
the Department of Energy funded Marine Earth Sciences Research Programme and I thank all staff, past and
present, particularly Dan Evans, Martyn Stoker, Dave Long, and Chris Evans for their help and encourage-
ment. Thanks are also due to Mrs Jane Sharp and Ms Jane Kyffin-Hughes for their excellent technical
assistance, to Miss Rosanna O’Gunleye and Mrs Janet Lines for their accurate and patient typing and to
colleagues in the Biostratigraphy Research Group, particularly Ms Diane Gregory. I thank Drs B. Owens,
M. S. Stoker, Mr D. A. Aldus and D. Long for their constructive criticism of an early draft of this paper.
This paper is published with permission from the Director, British Geological Survey (NERC).

REFERENCES

aksu, a. e. and mudie, p. J. 1985. Late Quaternary stratigraphy and paleoecology of northwest Labrador Sea.
Mar. Micropaleont. 9, 537-557.

arrhenius, G. 1952. Sediment cores from the East Pacific. Rep. Swed. Deep Sea Exped. (1947-1948), 5, 1-227.
barken, k. and dale, b. 1986. Dinoflagellate cysts in Upper Quaternary sediments from southwestern Norway
and potential correlations with the oceanic record. Boreas, 15, 185-190.
balch, w. M., reid, p. c. and surrey-gent, s. c. 1983. Spatial and temporal variability of dinoflagellate cyst
abundance in a tidal estuary. Can. J. Fish. Acjuat. Sci. 40 (Suppl. 1), 244-261.
beck, R. B., funnell, b. m. and lord, a. r. 1972. Correlation of Lower Pleistocene Crag at depth in Suffolk.
Geol. Mag. 109, 137-139.

berggren, w. a., Kent, D. v. and van couvering, j. a. 1985. The Neogene: Part 2 Neogene geochronology
and chronostratigraphy. In snelling, n. j. (ed. ). The Chronology of the Geological Record. Mem. Geol. Soc.
Loud. 10, 21 1 260.

binns, p. e., harland, r. and hughes, m. j. 1974. Glacial and post-glacial sedimentation in the Sea of the
Hebrides. Nature, Lond. 248, 751-754.

— mcQUILlin, r. and kenolty, n. 1974. The geology of the Sea of the Hebrides. Rep. Inst. geol. Sci. 73/14,
1 -43.

Bradford, m. r. and wall, d. a. 1984. The distribution of Recent organic-walled dinoflagellate cysts in the
Persian Gulf, Gulf of Oman, and northwestern Arabian Sea. Palaeontographica, B192, 16-84.

British geological survey. 1984. BGS boreholes 1983. Rep. Br. geol. Surv. 16 (11), I 15.
bujak, j. p. 1984. Cenozoic dinoflagellate cysts and acritarchs from the Bering Sea and northern Pacific, DSDP
Leg 19. Micropaleontology , 30, 180-212.

cameron, t. d. j., bonny, a. p., Gregory, d. m. and harland, r. 1984. Lower Pleistocene dinoflagellate cyst,
foraminiferal and pollen assemblages in four boreholes in the southern North Sea. Geol. Mag. 121, 85-97.

stoker, m. s. and long, d. 1987. The history of Quaternary sedimentation in the UK sector of the North
Sea Basin. J. geol. Soc. Lond. 144, 43-58.

caston, v. n. d. 1977. Quaternary deposits of the central North Sea. 1 . A new isopachyte map of the Quaternary
of the North Sea. Rep. Inst. geol. Sci. 77/11, 1-8.

cline, r. m. and hays, j. d. 1976. Investigation Late Quaternary paleoceanography and paleoclimatology.
Mem. geol. Soc. Am. 145, 1-464.

coope, g. r. 1977. Quaternary Coleoptera as aids in the interpretation of environmental history. In shotton,
f. w. (ed.). British Quaternary Studies, 55-68. Clarendon Press, Oxford.

CURRY, D., ADAMS, C. G., BOULTER, M. C., DILLEY, F. C., EAMES, F. E., FUNNELL, B. M. and WELLS, M. K. 1978. A
correlation of Tertiary rocks in the British Isles. Geol. Soc. Lond., Spec. Rept. 12, 1-72.

HARLAND: QUATERNARY DINOFL AGELLATE CYSTS

901

dale, b. 1976. Cyst formation, sedimentation, and preservation: factors affecting dinoflagellate assemblages
in Recent sediments from Trondheimsfjord, Norway, Rev. Palaeobot. Palynol. 22, 39 60.

1983. Dinoflagellate resting cysts: ‘benthic plankton’. In fryxell, g. a. (ed.,). Survival strategies of the
algae , 69 136. Cambridge University Press, Cambridge.

— 1985. Dinoflagellate cyst analysis of Upper Quaternary sediments in core GIK 1 5530-4 from the Skager-
rak. Norsk geol. Tidsskr. 65, 97 102.

davies, h. c., dobson, m. r. and Whittington, r. j. 1984. A revised seismic stratigraphy for Quaternary
deposits on the inner continental shelf west of Scotland between 55 30' N and 57 30' N. Boreas , 13, 49
66.

duplessy, j. c., delibrias, g., turon, j. l. and duprat, j. 1981 . Deglacial warming of the northeastern Atlantic
Ocean: correlation with the paleoclinratic evolution of the European continent. Palaeogeogr. Palaeoclimatol.
Palaeoecol. 35, 121 144.

ericson, d. b., broecker, w. s., kulp, j. L. and wollin, G. 1956. Late-Pleistocene climates and deep-sea
sediments. Sci. Bull. New York , 124, 385 389.

feyling-hanssen, r. w. 1981 . Foraminiferal indication of Eemian inter-glacial in the northern North Sea. Bull,
geol. Soc. Denmark , 29, 175 189.

1982. Foraminiferal zonalion of a boring in Quaternary deposits of the northern North Sea. Ibid. 31,
29-47.

Gregory, d. m. and harland, r. 1978. The late Quaternary climatostratigraphy of IGS Borehole SLN 75/33
and its application to the palaeooceanography of the north-central North Sea. Scott. J. Geol. 14, 147 155.

and wilkinson, i p. 1978. Palaeontology of a series of boreholes through the drift of the Firth of
Forth and Forth Approaches. In Thomson, m. e. IGS studies of the geology of the Firth of Forth and its
approaches. Rep. Inst. geol. Sci. 77/17 , 41 -48.

harland, r. 1973. Microplankton from boreholes in the lower reaches of the Firth of Clyde. In deegan,
c. E., kirby, r., rae, i. and floyd, r. The superficial deposits of the Firth of Clyde and its sea lochs. Ibid. 73/9.
36-39.

1 974. Quaternary organic-walled microplankton from Boreholes 7 1 /9 and 71/10. In binns, p. e., McQUILLIN,
r. and kenolty, n. The Geology of the Sea of the Hebrides , Ibid. 73/14, 37 39.

— 1977. Recent and late Quaternary (Flandrian and Devensian) dinoflagellate cysts from marine continental
shelf sediments around the British Isles. Palaeontographica , B164, 87 126.

— 1979. Dinoflagellate biostratigraphy of Neogene and Quaternary sediments at Holes 400/400A in the Bay
of Biscay (Deep Sea Drilling Project Feg 48). In montadert, l. et al. (eds.). Init. Repts D.S.D.P. 48, 531
545.

— 1982. A review of Recent and Quaternary organic-walled dinoflagellate cysts of the genus Protoperidinium.
Palaeontology , 25, 369-397.

— 1983. Distribution maps of Recent dinoflagellate cysts in bottom sediments from the North Atlantic
Ocean and adjacent seas. Ibid. 26, 321-387.

— 1984a. Quaternary dinoflagellate cysts from Holes 548 and 549A, Goban Spur (Deep Sea Drilling Project
Feg 80). In graciansky, p. c. de. et al. (eds.). Init. Repts D.S.D.P. 80, 761 766.

19846. Quaternary dinoflagellate cysts from Hole 552A, Rockall Plateau, Deep Sea Drilling Project Feg
81. In Roberts, d. G. et. al. (eds.). Ibid. 81, 541 546.

— 1984c. Recent and late Quaternary dinoflagellate cysts from the area of the Greenland Iceland Faeroe—
Scotland Ridge. J. micropalaeontol. 3, 95-108.

— 1987. Dinoflagellate cysts and the establishment of a Fate Quaternary climatostatigraphy, northwest UK
continental margin. INQUA . 1 2th Int. Cong.. Abs.. Res. Counc. Canada. Ottawa. 182.

-and sharp, j. 1986. Elongate Spiniferites cysts from North Atlantic bottom sediments. Palynology. 10,
25-34.

— Gregory, d. m., hughes, m. j. and wilkinson, i. p. 1978. A late Quaternary bio- and climatostratigraphy
for marine sediments in the north-central part of the North Sea. Boreas. 7, 91 96.

— reid, p. c., dobell, p. and norris, G. 1980. Recent and sub-Recent dinoflagellate cysts from the Beaufort
Sea, Canadian Arctic. Grana , 19, 21 I 225.

holmes, r. 1977. Quaternary deposits of the central North Sea, 5. The Quaternary geology of the UK sector
of the North Sea between 56° and 58 N. Rep. Inst. geol. Sci. 11/14, I 50.
hugfies, m. j., Gregory, d. m., harland, r. and wilkinson, i. p. 1977. Fate Quaternary Foraminifera and
dinoflagellate cysts from boreholes in the UK sector of the North Sea between 56 N and 58 N. In holmes, r.
Quaternary deposits of the central North Sea, 5. The Quaternary geology of the UK sector of the North
Sea between 56° and 58° N. Ibid. 36 46.

902

PALAEONTOLOGY, VOLUME 31

imbrie, j. 1985. A theoretical framework for the Pleistocene ice ages. J. geol. Soc. Lond. 142, 417 432.
institute OF geological sciences. 1974. I.G.S. Boreholes 1973. Rep. Inst. geol. Sci. 74/7, 1-23.
jansen, j. h. f. 1976. Late Pleistocene and Holocene history of the northern North Sea, based upon acoustic
reflection records. Neth. J. Sea Res. 10, 1-43.

1980. Holocene deposits in the northern North Sea: evidence of dynamic control of their mineral and
chemical composition? — a comment. Geologie Mijnb. 59, 179-180.

— doppert, j. w. c., hoogendoorn-toering, k., de jong, J. and spaink, G. 1979. Late Pleistocene and
Holocene deposits in the Witch and Fladen Ground area, northern North Sea. Neth. J. Sea Res. 13,
1-39.

jenkins, d. g., bowen, d. q., adams, c. G., shackleton, n. j. and brassell, s. c. 1985. The Neogene: Part 1.
In snelling, N. j. (ed.). The Chronology of the Geological Record. Mem. geol. Soc. Lond. 10, 199-210.
knudsen, k. l. 1985. Forminiferal stratigraphy of Quaternary deposits in the Roar, Skjold and Dan Fields,
central North Sea. Boreas 14, 31 1 324.

lewis, j., dodge, J. D. and tett, p. 1984. Cyst-theca relationships in some Protoperidinium species (Peridiniales)
from Scottish sea lochs. J. micropalaeontol. 3, 25-34.

long, d., bent, a., harland, r., Gregory, d. m., graham, d. k. and morton, a. c. 1986. Late Quaternary
palaeontology, sedimentology and geochemistry of a vibrocore from the Witch Ground Basin, central
North Sea. Mar. Geol. 73, 109-123.

mangerud, j., lie, s. e., furnes, it, Kristiansen, i. l. and lomo, l. 1984. A Younger Dryas ash bed in western
Norway, and its possible correlations with tephra in cores from the Norwegian Sea and the North Atlantic.
Quatern. Res. 21, 85-104.

matsuoka, k. 1984. Cyst and theca of Protoperidinium avellana (Meunier) Balech (Dinophyceae). Bull. Fac.
Lib. Arts , Nagasaki Univ. Nat. Sci. 25 (1), 37-47.

— 1985a. Cyst and thecate forms of Pyrophacus steinii (Schiller) Wall et Dale, 1971. Trans Proc. palaeont.
Soc. Japan , ns, 140, 240-262.

— 19856. Organic-walled dinoflagellate cysts from surface sediments of Nagasaki Bay and Senzaki Bay,
West Japan. Bull. Fac. Lib. Arts , Nagasaki Univ. Nat. Sci. 25 (2), 21-115.

— kobayashi, s. and iizuka, s. 1982. Cysts of Protoperidinium divarication (Meunier) Parke et Dodge,
1976 from surface sediments of Omura Bay, West Japan. Rev. Palaeobot. Palynol. 38, 109 118.

moore, p. D. and webb, j. a. 1978. An illustrated guide to pollen analysis. 133 pp. Hodder and Stoughton,
London.

mudie, p. j. and aksu, a. e. 1984. Palaeoclimate of Baffin Bay from 300,000-year record of foraminifera,
dinoflagellates and pollen. Nature, Lond. 312, 630-634.

— and short, s. k. 1985. Marine palynology of Baffin Bay. In Andrews, j. t. (ed.). Quaternary Studies of
Baffin Island, West Greenland and Baffin Bay, 263-308. Allen (George) & Unwin, London.

neves, R. and dale, b. 1963. A modified filtration system for palynological preparation. Nature, Lond. 187,
775-776.

pantin, h. m. 1978. Quaternary sediments from the north-east Irish Sea: Isle of Man to Cumbria. Bull. Geol.
Surv. Gt Br. 64, 1-43.

peacock, j. d. 1981. Scottish Late-Glacial marine deposits and their environmental significance. In neale, j.

and flenley, J. (eds.). The Quaternary in Britain, 222-236. Pergamon Press, Oxford.
reid, p. c. 1975. A regional sub-division of dinoflagellate cysts around the British Isles. New Phytol. 75, 589-
603.

— and downie, c. 1973. The age of the Bridlington Crag. Proc. Yorks, geol. Soc. 39, 315-318.

— and harland, r. 1977. Studies of Quaternary dinoflagellate cysts from the North Atlantic. In elsik,
w. c. (ed.). Contributions of Stratigraphic Palynology, Vol. 1, Cenozoic Palynology, Amer. Assoc. Strat.
Palynol., Cont., Series, 5A, 147 169. American Association of Stratigraphic Palynologists Foundation,
Houston.

ruddiman, w. F. and McIntyre, A. 1981. The North Atlantic Ocean during the last deglaciation. Palaeogeogr.
Palaeoclimatol. Palaeoecol. 35, 145-214.

scott, d. b., mudie, p. J., vilks, G. and younger, d. c. 1984. Latest Pleistocene Holocene paleoceanographic
trends on the continental margin of eastern Canada: foraminiferal, dinoflagellate and pollen evidence. Mar.
Micropaleontol. 9, 181 -218.

shackleton, n. j 1969. The last interglacial in the marine and terrestrial records. Proc. R. Soc 174B, 135-154.
shotton, f. w. 1973. General principles governing the subdivision of the Quaternary System. In mitchell,
G. F. et al. (eds.). A correlation of Quaternary deposits in the British Isles. Geol. Soc. Lond., Spec. Rept. 4,
1-7.

HARLAND: QUATERNARY DINOFLAGELL ATE CYSTS 903

skinner, A. c. and GREGORY, D. M. 1983. Quaternary stratigraphy in the northern North Sea. Boreas , 12,
145 152.

stabell, b. and thiede, ). (eds.) 1985. Upper Quaternary marine Skagerrak (NE North Sea) deposits:
stratigraphy and depositional environment. Norsk geol. Tidsskr. 65, i 149.
stoker, m. s., harland, r. and morton, a. c. 1987. Late Quaternary stratigraphy of the North Rockall
Trough and Faeroe- Shetland Channel, northwest UK continental margin. INQUA , 12th hit. Cong.. Ahs .,
Res. Counc. Canada , Ottawa , 270.

— long, d. and fyfe, j. a. 1985a. The Quaternary succession in the central North Sea. News/. Stratigr.
14, 119-128.

— 19856. A revised Quaternary stratigraphy for the central North Sea. Rep. Br. geol. Surv. 17

(2), I -35.

— skinner, a. c., fyfe, j. a. and long, d. 1983. Palaeomagnetic evidence for early Pleistocene in the central
and northern North Sea. Nature , Land. 304, 332-334.

Thomson, m. e. and eden, r. a 1977. Quaternary deposits of the central North Sea. 3. The Quaternary
sequence in the west-central North Sea. Rep. Inst. geol. Sci. 77/12, I -18.
turon, j.-l. 1980. Dinoflagelles et environnement climatique. Les kystes de dinoflagelles dans les sediments
Recents de l’Atlantique nord-oriental et leurs relations avec l’environnement oceanique. Application aux
depots Holocenes du Chenal de Rockall. Mem. Mus. Natn Hist. nat. Paris. B27, 269-282.

— 1981. Le palynoplancton dans l’environnement actual du l’Atlantique nord-oriental. Evolution climatique
et hydrologique dupuis le dernier maximum glaciare. (These Doct.) Univ. Bordeaux , 678, 1 313.

van der plas, l. and tobi, a. c. 1965. A chart forjudging the reliability of point counting results. Am. J.
Sci. 263, 87 90.

wall, D. 1970. Quaternary dinoflagellate micropalaeontology: 1959 to 1969. Proceedings of the North
American Paleontological Convention , G, 844 866.

— and dale, b. 1968a. Early Pleistocene dinoflagellates from The Royal Society Borehole at Ludham,
Norfolk. New Phytol. 67, 315-326.

— 1968 b. Modern dinoflagellate cysts and evolution of the Peridiniales. Micropaleontology , 14, 265

304.

— lohmann, g. p. and smith, w. k. 1977. The environmental and climatic distribution of dinoflagellate
cysts in modern marine sediments from regions in the North and South Atlantic Oceans and adjacent seas.
Mar. Micropaleontol. 2, 121-200.

west, r. G. 1961. Vegetational history of the Early Pleistocene of the Royal Society Borehole at Ludham,
Norfolk. Proc. R. Soc. 155B. 437-453.

— 1985. Climatic change in the Quaternary — evidence and ideas. J. geol. Soc. Lond. 142, 413-416.

ZIMMERMAN, H. B., SHACKLETON, N. J., BACKMAN, J., KENT, D. V., BALDAUF, J. G., KALTENBACK, A. J. and MORTON,
a. c. 1984. History of Plio-Pleistocene climate in the north-eastern Atlantic, Deep Sea Drilling Project
Hole 552A. In Roberts, d. g. et ai (eds.). I nit. Repts D.S.D.P. 81, 861 875.

Typescript received 1 September 1987
Revised typescript received 21 December 1987

REX HARLAND

Biostratigraphy Research Group
British Geological Survey
Keyworth

Nottingham NG12 5GG

'

J'

>

NOTES FOR AUTHORS

The journal Palaeontology is devoted to the publication of papers on all aspects of palaeontology. Review articles are
particularly welcome, and short papers can often be published rapidly. A high standard of illustration is a feature of the
journal. Four parts are published each year and are sent free to all members of the Association. Typescripts should conform in
style to those already published in this journal, and should be sent to Dr. Dianne Edwards, Department of Plant Science,
University College, P.O. Box 78, Cardiff CF1 1XL, who will supply detailed instructions for authors on request (these are
published in Palaeontology 1985, 28, pp. 793-800).

Special Papers in Palaeontology is a series of substantial separate works conforming to the style of Palaeontology.

SPECIAL PAPERS IN PALAEONTOLOGY

In addition to publishing Palaeontology the Association also publishes Special Papers in Palaeontology. Members may
subscribe to this by writing to the Membership Treasurer: the subscription rate for 1988 is £37-50 (U.S. $66) for Institutional
Members, and £ 1 8-50 (U.S. $32) for Ordinary and Student Members. A single copy of each Special Paper is available to
Ordinary and Student Members only, for their personal use, at a discount of 25% below the listed prices. Non-members
may obtain copies, but at the listed prices, from Marston Book Services, P.O. Box 87, Oxford 0X4 1LB, England.

RECENT PALAEONTOLOGICAL ASSOCIATION

PUBLICATIONS

Special Papers in Palaeontology

Numbers 1-20 are still in print and are available (post free) together with those listed below:

21. (for 1978): Devonian Ammonoids from the Appalachians and their bearing on International Zonation and Correlation,
by m. r. house. 70 pp., 12 text-figs ., 10 plates. Price £12 (U.S. $18).

22. (for 1978, published 1979): Curation of Palaeontological Collections. A joint Colloquium of the Palaeontological
Association and Geological Curators Group. Edited by m. g. bassett. 279 pp., 53 text-figs. Price £25 (U.S. $38).

23. (for 1979): The Devonian System. A Palaeontological Association International Symposium. Edited by m. r. house, c. t.
scrutton and m. g. bassett. 353 pp., 102 text-figs., 1 plate. Price £30 (U.S. $45).

24. (for 1980): Dinoflagellate Cysts and Acritarchs from the Eocene of Southern England, by i. p. bujak, c. downie, g. l.
Eaton and G. L. williams. 100 pp., 24 text-figs., 22 plates. Price £15 (U.S. $23).

25. (for 1980): Stereom Microstructure of the Echmoid Test, by a. b. smith. 81 pp., 20 text-figs., 23 plates. Price £15
(U.S. $23).

26. (for 1981): The Fine Structure of Graptolite Periderm, by P. R. crowther. 1 19 pp., 37 text-figs., 20 plates. Price £25
(U.S. $38).

27. (for 1981): Late Devonian Acritarchs from the Carnarvon Basin, Western Australia, by G. playford and R. S. dring.
78 pp., 10 text-figs., 19 plates. Price £15 (U.S. $23).

28. (for 1982): The Mammal Fauna of the Early Middle Pleistocene cavern infill site of Westbury-sub-Mendip, Somerset, by
m. j. bishop. 108 pp., 47 text-figs., 6 plates. Price £25 (U.S. $38).

29. (for 1982): Fossil Cichlid Fish of Africa, by j. a. h. van couvering. 103 pp., 35 text-figs., 10 plates. Price £30 (U.S. $45).

30. (for 1983): Trilobites and other early Arthropods. Edited by d. e. g. briggs and p. d. lane. 276 pp., 64 text-figs., 38 plates.
Price £40 (U.S. $60).

31. (for 1 984): Systematic palaeontology and stratigraphic distribution of ammonite faunas of the French Coniacian, by w. j.
Kennedy. 160 pp., 42 text-figs., 33 plates. Price £25 (U.S. $38).

32. (for 1984): Autecology of Silurian organisms. Edited by m. g. bassett and j. d. lawson. 295 pp., 75 text-figs., 13 plates.
Price £40 (U.S. $60).

33. (for 1985): Evolutionary Case Histories from the Fossil Record. Edited by j. c. w. cope and p. w. skelton. 202 pp., 80 text-
figs., 4 plates. Price £30 (U.S. $45_).

34. (for 1985): Review of the upper Silurian and lower Devonian articulate brachiopods of Podolia, by O. I. NIKIFOROVA,
t. l. modzalevskaya and m. G. bassett. 66 pp., 6 text-figs., 16 plates. Price £10 (U.S. $15).

35. (for 1986): Studies in palaeobotany and palynology in honour of N. F. Hughes. Edited by d. j. batten and
D. E. G. BRIGGS. 178 pp., 29 plates. Price £30 (U.S. $50).

36. (for 1986): Campanian and Maastrichtian ammonites from northern Aquitaine, France, by w. j. Kennedy. 145 pp., 43
text-figs., 23 plates. Price £20 (U.S. $35).

37. (for 1987): Biology and revised systematics of some late Mesozoic stromatoporoids, by rachel wood. 89 pp., 31 text-
figs., 7 plates. Price £20 (U.S. $35).

38. (for 1987): Taxonomy, evolution, and biostratigraphy of late Triassic-early Jurassic calcareous nannofossils, by
p. r. bown. 118 pp., 19 text-figs., 15 plates. Price £30 (U.S. $50).

Field Guides to Fossils

1. (1983): Fossil Plants of the London Clay, by m. e. collinson. 121 pp., 242 text-figs. Price £7-95 (U.S. $12).

2. ( 1987): Fossils of the Chalk, compiled by E. owen; edited by a. b. smith. 306 pp., 59 plates. Price £1 1-50.

Other Publications

1982. Atlas of the Burgess Shale. Edited by s. conway morris. 31 pp., 24 plates. Price £20 (U.S. $30).

1985. Atlas of Invertebrate Macrofossils. Edited by j. w. Murray. Published by Longman in collaboration with the
Palaeontological Association, xiii + 241 pp. Price £13-95. Available in the USA from Halsted Press at U.S. $24.95.

© The Palaeontological Association, 1988

Palaeontology

VOLUME 31 • PART 3

CONTENTS

Rare tetrapod remains from the late Triassic fissure infillings of Cromhall
Quarry, Avon

N. C. FRASER 567

Hypostomes and ventral cephalic sutures in Cambrian trilobites

H. B. WHITTINGTON 577

The enigmatic arthropod Duslia from the Ordovician of Czechoslovakia

i. chi.upaC 611

Upper Ordovician trilobites from the Zap Valley, south-east Turkey

w. T. dean and ZHOU zhiyi 621

A Silurian cephalopod genus with a reinforced frilled shell

S. STRIDSBERG 651

Palaeocorynid-type appendages in Upper Palaeozoic fenestellid Bryozoa

A. J. BANCROFT 665

Tremadoc trilobites from the Skiddaw Group in the English Lake District

A. W. A. RUSHTON 677

New material of the early tetrapod Acanthostega from the Upper Devonian
of East Greenland

j. a. clack 699

An extinct ‘swan-goose’ from the Pleistocene of Malta

E. M. NORTHCOTE 725

A new alga from the Carboniferous Frosterley Marble of northern England
G. F. ELLIOTT 741

The mosasaur Goronyosaurus from the Upper Cretaceous of Sokoto State,
Nigeria

T. SO LIAR 747

A new aeshnid dragonfly from the Lower Cretaceous of south-east England
E. A. JARZEMBOWSKI 763

Acanthodian fish remains from the Upper Silurian or Lower Devonian of
the Amazon Basin, Brazil

P. JANVIER and J. H. G. MELO 771

A middle Cambrian chelicerate from Mount Stephen, British Columbia

D. E. G. BRIGGS and D. COLLINS 779

Patterns of diversification and extinction in early Palaeozoic echinoderms

A. B. SMITH 799

The stratigraphic distribution and taxonomy of the trilobite Onnia in the
type Onnian Stage of the uppermost Caradoc

a. w. owen and}. K. ingham 829

A new capitosaurid amphibian from the early Triassic of Queensland and
the ontogeny of the capitosaur skull

A. A. WARREN and M. N. HUTCHINSON 857

Quaternary dinoflagellate cyst biostratigraphy of the North Sea

R. HARLAND 877

Printed in Great Britain at the University Printing House, Oxford

by David Stanford, Printer to the University ISSN 0031-0239

THE PALAEONTOLOGICAL ASSOCIATION

The Association was founded in 1957 to promote research in palaeontology and its allied sciences.

COUNCIL 1988-1989

President: Dr J. D. Hudson, Department of Geology, University of Leicester, Leicester LEI 7RH
Vice-Presidents: Dr L. B. Halstead, Department of Geology, University of Reading, Reading RG6 2AB
Dr P. W. Skelton, Department of Earth Sciences, The Open University, Milton Keynes MK7 6AA
Treasurer: Dr M. E. Collinson, Department of Biology, King's College, London W8 7AH
Membership Treasurer: Dr H. A. Armstrong, Department of Geology, University of Newcastle,

Newcastle upon Tyne NE1 7RU

Institutional Membership Treasurer: Dr A. W. Owen, Department of Geology, University of Dundee, Dundee DD1 4HN
Secretary: Dr P. Wallace, The Croft Barn, Church Street, East Hendred, Oxon 0X12 8LA
Circular Reporter: Dr D. Palmer, Department of Geology, Trinity College, Dublin 2
Marketing Manager: Dr V. P. Wright, Department of Geology, University of Bristol, Bristol BS8 1RJ
Public Relations Officer: Dr M. J. Benton, Department of Geology, The Queen’s University of Belfast, Belfast BT5 6FB

Editors

Dr M. J. Benton, Department of Geology, The Queen's University of Belfast, Belfast BT5 6FB
Dr J. E. Dalingwater, Department of Environmental Biology, University of Manchester, Manchester M13 9PL
Dr D. Edwards, Department of Plant Sciences, University College, Cardiff CF1 1XL
Dr C. R. C. Paul, Department of Earth Sciences, University of Liverpool, Liverpool L69 3BX
Dr P. A. Selden, Department of Extra-Mural Studies, University of Manchester, Manchester M13 9PL
Dr P. D. Taylor, Department of Palaeontology, British Museum (Natural History), London SW7 5BD

Other Members

Dr J. A. Crame, Cambridge Dr C. Hill, London

Dr G. B. Curry, Glasgow Dr E. A. Jarzembowski, Brighton

Dr R. A. Spicer, London

Overseas Representatives

Australia: Professor B. D. Webby, Department of Geology, The University, Sydney, N.S.W., 2006
Canada: Dr B. S. Norford, Institute of Sedimentary and Petroleum Geology, 3303-33rd Street NW., Calgary, Alberta
Japan: Dr I. Hayami, University Museum, University of Tokyo, Hongo 7-3-1, Bunkyo-Ku, Tokyo
New Zealand: Dr G. R. Stevens, New Zealand Geological Survey, P.O. Box 30368, Lower Hutt
U.S. A.: Dr R. J. Cuffey, Department of Geology, Pennsylvania State University, Pennsylvania 16802
Professor A. J. Rowell, Department of Geology, University of Kansas, Lawrence, Kansas 66045
Professor N. M. Savage, Department of Geology, University of Oregon, Eugene, Oregon 97403
South America: Dr O. A. Reig, Departamento de Ecologia, Universidad Simon Bolivar, Caracas 108, Venezuela

MEMBERSHIP

Membership is open to individuals and institutions on payment of the appropriate annual subscription. Rates for 1988 are:

Institutional membership
Ordinary membership .
Student membership
Retired membership

£50-00 (U.S. S79)
£2100 (U.S. $38)
£11-50 (U.S. $20)
£10-50 (U.S. $19)

There is no admission fee. Correspondence concerned with Institutional Membership should be addressed to Dr A. W. Owen,
Department of Geology, The University, Dundee DD1 4HN. Student members are persons receiving full-time instruction at
educational institutions recognized by the Council. On first applying for membership, an application form should be obtained
from the Membership Treasurer: Dr H. A. Armstrong, Department of Geology, University of Newcastle, Newcastle upon Tyne
NE1 7RU. Subscriptions cover one calendar year and are due each January; they should be sent to the Membership Treasurer.
All members who join for 1988 will receive Palaeontology, Volume 31, Parts 1-4. Back numbers still in print may be ordered
from Basil Blackwell, Journals Department, 108 Cowley Road, Oxford OX4 1JF, England.

Cover: The brachiopod Meristina obtusa (J. de C. Sowerby, 1823), a life position assemblage from the Much Wenlock
Limestone Formation, Abberley Hills, Hereford (Specimen no. BB52671, x 1). Photograph by Harry Taylor of the British

Museum (Natural History) Photographic Studio.

UPPER CAMBRIAN AND
BASAL ORDOVICIAN TRILOBITES FROM
WESTERN NEW SOUTH WALES

by B. D. WEBBY, WANG QIZHENG and K. J. MILLS

Abstract. Eleven trilobite species are described from the Upper Cambrian basal Ordovician succession
exposed to the south-eastern side of Koonenberry Mountain in western New South Wales. Included among
the forms are six new species, Rhaptagnostus leitclii , Pareuloma aculeatum , Pseudoyuepingia whitei , P. lata ,
Proceratopyge ocella , and Hysterolenus furcatus. The assemblages come from two stratigraphically distinct
horizons near the top of the Watties Bore Formation. The lower has the more diverse fauna with typical
Upper Cambrian elements such as Rhaptagnostus , Pseudoyuepingia , Proceratopyge , Hedinaspis , and ProsaukicP.
The upper horizon contains Hysterolenus usually taken as an indicator of a restricted early Tremadoc age.
There are no apparent lithological or physical breaks in the intervening barren, conformable, 100 m thick
siltstone and shale succession, and it probably spans the Cambrian Ordovician boundary. Both assemblages
are preserved in silty shaly beds, and are characteristic elements of a deeper, basinal, or slope-type biofacies.
Genera such as Pseudoyuepingia , Hedinaspis , Pareuloma , and Hysterolenus are not known from shallow
platform successions elsewhere in Australia but occur in equivalent biofacies of Chinese sequences. The
common occurrences suggest close zoogeographic links. Similar though less strong connections are suggested
with other circum-Pacific regions, in particular with Alaska and New Zealand.

Cambrian trilobites have been described from a number of localities and horizons in western
New South Wales (text-fig. 1), in particular from the late Early-Middle Cambrian successions of
the Mount Wright area by Opik (1967a, 1970, 1975a, b , 1979, 1982), Shergold (1969), and Jell
(1975) and from the Upper Cambrian (Mindyallan-ldamean) deposits of the Kayrunnera-Cupala
Creek region by Opik (19756) and Jell (in Powell et al. 1982). Opik (19756) listed a Mindyallan
fauna from Kayrunnera as including Agnostoglossa , Palaeodotes , Blackwelderia, Ascionepea,
Aulacodigma , and Meteoraspis , and Jell (in Powell et al. 1982) described an Idamean assemblage
from the upper part of the Cupala Creek Formation as comprising Pseudagnostus afif. idalis Opik,
19676, Stigmatoa tvsoni Opik, 1963, Aphelaspisl afif. australis Henderson, 1976, Notoaphelaspis
orthocephalis Jell, 1982, and Prismenaspisl sp.

Still younger assemblages are described herein from the Watties Bore Formation to the south-
east of Koonenberry Mountain (text-figs. 1 and 2). They comprise a latest Cambrian association
(locality 1 at grid reference 277-205, text-fig. 1 ) including Micragnostus sp., Rhaptagnostus leitclii
sp. nov., Pareuloma aculeatum sp. nov., Hedinaspis sp., Prosaukidl sp., Pseudoyuepingia whitei sp.
nov., P. lata sp. nov., Proceratopyge ocella sp. nov., and P. sp., and an earliest Ordovician
occurrence (locality 2 at grid ref. 276-206, see text-fig. 1) of Hysterolenus furcata sp. nov.

STRATIGRAPHIC RELATIONSHIPS

E. C. Leitch found the first trilobites in weakly cleaved green-grey shales near Watties Bore, at the south-
eastern end of Koonenberry Mountain (grid ref. 286-185, text-fig. 1), in May 1983. Subsequently one of the
authors (K.J.M.) made further discoveries especially al two localities on the lower, eastern slopes of
Koonenberry Mountain (grid refs. 277-205 and 276-206, text-fig. I). The mapping of this Koonenberry
Wonnaminta region, completed in 1986, established the following stratigraphic relationships. First, the
trilobite-bearing sequence was shown to have an exposed, unconformable base in Morden Creek. Secondly,

| Palaeontology, Vol. 31, Part 4, 1988, pp. 905 938, pis. 83-86.|

© The Palaeontological Association

906

PALAEONTOLOGY, VOLUME 31

[~XT': I"-.': - ■)

••.•.'KOONENBERRY GAP

text-fig. 1. Geological map of the Koonenberry-Wonnaminta area and inset locality maps of far western
New South Wales, and Australia, to show location of the trilobite collecting localities 1-5 and lines of section
A A, B B', and C C\ Note that the km2 grid is based on the universal grid presented on the 1:100000
orthophotomap no. 7336 (Wonnaminta), First Edn., 1978.

WEBBY ET AL.\ CAM BRO ORDOVICIAN TRILOBITES

907

WATTSES

1000-

FORMATION

warn

A-A'

"WONOMINTA

Koonenberry

Fault

■Fossil Loc. 2 A_ , . ,

- — • — — — G_ 6/0 boundary

Fossil 1008.1,3,4^

HORIZONTAL SCALE
0 1 2 3km

i— » f— —H

Shales & siltstones

| Carbonate-bearing
I beds & lenses

Very line grained .
massive a bedded
sandstones

Interbedded slltstonesf
a tine grained
sandstones

Fine grained
bedded sandstones

Medium grained
basal quartzite
with cross beds

0.

D

O

£C

O

<

ce

UJ

z

z

3

a

>

<

Unconformity
Tight folded llthlc 'WONOMINTA
sandstones a slates BEDS*

♦ Fosslliferous horizon

’♦□J^HFossil Loc. 5

B-B'

C-C'

text-fig. 2. Stratigraphic columns of sections through the Kayrunnera Group (Upper Cambrian -basal
Ordovician) at Boshy Creek (A-A') and at Morden Creek (B-B') and near Watties Bore (C C'). Note the
relationship between the three formations of the Kayrunnera Group, and the stratigraphical positions of the
collecting localities 1 -5 within the Watties Bore Formation. Note the position of the Cambrian-Ordovician

(G/0) boundary.

the lower part of the sequence was found to be laterally equivalent to the Upper Cambrian (Mindyallan)
trilobite-bearing beds recorded by Opik (1975fi) from Kayrunnera, some 16 km to the south-east. Thirdly
the new trilobite finds described here were established as coming from the uppermost part of the sequence,
from the upper part of the Watties Bore Formation (localities I and 2, text-figs. 1 and 2).

In the Koonenberry-Wonnaminta area the shale-dominated, trilobite-bearing sequence occupies a large,
lens-shaped sliver to the east and south-east of Koonenberry Mountain (text-fig. 1). It is bounded by near
vertical faults including the main line of the Koonenberry Fault south of Watties Bore (text-fig. 1), with only
the base of the sequence seen to overlie unconformably lithic sandstones of the ‘basement’ succession in the
vicinity of Morden Creek (grid ref. 362-132). These latter deposits are possibly of early-?middle Cambrian

908

PALAEONTOLOGY, VOLUME 31

age. In continuity with this large, faulted sliver is another to the south-east, extending off the mapped area
shown in text-fig. 1 , towards Kayrunnera station. The lower, more sandy part of the trilobite-bearing sequence
is best exposed near Kayrunnera, where it was first discovered and mapped by the Geological Survey of New
South Wales in 1963 (Rose et al. 1964; Rose 1974). Again the sequence was observed to overlie unconformably
the older basement. The Mindyallan trilobites identified by Opik (19756) come from this basal part of the
sequence. Brunker et al. (1971) referred informally to the sequence as the ‘Kayrunnera Beds’. It is here
proposed to formalize this name as the Kayrunnera Group and to incorporate all the Upper Cambrian-
basal Ordovician fossiliferous sequences from Kayrunnera to the eastern side of Koonenberry Mountain in
this unit.

Kayrunnera Group

The greatest thickness of the Kayrunnera Group, over 2000 m of dominant shale and siltstones, is preserved
in the Koonenberry-Wonnaminta area. The steeply dipping and west-facing sequence is essentially homoclinal,
although some relatively open folds have been found near the top of the sequence around grid ref. 310-145
(text-fig. 1 ). Bedding laminations, bands, and units are well preserved in most outcrops. The degree of
metamorphism is slight (chlorite zone) but a near- vertical slaty cleavage trending about 120° is a prominent
feature of the silty and shaly lithologies, and pencil cleavage results from the intersection of this cleavage
with bedding laminations. Although white quartz veins are an ubiquitous feature of the underlying basement
units, they are very rarely observed within the Kayrunnera Group. Over most of the mapped area of the
Kayrunnera Group shown in text-fig. I the scattered exposures of siltstones and shales are deeply weathered
to orange, yellow, and cream colours. Some of the better and fresher exposures occur in the higher ground
around Watties Bore and adjacent to the southern end of Koonenberry Mountain, where the freshest rocks
are green-grey.

On the basis of the detailed mapping the Kayrunnera Group can be divided into three formations as
follows:

Morden Formation. A thin ( 1 6 m) medium-grained pure quartzite forms a remarkably persistent basal unit
of the Kayrunnera Group over the 16 km length of mapped unconformity between Morden Creek (text-fig.
1) and Kayrunnera homestead. A calcareous cement may be found in places in the upper part of the unit,
and small to medium cross-beds are not uncommon and reveal a south-easterly current source. The formation
was first recognized, and a type section measured, in the bed of Morden Creek (grid ref. 362-132, text-figs.
1 and 2) where a 2 m thick quartzite bed can be seen to overlie unconformably isoclinally folded lithic
sandstones. This locality is designated as the type section of the Morden Formation. The underlying
sandstones are thought to be of early-?middle Cambrian age on the basis of lithological correlation with the
Copper Mine Range Beds (Pogson and Scheibner 1971), which contain sponge spicules and trace fossils
(Leitch et al. 1987) in an area near Cupala Creek about 40 km to the south-east.

Bosliy Formation. This formation consists of interbedded fine-grained sandstones and siltstones with some
calcarenites and limestone lenses. Some horizons are richly fossiliferous and the Mindyallan trilobites
identified by Opik (19756) come from this unit. The formation is named after Boshy Creek, 4 km south-east
of the type section on a tributary of JK Creek between grid ref. 410-089 to 409-089 (text-figs. I and 2) where
94-3 m of section is preserved. Another measured section is in Morden Creek (text-fig. 2) but here the Boshy
Formation is only 15-7 m thick. This section is very weathered; the lower 5-2 m consists of fine-grained
bedded quartz sandstones and the remaining 10-5 m consists of fine-grained well-bedded sandstone interbanded
with siltstones.

Watties Bore Formation. Some 2000 m of shales and siltstones, with a distinctive yellow-buff weathering
characteristic, conformably overlie the Boshy Formation. Some levels are well bedded with interleaved shales
and siltstones while other levels are more massive. The formation is named after Watties Bore (grid ref. 286-
185, text-fig. 1) where good exposures occur. The type section is represented along the line of section C-C'
on text-fig. I (grid refs. 321-167 to 304-148). Of the five fossil localities shown in text-fig. 1, assemblages from
localities I and 2 (grid refs. 277-205 and 276-206) are the best preserved and the basis for the present
descriptions (see above cited list of species). Others comprise occurrences of Pseudoyuepingia sp. and
Proceratopyge ? sp. first found by E. C. Leitch near Watties Bore (locality 3, text-fig. 1), and Pseudoyuepingia
whitei sp. nov. and P. lata sp. nov. from a creek section at grid ref. 308-147 (locality 4, text-fig. 1).

Some very fine-grained sandstones and siltstones interbedded in the shale sequence display occasional
cross-bedding and slumping indicating a south to north palaeoslope. Calcareous concretions occur within
some siltstone beds and these record flattening associated with tectonic deformation. Ellipsoidal concretions
up to 250 mm in length have been noted around grid ref. 349-141 (text-fig. I). Well bedded and laminated

WEBBY ET AL.: CAMBRO ORDOVICIAN TRILOBITES

909

impure shaly limestones are also found in some exposures, with a few containing indeterminate agnostid and
polymerid trilobite and brachiopod casts, such as at locality 5 (grid ref. 347-144, text-fig. 1).

Several limestone breccia lenses, representing channel deposits, are found within the shaly siltstone sequence
around grid ref. 281-202 (text-fig. 1). The largest lens is 10 m long and up to 1 m thick. It contains a mixture
of rounded to subangular and irregular limestone clasts to 200 mm in diameter and sub-rounded to subangular
lithic sandstone boulders to 100 mm diameter. The limestone clasts are of several lithological types and some
contain fossil fragments. Some clasts are micritic and others calcarenitic, with up to 50 % rounded and
polished quartz grains. The matrix of the breccia is a silty shaly limestone with abundant angular hard
siltstone fragments. Some Hmestone clasts contain simple protoconodonts of Upper Cambrian type.

Correlatives of the Kayrunnera Group. The 1000 m thick Cupala Creek Formation cropping out some 40 km
to the south-east (Powell et al. 1982) appears to represent an onshore (upslope) equivalent of the Kayrunnera
Group succession. It was derived from the south, and comprises an upwardly fining transgressive sequence
commencing in fluvial conglomerate and sandstone depositional phases and passing up through marginal
marine sandy to shallow marine silty deposition towards the top. The Idamean faunas recorded by Jell (in
Powell et al. 1982) are restricted to the upper part of the sequence. It seems likely that the shallow-marine
Morden and Boshy Formations of Mindyallan age are equivalent to the fluvial to marginal marine lower-
middle parts of the Cupala Creek Formation, and the deeper marine lower part of the Watties Bore Formation
is correlative with the shallow-marine upper part of the Cupala Creek Formation of Idamean age. The deeper
marine upper part of the Watties Bore Formation of latest Cambrian to earliest Ordovician age is not
represented by equivalent, preserved deposits in the Cupala Creek area.

AGE AND ZOOGEOGRAPHIC SIGNIFICANCE

Of the two stratigraphically distinct trilobite assemblages documented from the upper part of the
Watties Bore Formation, the lower, with its diverse association of Micragnostus sp., R. leitchi sp.
nov'., Pareuloma aculeatum sp. nov., Hedinaspis sp., Prosaukia! sp., saukiid gen. et sp. indet.?,
Pseudoyuepingia whitei sp. nov., P. lata sp. nov., Proceratopyge ocella sp. nov., and P. sp., is
characteristically an Upper Cambrian fauna. Hedinaspis and Pseudoyuepingia are genera with
ranges limited to the Upper Cambrian, Proceratopyge has a Middle-Upper Cambrian range, and
Pareuloma an Upper Cambrian to lowest Ordovician (Tremadoc) range. The upper horizon is only
about 100 m stratigraphically above the lower horizon, and contains Hysterolenus furcatus sp. nov.
This genus Hysterolenus is usually regarded as an indicator of the lowest part of the Ordovician
(Shergold 1988; see also later discussion of the genus). Judging from the uniformity of the green-
grey silty shale lithology at the two horizons and through the intervening sequence there is no
evidence for associated breaks or unconformable relationships. Consequently the lower diverse
fauna is at least post-Idamean, probably middle-Late Cambrian in age.

This Cambrian-Ordovician boundary succession with its accompanying faunas is most closely
comparable to a number of sections described from south-east China, as well as to sections in
parts of north-west and north China. For example, in the Duibian area of Jiangshan, eastern
Zhejiang Province, a Cambrian-Ordovician boundary section is recorded by Lu et al. (1984) as
exhibiting species of Hedinaspis , Pseudoyuepingia, and Proceratopyge in beds 1-2 of the Siyangshan
Formation (Hedinaspis subzone of the Lotagnostus punctatus Zone) and then about 45 m
stratigraphically higher, occurrences of Hysterolenus in the basal Yinchupu Formation (Hysterolenus
Zone). In this particular section the intervening sequence includes rich trilobite and cephalopod
faunas attributed to the Lophosaukia subzone, the Acaroceras-Antacaroceras Zone, and the
Lotagnostus hedini Zone of the latest Cambrian. None of these faunal elements have yet been
found in the 100 m thick intervening succession in western New South Wales. Similarly, in the
Cambrian-Ordovician boundary section through the Guotang Formation in Sandu county in
Guizhou Province, Yin et al. (1984) have reported species of Hedinaspis and Pseudoyuepingia
as occurring in the topmost beds of the Cambrian, less than 2 m below the first record of Rhab-
dinopora flabelliformis (the subspecies regularis ) and about 6-7 m below the first occurrence of
Hysterolenus in the basal Ordovician beds. In the Hangula region of Nei Monggol, north China,

910

PALAEONTOLOGY, VOLUME 31

a Cambrian-Ordovician boundary has been drawn by Lu et al. (1981) between the Hedinaspis -
Diceratopyge and the Hysterolenus assemblages.

Lu et al. (1984) have stressed the pattern of incomings of typical Early Ordovician graptolite
assemblages as occurring above the Hysterolenus Zone in south-east China but, like its occurrences
in Scandinavia in association with dendroid graptolites of the Dictyonema Shale (Bergstrom 1982),
Hysterolenus has also been recorded as mainly occurring with R. fiabelliformis ( s.l .) in Taoyuan of
north-west Hunan and Sandu of south-east Guizhou Provinces in south-east China (Lu et al.
1984). The conodont index Cordylodus proavus is recognized as appearing just before Hysterolenus
in the Cambrian-Ordovician boundary section in the Duibian area of western Zhejiang Province
(Lu et al. 1984), that is, in beds of the latest Cambrian (L. hedini Zone). Similar assignments of
C. proavus as spanning the boundary have been demonstrated in other Chinese sections, in north-
west Hunan Province (Peng 1984) and in north-east China (Zhou et al. 1984; Wang 1984).

There is little similarity between these Cambro-Ordovician trilobite assemblages from the deeper,
shaly, basinal Watties Bore Formation of western New South Wales and age equivalents from the
shallow carbonate successions of continental platform areas of Australia. The post-Idamean Late
Cambrian-earliest Ordovician interval has been subdivided into numerous zones based on sections
in the platform carbonates of the Georgina Basin of western Queensland (Jones et al. 1971;
Shergold 1975, 1980) but they cannot be applied in correlation of the western New South Wales
basinal deposits. As a member of the Rhaptagnostus convergens species group, the occurrence of
R. leitchi sp. nov., may suggest a pre-Payntonian age for the lower assemblage, that is, between
the Zones of Neoagnostus denticulatus and R. papilio of Shergold (1975). Also the presence of a
few saukiids may suggest a level in the upper pre-Payntonian or Payntonian, but a closer correlation
is not presently achievable. The Hysterolenus horizon is best assumed to approximate to a level
within the Datsonian of Jones et al. (1971).

The Watties Bore faunas of western New South Wales comprise several dominantly ‘Chinese’
genera such as Pseudoyuepingia and Hedinaspis , but also other ceratopygacean elements of more
general Asian and European affinities like Proceratopvge and Hysterolenus. They clearly have
closest connections within the South-east China Faunal Province of Lu et al. (1974, 1984), which
includes much of south-east China (geographical provinces of Zhejiang, Anhui, Hunan, and
Guizhou) and extends to north and north-west China (Nei Monggol, Qinghai, and Xinjiang), even
to southern Kazakhstan (Shergold 1988). However, what is referred to as the South-east China
Faunal Province is perhaps better viewed as an off-shelf or open-ocean facing biofacies whose
distribution, which is dominantly of ceratopygaceans, has a much wider geographical extent, being
recorded as peripheral to the shallower (colder?) olenid biofacies of the Baltic Faunal Province,
and to the shallower (warmer?) biofacies of North China Faunal Province type on the North
China Platform and Australian Platform (Shergold 1988).

This off-shelf biofacies with, for instance, its records of Hedinaspis , extends to include parts of
South Korea (Kobayashi 1966), the west coast of North America, particularly Alaska and Nevada
(Taylor 1976), to Australia including western New South Wales (described herein) and Tasmania
(Jago, in Shergold et al. 1985; Jago 1987), and to New Zealand (Wright and Cooper 1983).
Pseudoyuepingia has a more restricted distribution in south-east and north-west China, South
Korea, Tasmania, and western New South Wales, but earliest Ordovician Hysterolenus ( =
Ruapyge ) has a similarly wide spread of occurrences, in New South Wales, New Zealand, north-
north-west and south-east China, Kazakhstan, the Soviet Altai, Baltoscandia and, possibly, an
earlier occurrence in North Wales. Pareuloma is also represented mainly in the Upper Cambrian
of China and New South Wales, though there are other occurrences in Alaska, and an earliest
Ordovician record of the genus from Newfoundland.

Only the agnostid genera, the saukiids, and Proceratopvge are also known from platform
successions of Australia. However, the species of Proceratopvge from western Queensland
(Henderson 1976; Shergold 1982) are from older (Idamean) horizons and are morphologically
markedly distinct. Shergold (1988) has noted that the Australian Platform, extending to northern
Victoria Land, Antarctica, should be included within the North China Faunal Province. It is

WEBBY ET AL.: CAMBRO ORDOVICIAN TRILOBITES

91 1

dominated by saukiid and tsinaniid trilobites like the assemblages found in the platform areas of
western Queensland, central Australia, and the Mount Wright area of western New South Wales
(Shergold 1 97 1 <3; Shergold et al. 1982). These are on-shelf siliciclastic and carbonate associations
of north China type. In contrast the saukiids are poorly represented in the off-shelf assemblages
of the Watties Bore Formation in the Koonenberry-Wonnaminta area of western New South
Wales (text-fig. 1).

This Watties Bore section is important in providing the first documentation of the ceratopygacean-
dominated off-shelf biofacies of the South-east China Faunal Province through the Cambrian-
Ordovician boundary interval of the Australian region. Elements of this biofacies have potential
for correlation in fold-belt regions of the circum- Pacific and in parts of central and south-eastern
Asia (particularly in China, Kazakhstan, and the Soviet Altai).

SYSTEMATIC PALAEONTOLOGY

Type specimens are housed in the palaeontology collection of the Department of Geology and Geophysics,
University of Sydney, and have the prefix SUP. Two of the authors (B. D. W. and W. Q.) are responsible for
the trilobite descriptions.

Order miomera Jaekel, 1909
Suborder agnostina Salter, 1 864
Family agnostidae MlCoy, 1849
Subfamily agnostinae M‘Coy, 1849
Genus micragnostus Howell, 1935

Type species. Agnostus calvus Lake, 1906.

Discussion. Fortey (1980) clarified relationships between the Upper Cambrian-Lower Ordovician
agnostid genera Micragnostus Howell, 1935 and Geragnostus Howell, 1935. He established that a
number of North American and Chinese species assigned previously to Geragnostus should be
referred to Micragnostus. The genus Micragnostus is regarded by Fortey (1980) as belonging in a
conservative plexus with Upper Cambrian Agnostus Brongniart, 1822 and Homagnostus Howell,
1935, with Micragnostus seen as likely to be in direct line of descent from Homagnostus (see also
Robison and Pantoja-Alor 1968).

Micragnostus sp.

Text-fig. 3a, b

Material. Two specimens (SUP 48900-48901) from the lower horizon (locality 1) in the upper part of the
Watties Bore Formation on the eastern side of Koonenberry Mountain, western New South Wales.

Description. Specimens partially damaged internal moulds attaining length of about 6 mm (sag.). Cephalon
gently convex, with its maximum width near mid-length. Glabella subcylindrical in outline with slight forward
taper, defined by deep, broad axial furrows; approximately 0-60 of total cephalic length (sag.); deep transverse
glabellar furrow divides small anterior lobe from larger posterior lobe; the latter with faint median glabellar
node situated just behind mid-point (sag.). Small, triangular basal lobes. Cheeks moderately convex, with no
trace of median furrow on preglabellar field; outlined by deep and very wide anterior and lateral border
furrow; posterior border furrow much narrower (exsag.).

Thorax of relatively narrow (sag.) segments; axis as broad (tr.) as glabellar base; axial ring broad, evenly
divided into convex median, and lateral lobes. Pleura relatively narrow (tr.), the second being longer (tr.)
than the first.

Pygidium moderately convex, subquadrate with broadly rounded posterior margin. Axis relatively broad
(tr.) and long, about 0-5 of total width and 0-75 of pygidial length (sag.). First axial ring divided into a pair
of lateral lobes, each with transversely ovoid outline, and a median area which is a forward, tongue-like
extension of larger triangular median lobe (which includes second axial ring); only broken base of median

912

PALAEONTOLOGY, VOLUME 31

text-fig. 3. A-B, Micragnostus sp., Watties Bore Formation, uppermost Cambrian, x 10. a, internal mould
of SUP 48901 showing cephalon detached from thorax and pygidium; b, internal mould of cephalon, thorax,
and incomplete pygidium, SUP 48900. oh, Rhaptcignostus leitchi sp. nov., Watties Bore Formation, uppermost
Cambrian, c, internal mould of incomplete cephalon of paratype, SUP 48909, x 5; d, internal mould of
pygidium of paratype, SUP 48908, x8; e, internal mould of complete dorsal exoskeleton of paratype, SUP
48904, x 5; f, internal mould of complete exoskeleton of holotype, SUP 48905, x 7; G, internal mould of
cephalon and incomplete thorax of paratype, SUP 48902, x 5; h, internal mould of thorax of paratype, SUP
48906, x 8; i, Rhaptagnostus leitchi sp. nov.?, Watties Bore Formation, uppermost Cambrian; internal mould
of complete dorsal exoskeleton of specimen, SUP 48910, x 5.

tubercle vaguely shown towards rear margin of median lobe. Relatively large terminal piece, about twice
sagittal length of anterior two axial segments. Pleural fields narrow, slightly wider anterolaterally.

Remarks. This species is closely similar to M. intermedins (Palmer 1968) from the Upper Cambrian
of Alaska and Tremadoc of Mexico (Robison and Pantoja-Alor 1968), differing only in details
such as lack of a median furrow on the preglabellar field and apparent lack of posterolateral
border spines on the pygidium. Another closely related species is assigned to Homagnostus , as H.

WEBBY ET A L.: CAMBRO ORDOVICIAN TRILOBITES

913

zhuangliensis Qian, 1985 from the latest Cambrian Tangcun Formation of Jingxian, southern
Anhui Province, China. It is described as lacking a preglabellar median furrow and exhibiting at
least in one pygidium (Qian 1985, pi. 1, fig. 1 1) a less expanded pygidial axis than typically found
in Homagnostus. The species should instead be assigned to the genus Micragnostus. It only differs
from M. sp. in showing a slightly more elongated (sag.) cephalon and having the faint median
tubercle nearer to the mid-point (sag.) on the posterior lobe of the glabella.

Of the Australian species of the genus, it most resembles M. cf. intermedins (Palmer 1968) from
the Upper Cambrian Chatsworth Limestone of Black Mountain, western Queensland (Shergold
1975), but has a broader cephalic border furrow and lacks the faint median preglabellar furrow.
It also differs from M. acrolebes (Shergold 19716) from the Upper Cambrian Gola Beds of western
Queensland in lacking traces of anterolateral glabellar lobes, and in having a pygidium with
relatively wider axis, narrower border, and lacking posterolateral border spines. M. Iioeki
(Kobayashi 1939) from the Digger Island Formation (Lower Tremadoc) of Victoria (Jell 1985)
has a relatively wider glabella, a trace of a medium preglabellar furrow adjacent to the glabella,
prominent median node not far behind transglabellar furrow, and a pygidium with a relatively
narrow axis, wider posterolateral border, and well-developed marginal spines.

Family diplagnostidae Whitehouse, 1936, emend. Opik, 19676
Subfamily pseudagnostinae Whitehouse, 1936
Genus rhaptagnostus Whitehouse, 1936

Type species. Agnostus cyclopygeformis Sun, 1924; designated by Whitehouse 1936.

Discussion. Shergold (1977, 1980) has discussed the status and subdivision of this pseudagnostinid
genus. He recognized two species groups defined by R. convergens (Palmer 1955) and R. clarki
(Kobayashi 1935). Both have widespread distribution in Upper Cambrian deposits of Australia,
Asia (especially China), and North America. In western Queensland representatives of the
convergens group are recorded from the pre-Payntonian N. denticulatus and R. papilio Assemblage
Zones of Shergold (1975) and members of the clarki group from the pre-Payntonian R. clarki
maximus to Payntonian N. quasibilobus-Tsinania nomas Assemblage Zones of Shergold (1975).

Rhaptagnostus leitchi sp. nov.

Text-fig. 3c-h

Material. Holotype (SUP 48905) and seven paratypes (SUP 48902-48904, 48906 48909) from the lower
horizon (locality 1) in the upper part of the Watties Bore Formation, eastern side of Koonenberry Mountain,
western New South Wales.

Etymology. After Dr E. C. Leitch who found the first trilobite sample in the area near Watties Bore in 1983.

Diagnosis. Member of the R. convergens species group (Shergold 1977, 1980) with a large and
attenuated, bell-shaped glabella, and long (sag.) preglabellar field.

Description. Dorsal exoskeleton of mature specimens of relatively large size, from 12 to 15 mm in length and
usually about 5 mm in width. All the material somewhat poorly preserved and compressed; a few specimens
also tend to be a little distorted. Cephalon widest (tr.) along a transverse line just behind axial glabellar node.
Glabella with bell-shaped outline, bounded by variably preserved, narrow axial furrows; about two-thirds
length (sag.) of cephalon and nearly half maximum width of cephalon. Ill-defined anterior lobe of glabella
occupies about one-third total glabellar length; a pair of oval-shaped anterolateral lobes separated from
anterior lobe by faint V-shaped furrow and bisected adaxially by prominent ridge-like, backwardly directed
axial glabellar node; weakly defined, large posterior lobe and a pair of moderate-sized triangular basal lobes
at rear. Median preglabellar furrow almost continuous sagittally to anterior border furrow; preglabellar field
relatively long (sag.), extending backwards and outwards into broad, flattened cheek regions; moderately
deep, continuous cephalic border furrow, defining raised, narrow cephalic border.

First thoracic segment slightly better developed and longer (sag.); axis varies in width (tr ), owing to mainly

914

PALAEONTOLOGY, VOLUME 31

zigzag course of axial furrow. Articulating furrow of first segment exhibits raised, median axial bar (text-fig.
3h). Pleura relatively narrow, with sharply rounded, backwardly turned tips; pleural furrow of second
segment, in contrast to first segment, is placed close to anterior margin.

Pygidial axis lobate, with first two axial segments occupying about 04 of pygidial length (sag.), and axis
about 0-3 of total width (tr.) at anterior margin; transverse furrow between first and second axial segments
poorly defined, but with strong, raised axial node extending backwards and expanding slightly from point
of origin near rear edge of first segment. Large third segment (or deuterolobe) not clearly outlined. Relatively
deeply grooved (deliquiate of Shergold 1975) border furrows; border widening backwards into pair of very
small posterolateral spines, then of more even width around posterior margin.

Remarks. Two additional specimens (SUP 48910 and 48911) of Rhaptagnostus from the same
locality and horizon show the closest relationships, but differ in exhibiting deeper and broader
axial and marginal furrows (text-fig. 3i). They may indeed represent the less effaced members of
the species, but for the present should only be included tentatively in the species.

Order ptychopariida Swinnerton, 1915
Superfamily ptychopariacea Matthew, 1887
Family ptychopariidae Matthew, 1887
Subfamily eulominae Kobayashi, 1955
Genus pareuloma Rasetti, 1954

Type species. Pareuloma brachymetopa Rasetti, 1954.

Discussion. The eulominid genera are known to extend as a group from Franconian to Arenig age
(Fortey 1983). Type species of the genus Pareuloma , P. brachymetopa , is recorded by Rasetti (1954)
as coming from Cap des Rosiers, Quebec and Broom Point, Newfoundland. At Broom Point the
type species and another, P. impunctata Rasetti, 1954, were apparently collected from the interval
associated with occurrences of Radiograptus and D . flabelliforme close to the base of the Tremadoc
(Fortey et al. 1982, fig. 1). Other species of Pareuloma have been recorded from the Upper
Cambrian (Upper Franconian) beds of east-central Alaska (Palmer 1968), from the Upper
Cambrian of Xinjiang and Qinghai provinces, China (Zhu 1979; Zhang 1981; Xiang and Zhang
1985), and from the Lower Ordovician of Salair in the Altai Sayan mountain region of the Soviet
Union (Naletov and Sidorenko 1970).

Most authors have regarded Pareuloma as having separate generic status, but Sdzuy (1958)
suggested it might be viewed as a subgenus of Euloma Angelin, 1854. As originally diagnosed by
Rasetti (1954) the genus Pareuloma is distinguished by having a relatively smaller glabella with
correspondingly wider fixed cheeks, and the presence of a much smaller, more anteriorly placed
pair of palpebral lobes. The genus Sanduspis Chien, 1961 (type species, S. gracilis Chien, 1961)
from the Upper Cambrian Sandu Formation of Guizhou Province— see also Yin and Li (also spelt
Lee) 1978, p. 453, pi. 158, fig. 10— has similar features, only differing in its smaller size (possibly
as an immature stage of growth), relatively larger glabella and more rounded, almost sharply
rounded anterior margin, but these differences may not be truly diagnostic of a separate genus.
Indeed, it may be best to regard Sanduspis tentatively as a junior synonym of Pareuloma. Another
Chinese genus which appears to be quite closely related is Chekiangaspis Lu (type species C.
chekiangensis, Lu). Lu et al. (1965, p. 178) have attributed this genus to a publication by Lu in
1960 but apparently the first description and illustrations in print are in Chien (1961, p. 103, pi.
4, figs. 11 and 12; pi. 5, figs. 8 and 9). This form, which is recorded by Yin and Li (1978, p. 476,
pi. 162, fig. 4) from the Upper Cambrian Xiyangshan Formation of Jiangshan and Changshan,
Zhejiang Province, has similar thoracic and pygidial features but differs in exhibiting a more
transversely expanded cephalon, different proportions between fixed and free cheeks and less
prominent lateral glabellar furrows. The fixed cheeks are narrower (tr.) and the eye ridges
inconspicuous. The free cheeks are expanded anterolaterally and prolonged posterolaterally into
large outwardly and backwardly directed genal spines.

Proteuloma Sdzuy, 1958 (type species, Conocephalites geinitzi Barrande, 1868) may also be

WEBBY ET AL.\ CAMBRO ORDOVICIAN TRILOBITES

915

compared with Pareuloma, especially since Xiang and Zhang (1985) have recently reassigned a
number of species originally grouped in Pareuloma to this genus. The glabella of Pareuloma is,
however, relatively much shorter, about half the length of the cranidium, the palpebral lobes are
further forward, placed at the level of the anterior end of the glabella, the preglabellar field may,
but does not always, show a gently raised medium boss, the posterior border furrow is deeper,
and the pygidium more distinctly multisegmented.

Pareuloma aculeatum sp. nov.

Plate 83, figs. 113

Material. Holotype (SUP 48913) and fifteen paratypes (SUP 48912, 48914 48928) from the lower horizon
(locality 1) in the upper part of the Watties Bore Formation, eastern side of Koonenberry Mountain, western
New South Wales.

Etymology. Latin aculeatus , spine-like, referring to long medial spines on the occipital ring and axis of the
thorax.

Diagnosis. Species of Pareuloma with moderately elongate (sag.), forwardly tapering glabella,
relatively narrow (sag.) preglabellar area with poorly differentiated median boss, small palpebral
lobes, large macrospine on occipital ring, transversely elongated, somewhat flattened triangular
pygidium with maximum width at level of first axial ring, terminal piece close to posterior border,
weakly furrowed pleural fields, and a fine granulation.

Description. Exoskeleton of moderate size, usually from 20 40 mm in length (sag.), elongate-elliptical in
dorsal outline and gently convex. Glabella with maximum width at level of occipital ring, being between 0-6
and 10 of glabellar length (sag.); width across base of glabella about one-quarter maximum width of
cranidium. Two pairs of short, notch-like lateral glabellar furrows IP and 2P at sides of glabella; IP furrows
more continuous, directed backwards and inwards, about half-way from occipital to 2P furrows; 2P furrow
set approximately opposite rear end of palpebral lobe; trace of a 3P furrow seen in a few cranidia, placed
near anterolateral corner of glabella. Occipital ring widening (exsag. and sag.) adaxially and posteriorly into
base of large, straight, obliquely backwardly directed median macrospine, with separate median node set
directly in front of spine. Preglabellar area extending to 0-6 of glabellar length (sag.); differentiated by deep
and broad, anterior border furrow into extended gently convex (sag. and exsag.) area of preglabellar field
and more sharply convex (sag. and exsag.) anterior border; small cranidia occasionally show small pits in
anterior border furrow; median part of preglabellar field exhibits slighty updomed median boss.

Small arcuate palpebral lobe, placed opposite glabellar lobe 3P, with associated palpebral furrow extending
into gently curved, forwardly convex eye ridge. Postocular cheek, large, and gently convex with maximum
width greater than that of glabella; deep and wide posterior border furrow separates narrow border, which
widens (exsag.) laterally, and then is deflected forwards and downwards posterolaterally. Course of preocular
facial suture only slightly divergent in front of palpebral lobe, then curves sharply inwards along anterior
edge of border. External surface covered with fine, close-spaced granules and scattered coarser granules
especially in posteromedian corner of postocular cheek.

Free cheek relatively narrow (tr.), with lateral border evenly curved into long, backwardly, and slightly
outwardly directed genal spine. Lateral border furrow deep and broad. Doublure, rostral plate, and hypostoma
unknown.

Thorax of fifteen segments, approximately as wide as long; first six segments of similar length (tr.), then
tapering progressively posteriorly. Axial rings of fourth to ninth segments have large, straight median
macrospines, directed obliquely backwards and slightly upwards; traces of a small median tubercle may be
seen on axial rings of first to third segments; transverse, slot-like apodemal pits set just inside deep axial
furrows. Pleurae flattened, with first six pairs, especially the first three, exhibiting more sharply pointed,
backwardly deflected spines, and more prominent triangular facets; the posterior pairs have shorter (tr.),
more bluntly deflected tips. Pleural furrows broad (exsag.) and deep, with straight transverse course except
for slight backward curvature of abaxial ends; usually placed towards centre (exsag.) of segment; tend to be
pinched out well inside lateral margin of first five segments, but almost extend to tips of more posteriorly
placed segments. Rows of coarse granules extend along anterior and posterior margins of pleural segments;
a finer granulation covers entire surface of thorax; macrospines also show ornamentation of longitudinal
furrows and fine granulation.

916

PALAEONTOLOGY, VOLUME 31

Pygidium small, subtriangular; about one-tenth of total length (sag.) of exoskeleton (excluding macrospines);
length/width ratio varying from 0-3 to 0-4. Axis about one-quarter anterior width of pygidium, almost
reaching posterior margin, and comprising up to four axial rings and a terminal piece; defined by shallow
axial furrows which converge and weaken posteriorly. Pleural fields flat, with two pairs of broad (exsag.),
shallow, transverse pleural furrows, the second pair being much less distinct. Border narrow with fine granules
aligned in rows of wavy lines subparallel to margin. Surface ornamentation of fine and scattered coarser
granules.

Remarks. P. aculeatum sp. nov., differs from the type species. P. brachymetopa Rasetti, 1954 from
the basal Ordovician of Quebec and Newfoundland in having a slightly narrower (sag., exsag. and
tr.) and less well differentiated, trilobed, preglabellar field, an occipital ring with large, backwardly
directed, median macrospine, and flatter pygidium with less strongly furrowed pleural fields. The
other Canadian species, P. impunctata Rasetti, 1954, has a poorly developed, unpitted anterior
border furrow and relatively wider (sag. and exsag.) anterior border. Among the Upper Cambrian
forms, the species P. spinosa Palmer, 1968 from the upper Franconian succession of east-central
Alaska has the closest relationship. It has a closely similar cranidium, only differing from P.
aculeatum in exhibiting larger palpebral lobes and a coarser external surface granulation. The
pygidium is also comparable except that the terminal piece of the axis does not quite reach the
posterior border as in P. aculeatum.

Of the Upper Cambrian species of Pareuloma from China only P. huochengensis Zhang, 1981
from the Guozigou Formation of Guozigou, Huoching County, Xinjiang Autonomous Region,
shows a close resemblance. However, it has larger palpebral lobes, lacks a large macrospine on
the occipital ring and has a less markedly transverse elongated, triangular-shaped pygidium with
the maximum width of the pygidium behind the anterior margin, at the level of the second axial
ring. Another Chinese species, P. qinghaiensis Zhu, 1979, from the Lindaogou Group of
Angshidogou, southern side of Nidanshan mountain, in the Lajishan region of Hualong County,
Qinghai Province, has a much broader (sag. and exsag.) preglabellar field, markedly shorter (sag.)
and a more quadrate-shaped glabella. Xiang and Zhang (1985) have recognized other species which
seem to be closely related, but have chosen to assign them to the genus Proteuloma (see previous
discussion). Their main justification for reassigning such forms as P. houchengensis Zhang, 1981
and part of P. spinosa Palmer, 1968 to Proteuloma is apparently that they do not exhibit a median
boss on the preglabellar field. Otherwise they are closely similar. Indeed, it seems that the
subdivision is excessively arbitrary, especially Palmer’s P. spinosa being split into two separate
genera and species.

EXPLANATION OF PLATE 83

Figs. 1 13. Pareuloma aculeatum sp. nov., Watties Bore Formation, uppermost Cambrian. 1, latex cast of
external mould of cranidium, thorax and pygidium of holotype, SUP 48913, x 4. 2, latex cast of external
mould of cranidium, thorax, and pygidium of paratype, SUP 48915, x 5. 3, latex cast of external mould
of incomplete dorsal exoskeleton of paratype, SUP 48916, x 5. 4, latex cast of external mould showing
part of the cranidium attached to a complete thorax and pygidium, paratype, SUP 48921, x3. 5, latex
cast of external mould of incomplete thorax and pygidium of paratype, SUP 48919, x 2. 6 and 7, latex
cast of external mould of cranidium of paratype, SUP 48922. 6, detail of granulation in the vicinity of
the eye ridge, x 8. 7, general dorsal view, x 4. 8, internal mould of almost complete dorsal exoskeleton
in meraspid stage, paratype, SUP 48917, x 8. 9, internal mould of near complete dorsal exoskelcton of
late meraspid stage, paratype, SUP 48918, x 6. 10, latex cast of external mould of incomplete thorax and
pygidium of paratype, SUP 48925, x 2. 11, internal mould of cephalon and incomplete thorax of paratype,
SUP 48920, x 3. 12, internal mould of incomplete cranidium and thorax of paratype SUP 48914, x 3.

13, latex cast of external mould of incomplete thorax and pygidium of paratype, SUP 48924, showing large
median macrospines, x 3.

PLATE 83

. -

im* , :

|£- srsfv

■-:■.■ .■

'.f'-' ; *>s.%

£S$)i ‘w-'v* ^ : f;{i ■'*)-

' ^ -s-r^4£-, - ; \?^V-
.y M ■■' <-

■

111PI

-^!j*rt.-;T]

WEBBY, WANG and MILLS, Pareuloma

918

PALAEONTOLOGY, VOLUME 31

text-fig. 4. a, Prosaukia’! sp., Watties Bore Formation, uppermost Cambrian. Internal mould of cranidium
of specimen, SUP 48930, x 10; b, saukiid? gen. et sp. indet., Watties Bore Formation, uppermost Cambrian.
Internal mould of incomplete cranidium and thorax of specimen, SUP 48934, x 5; c, Hedinaspis sp., Watties
Bore Formation, uppermost Cambrian. Internal mould of incomplete thorax of specimen, SUP 49938, x 3;
d, Hedinaspis sp., Watties Bore Formation, uppermost Cambrian. Internal mould of cephalon and incomplete
thorax of early meraspid stage, specimen SUP 49935, x 10; E, Hedinaspis ? sp., Watties Bore Formation,
uppermost Cambrian. Internal mould of incomplete cephalon and thorax of ?late meraspid stage, specimen

SUP 49933, x 10.

Superfamily dikelocephalacea Miller, 1889
Family saukiidae Ulrich and Resser, 1930
Genus prosaukia Ulrich and Resser, 1933

Type species. Dikelocephalus misa Hall, 1863.

Prosaukia! sp.

Text-fig. 4a

Material. One cranidium (SUP 48930) from the lower horizon (locality 1) in the upper part of the Watties
Bore Formation, eastern side of Koonenberry Mountain, western New South Wales.

Description. Small cranidium with subquadrate outline except for slightly extended posterolateral extremities.
Glabella rectangular with rounded anterior margin; maximum width 0-5 of sagittal glabellar length. Two
(possibly three) pairs of transglabcllar furrows; first pair deeply impressed abaxially, deflected backwards
and inwards at about 45° to exsagittal line, to mid-point (tr.) on occipital furrow thus delimiting pair of
triangular IP lobes, but also with gently inward curving more weakly impressed ‘transglabellar’ branch,
isolating small, depressed, median triangular area. Second pair shorter, deeply and backwardly directed near
axial furrows but weakening into very gently curved depression adaxially, situated just in front of glabellar
mid-length. Possible third pair seen as faint nick on glabellar surface just in from axial furrow, about half-
way from 2P furrow to frontal glabellar margin. Occipital ring slightly wider (tr.) than rest of glabella.

WEBBY ET AL CAM BRO ORDOVICIAN TRILOBITES

919

narrowing abaxially; trace of median node towards posterior margin may be base of small nuchial spine.
Anterior border furrow broad and deep, separating very narrow rim-like anterior border from wider (sag.
and exsag.), convex preglabellar field; the latter broadens into relatively narrow (tr.) preocular fixed cheek,
about one-third width of glabella. Palpebral lobe of moderate size and narrow kidney-shaped outline.
Postocular fixed cheek broader (tr.), triangular in outline; broad and deep posterior border furrow delimits
convex, outwardly sloping border. Facial suture has almost parallel to slightly divergent preocular course,
and outward and backwardly curving postocular path. Only vague impression of granulose ornamentation
seen.

Remarks. This species is allied to the genus Prosaukia Ulrich and Resser, 1933 because, while it
has a similar glabella character, preglabellar field, and anterior border, it also shows some
differences, such as the presence of a pair of triangular lateral glabellar IP lobes instead of the
more typical development of a rectangular, transglabellar IP lobe, and possibly it also has a nuchal
node. Of the various described Australian species of Prosaukia and Prosaukia'} it seems to bear
closest resemblances to P. sp. A of Shergold (1975) from the upper Upper Cambrian Chatsworth
Limestone of western Queensland in exhibiting traces of a lateral glabellar furrow 3P, and
moderately sized palpebral lobes, and in lacking eye ridges. However, the Chatsworth species has
a relatively wider (tr.) glabella, a more typical rectangular transglabellar IP lobe and no trace of
a nuchal node. Another species with similar features is SaukiaP. aojii Kobayashi 1933? (see Lu et
al. 1965, p. 440, pi. 86, fig. 5) from the Fengshan Formation (upper Upper Cambrian) of Baijiashan,
Wuhujui, Liaoning Province, China, but this differs in having a less well-differentiated anterior
border furrow and a slightly wider glabella with less markedly V-shaped, inwardly and backwardly
directed glabellar furrows IP.

Family saukiidae Ulrich and Resser 1930?

Saukiid? gen. et sp. indet.

Text-fig. 4b

Material. One incomplete cranidium and thorax (SUP 49934) from the lower horizon (locality 1 ) in the
upper part of the Watties Bore Formation, eastern side of Koonenberry Mountain, western New South
Wales.

Description. Internal cast of small, imperfectly preserved cranidium, thorax of up to eleven segments and
tiny, displaced, triangular pygidium. Cranidium with subquadrate glabella, rounding anteriorly; of almost
equal glabellar width (tr.) and length (sag.). Two pairs of broad, shallow transglabellar furrows define weakly
raised transversely elongate IP and 2P lobes. Occipital ring narrow (sag.); right side partially underridden
by first thoracic segment and consequently broken away. Preglabellar area very narrow (sag. and exsag.), in
continuity with narrow (tr.) fixed cheek opposite palpebral lobe; widening posterolaterally into triangular
postocular region. Deep posterior border furrow delimits narrow border.

Thorax with broad axis, tapering markedly backwards. Pleurae with deep, broad pleural furrows extending
diagonally from anteromedian corner almost to posterolateral margin, dying out inside lateral extremities.
Surface covered by fine granules.

Pygidium very small, transverse, with sharply tapering axis; three axial rings; pleural lobes smooth except
for pleural furrow on first segment.

Remarks. This species is difficult to classify. It is referred tentatively to the saukiids because it
has a low, subquadrate glabella with clearly differentiated transglabellar lobes IP and 2P, a
narrow preglabellar area, and preocular fixed cheek region, and a rapidly backwardly tapering
thoracic axis. However, it has slightly wider (tr.) posterolateral limbs (that is, postocular
regions of the fixed cheek) than are typically represented in this group, and this may suggest
that, alternatively, a closer relationship with ptychaspidids (Ptychaspis Hall, 1863) or elviniids
( Cltariocephalus Hall, 1863). The presence of a tiny, transversely elongated subtriangular pygidium
with few axial rings in a sharply backwardly narrowing axis is more typically seen in some
elviniid genera.

920

PALAEONTOLOGY, VOLUME 31

Suborder asaphina Salter, 1864
Superfainily ceratopygacea Linnarsson, 1869
Family ceratopygidae Linnarsson, 1869
Subway iwayaspidinae Kobayashi, 1962
Genus pseudoyuepingia Chien, 1961

Type species. Pseudoyuepingia modes ta Chien, 1961.

Emended diagnosis. Glabella parallel-sided to slightly forwardly tapering, with median glabellar
tubercle placed in front of occipital ring at about twice its length (sag.); up to four pairs of poorly
defined lateral glabellar furrows; more distinct backwardly arched occipital furrow but not
continuous laterally into axial furrows; palpebral lobes of moderate size, may be near to or up to
one-half glabellar width away from axial furrows; preglabellar held clearly differentiated from
narrow anterior border; free cheek with lateral border prolonged into genal spine; thorax of eight
or nine segments; pygidium varies from relatively smooth, less prominently segmented forms to
those with up to eight axial rings, five pleural and interpleural furrows, and broad concave posterior
border.

Discussion. This diagnosis is modified from that proposed by Jago (1987) to accommodate features
such as the presence of a median glabellar tubercle, the palpebral lobes sometimes set well away
from the glabella, the pleurae of the anterior thoracic segments not markedly spinose, and the
pygidium varying between different species from relatively smooth to having well-segmented pleural
areas.

The genus Pseudoyuepingia Chien, 1961 has been assigned to the Ceratopygidae (Iwayaspidinae)
by Kobayashi (1962) and Shergold (1980), and to the Asaphidae (Niobinae) by Lu et al. (1965),
Qiu et a/. (1983) and Xiang and Zhang (1985), in consequence of its morphologically intermediate
position between the two groups. Closely related is the genus Iwayaspis Kobayashi, 1962, regarded
by Shergold (1980) as having separate status from Pseudoyuepingia. However, the morphology of
the type species of Iwayaspis , I. asaphoides Kobayashi, falls well within the range of variability of
known species of Pseudoyuepingia , and is accommodated within the emended diagnosis given
above. Indeed, apart from being markedly more slender, it is quite similar to the second New
South Wales species of Pseudoyuepingia (P. lata sp. nov.) described herein. Following Qiu et al.
(1983, p. 207), the genus Iwayaspis is therefore best viewed as a junior synonym of Pseudoyuepingia.
But it seems preferable to adopt Shergold’s classification of Pseudoyuepingia as a ceratopygacean
of the subfamily Iwayaspidinae.

Pseudoyuepingia has a widespread distribution in Upper Cambrian successions of north-west,
southern, and eastern China (from the Xinjiang Uighur Autonomous Region and from Guizhou,
Hunan, Anhui, and Zhejiang Provinces), South Korea, and western New South Wales, Australia.

The only closely related Australian forms are Cennatops Shergold, 1980 from the post-Idamean

EXPLANATION OF PLATE 84

Figs. 1-10. Pseudoyuepingia whitei sp. nov., Watties Bore Formation, uppermost Cambrian. 1, internal mould
of cranidium, thorax, and pygidium of holotype, SUP 48928, x 4. Note small specimen of Pareuloma
aculeatum sp. nov., at top. 2, internal mould of cranidium and incomplete thorax of paratype, SUP 48931,
x 4. 3, internal mould of incomplete thorax and pygidium of paratype, SUP 48936, x4. 4, internal
mould of enlarged part of thorax and pygidium, paratype, SUP 48940, x 8. 5, internal mould of free
cheek, paratype, SUP 48944, x 5. 6, latex cast of external mould of partially complete dorsal exoskeleton,
paratype, SUP 48935, x 3. 7, internal mould of near complete dorsal exoskeleton, paratype, SUP 48929,
x 5. 8, latex cast of external mould of ventral side of cephalic doublure and hypostoma, paratype, SUP
48933, x 9. 9, latex cast of external mould of damaged, near complete dorsal exoskeleton of paratype,
SUP 48941, x 3. 10, internal mould of fragmentary cranidium, complete thorax and pygidium, paratype,
SUP 48932, x 3.

PLATE 84

WEBBY, WANG and MILLS, Pseudoyuepingia

922

PALAEONTOLOGY, VOLUME 31

part of the Late Cambrian Chatsworth Limestone of western Queensland and a possible species
of Yuepingia Lu, 1956 from a similar level in the Georgina Limestone also in western Queensland
(Henderson 1976). The genus Cermatops differs in having smaller palpebral lobes, less distinct eye
ridges, and a pygidium with strongly developed postaxial ridge and very gently rounded anterolateral
corners. Yuepingia Lu, 1956 is based on type species Y. niobiformis from the Upper Cambrian of
southern China, and is distinguished by its relatively larger, more elongate, forwardly tapering
glabella, weakly developed occipital furrow, narrow (tr.), poorly differentiated preglabellar area,
and much larger palpebral lobes. Psiloyuepingia Qian and Qiu (in Qiu et al. 1983, p. 208) based
on type species P. cylindrica from the Upper Cambrian of Anhui Province, eastern China, is
another which may be compared but differs from Pseudoyuepingia in exhibiting a more elongate,
parallel-sided glabella, larger palpebral lobes, and outwardly diverging preocular facial suture.

Pseudoyuepingia whitei sp. nov.

Plate 84, figs. 110

Material. Holotype (SUP 48928) and fifteen paratypes (SUP 48929, 48931-48944) from the lower horizon
(locality 1 ) in the upper part of the Watties Bore Formation, eastern side of Koonenberry Mountain, western
New South Wales.

Etymology. After Mr Alan White of Wonnaminta Station.

Diagnosis. Species of Pseudoyuepingia with a moderately short (sag.) preglabellar field of similar
length (sag.) to the anterior border, moderately wide fixed cheeks with palpebral lobes about one-
third glabellar width from the axial furrow and eye ridges, an incompletely differentiated occipital
ring, eight thoracic segments, and a relatively smooth weakly segmented pygidium with up to four
axial rings, and a pleural field with only one pair of pleural furrows.

Description. Exoskeleton elliptical in dorsal outline, usually a little less than 20 mm in length. Most of the
material is flattened which does not greatly alter proportions but a few specimens have been tectonically
distorted, thus altering proportions. Glabella with maximum width at level of occipital ring, tapering gently
forwards and rounded anteriorly. Four pairs of rather ill-defined lateral glabellar furrows; IP somewhat
crescentic with concave side facing outwards, seemingly dividing glabella into three roughly equal parts — a
median, and a pair of lateral glabellar lobes; 2P, 3P, and 4P are much fainter, inwardly and backwardly
directed impressions near the axial furrows; 2P is opposite mid-length of palpebral lobe; 3P seemingly near
opposite eye ridge, and 4P close to anterolateral corner of glabella. Occipital ring not well differentiated
because of the discontinuous, weakly developed occipital furrow. Faint median glabellar tubercle developed
in front of occipital furrow. Preglabellar area more or less equally divided (sag.) by conspicuous, broad
anterior border furrow into anterior border and preglabellar field.

Palpebral lobes of moderate size, and situated at mid-length of cranidium, about one-third glabellar width
from axial furrows. Eye ridge distinct, running from axial furrow towards rather poorly defined palpebral
rim. Large L-shaped postocular area with a deep and broad posterior border furrow separating a narrow
(exsag.) convex posterior border. Preocular facial suture runs in parallel-sided to very gently, outwardly
curving arc to intersection with anterior border, then converges sharply inwards along rim of border.
Postocular facial suture arcuate, diverging most sharply behind palpebral lobe.

Free cheek with relatively short genal spine, extending backwards to second or third thoracic segment.
Deep posterior border furrow dies out approaching base of genal spine; anterior and lateral border furrow
deep and relatively broad, also dying out posterolaterally. Narrow, raised lateral border broadens (tr.) into
genal spine, this latter developing an associated longitudinal groove. Doublure broad beneath anterior border
but narrows into lateral border; with up to fifteen terrace lines running subparallel to margin.

Only one very poorly preserved and deformed hypostoma has been found; it is weakly convex, generally
ovate in outline and with very vague differentiation into larger rounded anterior and smaller transversely
elliptical posterior lobes. Anterior wing with subtriangular form narrowing anteriorly. Lateral border extends
from about mid-length of median body into broad posterior border with sharply V-shaped notched posterior
margin.

Thorax of eight segments, with axis occupying between one-quarter and one-third total width. Axial rings
of uniform width (sag.) and defined by deep axial and articulating furrows, with small apodemal pits at their

WEBBY ET AL.\ CAM BRO-ORDOVICIAN TRILOBITES

923

j unctions. Pleurae more or less transversely aligned but with anterior segments more strongly faceted and
outwardly bluntly pointed; posteriorly, pleural segments more expanded, blade-like, and backwardly deflected
into pointed tips, with conspicuous, circular Panderian openings on middle part of doublure. Pleural furrows
broad, deep, and transverse but beyond fulcrum they narrow and become more diagonally directed, dying
out near inner edge of doublure.

Pygidial axis subdivided by ring furrows into three, possibly four, axial rings and a semicircular-shaped
terminal piece. Some specimens also show weakly developed, triangular-shaped postaxial ridge extending
beyond terminal piece almost to posterior margin. Pleural field flattened and relatively smooth, with only
the first pair of pleural furrows developed. Posterior and lateral borders not clearly differentiated from rest
of smooth, flattened pleural field. Doublure extends in to tip of terminal piece and then runs in gentle curve
towards anterolateral corner of pygidium; with about twelve terrace lines subparallcl to margin.

Remarks. P. whitei sp. nov. is similar to P. zhejiangensis Lu and Lin, 1980 from the Upper
Cambrian Xiyangshan Formation of Changshan and Jiangshan in Zhejiang Province of eastern
China, in having only eight rather than nine thoracic segments, and a wider (tr.) area of fixed
cheek at the level of the palpebral lobes, at least one-third of glabellar width. However, the Chinese
species differs in exhibiting a more parallel-sided glabella, larger palpebral lobes, a relatively slightly
narrower thoracic and pygidial axis, and more clearly defined segmentation of the pygidium. P.
distincta Xiang and Zhang, 1985 from the uppermost zone in the Guozigou Formation (upper
Upper Cambrian) of the western part of northern Tianshan, Xinjiang, north-west China, is also
similar but exhibits a more parallel-sided glabella, markedly diverging preocular facial sutures,
and a narrower anterior border with, immediately in front of the glabella, no clearly differentiated
intervening preglabellar field. P. whitei may also be compared with type species P. modesta Chien,
1961 (see Lu et al. 1965) from the Upper Cambrian Sandu Formation of Sandu, Guizhou Province,
southern China, in showing a gently forwardly tapering glabella, weakly developed lateral glabellar
furrows (up to three or four pairs), a median glabellar tubercle towards the rear of the glabella,
median-sized palpebral lobes with weak eye ridges crossing an area of fixed cheek which is at least
one-third of glabellar width (tr.), an anterior border and preglabellar field of subequal width (sag.),
and a comparatively similar pygidium. In contrast the glabella, the thoracic axis and pygidial axis
of P. whitei are comparatively broader (tr.), the posterior border of the cranidium is narrower
(exsag.), the posterior margin of the hypostoma is more distinctly notched, the thorax exhibits
only eight segments with pleural extremities more backwardly deflected (hook-like), and Panderian
openings are more conspicuous on the doublure.

The Idamean (late Cambrian) species of Pseudoyuepingia , P. vanensis Jago 1987, from the
Singing Creek Formation of the Denison Range, south-west Tasmania, exhibits a similarly
short (sag.) preglabellar field, but differs in having a more effaced and narrower (tr.) glabella
and palpebral lobes set closer to the glabella, and it apparently lacks eye ridges, and has nine
thoracic segments.

Pseudoyuepingia lata sp. nov.

Text-fig. 5a e

Material. Holotype (SUP 48947) and six paratypes (SUP 48945, 48948-48949, 49900-49902) from the lower
horizon (locality 1 ) in the upper part of the Watties Bore Formation, eastern side of Koonenberry Mountain,
western New South Wales.

Etymology. Latin, latus, broad, alluding to the wider (tr.) and longer (sag.) preglabellar field.

Diagnosis. Species of Pseudoyuepingia with a long (sag.) and wide (tr.) preglabellar field, long (tr.)
and conspicuous eye ridges, and palpebral lobes with associated wide fixed cheeks (more than one-
half glabellar width), a well-differentiated occipital ring, a thorax of nine segments with a relatively
narrow (tr.) axis, and a pygidium with markedly more segmented axis (up to eight axial rings) and
pleural lobes (up to five pleural and interpleural furrows).

924

PALAEONTOLOGY, VOLUME 31

Comparative description. The exoskeleton has a flattened, elliptical dorsal outline with a length/width ratio
varying dependent on the degree of transverse or longitudinal extension (or compression), from 0-6 to 0-8
(as compared with P. whitei which has a length/width ratio of from 0-5 to 0-7). The proportions between
cephalon and thorax are also slightly different because the thorax with its nine segments tends to be relatively
slightly longer (sag.) than the cephalon. Glabella is slightly less markedly tapering forwards, with only
apparently three pairs of rather ill-defined lateral glabellar furrows. 1 P is developed as inward and backwardly
directed impression, 2P and 3P as much shorter and less well-formed structures, the 3P furrows being situated
adjacent to the eye ridges. A glabellar tubercle is present on the mid-line between the IP furrows. The
occipital ring is well defined by occipital furrow, though it is not completely continuous into the axial furrows.
The preglabellar furrow bounds the frontal part of the glabella and is less deeply impressed than the axial
furrows. A pair of small, pit-like fossulae lie on the axial furrows at anterolateral corners of the glabella.
The gently convex preglabellar field is about twice as wide (sag.) as the raised, rim-like anterior border, and
is more extended laterally (tr.). The width (tr.) across the fixed cheek at the mid-level (exsag.) of the palpebral
lobe is more than half the glabellar width. The palpebral lobes are of moderate size, with a slightly raised,
well-formed, crescentic palpebral rim which extends into the conspicuous eye ridge.

text-fig. 5. A-E, P seudoyuepingia lata sp. nov., Watties Bore Formation, uppermost Cambrian. A, latex cast
of external mould of cranidium, thorax, and pygidium of holotype, SUP 48947, x 4; b, internal mould of
incomplete thorax and pygidium of paratype, SUP 48949, x 5; c, latex cast of external mould of incomplete
cranidium and thorax of meraspid stage, paratype, SUP 49901, x 6; d, latex cast of external mould of
pygidium of paratype, SUP 49900, x 3; e, internal mould of incomplete cranidium, thorax, and pygidium of

paratype, SUP 48948, x 3.

The thorax is of nine segments. The axis occupies from between one-fifth and one-quarter of the width of
the thorax. Pleural lobes are flattened and exhibit transversely aligned pleurae with backwardly deflected
pointed pleural ends. The pleural furrows are broad and shallow, becoming deeper and directed more
diagonally behind the fulcrum. Panderian openings may be seen on the doublure.

The pygidium has a moderately convex, narrow (tr.) axis, with up to seven axial rings, and a small
semicircular terminal piece. The pleural fields exhibit up to five pairs of pleural and interpleural furrows, and
a broad, smooth, slightly concave posterior border, only interrupted by the extension behind the axis of a
weakly raised postaxial ridge.

WEBBY ET AL.: CAM BRO ORDOVICIAN TRILOBITES

925

Remarks. The differences between these two species of Pseudoyuepingia are quite considerable yet
they occupy the same horizon at the particular collecting locality. This suggests they may be sexual
dimorphs of the one species. Whittington (1965) has similarly noted this possibility in two species
of the genus Niobe Angelin, 1851 (members of the subfamily Niobinae Jaanusson, 1959) from the
Middle Ordovician of Newfoundland.

Of the more closely comparable East Asian species of Pseudoyuepingia , P. zhejiangensis Lu and
Lin, 1980 has a more parallel-sided glabella, a shorter (sag.) and narrower (tr.) preglabellar field,
and only eight thoracic segments, the type species, P. modesta Chien, 1961, has a relatively narrower
glabella, narrower area of fixed cheek between palpebral lobes with shorter (tr.) less conspicuous
eye ridges, less extended (sag.) preglabellar field, and less markedly segmented pygidium, and P.
asaphoides (Kobayashi 1962) from the lower Upper Cambrian succession of the southern slopes
of Mount Sambang-san, east of Seto, Puk-myon, South Korea, is overall a more slender (tr.) form
with a narrower, more parallel-sided glabella and very faint, short (tr.) eye ridges on a narrow
area of fixed cheek. An Alaskan species (thorax and pygidia only) from the Franconian 1 level of
the Upper Cambrian, identified by Palmer (1968) as P. cf. asaphoides (Kobayashi 1962), shows a
similar thorax of nine segments and pygidium but without associated cranidia cannot be closely
identified with P. lata.

Other Chinese species like P. aspinosa Qian, 1983 (in Qiu et al. 1983) from the Qingkeng
Formation (middle Upper Cambrian) of Qingkeng, Qingyang, Anhui Province, P. laochatianensis
Yang (MS) (in Zhou et al. 1977; Yang 1978) from the lower Upper Cambrian of western Hunan
Province, and P. I. kontianwuensis Qiu, 1983 (in Qiu et at. 1983) from the Tuanshan Formation
(lower Upper Cambrian) of Huamiao, Guichi, also from Anhui Province, are characteristically
small, slender forms, each with an elongated, parallel-sided glabella, and a prominent, gently raised
median preglabellar ridge extending longitudinally from frontal margin of the glabella to the
anterior border.

Subfamily proceratopyginae Wallerius, 1895
Genus proceratopyge Wallerius, 1895

Type species. Proceratopyge conifrons Wallerius, 1895.

Discussion. Proceratopyge has a widespread distribution in the Middle-Upper Cambrian of Europe
and the Upper Cambrian of the USSR, China, Alaska, Australia, and Antarctica (Shergold 1982).
In China (Lu and Lin 1980) and Kazakhstan (Apollonov et al. 1984) Proceratopyge is recorded
from the upper part of the Upper Cambrian. Rushton (1983) listed some forty-three named species
of the genus, and an additional six species have recently been added to this list by Xiang and
Zhang (1985) from the Upper Cambrian successions of the northern Tianshan, Xinjiang, north-
western China. Of the Australian Upper Cambrian species described previously by Whitehouse
(1939), Opik (1963), Henderson (1976), Shergold (1982), and Jago (1987), there are two species P.
nectans Whitehouse and P. cryptica Henderson from the early Idamean and one species, P. lata
Whitehouse from the late Idamean of western Queensland and P. gordonensis Jago from the
Idamean of Tasmania. These occurrences are from substantially older Upper Cambrian deposits
than the New South Wales record of P. ocella sp. nov. described herein.

Jago (1987) has recently recommended that the species of the genus Proceratopyge should
be split into two broad groups based on various cranidial features. The New South Wales
species belongs to the first group, comprising species with small palpebral lobes placed well
forwards, large posterolateral limbs and preocular sections of the facial suture which diverge
only slightly. In contrast all the described Idamean species from Queensland and Tasmania
belong to Jago’s second grouping, that is, they are forms with larger, more centroposteriorly
placed, crescent-like palpebral lobes, ‘strap-likc’ posterolateral limbs, and a sharply diverging
preocular facial suture.

926

PALAEONTOLOGY, VOLUME 31

Proceratopyge ocella sp. nov.

Plate 85, figs. 1-10

Material. Holotype (SUP 49922) and eleven paratypes (SUP 49921, 49923-49931, 49937) from the lower
horizon (locality 1) in (he upper part of the Watties Bore Formation, eastern side of Koonenberry Mountain,
western New South Wales.

Etymology. Latin, ocellus , a little eye, referring to the relatively small palpebral lobes.

Diagnosis. Species of Proceratopyge (first group of Jago 1987) with faint but clearly defined lateral
glabellar furrows in front of IP, a flattened anterior border, relatively small palpebral lobes placed
just in front of glabellar mid-length, diverging preocular facial suture, up to 20° away from the
exsagittal line, a relatively wide thoracic axis, and a pygidium with up to nine clearly defined axial
and pleural segments, a wide, flattened posterolateral border and a moderately gently rounded
anterolateral angle.

Description. Moderately large, flattened to gently convex, exoskeleton with a length (sag.) of up to 80 mm.
Much of the material of this species is flattened but this does not markedly alter proportions. Glabella, apart
from slight narrowing opposite IP furrows, tapers gently forwards to its rounded anterior margin. Four pairs
of lateral glabellar furrows; IP developed as deeper, backwardly and inwardly curved depressions set well in
from axial furrow, and just behind glabellar mid-length (sag.); 2P much more faintly impressed backward
and inwardly directed impressions just in front of glabellar mid-length; 3P a faint, inward and forwardly
directed slit-like depression, also well inside axial furrow; 4P only a little further forward and close to axial
furrow, almost opposite eye ridge. Small median tubercle faintly developed near mid-length of IP. Occipital
furrow shallows medially but deepening laterally; deepest on exsagittal line of lateral glabellar furrows IP
3P; not extending into axial furrows. Anterior border and preglabellar field subcqual in width (sag. and
exsag.); flattened and only weakly differentiated by broad, very shallow anterior border furrow. Eye ridge
short, extending into well-defined, crescentic rim of palpebral lobes, placed just in front of glabellar mid-
length (sag.) and between one-half and one-third glabellar width from axial furrows. Postocular cheeks
triangular, with very shallow posterior border furrow weakly delimiting narrow posterior border. Preocular
facial suture diverges at about 15-20° to sagittal line, then inward on to anterior margin. Postocular facial
suture diverges sharply behind palpebral lobes, then in gentle sigmoidal course to posterior margin.

Broad cephalic doublure with its numerous concentrically aligned terrace lines and hypostoma shown in
one specimen (PI. 85, fig. 6); median suture not apparently developed as free cheeks are conjoined; preocular
facial suture seems to be impressed on ventral doublure. Rostral plate unknown. Hypostoma has tongue-
shaped outline; widest near mid-length (sag.). Ovate moderate convex median body divided by median furrow
into large, rounded anterior and smaller, transversely elongated, crescentic, posterior lobe. A pair of prominent
raised maculae on median furrow in continuity with lateral border furrow. Anterior wings large, triangular,
directed outwards; no anterior border; narrow lateral border commences opposite hypostomal mid-length
(sag.) and appears to extend backwards into crescentic posterior lobe; sharp angle between lateral and
posterolateral margin; posterolateral border furrow separates very narrow, raised border from posterior lobe;

EXPLANATION OF PLATE 85

Figs. 110. Proceratopyge ocella sp. nov., Watties Bore Formation, uppermost Cambrian. 1, latex cast of
external mould of cranidium and thorax of holotype, SUP 49922, x 1-5. 2, latex cast of external mould
of incomplete cranidium, thorax, and pygidium of paratype, SUP 49923, x 1. 3, latex cast of external
mould of cranidium and thorax of paratype, SUP 49926 (designated specimen at top), x 2-5. 4, latex cast
of external mould of incomplete dorsal exoskeleton of paratype, SUP 49925, x2. 5, internal mould of
incomplete thorax and pygidium of meraspid stage, paratype, SUP 49931, x6. 6, internal mould of

cephalic doublure and hypostoma of paratype, SUP 49937, x4. 7, internal mould of hypostoma of
paratype, SUP 49930, x 3. 8, internal mould of incomplete cranidium of paratype, SUP 49921, x 3. 9,
internal mould of cranidium of paratype, SUP 49924, x 3. 10, latex cast of external mould of pygidium
of paratype, SUP 49927, x 1-5.

Fig. 1 1 . Proceratopyge sp., Watties Bore Formation, uppermost Cambrian; internal mould of incomplete
thorax and pygidium of specimen, SUP 49932, x 2.

PLATE 85

WEBBY, WANG and MILLS, Proceratopvge

928

PALAEONTOLOGY, VOLUME 31

another sharp angle between posterolateral and posterior margin; also a weakly developed median notch.
Anterior lobe has an ornamentation of concentrically arranged anastomosing terrace lines; other parts of
hypostoma show terrace lines running parallel to margins.

Thorax of nine segments with almost parallel-sided axis. Pleural lobes flattened; individual pleurae
transversely aligned, then curved backwards towards bluntly pointed tips; pleural furrows run in a slightly
sigmoidal course, deepening towards the fulcrum, then weakening to die out inside pleural tips; terrace lines
run subparallel to backwardly curving outer ends. Small rounded Panderian structures occur on doublure
of outer part of pleurae. Inner margin of doublure has scalloped appearance, with associated terrace lines
aligned subparallel to lateral margins.

Pygidium with axis slightly tapering backwards, and consisting of up to nine rings and semicircular terminal
piece just inside posterior border. Up to nine pairs of pleural segments, the first prolonged into a long
backwardly directed pleural spine. First pleural furrow curves in arc on to pleural spine, and second also
extends on to flattened posterolateral border; the remaining pleural, and the interpleural, furrows do not
extend on to border. Weakly developed postaxial ridge may also extend on to border. Doublure broad,
widening anterolaterally away from postaxial ridge; terrace lines more or less subparallel to sigmoidally
aligned inside posterolateral margins, and form in an acutely V-shaped pattern along large pleural spines.

Remarks. P. ocella sp. nov. belongs most closely to Proceratopyge ( Lopnorites ), based on the type
species P. (L.) rectispinata Troedsson, 1937 from eastern Tianshan, Xinjiang, north-western China,
in having eye ridges, a subparallel-sided glabella and six or more pygidial axial rings. Henderson
(1976) pointed out the difficulties of recognizing such morphological features as consistently
characterizing this particular subgenus, and recommended against its adoption as a valid subgenus.
The species P. rectispinata , as described by Lu et at. (1965) and Palmer (1968) from China and
Alaska, is similar to P. ocella except in having its palpebral lobes placed a little further forward
on the cheek region, less clearly defined lateral glabellar furrows in front of 1 P, a relatively narrower
thoracic axis, and a less segmented pygidium with narrower border. P. copiosa Xiang and Zhang,
1985, also from the Tianshan region of Xinjiang, similarly resembles P. ocella but for the preocular
facial sutures which are parallel (not diverging), the anterior border more conspicuously upraised
and the pygidium with narrower posterior border, and more sharply rounded anterolateral angle.
P. constrict a Lu, 1964, recently assigned to another subgenus, Sinoprocer atopy ge Lu and Lin, 1980,
from the upper part of the Upper Cambrian in the Wujiajian section of Jiangshan County in
western Zhejiang Province, is a third Chinese species with resemblances to P. ocella but it has a
more regularly parallel-sided, almost quadrate-shaped glabella, larger crescent-shaped palpebral
lobes, and a subtriangular shaped pygidium.

One additional specimen (SUP 49932) of Proceratopyge from the same locality and horizon in
the Watties Bore Formation may represent a second species. However, it is only represented by
an incomplete thorax and pygidium (PI. 84, fig. 11). In contrast to P. ocella , the pygidium is
relatively shorter (sag.) and the axis only exhibits four axial rings and a terminal piece.

Genus hedinaspis Troedsson, 1951
Type species. Hedinia regalis Troedsson, 1937.

Discussion. This genus has a widespread occurrence in the Upper Cambrian of Asia, especially
China (Zhejiang and Guizhou Provinces and Xinjiang Uighur Autonomous Region), Kazakhstan
(Ergaliev 1983u), Alaska and the western United States, western New South Wales, Tasmania,
and New Zealand. It is characteristic of Taylor’s (1976) basinal biofacies, spanning the mid-
Franconian to mid-Trempealeanan interval of central Nevada. The genus is known from horizons
in the topmost part of the Upper Cambrian of western Zhejiang Province (Lu and Lin 1980) and
Xinjiang Uighur Autonomous Region (Xiang and Zhang 1985) of China, and from New Zealand
(Wright and Cooper 1983). It has also been found in a ’possible correlate’ of the Climie Formation
of late Late Cambrian age in Tasmania (Jago, in Shergold et at. 1985). In a direct line of descent
from Hedinaspis is the genus Neohedinaspis Xiang and Zhang, 1984 (type species, N. xinjiangensis
Xiang and Zhang, 1984) from the Tremadoc Sayram Formation of northern Tienshan, Xinjiang.

WEBBY ET AL.: CAMBRO ORDOVICIAN TRILOBITES

929

It differs in having a shorter and broader glabella, shorter eye ridges, short marginal pygidial
spines, and lacks a preglabellar held.

Hedinaspis sp.

Text-fig. 4c

Material. One incomplete thorax (SUP 49938) from the lower horizon (locality I ) in the upper part of the
Watties Bore Formation, eastern side of Koonenberry Mountain, western New South Wales.

Description. Ten segments of moderately sized, flattened thorax (probably posterior portion), showing
backward tapering narrow (tr.) axis between one-fifth and one-sixth of total thoracic width; maximum width
of specimen is 32 mm. Axial furrow moderately deep and scalloped around outwardly convex axial rings,
with very gently backwardly arched, broad articulating furrows intersecting axial furrows at junctions between
scallops. Axial rings of almost similar width (sag. and exsag.) posteriorly.

Remarks. This genus has a distinctive thorax allowing this flattened, partially complete specimen
to be referred to it. However, it must be left in open nomenclature until less fragmentary material
is found.

Two additional specimens from the same locality and horizon may also be allied to Hedinaspis ,
possibly to this same species. The first (SUP 49935) consists of an internal mould of an immature
(early meraspis) stage (text-fig. 4d). Maximum width of the parallel-sided glabella is 0-6 of sagittal
length. Faint impressions of three pairs of lateral glabellar furrows are developed just in from axial
furrows. Occipital ring narrows (exsag.) laterally. Preglabellar field is relatively broad (sag.) and
differentiated from narrow (sag.), raised anterior border. Fixed cheek is broad, with small to
moderate sized, palpebral lobe, and narrow (exsag.) ridge-like posterior border. Free cheek is
relatively narrow (tr.) with prolongation into slender genal spine. Thorax has a relatively narrow
(tr.), gently convex axis and flattened pleural lobes, the pleurae with furrows and spine-like
extremities. Indeed, the specimen has the typical features of the meraspis stage of Hedinaspis
described by Taylor (1976), and is consequently attributed to it.

The second specimen (SUP 49933) is less confidently assigned to Hedinaspis. This single,
incomplete, somewhat damaged cephalon and partial thorax (text-fig. 4e) has a maximum width
of about 6 mm, and consequently probably also represents an immature (?late meraspid) stage.
Glabella is almost parallel-sided with three pairs of rounded to transverse slot-like lateral glabellar
furrows impressed on its outer slopes. Faint, tiny median node is seen in external mould between
IP and 2P furrows. Occipital ring has a markedly crescentic outline, possibly with a small median
node. Preglabellar field is only slightly less than 0-2 of the total glabellar length (sag.) and not
clearly showing anterior border. Fixed cheeks form gently convex subtriangular areas with poorly
developed palpebral lobes. Free cheeks damaged by crushing but clearly with attenuation into
genal spine. Thorax of at least four segments, with relatively broad axis and flattened pleural
regions. Pleurae exhibit deep transverse pleural furrows, narrowing and posteriorly placed beyond
the fulcrum, narrow articulating facets on anterolateral edges and slightly backwardly directed,
and rather spine-like pleural tips. In summary the specimen shows a number of features, such as
more prominent anterolateral facets on pleurae, an axis more than one-quarter of the total thoracic
width, poorly formed palpebral lobes, and lack of eye ridges, which do not seem to be typical of
the genus. Consequently, it is only doubtfully assigned to Hedinaspis.

Subfamily ceratopyginae Linnarsson, 1869
Genus hysterolenus Moberg, 1898

Type species. Hysterolenus toernquisti Moberg, 1898.

Discussion. The ceratopyginid genus Hysterolenus Moberg has until comparatively recently been
viewed as having a restricted early Tremadoc age. In southern Sweden the Hysterolenus fauna with
its type species H. toernquisti , is confined to the Dictyonema Shale (Bergstrom 1982). It comes

930

PALAEONTOLOGY, VOLUME 31

from similar horizons in Kazakhstan (Nikitin et al. 1986), though Ergaliev (1983a) has also
recorded the genus from what he regards as the higher part of the early Tremadoc, in strata of
the Bol’shoy, Karatau, and Ulutau regions, and from the Altai-Sayan mountain region (Petrunina
et al. 1984). Hysterolenus has been widely reported from early Tremadoc stratigraphic levels in
northern and south-east China (Lai 1984). H. tenuispinus Lu and Zhou and H. oblongus Lisogor
have been recorded from the Hangula region, W. Nei Monggol (Lu et al. 1981); H. oblongus from
the Sayram Lormation of the western part of northern Tianshan, Xinjiang Province (Xiang and
Zhang 1984); and H. asiaticus Lu from the Yinchupu Lormation of Changshan, western Zhejiang
Province (Lu and Lin 1980). The range of this latter is the basis for the biostratigraphic subdivisions,
the Hysterolenus Zone and the Onychopyge-Hysterolenus Assemblage Zone, used to correlate early
Tremadoc successions in the Jiangnan ‘shelf margin’ region of south-east China (Lu et al. 1983,
1984; Peng 1983, 1984). Palaeogeographically the Hysterolenus occurrences are restricted to basin
margin-type deposits to the north of the Tarim and North China Platforms, and to the south-east
of the Yangzi Platform (Lai 1984).

Rushton (1982) has raised the possibility of Hysterolenus first appearing in the late Cambrian
by finding an occurrence in the Bryn-Ilin-fawr section of North Wales, 22 m below the first record
of Dictyonema. This appearance of Dictyonema is regarded by Rushton (1982) as indicating the
base of the Tremadoc in Wales. However, the Welsh occurrence, while considered by Rushton
(1982) to belong to the late Cambrian Acerocare Zone, is associated with forms traditionally
characteristic of the Tremadoc such as Niobella homfrayi homfrayi, Parabolina ( Neoparabolina )
frequens , Beltella nodifer, and Shumardia alata (Rushton 1982). Alternatively, the Welsh species
which is so far based on only one pygidium may like a Hysterolenus- type pygidium from the late
Cambrian of China (Lu et al. 1965, pi. 116, fig. 7), represent a different genus. Owing to these
lingering doubts relating to the identification and age of the Welsh material, it seems therefore
that the genus Hysterolenus should continue to be regarded as one of the most useful, restricted
early Tremadoc index fossils, apparently throughout its entire European, Asian, and Australasian
geographic range.

The genus Ruapyge was erected by Wright (1979) with R. hectori (Reed 1926) as type species.
Wright's descriptions were based on at least one of Reed’s specimens, a pygidium (see Reed 1926,
pi. 17, fig. 2c, and Kobayashi 1941, pi. 20, figs. 1 — 1") designated as lectotype, and new collections
made by him from the original type locality at Mount Patriarch, in the South Island of New
Zealand. All the material, including Reed’s type specimens, are poorly preserved and distorted,
and consequently difficult to interpret. Wright (1979) noted the close morphological resemblances
of Ruapyge to Hysterolenus but claimed that Ruapyge differed in having only three (instead of
four) pairs of lateral glabellar furrows and up to eight (rather than from eight to ten) pygidial
axial rings. The glabellar regions of Wright’s illustrated specimens are badly distorted with much
of the detail having been obliterated. Indeed, it is difficult to identify in any of his photographic
illustrations of the material (Wright 1979, pis. 1 and 2), the same patterns of 2P and 3P furrows
he has shown in his reconstruction of R. hectori (Wright 1979, fig. 2). He does not depict a 4P

EXPLANATION OF PLATE 86

Figs. I 11. Hysterolenus furcatus sp. nov., Watties Bore Formation, basal Ordovician. 1, latex cast of external
mould of free cheek of paratype SUP 49907, x 2. 2, latex cast of external mould of cranidium and
incomplete thorax of holotype, SUP 49903, x 2. 3, latex cast of external mould of cranidium of paratype,
SUP 49906, x 3. 4, latex cast of external mould of cranidium and incomplete thorax of paratype, SUP
49905, x 2. 5, internal mould of pygidium of paratype, SUP 49916, x 2-5. 6, internal mould of thoracic
segments of paratype, SUP 49913, x 2. 7, internal mould of incomplete cranidium of paratype, SUP
49918, x2-5. 8, internal mould of pygidium of paratype, SUP 49919 (designated specimen to left side),
x 3. 9, latex cast of external mould of pygidium of paratype, SUP 49908, x 2. 10, internal mould of

incomplete cranidium of paratype, SUP 49904, x 2. 11, internal mould of pygidium of paratype, SUP

49910, x 2.

PLATE 86

WEBBY, WANG and MILLS, Hysterolenus

932

PALAEONTOLOGY, VOLUME 31

furrow yet there seems to be some evidence of one in his plate 2b, with a pair of small forwardly
and inwardly directed slits running off the axial furrows in front of the eye ridges as in Hysterolenus.
R. hectori has the same type of subtriangular pygidium with narrowly tapering axis, even traces
of a postaxial ridge, and the long, slender, backwardly directed pleural spines from the second
segment, as in Hysterolenus. The lesser number of axial rings may be an expression of the poor
state of preservation, with especially some of the smaller rings in the posterior part of the pygidial
axis being selectively destroyed in the deformation. Consequently, we regard Ruapyge as a subjective
junior synonym of Hysterolenus.

Hysterolenus furcatus sp. nov.

Plate 86, figs. I 1 1

Material. Holotype (SUP 49903) and seventeen paratypes (SUP 49904-49920) from the upper horizon (locality
2) in the uppermost part of the Watties Bore Formation, eastern side of Koonenberry Mountain, western
New South Wales.

Etymology. Latin .furcatus, forked, alluding to the bifurcation of the lateral glabellar furrow IP.

Diagnosis. Species of Hysterolenus with a relatively elongate, slightly forwardly tapering glabella,
a conspicuous, deeply indented, long and forked lateral glabellar furrow IP, a moderately
extended preglabellar area, an almost continuous, well-defined occipital furrow, and an elongated,
subtriangular-shaped pygidium with relatively slender axis of nine to eleven axial rings, pleural
field of five to seven ribs, and long, slender marginal spines.

Description. Proportions of cranidium vary because of distortion from just less than twice as wide as long to
slightly wider than long. Glabella tapers gently forwards from maximum width near base, though with slight
outward bulge of 4L lobe; maximum width of glabella varies from 0-5 to 0-9 of sagittal length (including
occipital ring); undistorted maximum glabellar width about 0-65 length; four pairs of lateral glabellar furrows;
IP, 2P, and 3P confined well away from axial furrows; IP most conspicuous, deep, sigmoidally curved
depression, directed mainly backwards and slightly inwards, and with short, laterally directed fork on outer
side; 2P and 3P transversely aligned to rounded slots, 2P near the mid-length of glabella opposite palpebral
lobes, and 3P slightly less conspicuously developed in front of palpebral lobes; 4P is a small slit-like furrow
lying just in front of 3P, but close to axial furrow. Small median tubercle seen near mid-length of IP in some
external moulds. Occipital ring bounded by occipital furrow which deepens laterally into apodemal pits, but
is not continuous into axial furrow; from apodemal pits a pair of branch furrows bifurcate backwards and
inwards across occipital ring. Axial furrows at anterolateral corners of glabella exhibit deep, slit-like fossulae.
Preglabellar field gently concave, extending to 025 of glabellar length (sag.) and widening abaxially; anterior
border furrow separates brim-like, laterally tapering anterior border from rest of preglabellar area. Palpebral
lobes small, crescentic, only slightly elevated and placed near mid-length of glabella, extending into short,
weakly developed, eye ridges. Posterior border of uniform width (exsag.). Preocular facial suture diverges at
between 25 to 35° to exsagittal line, then curves adaxially beneath rim of anterior border. Postocular suture
diverges sharply behind palpebral lobes then more gently to posterior margin.

Free cheeks broad, gently convex, and with narrow, rim-like lateral border in continuity with long, slender
genal spines. Lateral border furrow broad and shallow, dying out towards base of genal spine; posterior
border and furrow not clearly differentiated. Genal caeca of fine radiating and forking lines running across
cheek from near base of eye. Doublure broad, with up to ten terrace lines. Rostrum and hypostoma unknown.

Thorax with up to six segments; possibly maximum number for the species. Axis relatively narrow (tr.),
occupying about one-fifth the width of thorax. Anterolateral slopes of axial ring notched by a pair of
apodemal pits set adjacent to articulating furrow, well inside axial furrow. Pleura crossed by diagonally
directed pleural furrow beginning as narrow groove close to anterior margin, widening abaxially, but then
narrowing again to die out on prolongation into blunt pleural spine. Axial and pleural furrows, and a
transverse, posteriorly placed furrow outline gently raised adaxial pleural lobe.

Pygidium large, with length/width ratio varying between 0-5 and 0-8 on available, mainly deformed,
material. Axis narrow (tr.), about 0T5 of maximum pygidial width, and tapering gently backwards, with
nine to eleven axial rings and a postaxial ridge extending across border on to posterior extremity. Pleural
fields with seven pairs of pleural ribs, the last two being poorly developed. Pleural furrows, at least the first
five, extend obliquely across pleural ribs as deep and wide grooves, narrowing adaxially and abaxially.

WEBBY ET AL.: CAMBRO ORDOVICIAN TRILOBITES

933

Interpleural furrows also well developed as narrower, sharp grooves, outlining clearly the first five pleural
ribs; more posteriorly placed, more closely parallel to ad jacent pleural furrows. Pair of long, slender, marginal
spines issue from second pygidial segment and extends backward beyond the posterior margin of pygidium,
to at least half its sagittal length; marginal spine has at least one longitudinal furrow and faint longi-
tudinally orientated terrace lines. Smooth curvature of relatively narrow posterior border only inter-
rupted by intersection of marginal spines and postaxial ridge. Doublure broad, evenly curved with up to
fifteen terrace lines.

Remarks. Compared with other species of Hysterolenus , H. furcatus sp. nov. is apparently most
closely related to the Baltoscandian type species, H. toernquisti Moberg, 1898 and to the Chinese
H. tenuispinus Lu and Zhou, 1981 (in Lu et al. 1981). It differs from H. toernquisti in having a
more elongate (exsag.) and more conspicuously forked lateral glabellar furrow IP, more marked
occipital furrow and longer, backwardly directed pleural spines. It may be distinguished from H.
tenuispinus by exhibiting a relatively longer (sag. and exsag.) preglabellar area, a more conspicuous
and adaxially more continuous occipital furrow, and a slightly more elongated (exsag.) backwardly
directed arm of the forked IP furrow.

Other species seem to be more markedly different. For instance, H. asiaticus Lu, 1959 (see Lu
et al. 1965; Lu and Lin 1980, 1984) has a relatively broader (tr.), almost parallel-sided glabella
and a relatively more transverse pygidium, with fewer axial rings in a broader (tr.) and shorter
(sag.) terminally more abruptly tapering axis, and fewer pleural ribs. H. oblongus Lisogor, 1961
has a similar glabellar shape but the IP furrows are set relatively further in towards the median
node, and the pygidium appears to exhibit fewer axial rings. H. sarysaiensis Ergaliev, 1983/?, also
from the early Tremadoc of Kazakhstan, is only based on one incomplete pygidium, and similarly
has fewer (seven or eight) pygidial axial rings. H. hectori (Reed 1926), from the early Tremadoc
of New Zealand (Wright 1979), despite its highly deformed and poorly preserved nature, seems
most closely to resemble H. asiaticus in having a broad, almost parallel-sided glabella, a relatively
short preglabellar field, and a transversely extended pygidium with a broad, blunt, less segmented
axis.

Acknowledgements . This study has been supported by funds from the Australian Research Grants Committee
(A.R.G.S. grant no. E82/15297). Thanks are extended to the White family of Wonnaminta station for
providing support and encouragement in the field, and to Dr J. H. Shergold (Bureau of Mineral Resources,
Canberra) and other anonymous referees for reviewing the manuscript and offering constructively useful
suggestions.

REFERENCES

angelin, n. p. 1851 -1854. Palaeontologica Scandinavica. Pars. I Crustacea formationis transitionis, 1 24, pis.

1 24; Pars. II, pp. i ix, 25 92, pis. 25 41. Academiae Rcgiae Scientarium Suecanae (Holmiae), Lund.
apollonov, m. k., chugaeva, m. n. and dubinina, s. v. 1984. Trilobites and conodonts from the Batyrbay
section ( uppermost Cambrian- Lower Ordovician ) in Malyi Karatau Range. Atlas of palaeontological
plates, I 48, pis. 1 32. Academy of Sciences of the Kazakh SSR, Order of the Red Banner of Labour,
K. I. Satpaev Institute of Geological Sciences, ‘Nauka’ Kazakh SSR Publishing House, Alma-Ata. [In
Russian.]

barrande, j. 1868. Silurische Fauna aus der Umgebung von Hof in Bayern. Neues Jb. Miner. 641-696,
pis. 6-7.

bergstrom, j. 1982. Scania. In bruton, d. l. and williams, s. h. (eds.). Field excursion guide. IV Int. Symp.

Ordovician System, Paleont. Contr. Univ. Oslo , 279, 184 197.
brongniart, a. 1822. Les Trilobites. In brongniart, a. and desmarest, a. g. 1822. Histoire naturelle des
Crustaces fossiles sous les rapports zoologiques et geologiques, 1-58, pis. 1-6. F. G. Levrault, Paris.
brunker, r. l., offenberg, a. and rose, g. 1971. Koonenberry, 1:500000 Geological Series Sheet. Department
of Mines, NSW.

chien yi-yuan. 1961. Cambrian trilobites from Sandu and Duyun, southern Kweichow. Acta palaeont. sin.
9 (2), 91-139, pis. 1-5. [In Chinese with English summary.]

934

PALAEONTOLOGY, VOLUME 31

ergaliev, G. k. 1983a. The Cambrian-Ordovician boundary in southern Kazakhstan and Ulutau. In
apollonov, m. K., bandeletov, s. M. and ivshin, N. K. (eds.). The Lower Palaeozoic Stratigraphy and
Palaeontology of Kazakhstan , 6-16. Academy of Sciences of the Kazakh SSR, Order of the Red Banner
of Labour, K. I. Satpaev Institute of Geological Sciences, ‘Nauka' Kazakh SSR Publishing House, Alma-
Ata. [In Russian.]

19836. Certain Upper Cambrian and Lower Ordovician trilobites of High Karatau and Ulutau. In
APOLLONOV, M. K., bandaletov, s. M. and ivshin, N. K. (eds. ). The Lower Palaeozoic Stratigraphy and
Palaeontology of Kazakhstan, 35 -66, pis. I 6. Academy of Sciences of the Kazakh SSR, Order of the Red
Banner of Labour, K. I. Satpaev Institute of Geological Sciences, ‘Nauka’ Kazakh SSR Publishing House,
Alma-Ata. [In Russian.]

fortey, r. a. 1980. The Ordovician trilobites of Spitzbergen. III. Remaining trilobites of the Valhallfonna
Formation. Skr. norsk Polarinst. 171, 1-113, pis. 1-25.

1983. Cambrian-Ordovician trilobites from the boundary beds in western Newfoundland and their
phylogenetic significance. In briggs, d. e. g. and lane, p. d. (eds.). Trilobites and other early arthro-
pods: papers in honour of Professor H. B. Whittington, F.R.S. Spec. Pap. Palaeont. 30, 179-211,
pis. 23-27.

— landing, e. and skevington, d. 1982. Cambrian-Ordovician boundary sections in the Cow Head
Group, western Newfoundland. In bassett, m. g. and dean, w. t. (eds.). The Cambrian-Ordovician
boundary: sections , fossil distributions and correlations , 95-129. National Museum of Wales, Geological
Series, no. 3, Cardiff.

hall, j. 1863. Preliminary notice of the fauna of the Potsdam Sandstone. New York State Cabinet , Nat.

History, Albany. 16th Ann. Rep., Appendix D, 119-184.
henderson, r. a. 1976. Idamean (early Upper Cambrian) trilobites from north-western Queensland, Australia.
Palaeontology, 19, 325-364, pis. 47-51.

howell, b. f. 1935. Some New Brunswick Cambrian agnostians. Bull. Wagner Inst. Sci. Philad. 10, 13-16,
pi. 1.

jaanusson, v. 1959. Asaphidae. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part O. Arthropoda
I, 334-355. Geological Society of America and University of Kansas Press, Boulder, Colorado and
Lawrence, Kansas.

jaekel, o. 1909. Ueber die Agnostiden. Z. dt. geol. Ges. 61, 380 401.

jago, J. b. 1987. Idamean (Late Cambrian) trilobites from the Denison Range, south-west Tasmania.
Palaeontology, 30, 207-231.

jell, p. a. 1975. Australian Middle Cambrian eodiscoids with a review of the superfamily. Palaeontographica,
A 150, 1-97, pis. 1-29.

— 1985. Tremadoc trilobites of the Digger Island Formation, Waratah Bay, Victoria. Mem. Mus. Viet.
46, 53-88, pis. 19-33.

jones, p. J., shergold, j. h. and druce, E. c. 1971. Late Cambrian and Early Ordovician stages in western
Queensland. J. geol. Soc. Aust. 18, 1-32.

kobayashi, t. 1933. Upper Cambrian of the Wuhutsui Basin, Liaotung, with special reference to the limit
of the Chaumitien (or Upper Cambrian) of eastern Asia, and its subdivision. Jap. Jl Geol. Geogr. 1 1
(1-2), 55- 155, pis. 9 15.

1935. The Briscoia fauna of the late Upper Cambrian in Alaska with descriptions of a few Upper
Cambrian trilobites from Montana and Nevada. Ibid. 12 (3 4), 39-57, pis. 8 10.

1939. On the agnostids (Part 1). Jour. Fac. Sci. Univ. Tokyo, ser. 2, 5, 69-198.

— 1941. On the occurrence of Taihungshania, a characteristic Arenigian trilobite, in New Zealand. Jap. J.
Geol. Geogr. 17, 195-201.

— 1955. The Ordovician fossils from the McKay Group in British Columbia, western Canada, with a note
on the Early Ordovician palaeogeography. J. Fac. Sci. Tokyo Univ. sec. 2, 9, 355-493, pis. 1-9.

1962. The Cambro-Ordovician formations and faunas of South Korea, part IX. Palaeontology VIII.
The Machari fauna. Ibid. 14, 1-152, pis. 112.

— 1966. The Cambro-Ordovician formations and faunas of South Korea, Part 10. Stratigraphy of the
Chosen Group in Korea and South Manchuria and its relation to the Cambro Ordovician formations of
other areas. Section A. The Chosen Group of South Korea. Ibid. 16, 1-84.

lai caigen. 1984. The Tremadoc Series in China. Aust. Jl Earth Sci. 31, 1-6.

lake, p. 1906. A monograph of the British Cambrian trilobites. Palaeontogr. Soc. [Monogr.], Pi. 1, 1-28, pis.
1-2.

leitch, e. c., webby, b. d., mills, k. j. and kolbe, p. 1987. Terranes of the Wonominta Block, far western

WEBBY ET AL.: CAMBRO ORDOVICIAN TRILOBITES

935

New South Wales. In leitch, e. c. and scheibner, e. (eds.). Terrane accretion and orogenic belts , 31-37.
American Geophysical Union, Washington D.C.

linnarsson, j. G. o. 1869. Om Vestergotlands cambriska och siluriska, flagringar. Handl. K. svenska
Vetenskapsakad. Stockholm , 8, (2), 3-89.

lisogor, k. a. 1961. Tremadocian trilobites and their associates in the deposits of the Kendyktas. Trans.
Geol. Inst., Moscow, IS, 55-91, pis. 1-4. [In Russian ]

lu yanhao. 1956. An Upper Cambrian trilobite faunule from eastern Kueichou. Acta palaeont. sin. 4 (3),
365-380, pi. 1. [In Chinese.]

— 1959. Guide-book, Excursion to W. Zhejiang, 1-28, pis. 1 5. [In Chinese.] [Not seen.]

— 1964. Trilobita ( partim ). In wang yu (ed.). Index fossils of South China, 32-35, 42-47. Science Press,
Beijing. [In Chinese.] [Not seen.]

— and lin huanling. 1980. Cambro Ordovician boundary in western Zhejiang and the trilobites contained
therein. Acta palaeont. sin. 19 (2), 118-135, pis. 1-3. [In Chinese with English summary.]

- 1984. Late late Cambrian and earliest Ordovician trilobites of Jiangshan-Changshan area, Zhejiang.
In Stratigraphy and palaeontology of systemic boundaries in China. Cambrian-Ordovician boundary, 1,
45 143, pis. 1 19. Anhui Science and Technology Publishing House, Hefei.

— chang wentang, chu chaoling, chien yiyuan and hsiang li wen. 1965. Trilobites of China, vol. 1,
1-362, pis. 1-66; vol. 2, 363-766, pis. 67-135. Science Press, Beijing. [In Chinese ]

— chu chaoling, chien yiyuan, lin huanling, chow tseyi and yuan kexing. 1974. Bio-environmental
control hypothesis and its application to the Cambrian biostratigraphy and palaeozoology. Mem. Nanjing
Inst. Geol. Palaeont. 5, 27-116, pis. 1-4.

— lin huanling, han nairen, li luozhao and ju tianyin. 1983. Cambrian-Ordovician boundary of
Jiangshan-Changshan area. In Papers for the Symposium on the Cambrian- Ordovician and Ordovician-
Silurian Boundaries, Nanjing, China , October 1983, I -5. Nanjing Institute of Geology and Palaeontology,
Nanjing.

1984. On the Cambrian-Ordovician boundary of the Jiangshan-Changshan area,
W. Zhejiang. In Stratigraphy and palaeontology of system boundaries in China. Cambrian-Ordovician
boundary, I, 9-44. Anhui Science and Technology Publishing House, Hefei.

— zhou zhiyi and zhou zhigiang. 1981. Cambrian-Ordovician boundary and their related trilobites in
the Hangula region, W. Nei Monggol. Bull. Chin. Acad. Geol. Sci. ser. 10, 2(1), 1-19, pis. 1-3. [In Chinese
with English summary.]

m‘coy, f. 1849. On the classification of some British fossil Crustacea, with notices of new forms in the
University collection at Cambridge. Ann. Mag. nat. Hist. ser. 2, 4, 161 179, 392-414.
matthew, G. f. 1887. Illustrations of the fauna of the St. John Group, No. IV, pt. I, Description of new
species of Paradoxides (Paradoxides regina). Pt. II, The smaller trilobites with eyes (Ptychoparidae and
Ellipsocephalidae). Trans. Proc. R. Soc. Can. 5, I 15-166, pis. 1-2.
miller, s. A. 1889. North America geology and paleontology for the use of amateurs , students and scientists,
718 pp., 1265 figs. Cincinnati, Ohio.

moberg, j. c. 1898. Om Acerocarezonen. Geol. For. Stockh. Fbrh. 20, 197-290, pis. 10 14.
naletov, b. f. and Sidorenko, t. f. 1970. Early Ordovician volcanic assemblages of Salair Ridge. Geologiya
Geofiz. Novosibirsk 1970 (5), 72-78. [In Russian.]

NIKITIN, I. F., APOLLONOV, M. K., TZAI, D. T., KOROLJOV, V. G., KIM, A. I., ERINA, M. V., LARIN, N. M. and GOLIKOV,
a. n. 1986. The Ordovician System in Kazakhstan and Middle Asia. Correlation charts and explanatory
notes. Int. Union Geol. Sci. Pub!. 21, 1-34.

opik, a. a. 1963. Early Upper Cambrian fossils from Queensland, Bull. Bur. Miner. Resour. Aust. 64, 1-133,
pis. 1-9.

— 1967a. The Ordian Stage of the Cambrian and its Australian Metadoxididae. Ibid. 92, 133 169.

19676. The Mindyallan fauna of North-Western Queensland. Ibid. 74 (2 vols.), 1 404, 1-167,

pis. 1-67.

— 1970. Redlichia of the Ordian (Cambrian) of northern Australia and New South Wales. Ibid. 114,
1-66, pis. I 14.

— 1975a. Templetonian and Ordian xystridurid trilobites of Australia. Ibid. 121, 1 84, pis. 1-32.

— 19756. Cymbric Vale fauna of New South Wales and Early Cambrian biostratigraphy. Ibid. 159, 1-78,
pis. 1 -7.

— 1979. Middle Cambrian agnostids: systematics and biostratigraphy. Ibid. 172 (2 vols.), 1-188,
pis. 1 -67.

936

PALAEONTOLOGY, VOLUME 31

opik, a. a. 1982. Dolichometopid trilobites of Queensland, Northern Territory and New South Wales. Ibid.
175, 1-85, pis. 1-32.

palmer, a. r. 1955. Upper Cambrian Agnostidae of the Eureka District, Nevada. J. Paleont. 29, 86-101, pis.
19-20.

— 1968. Cambrian trilobites of east-central Alaska. US Geol. Surv ., Prof. Paper, 559-B, 1-115, pis. I 13.
peng shanchi. 1983. Cambrian-Ordovician boundary in the Cili Taoyuan border area, northwestern Hunan.

In Papers for the Symposium on the Cambrian-Ordovician and Ordovician- Silurian boundaries, Nanjing,
China 1983 , 44 52, pis. I 3. Nanjing Institute of Geology and Palaeontology, Nanjing.

— 1984. Cambrian Ordovician boundary in the Cili-Taoyuan border area, northwestern Hunan with
descriptions of the relative trilobites. In Stratigraphy and Palaeontology of systemic boundaries in China.
Cambrian Ordovician boundary. 1, 285-405, pis. I 18. Anhui Science and Technology Publishing House,
Hefei.

PETRUNINA, Z. E., SENNIKOV, N. V., ERMIKOV, V. D., ZEIFERT, L. L., KRIVCHIKOV, A. V. and PUZYREV, A. A. 1984.
Lower Ordovician Stratigraphy of the Gorny Altai. Trudy Inst. Geol. Geofiz., sib. Otd. Akad. Nauk SSSR,
565, 3-33. [In Russian.]

pocock, k. j. 1974. Estaingia , a new trilobite genus from the Lower Cambrian of South Australia.
Palaeontology, 7, 458-471.

pogson, d. j. and scheibner, e. 1971. Pre-Upper Cambrian sediments east of Copper Mine Range, New
South Wales. NSW geol. Surv. Quart. Notes, 4, 3-8.

powell, c. mca., neef, G., crane, d., jell, p. and percival, i. G. 1982. Significance of Late Cambrian
(Idamean) fossils in the Cupala Creek Lormation, northwestern New South Wales. Proc. Linn. Soc. NSW,
106, 127-150.

qian yiyuan. 1985. Late late Cambrian trilobites from the Tangcun Lormation of Jingxian, southern Anhui.
Palaeontologia Cathayana, 2, 137-167.

QIU HONGAN, LU YANHAO, ZHU ZHAOLING, BI DECHANG, LIN TIANRUI, ZHOU ZHIYI, ZHANG QUANZHONG and
qian yiyuan. 1983. Trilobita. In Palaeontological Atlas of East China. Volume of Early Palaeozoic, 28-254,
pis. I I 88. Geological Publishing House, Beijing. [In Chinese.]
rasetti, f. 1954. Early Ordovician trilobite faunules from Quebec and Newfoundland. J. Paleont. 28, 58 1
587.

1966. Revision of the North American species of the Cambrian trilobite genus Pagetia. Ibid. 40, 502-
511.

reed, f. r. c. 1926. New trilobites from the Ordovician beds of New Zealand. Trans. R. Soc. NZ, 57, 310-
314.

robison, r. a. and pantoja-alor, j. 1968. Tremadocian trilobites from the Nochixtlan region, Oaxaca,
Mexico. J. Paleont. 42, 767-800, pis. 97-104.

rose, G. 1974. Explanatory notes on the White Cliffs 1:250000 Geological Sheet. Geol. Surv. NSW, Sydney,
1-48.

— louden, a. G. and o’connell, p. 1964. White Cliffs Sheet SH 54-12, 1:250000 Geological Series (1st
edn.). Geol. Surv. NSW, Sydney.

rushton, a. w. a. 1982. The biostratigraphy and correlation of the Merioneth Tremadoc Series boun-
dary in North Wales. In bassett, m. g. and dean, w. t. (eds.). The Cambrian-Ordovician boundary;
sections, fossil distributions and correlations, 41 59. National Museum of Wales Geological Series, no. 3,
Cardiff.

— 1983. Trilobites from the Upper Cambrian Olenus Zone in central England. In briggs, d. e. g. and
lane, p. d. (eds.). Trilobites and other early arthropods: papers in honour of Professor H. B. Whittington,
L.R.S. Spec. Pap. Palaeont. 30, 107-139, pis. 14-19.

salter, J. w. 1864. A monograph of the British trilobites. Palaeontogr. Soc. [Monogr.], 1 80, pis. I 6.
sdzuy, K. 1958. Lossilien aus dem Tremadoc der Montagne Noire. Senckenberg. leth. 39 (3/4), 255-288, pis.
1 3.

shergold, j. h. 1969. Oryctocephalidae (Trilobita: Middle Cambrian) of Australia. Bull. Bur. Miner. Resour.
Aust. 104, I 66, pis. I 12.

I971u. Resume of data on the base of the Ordovician in northern and central Australia. In Colloque
ordovicien-silurien, Brest, Septembre 1971. Mem. Bur. Rech. geol. minier. 73, 391 402.

19716. Late Upper Cambrian trilobites from the Gola Beds, western Queensland. Bull. Bur. Min.
Resour. Aust. 112, 1-87, pis. 1 19.

1975. Late Cambrian and early Ordovician trilobites from the Burke River Structural Belt, western
Queensland. Ibid. 153 (2 vols.), 1-251, pis. 1-58.

WEBBY ET AL.\ CAMBRO ORDOVICIAN TRILOBITES

937

1977. Classification of the trilobite Pseudagnostus. Palaeontology , 20, 69-100, pis. 15 16.

1980. Late Cambrian trilobites from the Chatsworth Limestone, western Queensland. Bull. Bur. Miner.
Resow. Aust. 186, 1 111, pis. 1-35.

1982. Idanrean (Late Cambrian) trilobites, Burke River Structural Belt, western Queensland. Ibid. 187,

1 69, pis. 117.

1988. Review of trilobite biofacies distributions at the Cambrian Ordovician boundary. Geol. Mag.
125, 363-380.

— cooper, r. a., druce, E. c. and webby, b. d. 1982. Synopsis of selected sections at the Cambrian-
Ordovician boundary in Australia, New Zealand and Antarctica. In bassett, m. g. and dean, w. t. (eds.).
The Cambrian-Ordovician boundary; sections, fossil distributions and correlations , 211-227. National
Museum of Wales Geological Series, no. 3, Cardiff.

jago, J., cooper, R. a. and laurie, j. 1985. The Cambrian System in Australia, Antarctica and New
Zealand. Correlation charts and explanatory notes. Int. Union Geol. Sci ., Publ. 19, 1 85.
sun yunchu. 1924. Contribution to the Cambrian faunas of China. Palaeont. sin. Ser. B , 1, Fasc. 4, 1-109,
pis. 1 5.

swinnerton, h. h. 1915. Suggestions for the revised classification of trilobites. Geol. Mag. 2, 487 496,
538-545.

taylor, m. e. 1976. Indigenous and redeposited trilobites from late Cambrian basinal environments of central
Nevada. J. Paleont. 50, 668-700, pis. I 3.

troedsson, g. t. 1937. On the Cambro-Ordovician faunas of western Qurug-Tagh, eastern Tienshan. In
Report of the scientific expedition to the north-western provinces of China under the leadership of Dr.
Sren Hedin. The Sino-Swedish Expedition Publ. 4. V. Invertebrate Palaeontology, 1. Palaeont. sin. ns B ,

2 (whole ser. 106), 1-74, pis. 1 10.

- 1951. Hedinaspis , new name for Hedinia Troedsson, non Navas. Geol. For. Stockh. Forli. 73, 695.
ulrich, e. o. and resser, c. e. 1930. The Cambrian of the Upper Mississippi Valley, Pt. I, Trilobita;

Dikelocephalinae and Osceolinae. Bull. Milwaukee publ. Mus. 12 (1), 1 122, pis. 1 23.

1933. The Cambrian of the Upper Mississippi Valley, Pt. 2, Trilobita; Saukiinae. Ibid. 12 (2), 123-
306, pis. 24-45.

wallerius, i. d. 1895. Undersokninga ofver zonen med Agnostus laevigatus i Vestergotlands samtliga
Paradoxides-fager. Akademisk afhandl., iii, 1-73, pi. I. Lund.
wang zhihao. 1984. Late Cambrian and early Ordovician conodonts from north and northeast China with
comments on the Cambrian Ordovician boundary. In Stratigraphy and palaeontology of systemic boundaries
in China. Cambrian Ordovician boundary , 2, 195 -258, pis. 1 14. Anhui Science and Technology Publishing
House, Hefei.

whitehouse, f. w. 1936. The Cambrian faunas of northeastern Australia. Part 1— Stratigraphic outline. Part
2— Trilobita (Miomera). Mem. Qd. Mus. 11 (1), 59-112, pis. 8-10.

1939. The Cambrian faunas of northeastern Australia. Part 3 — The polymerid trilobites (with supplement
no. 1). Ibid. 11, 179-282, pis. 19-25.

Whittington, h. b. 1965. Trilobites of the Ordovician Table Head Formation, Western Newfoundland. Bull.
Mus. comp. Zool. Harv. 132 (4), 275-442, pis. 1 68.

wright, a. j. 1979. Evaluation of a New Zealand Tremadocian trilobite. Geol. Mag. 116, 353-364,
pis. 1-2.

and cooper, r. a. 1983. Cambrian Ordovician boundary at Mount Patriarch, New Zealand. In Papers
for the Symposium on the Cambrian-Ordovician and Ordovician- Silurian boundaries, Nanjing , China, October
1983, 62-63. Nanjing Institute for Geology and Palaeontology, Nanjing.
xiang liwen and zhang tairong. 1984. Tremadocian trilobites from the western part of northern Tianshan,
Xinjiang. Acta palaeont. sin. 23 (4), 399-409, pis. I 3. [In Chinese with English summary.]

1985. Stratigraphy and trilobite faunas of the Cambrian in the western part of northern Tienshan,
Xinjiang. Ministry of Geol. and Mineral Resour., Geol. Mem. Ser. 2, no. 4, i-ix, 1-243, pis. 1-52. Geological
Publishing House, Beijing. [In Chinese with English summary ]
yang jialu. 1978. Middle and Upper Cambrian trilobites of western Hunan and eastern Guizhou. Prof. Pap.

Stratigraphy and Palaeontology, 4, 1-82, pis. 1 13. [In Chinese with English summary.]
yin gongzheng and Li SHANJi. 1978. Trilobita. In Palaeontological Atlas of south-west China. Guizhou
Volume, Pt. 1 (Cambrian- Devonian), 385-594, pis. 144-192. Geological Publishing House, Beijing.
[In Chinese.]

938

PALAEONTOLOGY, VOLUME 31

yin gongzheng, gong lianzan, cai ying and jiao huiliang. 1984. On the Cambrian-Ordovician boundary
in Guizhou of China. Scientific Papers on Geology for International Exchange , vol. I (for 27th Int. Geol.
Congr .), 25-34, pis. 1-2. Geological Publishing House, Beijing. [In Chinese with English summary.]
zhang tairong. 1981. Trilobita. In Palaeontological Atlas of Northwestern China, Xinjiang volume, Pt. I
(later Proterozoic- Early Palaeozoic), 134-213, pis. 54-79. Geological Publishing House, Beijing. [In
Chinese.]

zhou tianmei, liu yiren, meng xiansong and sun zhenhua. 1977. Trilobita. In Palaeontological Atlas of
Central and Southern China. Vol. 1 (Early Palaeozoic), 104-266, pis. 36-81. Geological Publishing House,
Beijing. [In Chinese.]

zhou zhiyi, wang zhihao, zhang junming and lin yaokun. 1984. Cambrian-Ordovician boundary sections
and the proposed candidates for stratotype in North and Northeast China. In Stratigraphy and palaeontology
of systemic boundaries in China. Cambrian-Ordovician boundary, 2, I 57, pis. 1-3. Anhui Science and
Technology Publishing House, Hefei.

zhu zhaoling. 1979. Superfamily Ptychopariacea (Trilobita). In zhu zhaoling, lin huanling and zhang
zhiheng. Palaeontological Atlas Northwest China, Qinghai Volume, Pt. II, 81 116, pis. 35 45. Nanjing
Institute of Geology and Palaeontology and Qinghai Institute of Geology. Geology Publishing House,
Beijing. [In Chinese.]

B. D. WEBBY and K. J. MILLS

Department of Geology and Geophysics
University of Sydney
Sydney, N.S.W. 2006
Australia

WANG QIZHENG

Department of Geology
Hebei Institute of Geology
Xuanhua County

Typescript received 26 June 1987 Hebei Province

Revised typescript received 6 December 1987 People’s Republic of China

PARASITISM OF ORDOVICIAN BRYOZOANS
AND THE ORIGIN OF PS EU DO BO R 1 NGS

by t. j. palmer and m. a. wilson

Abstract. Upper Ordovician trepostome bryozoans from the vicinity of Cincinnati, Ohio, USA, contain
trace fossils that resulted from the overgrowth by the bryozoan of soft-bodied parasites that settled on
the living colony. The resulting structures (pseudoborings) superficially resemble borings, and the term
‘bioclaustration’ is introduced to describe the process. The pseudoboring consists of groups or rows of sub-
circular pits, connected by tunnels that were formed by the roofing-over of adventitious stolons by localized
bryozoan growth. The structure reflects the external morphology of the parasite, and is named Catellocaula
vallata ichnogen. and ichnosp. nov. A hydroid or colonial ascidiacian tunicate is suggested as the perpetrator.

The study of trace fossils in the Upper Ordovician rocks in the vicinity of Cincinnati, Ohio, where
minimal diagenetic overprinting and exquisite preservation rival anything that can be found in the
European Mesozoic, has largely concentrated on burrows and trails (Osgood 1970). Although the
hard substrate trace fossils have received passing mention from a number of workers (Palmer 1982;
Wilson 1985), there have been no detailed studies of the borings that are found abundantly in
organic and inorganic hard substrates.

By far the most common of these borings is Typanites , which is found in the massive skeletons
of bryozoans, corals, and stromatoporoids; in the thin shells of molluscs and brachiopods; and in
cobbles and hardgrounds. Trypanites , which undoubtedly represents the dwelling tubes of a variety
of filter-feeding worms, is extensively known from other Ordovician rocks throughout North
America and Europe (Kobluk et al. 1978). Of far more limited geographic extent, apparently
limited to the Lower Cincinnatian of the type area, is the groove-shaped boring first described by
Pojeta and Palmer (1976) and ascribed to the rasping activity of the modiomorphid bivalve
Corallidomus scobina. These borings, named Petroxestes pera by Wilson and Palmer (1988), occur
solitarily or as aggregated clusters in cobbles, hardgrounds, and massive skeletons.

But if borings have received short shrift relative to soft-sediment trace fossils in the Upper
Ordovician, how much more so has the second class of hard-substrate trace fossil, formed by
biological infestation of a living host that subsequently adapted its growth to enclose and isolate
the infester. Such embedment structures are generally acknowledged to be a class of trace fossil
(Muller 1962; Bromley 1970; Conway Morris 1980; Ekdale et al. 1984), but are easily mistaken
for borings because they end up as holes in the skeleton of the host. The walls and rims of such
holes must be closely examined for signs that the skeletal elements and growth lamellae of the host
are distorted around the hole, rather than cut by it. Only thus can such pseudoborings be
distinguished from true borings. Bromley (1970) discussed several examples of such embedment
structures, and pointed out that in some cases, elements of both embedment and boring can be
seen in the same structure. Borers, for example, may break through the inner surface of the shell
of a living bivalve, and cause it to cover the intrusion with a blister of carbonate, secreted by the
outside face of the mantle. Similarly, embedded parasites may enlarge their holes by boring, in
order to accommodate growth or erosion.

The process of embedment of a soft-bodied infesting organism by skeletal growth of its host is
called by us ‘bioclaustration’ (biologically claustrated, or enclosed behind a wall, cloister, or
rampart). The unequivocal example of this process that forms the subject of this paper is the
earliest yet described in detail, and the only example so far elucidated that involves fossil
bryozoans.

IPalaeontoIogy, Vol. 31, Part 4, 1988, pp. 939-949, pi. 87.)

© The Palaeontological Association

940

PALAEONTOLOGY, VOLUME 31

Recognition of bioclaustration in the fossil record requires the host organism to be skeletonized.
The infester, probably parasitic but conceivably mutualistic (see discussion below) is, by definition,
soft-bodied. Comparable growth interactions may take place between two skeletized taxa, to
produce skeletal intergrowths (e.g. between Palaeozoic stromatoporoids and corals — see Kershaw
1987; Mistiaen 1984) but we do not regard such interactions as examples of bioclaustration, even
if the infester is much smaller than the host and locally embedded in it, rather than inextricably
intergrown. An example of this situation is provided by tube-secreting worms that embed within
living coral and lengthen to keep pace with its growing surface, or by cornulitids that settled on
Silurian crinoid stems and became embedded by excessive calcite secretion (Franzen 1984).
Bioclaustration structures, in contrast, are trace fossils and are recognized only by the disturbance
caused to the growth of the host.

Bioclaustration is not to be confused with bioimmuration. The latter involves two sessile
organisms, one soft-bodied and one calcified, growing alongside one another. Crowding may result
in the skeletized neighbour overgrowing the other, and moulding its attachment surface over the
soft-bodied competitor. Bioimmuration thus demonstrates chance competition for space, not a
response to an interaction that is of one of the partner’s seeking.

EXAMPLES OF BIOCLAUSTRATION

Reports of bioclaustration in the fossil record are few and far between, but span the Phanerozoic.
Scrutton (1975) reported Jurassic, Cretaceous, and Tertiary serpulid worm tubes that claustrate
the stoloniferous hydroid Protulophila gestroi Rovereto. Scrutton speculated that the relationship
could have been of mutual advantage, the worm conferring both substrate and an increased supply
of suspended food particles, and the hydroid offering the protection of its nematocysts.

There is a more extensive literature on the formation of gall-like structures in echinoderms,
caused by an irritating infester leading to secretion of adventitious stereom in an attempt to isolate
the irritant. Franzen (1974) and Brett (1978) have reviewed such examples and added further data
on Silurian and Devonian crinoid infestation. Some examples demonstrate a response to encrustation
by shelly fossils (cornulitids, crinoids, bryozoans, forams) but others show only pits or cavities
within the swellings and appear to represent bioclaustration.

Bromley (1970, p. 50) has reviewed examples of embedment in the fossil record, and has noted
that some holes traditionally ascribed to borings show distributions and morphologies more in
keeping with bioclaustration structures. Chatterton (1975) has described bioclaustration by
Devonian spiriferids of a soft-bodied filter-feeder that settled on the growing valve margin and
extended its feeding crown into the inhalant feeding currents within the mantle cavity of the host.
The brachiopod responded by secreting a cylinder of shell material around it, now preserved as a
small, calcite, inwardly projecting chimney. This relationship did not involve penetration through
the shell by a borer that encountered protective secretion only when it broke through to the inside
of the shell and irritated the living mantle surface. There is an extensive literature on this latter
phenomenon, with many recent examples that have commercial implications in shell-fishery (see
references quoted in Boucot 1981). We regard such cases as modified borings, not examples of
bioclaustration, and the resultant traces can usually be ascribed to the same ichnotaxa as examples
of the same borings that do not break through the shell and irritate the host (Bromley 1970). The
results of bioclaustration, in contrast, constitute their own category of trace fossil and require their
own ichnotaxonomy. Incidentally, of course, the holes resulting from this type of embedment
accurately reflect the external shape of the infesting organism and may point to its zoological
affinities.

INTRASPECIFIC RELATIONSHIPS

The recognition of the precise nature of an interaction between species in the fossil record
is difficult. In living organisms, recognition of parasitic, as opposed to protoco-operative or

PALMER AND WILSON: PARASITISM OF ORDOVICIAN BRYOZOANS

941

mutualistic interactions, can be made by assessment of growth rates and population dynamics in
associated versus independent species (e.g. Osman and Haugsness 1981). Such options are not
open to palaeontologists, whose assessment of the cost-benefit analysis must be based on limited
observation tempered with common-sense. Subtle details of an interaction cannot be observed,
and should not be surmised if parsimony is to be maintained. However, some general principles
apply.

Any interaction that promotes a growth response in one of the parties is energetically costly.
An infestation that eliminates some of the members in a colonial organism further reduces food
intake and limits fecundity. In such cases, the infester is presumably advantaged because it is the
infester that initiates the contact. Such relationships should be regarded as being of a +/— nature,
and hence parasitic, unless the advantages conferred upon the host outweigh these disadvantages.
We might reasonably expect that examples of a particular pairwise interaction would be more
common if of mutual benefit, than if only to the advantage of one of the parties involved. This is
because selection may be expected to favour the attraction and conjoining of the two species
involved. The end evolutionary result of such cases is mutualism, in which the interaction is
obligatory for both parties. This is not to say that heavy infestation of a host by a parasite may
not occur in some host populations, but it is not unreasonable to infer that low levels of infestations,
in which there is clear evidence of morphologic damage to the host, are more likely to represent
examples of parasitism than a +/+ interaction. The association that forms the subject of this
paper is only seen in a few percent of the individuals of the species of bryozoan involved. The
advantages that accrued to those few individuals may have outweighed the disadvantages, but we
think it is much more likely that this was not so, and that we are dealing with a case of
ectoparasitism. Our vocabulary in the following section will reflect this belief.

INFESTATION OF CINCINNATIAN BRYOZOA
Borings and pseudoborings

The Upper Ordovician rocks that occur around Cincinnati in south-west Ohio, USA, and in the
adjacent parts of the neighbouring states of Indiana and Kentucky, consist of interbedded soft
silty and bioclastic limestones. Aragonitic taxa have been dissolved out in both lithologies, but
skeletons of original calcite are vitually unaltered. Amongst the calcitic groups, bryozoans weather
out of the sections in great profusion, and can be collected in large numbers. Many zoaria show
signs of boring by the worms that produced Trypanites, usually as post-mortem colonization. A
few specimens indicate infestation of the living bryozoan (as evidenced by a growth response of
the adjacent zooecia). Trypanites occurs as circular holes, up to 2 mm across, penetrating the
bryozoan skeleton. Many such holes may occur on a single fragment.

In the Kope Formation at the base of the Upper Ordovician sequence, bryozoans of the genera
Amplexopora and Peronopora contain a different structure that looks to the unwary eye like an
array of equispaced Trypanites that differ from the norm by the fact that their inner margins are
slightly crenulate to stellate. However, when a recently collected specimen of A. persimilis Nickles,
1905 (from Mr B. Bodenbender) was examined closely, the pits were seen to be part of a single
structure. This was suggested by their regular spacing (2-3 mm apart), and the fact that in the
outer parts of the array the pits define straight or gently curving lines, four or five pits long,
terminating in an elongate shallow groove. The integral nature of the pit array was confirmed by
sectioning, which revealed buried tunnels that join the bases of the adjacent pits in each line. The
crenulate margins of the pits are formed by the walls of the zooecia that surround them (PI. 87,
fig. 1). They may be somewhat thickened and, in well-preserved specimens, they are raised slightly
above the surface of the surrounding zoarium (text-fig. 1a). This feature suggests that the holes
are, at least in part, pseudoborings that represent reaction by the bryozoan. In contrast, Trypanites
that are inferred to have been excavated post mortem exhibit sharp truncation of the zooecia and
do not show raised rims (text-fig. 1b).

942

PALAEONTOLOGY, VOLUME 31

text-fig. 1. Difference between bioclaustration structures and borings in Ordovician trepostomes from the
Upper Ordovician, Kope Formation, near Cincinnati, Ohio, USA. a, pits of Catellocaula vallata ichnogen.
and ichnosp. nov., formed by bioclaustrating growth of host bryozoan, showing pit margin and slightly
thickened raised reaction rim; note that the zooecia adjacent to the pit are not truncated. USNM 419444,
x 13. b, borings ( Trypanites ), showing truncation of zooecia. USNM 419476, x 6.

text-fig. 2. Catellocaula vallata ichnogen. and ichnosp. nov. in Ordovician trepostomes from the Upper
Ordovician, Kope Formation, near Cincinnati, Ohio, USA. a, part of USNM 419444, in Amplexopora
persimilis , showing three lines of pits each terminating distally in a groove, x 2-9. b, part of USNM 419462,
in Peronopora sp., showing pits with crenulate margins, x 4-4.

Formation of the pseudoborings by bioclaustration

The pseudoborings consist of four interconnected elements: pits, grooves, galleries, and tunnels.
Pits and grooves are visible as holes or indentations in the exterior surface of the bryozoan (text-
fig. 2). Galleries and tunnels respectively represent pits and grooves that have been roofed over by
bryozoan growth, and are only seen in cut sections. The soft tissue of the parasite was originally
continuous throughout the four structures, each of which represents a unique combination of
interaction between bryozoan growth pattern and different parts of the parasite’s body.

The floors of all four of the structures that constitute the pseudoboring are located at the same
level within the bryozoan zoarium, and invariably mark a growth interruption. These interruptions
are usually interpreted as having been caused by local damage to the surface and cessation of
growth of the bryozoan, with rupture of exterior membranous colony walls (Boardman 1983,
p. 129). They are easily recognized by the thickened zooecial walls immediately below the interrup-

PALMER AND WILSON: PARASITISM OF ORDOVICIAN BRYOZOANS

943

text-fig. 3. Catellocau/a vallata ichnogen. and ichnosp. nov., showing grooves which terminate lines of pits.
A, USNM 419449. Bifurcating groove, x 4-6. b, USNM 419444, showing constriction (arrowed) where groove
walls roof over to isolate new pit proximally, x 1 1-5.

tion, upon which the pseudoboring sits. In adjacent regions, renewed growth of the zooecia above
the interruption shows up as a zone marked by thin zooecial walls that sit upon, and contrast
markedly with, the thick walls below (PI. 87, fig. 2). Whether the damage that initiates these intra-
colony overgrowth surfaces is external and merely exploited by the parasite, or whether the parasite
causes the damage in the first place, is discussed below.

Pits. Pits are c. 2 mm in diameter and c. 1-2 mm deep. Their floors are formed by the zoarial
surface below the intra-colony overgrowth surface, and their walls by the walls of the zooecia that
grew up around them and which become more thickened upwards. These walls are vertical, or
slightly divergent upwards as adjacent zooecia lean away from the pit centre (PI. 87, fig. 3).
Examples in Amplexopora show that the walls are slightly fluted vertically and, when well preserved,
rise just above the surface of the surrounding zoarium to form a reaction rim (text-fig. 1 a). The
fluting on the walls gives this rim a crenulated rather than a perfectly circular outline. Examples
in Peronopora show much better development of the crenulations, so that the pits become stellate
(text-fig. 2b).

In the older (more central) parts of mature colonies, the pits are more or less equispaced, their
centres 2-3 mm apart. Towards the periphery they line up in straight or gently curving rows, two
to five pits long (text-fig. 2). The rows usually terminate in grooves. The older pits in the central
parts tend to be deeper than those towards the edges as a result of upward growth of the
surrounding zooecia.

Grooves. Grooves have the same width as pits, but are up to c. 10 mm long (text-figs. 2a and 3).
They are straight to gently curving and may bifurcate. They are deepest at their proximal ends (c.
1 mm) and shallow distally, becoming flush with the exterior surface of the zoarium. Where deepest,
their walls are thickened to produce a reaction rim. Grooves develop into lines of pits by localized
ingrowth of bryozoan zooecia on either side. These ingrowths roof over the top of the groove,
meeting one another and thereby isolating the proximal end of the groove as a new pit. Some
grooves show the start of this process as a constriction c. 2 mm from the proximal end (text-fig.
3b).

Tunnels. Tunnels join adjacent pits along a single line, and their floors lie at the same level along
an intra-colony overgrowth surface as the pit floors. They developed by local encroachment of the
bryozoan across grooves, thus pinching off new pits that retained a soft-tissue connection with
the truncated groove (and with each other) via the tunnels (PI. 87, fig. 4). Some pits have such
connections with two distal neighbours and represent overgrowth of a bifurcating groove. This
method of tunnel formation is clearly critical to the recognition that these are bioclaustration
structures rather than borings. Sections through tunnels show that the roofs are formed by oblique
lateral walls of adjacent zooecia which spread out by lateral budding from those adjacent to the

944

PALAEONTOLOGY, VOLUME 31

groove margin. They are not truncated (PI. 87, fig. 5). Although such ingrowth seems to take place
from both sides of the groove simultaneously, no obvious suture is formed where the two sides
meet.

Tunnels are often filled with mud matrix, as are pits and some of the zooecial living chambers
immediately below the intra-colony overgrowth surfaces within the bryozoan. However, a few
tunnels seem to have become occluded by a meshwork of curved diaphragms which together form
a plug (PI. 87, fig. 6). The plugs kept out mud, and the spaces between the diaphragms are now
filled with large equant crystals of calcite spar. We discuss the origin of these tunnel diaphragms
below.

Galleries. A few specimes show pits that have been roofed over by an encroachment process similar
to that which gives rise to the tunnels. Such galleries are encountered in sections, or rarely
indentified on the surface where their roofs of delicate, oblique zooecia have been crushed and
impressed into the underlying space.

Host-parasite interactions

Pseudoboring lining and tunnel diaphragms. The lining of the pseudoboring is marked by a thin
membrane that is continuous over the inner surface of the tunnels and pits in well-preserved
specimens (PI. 87, fig. 7). We conclude that it was originally present in grooves and galleries as
well, but we have not seen enough sections through unabraded examples of these structures to be
sure. The membrane marks the original outer surface of the parasite. It is thin (considerably
thinner than zooecial walls and diaphragms) and, in acetate peels, shows minute irregular brown
blobs along its length that are probably remnant organic material or oxidized pyrite. It drapes
over the upstanding walls of the zooecia beneath (PI. 87, figs. 7 and 8). Where stretched across
the apertures of these zooecia, it appears to have prevented access of sediment into their lumina.
Sediment-filled zooecia have lost this coating membrane. On tunnel roofs the membrane lies against
the outside of the oblique walls of the overlying zooecia.

Within tunnels the curving diaphragms of the tunnel plugs insert on to the surface of the
membrane (PI. 87, fig. 8) and appear to be made of the same material. The diagenetic calcite that
fills the spaces within the plug consist of one or a few large crystals, rather than drusy calcite
which typically fills the spaces within the bryozoan zoarium. Drusy texture is controlled by the
presence of seed crystals in the walls, upon which the precipitating cement can initiate. Absence
of this texture within the plugs suggests that the diaphragms, unlike the bryozoan skeletal tissue,
are not of an original calcite composition.

EXPLANATION OF PLATE 87

Figs. 1-8. Catellocaula vallata ichnogen. and ichnosp. nov. USNM 419461. Acetate peels of cut and polished
surfaces through Amplexopora persimilis to show relationships between the bryozoan skeleton and the
parasite, Newport Shopping Center, Ky., USA, Kope Formation, Edenian. 1, tangential section through
pit, showing that pit wall is formed by zooecial walls (arrowed), x 48. 2, longitudinal vertical section
through tunnel, showing that tunnel floor is not bioerosive, but sits upon thickened zooecial walls along
an intra-colony overgrowth surface, x 108. 3, transverse vertical section through zoarium between two
pits (upper right and left) showing deflection of growth of adjacent zooecia away from pits, x48. 4,
longitudinal vertical section through tunnel between adjacent pits along a line; right-hand pit is filled with
dark sediment, x 46. 5, close-up of fig. 4 showing that roof of tunnel is formed by walls of oblique zooecia
(arrowed) that overgrow from the sides, x 112. 6, longitudinal tangential section through tunnel showing
curved diaphragms of the tunnel plug, x48. 7, oblique vertical section through tunnel showing dark

bounding membrane (arrowed) draping over upstanding zooecial walls on tunnel floor, and overgrown by
oblique zooecia of tunnel roof, x 105. 8, close-up of fig. 6 showing thin dark tissue of tunnel diaphragm
(upper arrow) joining bounding membrane that lines tunnel wall (lower arrow), x 134.

PLATE 87

PALMER and WILSON, Catellocaula

946

PALAEONTOLOGY, VOLUME 31

There are two possible origins for this organic membrane. It may represent remnants of the
cuticle of the bryozoan. Trepostome cuticular appearance is poorly documented, but similar
structures have been described and illustrated by Boardman (1973). If it is of bryozoan origin,
then it might also be expected to be visible over the external surface of the zoarium, or covering
zooecia that have mounded up around the mouths of Trypanites borings that were excavated while
the colony was still alive. Our research for the membrane in these circumstances, though not
exhaustive, has been unsuccessful. The alternative is that the membrane was laid down as an outer
integument by the outer surface of the parasite. The use of a meshwork of membrane material to
form the tunnel plugs seems to us somewhat more in keeping with this second explanation. The
tunnels originally contained stolon tissue that was important in vegetative growth and reproduction
of the parasite (see below), but not necessary for everyday function. Once their purpose had been
achieved, they could be infilled.

Settlement and growth of the parasite. Initial infestation of the bryozoan surface occurred at a
single point. Some examples show that the bryozoan could respond rapidly, resulting in a single
pit. There is some indication that these single pits coincide with the position of maculae on the
bryozoan surface, but we have not seen enough unequivocal examples to be sure. These cases
showing an immediate response of the bryozoan, claustrating the newly settled parasite before it
had time to grow, suggest that the adjacent zooecia were alive and able to respond rapidly. This
supports a contention that the parasite settled on a live area of the zooecium, and gained access
to the host by its own activities. If it had settled on a larger expanse of dead zoarium, claustration
could not have commenced until living bryozoan tissue had invaded from the edges of the dead
area, and the distal parts of the infester are likely to have had time to grow and to have been
claustrated before the central part.

Having become established, the parasite sent out ribbon-like stolons of adventitious tissue
radially in several (usually three or four) directions. As these grew distally, away from the ancestral
pit, they came to lie in grooves as bryozoan zooecia grew up to flank them. Proximal ends of
grooves are deeper than distal ends because the parasite tissue within them was older, and more
zooecial growth had occurred around them. As the grooves elongated distally, their proximal ends
became pinched off into pits in the manner described above. This pattern of deeper claustration
in the older, more central parts of the infestation, becoming shallower outwards in all directions,
is critical to support the contention that infestation took place on a live zoarium and that adjacent
zooids were immediately stimulated to claustrate. As the stolons radiated, they overgrew and killed
zooids in their path, whilst adjacent ones were stimulated to grow up around the invader. The
pattern of stimulation to claustrate therefore proceeded centrifugally. But if overgrowth were
effected by an advancing wall of zooecia proceeding inwards from the perimeter or from one side
of a damaged area, then the pattern of claustration would have proceeded centripetally or sideways
across the infester.

As the initial stolons radiated and diverged from the centre, they branched so as to utilize space
efficiently. Thus grooves and lines of pits also branch. In mature infestations, individual stolons
can only be distinguished round the edge. The centre appears to be a mass of equispaced pits, but
only those laid down on the same stolon are connected by tunnels.

SYSTEMATIC PALAEONTOLOGY

The arrays of holes that we describe here are the result of modification of the growth pattern of
a bryozoan by the presence of a soft-bodied parasite. We choose to regard such bioclaustration
structures as trace fossils because others have done so before us (Bromley 1970; Muller 1962), and
because they may share similarities and intergrade with borings. However, they differ from the
popular perception of trace fossils as indicators of animal behaviour. Other dwelling-structures
require work to have been perpetrated by the constructor in the form of boring or burrowing
activity. Bioclaustration structures result from the mere existence of the infester, coupled with
modification of the growth behaviour of the host. The end result is likely to mimic accurately the

PALMER AND WILSON: PARASITISM OF ORDOVICIAN BRYOZOANS

947

text-fig. 4. Camera lucida sketch of pits and grooves of holotype
of Catellocaula vallata ichnogen. and ichnosp. nov. in external surface
of Amplexopora persimilis. USNM 419443, scale in mm.

^ if

external shape of the parasite, and could be thought of as a biologically formed external mould
of its body. Many borings also have this property (clionid sponges, acrothoracican barnacles,
clenostome bryozoans, thallophytes), but are now regarded as trace fossils (Bromley 1970).

Ichnogenus catellocaula nov.

Type species. Catellocaula vallata ichnosp. nov.

Derivation of name. Latin: catella = little chain; caula = hole.

Diagnosis. Bioclaustration structure in bryozoans, consisting of a group of pits sunk into the
surface of the zoarium. Pits c. 2 mm diameter, up to c. 2 mm deep; in plan view pit mouth
subcircular to oval with slightly to strongly fluted edges; pit walls may extend up above bryozoan
surface to form low thickened rim around pit mouth. Mature specimens consist of arrays of up
to thirty or more such pits; in centre of array, pits spaced evenly, c. 2-3 mm apart; towards
periphery, pits lie equispaced along straight or gently curving lines, each often terminating in a
groove, c. 2 mm wide, several millimetres long; groove shallows distally so that outer end merges
imperceptibly with surface of surrounding zoarium. Floors of adjacent pits along line joined by
tunnels, c. 2 mm wide, 0-5 mm high. Lines increase in number by bifurcation.

Catellocaula vallata ichnosp. nov.

Plate 87; text-figs. 1 4

Type material. Holotype: USNM 419443 (text-fig. 4); paratypes: USNM 419444 419462. Number prefix
USNM refers to collections of United States National Museum, Smithsonian Institution, Washington DC,
where all material is housed.

Additional material. Probable additional examples of C. vallata occur in poorly silicified ? Peronopora (which
cannot therefore be sectioned to confirm the identification), from the Clays Ferry Formation near Lexington,
Kentucky (USNM 419463-419473).

Type locality. Original label on the holotype states it was collected from the ‘Eden (McMicken)’ of Newport,
Kentucky, USA. This is equivalent in modern nomenclature to the upper part of the Kope Formation (Weir
et al. 1984).

Derivation of name. Latin: vallatum = surrounded with a rampart.

Occurrence. Kope Formation (Edenian = Caradocian, Upper Ordovician); widespread in the vicinity of
Cincinnati, Ohio, USA.

Diagnosis. As for genus.

948

PALAEONTOLOGY, VOLUME 31

ZOOLOGICAL INTERPRETATION

The soft-bodied organism that provoked the bioclaustration response of the trepostome bryozoans
was a sessile, stoloniferous, colonial form. The scalloped margin of the pits may indicate that the
larger portions of the colony were lobed. The colony could apparently survive the effects of partial
envelopment by the bryozoan zooecia, and it is found on all sides of erect zoaria, so it was probably
not photosynthetic.

Two Recent groups of organisms may provide models for a palaeobiological reconstruction of
this bryozoan parasite. Hydroids (Phylum Cnidaria) sometimes produce horizontal, root-like
stolons, termed hydrorhizae, from which arise single upright polyps or branches of polyps. Most
colonial hydroid stolons are covered by a non-living chitinous envelope called the perisarc (Barnes
1987), but are much smaller than those described here. Ascidiacian tunicates (Subphylum
Urochordata) also include stoloniferous colonial forms most notably the living genus Perophora.
These tunicates are covered by a cellulose-rich tissue (the tunic). The scalloped pit margins of C.
Valletta strongly evoke the image of compound ascidiacians, especially the living Botryllus (see
Abbott and Newberry 1980) and the fossil Palaeobotryllus taylori, preserved as phosphatic
microfossils in the Upper Cambrian of Nevada (Muller 1977). Both shape and size of these forms
correspond to the pseudoborings we describe here, and we favour a tunicate origin for C. vallata.

Acknowledgements. We thank Brian Bodenbender for field assistance, and Fred Collier and Dr John Pojeta,
Jr., for help in obtaining and cataloguing specimens. We especially thank the administrators of the College
of Wooster for the hospitality and support given to T. J.P. during a 1987 summer visit.

REFERENCES

abbott, d. p. and newberry, a. t. 1980. Urochordata: the tunicates. In morris, r. h., abbott, d. p. and
haderlie, E. c. (eds.). Intertidal invertebrates of California , 177-226, Stanford University Press, California.
barnes, R. d. 1987. Invertebrate zoology (5th edn.), 893 pp. Holt, Rinehart and Winston, New York.
boardman, r. s. 1973. Body walls and attachment organs in some Recent Cyclostomes and Paleozoic
Trepostomes. In larwood, g. p. (ed.). Living and fossil Bryozoa, 231-246. Academic Press, London.

— 1983. General features of the Class Stenolemata. In robison, r. a. (ed.). Treatise on invertebrate
paleontology. Part G. Bryozoa (revised), G49-G137. Geological Society of America and University of
Kansas Press, Boulder, Colorado, and Lawrence, Kansas.
boucot, a. J. 1981. Principles of benthic marine paleoecology, 463 pp. Academic Press, New York, London.
brett, c. e. 1978. Host specific pit-forming epizoans on Silurian crinoids. Lethaia, 11, 17-232.
bromley, r. G. 1970. Borings as trace fossils and Entobia cretacea as an example, 49-90. In crimes, t. p. and
harper, j. c. (eds.). Trace fossils. Geol. J. Spec. Issue , 3, 1-547.
chatterton, b. d. e. 1975. A commensal relationship between a small filter-feeding organism and Australian
spiriferid brachiopods. Paleobiology, 1, 371-378.
conway morris, s. 1980. Parasites and the fossil record. Parasitology , 82, 489-509.

ekdale, A. a., bromley, r. G. and Pemberton, s. G. 1984. Ichnology: the use of trace fossils in sedimentology
and stratigraphy. Soc. econ. Paleont. Mineral., Short Course , 15, 1-317.
franzen, c. 1984. Epizoans on Silurian-Devonian crinoids. Lethaia, 10, 287-301.
kershaw, s. 1987. Stromatoporoid-coral intergrowths in a Silurian biostrome. Ibid., 20, 371-380.
kobluk, d. r., james, n. p. and Pemberton, s. G. 1978. Initial diversification of macroboring ichnofossils and
exploitation of the macroboring niche in the Lower Paleozoic. Paleobiology, 4, 163-170.
misti aen, N. 1984. Comments on the caunopore tubes: stratigraphic distribution and microstructure.
Palaeontogr. am. 54, 501-508.

muller, A. H. 1962. Zur Ichnologie, Taxiologie and Okologie fossiler Tiere. Freiberger ForschHft. 151, 5-49.
muller, k. j. 1977. Palaeobotryllus from the Upper Cambrian of Nevada— a probable ascidian. Lethaia , 10,
107-118.

nickles, j. m. 1905. The Upper Ordovician rocks of Kentucky and their Bryozoa. Bull. Ky geol. Surv. 5, 1-
64.

osgood, r. G. 1970. Trace fossils of the Cincinnati area. Palaeontogr. am. 6, 280 444.

PALMER AND WILSON: PARASITISM OF ORDOVICIAN BRYOZOANS

949

osman, r. w. and haugsness, j. a. 1981. Mutualism among sessile invertebrates: a mediator of competition
and predation. Science, NY, 211, 846-848.

palmer, t. j. 1982. Cambrian to Cretaceous changes in hardground communities. Letliaia, 15, 309 323.
pojeta, J., jr. and palmer, t. j. 1976. The origin of rock boring in mytilacean pelecypods. Alcheringa, 1,

scrutton, c. t., 1975. Hydroid-serpulid symbiosis in the Mesozoic and Tertiary. Palaeontology, 18, 255-274.
weir, g. w., peterson, w. l. and swadley, w. c. 1984. Lithostratigraphy of Upper Ordovician strata exposed
in Kentucky. US. geol. Surv. Prof. Pap. 1151 -E, 1-121.
wilson, m. a. 1985. Disturbance and ecologic succession in an Upper Ordovician cobble-dwelling hardground
fauna. Science, NY, 228, 575-577.

— and palmer, t. j. 1988. Nomenclature of a bivalve boring from the Upper Ordovician of the mid-
western United States. J. Paleont. 62, 306-308.

167-179.

T. J. PALMER

Department of Geology
University College of Wales

Aberystwyth

Dyfed SY23 2DB

Wales, UK

M. A. WILSON

Typescript received 2 July 1987

Revised typescript received 10 January 1988

Department of Geology
College of Wooster
Wooster

Ohio 44691, USA

A WEIGELTISAURID REPTILE FROM THE
LOWER TRIASSIC OF BRITISH COLUMBIA

by DONALD BRINKMAN

Abstract. The skull of a new weigeltisaurid reptile, Wapitisaurus problematicus gen. et sp. nov., from the
Lower Triassic Vega-Phoroso Member of the Sulphur Mountain Formation is described. It shares with
Coelurosauravus, the only other known weigeltisaurid, the presence of an incomplete lower temporal arcade,
a jugal with reduced postorbital process, and a squamosal crest ornamented with tooth-like projections. It
differs from Coelurosauravus in its large size and in the structure and implantation of the teeth.

While marine reptile faunas of the Upper Permian and Middle Triassic are well known, those of
the Lower Triassic are very incompletely understood, so that the discovery of marine reptiles in the
Lower Triassic Vega-Phoroso Member of the Sulphur Mountain Formation greatly increases our
understanding. These beds have been known for their abundant and well-preserved vertebrate fauna
since 1949 (Laudon et al. 1949). The fish fauna has been described by Schaeffer and Mangus (1976)
and Neuman (1986), and reptile remains were noted by Schaeffer and Mangus, but only recently
has diagnostic material been collected. This includes ichthyosaur remains, currently being studied
by J. Callaway and D. Brinkman, and the skull of a peculiar reptile described here.

Geological occurrence

The Vega-Phoroso Member of the Sulphur Mountain Formation (Gibson 1972, 1975) consists of
flaggy weathering shale at its base that intertongues with, and is overlain by, a sequence of rusty
brown siltstones. This member is interpreted as having been deposited in a restricted, relatively
deep-water environment, although some evidence indicates that at times deposition may have been
above active wave base (Gibson 1975). The Vega-Phoroso Member is entirely Lower Triassic in
age. It ranges from the Griesbachian to the Spathian with most collections being dated as Smithian
largely on the basis of pelecypods. The specimen described here was found in a scree slope derived
from the siltstone facies of the Member, but the position of the exposure from which the scree
originated could not be determined. Thus the exact age of the specimen is uncertain, although a
Smithian age is likely.

SYSTEMATIC PALAEONTOLOGY

Class REPTILIA
Subclass DIAPSIDA

Family weigeltisauridae Kuhn, 1939
Genus wapitisaurus gen. nov.

Type species. Wapitisaurus problematicus sp. nov.

Etymology. Refers to Wapiti Lake, a large lake about 4 km north of the type locality.

Diagnosis. Differs from Coelurosauravus in its large size, subthecodont tooth implantation, and
presence of few, short, laterally compressed teeth that are about as wide at their base as they are
high. The postcranial skeleton is unknown.

| Palaeontology, Vol. 31, Part 4, 1988, pp. 951 955. |

© The Palaeontological Association

952

PALAEONTOLOGY, VOLUME 31

sq

text-fig. 1. The type specimen of Wapitisaurus problematicus gen. et sp. nov. (TMP 86.153.14). Abbreviations:
fr, frontal; ju, jugal; pa, parietal; po fr, postfrontal; po o, postorbital; pt, pterygoid; sq, squamosal; st,

supratemporal.

Wapitisaurus problematicus sp. nov.

Text-fig. 1

Etymology. Named for the taxonomic and anatomical problems raised by the type specimen.

Holotype. Tyrrell Museum of Palaeontology, specimen number TMP 86.1 53. 14. Partial skull seen in left lateral
view, lacking the maxilla and premaxilla. The left pterygoid and left lower jaw are preserved below the skull.

Horizon and locality. From the Vega-Phoroso Member of the Sulphur Mountain Formation. Type locality:
UTM 647,000 E., 6045000 N., Zone 10, map 93 1/10. Near Wapiti lake, British Columbia, Canada.

Specific diagnosis. As for the genus.

DESCRIPTION

The general proportions of the skull (text-fig. 1 ) are clear: the orbit is large, the postorbital region is slightly
shorter than the diameter of the orbit, and the ventral margin of the skull sweeps upward posterior to the
orbit. In general, these proportions are similar to those of Coelurosauravus as reconstructed by Evans and
Flaubold (1987), although the postorbital region is shorter relative to the length of the orbit than in that
genus.

The postorbital region is nearly completely preserved on the left side of the skull and the squamosal,
postorbital, postfrontal, frontal, and jugal remain in articulation. A fragment of bone preserved in the position
of the parietal may represent a part of that element. Most of these bones are represented by impressions of
the internal surface of the bones or by broken bone surface, but part of the external surface of the postorbital
and squamosal is preserved. An element with tooth-like ornamentation is visible within the upper temporal
opening. This is either the right squamosal or a supratemporal.

BRINKMAN: LOWER TRIASSIC WEIGELTI S AU R I D

953

The arrangement of the bones forming the postorbital region is much like that of Coelurosauravus. The
frontal forms much of the orbital margin. The postfrontal is a small crescent-shaped bone extending along
the margin of the orbit between the frontal and postorbital. The postorbital forms the posterior margin of the
orbit and contacts the squamosal and jugal ventrally. The posterior edge of the postorbital is incompletely
preserved, but it must have been large and generally triangular in shape. The squamosal forms the ventral
margin of the postorbital region of the skull. As in Coelurosauravus, it sweeps upwards from the ventral edge
of the orbit giving the postorbital region a crest-like aspect. Also, as in Coelurosauravus , the ventral edge of
the squamosal is ornamented by small, irregular tooth-like projections, of which six are present on the
preserved part of the bone. The ornamentations on the element visible through the upper temporal opening
do not match those of the left squamosal. Thus this element may represent a supratemporal which Evans and
Haubold (1987) have shown to be ornamented also in Coelurosauravus. The posterior end of the jugal is
preserved, and the contact of the jugal with the postorbital can be identified, but the contact with the squamosal
is obscured. The jugal extends anteriorly from the postorbital as a narrow bar below the orbit. A posterior
process is not present.

Impressions of some of the bones of the face are present. These show that the eye was large and bordered
anteriorly by a thickened ridge. The identity of the bones in this area and the position of sutures is, however,
uncertain.

A left pterygoid is preserved below the skull. Numerous conical, recurved teeth are present on the transverse
flange region of the bone. They increase in size towards the lateral edge of the bone. They are not organized
into distinct rows or tooth patches, but form a uniform covering over the entire surface of the preserved
portion of the transverse flange.

Most of the left lower jaw is present, only the tip of the dentary and the lower edge of the postdentary
being missing. The dentary is represented by impression and by broken bone surface, and the postdentary
region by impression and by the lateral surface of its posterior end. The sutural contact between the dentary
and postdentary regions can be clearly identified. No sutures can be recognized in the postdentary region.

The dentary is a rather slender bone bearing ten teeth. The posterior teeth are nearly completely preserved.
These are broad-based and laterally compressed, about as high as they are wide, and with sharp conical tips.
The anterior teeth are represented by impressions in the matrix. They are shorter, smaller, and tend to be
more conical. The base of the teeth extends into the body of the bone, at least in the case of the most posterior
two teeth, indicating that tooth implantation is subthecodont. The most posterior tooth is located well anterior
to the posterior end of the dentary.

The dorsal margin of the postdentary region sweeps upward, corresponding to the upward sweep of the
ventral edge of the postorbital region of the skull. The preserved portion of the lower jaw extends to the
region where an articular would be expected. A swelling of the bone in this area may represent the lateral
expression of the articular. If correctly identified, this indicates that the jaw joint was located relatively further
posteriorly than in Coelurosauravus , which Evans and Haubold (1987) have shown to be located just posterior
to the orbit.

RELATIONSHIPS

The Weigeltisauridae (Coelurosauravidae of Evans, 1982), most recently reviewed by Carroll ( 1978),
Evans (1982), and Evans and Haubold (1987), are a family of small lizard-like primitive diapsids
represented by one genus, Coelurosauravus , from the Upper Permian of Europe and Madagascar.
It has a number of derived features of the cranial and postcranial skeleton, the most striking of
which is the elongation of the ribs to form a gliding structure. Derived features of the skull listed
by Evans (1982) and Evans and Haubold (1987) are: pleurodont or subpleurodont dentition,
ornamented squamosal and supratemporal, long straight postparietal processes, incomplete lower
temporal arcade, and jugal with reduced posterior process. Wapitisaurus shares with Coelurosaura-
vus the following traits: ornamented squamosal, incomplete lower temporal arcade, jugal with
reduced posterior process, and lacrimal small or absent. In addition, the proportions of the skull of
Wapitisaurus are similar to those of Coelurosauravus'. the orbit is large and the ventral margin of
the postorbital region slopes upward from the ventral margin of the orbit.

There are, however, a number of features in which Wapitisaurus is different from Coelurosauravus
which bring this assignment into question. One of these is its large size. The kind of gliding
adaptations seen in Coelurosauravus may well have an upper size limit, raising the possibility that

954

PALAEONTOLOGY, VOLUME 31

Wapitisaurus did not have similar adaptations. However, by analogy with agamids, the absence of
gliding adaptations in Wapitisaurus would not prevent these two genera being considered members
of a single family, since the Agamidae contains genera that are gliding and genera that have a
normally constructed postcranial skeleton.

A second difference between Wapitisaurus and Coelurosauravus is in the structure of the teeth.
Those of Coelurosauravus are small, numerous, and conical, presumably a primitive condition, and,
as interpreted by Evans and Haubold (1987), have a pleurodont or subpleurodont implantation.
Those of Wapitisaurus have a subthecodont implantation and are derived in being few in number,
and in that the posterior teeth are stoutly constructed.

Teeth like those of Wapitisaurus are also seen in two groups of marine reptiles from the Triassic,
the Thalattosauria and the Ichthyopterygia. Thus an alternative to the hypothesis that Wapitisaurus
is related to Coelurosauravus is that it is a member of one of these groups.

The Thalattosauria is a group known from the Middle Triassic (Merriam 1905; Peyer 1936;
Rieppel 1987). They differ from Wapitisaurus and Coelurosauravus in the structure of the postorbital
region of the skull. In the thalattosaurs, the upper temporal opening has been reduced or lost, and
a large lower temporal opening is present (Rieppel 1987). This contrasts with the condition in
Wapitisaurus and Coelurosauravus where the lower temporal opening has been lost and the postorbi-
tal region is relatively short. Thus the hypothesis that Wapitisaurus and thalattosaurs are related is
not corroborated by other features in the structure of the skull.

The second group of Triassic marine reptiles that have a dental arrangement like that of Wapiti-
saurus are the ichthyosaurs. Primitive ichthyosaurs such as Grippa (Mazin 1981) are similar to
Wapitisaurus in that the posterior teeth are blunt, crushing teeth and the anterior teeth are conical.
Wapitisaurus, Coelurosauravus, and primitive ichthyosaurs are also similar in that the orbit is large,
the lower temporal bar has been lost, the jugal is without a posterior process, and the cheek
region has been shortened. Using primitive diapsids such as Petrolacosaurus (Reisz 1981) and
Acerodontosaurus (Currie 1980) as outgroups, these can be interpreted as derived features. However,
the postorbital region of the skulls of Wapitisaurus and Coelurosauravus is very different from that
of ichthyosaurs. In Coelurosauravus the quadratojugal is small and the supratemporal is a large
element located behind the upper temporal opening. In ichthyosaurs the quadratojugal is large, the
squamosal forms the posterior border of the upper temporal opening, and a supratemporal is absent
(Romer 1968; McGowan 1973). Assuming that the homologies of the temporal bones are correctly
interpreted, a phylogenetic relationship between ichthyosaurs and Coelurosauravus is unlikely.
Wapitisaurus, as interpreted here, is similar to Coelurosauravus in preserved portions of the
postorbital region, so the similarities in the structure of the teeth of Wapitisaurus and ichthyosaurs
are best interpreted as parallel developments. Thus at present, a relationship between Coelurosaura-
vus and Wapitisaurus is considered the most strongly supported hypothesis of relationships.

Acknowledgements. I thank Drs S. E. Evans, R. L. Carroll, H. Sues, and P. J. Currie, who read earlier drafts
of this paper and made many comments leading to its improvement. Text-fig. 1 was drawn by Donna Sloan
of the Tyrrell Museum. Able field assistance leading to the discovery of this specimen was provided by Paul
Neilsen and Avis Schelski.

REFERENCES

carroll, r. l. 1978. Permo-Triassic ‘lizards’ from the Karoo system. Part 2: A gliding reptile from the Upper
Permian of Madagascar. Palaeont. afr. 21, 143 159.

currie, p. j. 1980. A new younginid (Reptilia: Eosuchia) from the Upper Permian of Madagascar. Can. J.
Earth Sci. 17, 500-511. '

evans, s. e. 1982. The gliding reptiles of the Upper Permian. Zool. J. Linn. Soc. 76, 97-123.

— and haubold, H. 1987. A review of the Upper Permian genera Coelurosauravus, Weigeltisaurus and
Gracilisaurus (Reptilia: Diapsida). Ibid. 90, 275-303.

gibson, d. w. 1972. Triassic stratigraphy of the Pine Pass-Smoky River area. Rocky Mountain foothills and
front ranges of British Columbia and Alberta. Pap. geol. Surv. Can. 71-30, 108 pp.

BRINKMAN: LOWER TRIASSIC WEIGELTISAU RID

955

1975. Triassic rocks of the Rocky Mountain foothills and front ranges of northeastern British Columbia
and west-central Alberta. Bull. geol. Surv. Can. 247, 1-61.
kuhn, o. 1939. Schadelbau und systematische Stellung von Weigeltisaurus. Paldont. Z. 21, 163 167.

LAUDON, L. R., DEIDRICK, E., GREY, E., HAMILTON, W. B., LEWIS, P. J., McBEE, W., SPRENG, A. C. and STONEBURNER,
r. 1949. Devonian and Mississippian stratigraphy. Wapiti Lake area, British Columbia, Canada. Bull. Am.
Ass. petrol. Geol. 33, 1502-1552.

McGowan, c. 1973. The cranial morphology of the lower Liassic latipinnate ichthyosaurs of England. Bull.
Br. Mus. nat. Hist. (Geol.), 24, 1 109.

mazin, j. m. 1981. Grippia longirostris Wiman, 1929, un Ichthyopterygia primitif du Trias inferieur du
Spitsberg. Bull. Mus. natn. Hist, nat., Paris, (4) 3, (c), 317-340.
merriam, j. c. 1905. The Thalattosauria, a group of marine reptiles from the Triassic of California. Mem.
Calif. Acad. Sci. 5, 1-38.

neuman, a. 1986. Fossil fishes of the families Perleididae and Parasemionotidae from the Lower Triassic
Sulphur Mountain Formation of western Canada. M.Sc. thesis (unpublished). University of Alberta.
peyer, b. 1936. Die Triasfauna der Tessiner Kalkalpen. X. Clarazia schinzi nov. gen. nov. spec. Abh. schweiz.
paldont. Ges. 57, 1 61.

rieppel, o. 1987. Clarazia and Hescheleria: A re-investigation of two problematical reptiles from the Middle
Triassic of Monte San Giorgio (Switzerland). Palaeontographica A, 195, 101-129.
reisz, r. 1981. A diapsid reptile from the Pennsylvanian of Kansas. Univ. Kansas Spec. Pubis. Mus. nat. Hist.
7, 1-74.

romer, a. s. 1968. An ichthyosaur skull from the Cretaceous of Wyoming. Contr. Geology, 7, 27 -41.
Schaeffer, b. and mangus, m. 1976. An early Triassic fish assemblage from British Columbia. Bull. Am. Mus.
nat. Hist. 156, 519 563.

DONALD BRINKMAN

Typescript received 4 August 1987
Revised typescript received 4 January 1988

Tyrrell Museum of Palaeontology
Box 7500, Drumheller
Alberta T0J 0Y0, Canada

THE UPPER PERMIAN REPTILE ADELOSAURUS

FROM DURHAM

by SUSAN E. EVANS

Abstract. The Upper Permian reptile Adelosaurus from the Marl Slate of Durham, England, is redescribed
and compared with contemporary genera. The study confirms Watson’s (1914) conclusion that Adelosaurus is
generically distinct from Protorosaurus to which it was originally referred. The skeleton seems immature, and
shows a combination of primitive and derived character states. Amongst the latter, are the possession of a
strong humerus with little proximal or distal expansion, and of a slender sigmoidal femur and triangular
ilium, character states shared with diapsids. In the absence of the skull and ankle, however, this classification
remains tentative. Adelosaurus adds a fifth, probably terrestrial, component to the Kupferschiefer/Marl Slate
reptilian assemblage which currently includes a glider, Coelurosauravus , the long-necked, perhaps semi-aquatic,
Protorosaurus and, from German deposits only, a parieasaur, and the enigmatic Nothosauravus.

In the last decade, there has been a resurgence of interest in early diapsid reptiles, particularly with
respect to their phylogenetic relationships. The earliest known diapsid, Petrolacosaurus has been
shown to have affinities both to protorothyrid captorhinomorphs (Reisz 1981; Heaton and Reisz
1986) and to the enigmatic A raeoscelis (Reisz el al. 1984). Together, Petrolacosaurus and Araeoscelis
form the diapsid group Araeoscelidia. Most of our information about these diapsids comes from
Upper Carboniferous and Lower Permian deposits in northern Pangaea, while the bulk of our
knowledge of Upper Permian diapsids, amongst which the ancestors of Mesozoic and Cenozoic
groups are usually sought, is from southern Pangaea— most notably from deposits in Madagascar
and South Africa. Relatively little is known of contemporary diapsid faunas in northern Pangaea.
However, the Kupferschiefer/Marl Slate deposits of northern Germany and Britain provide at least
a weak link between the northern and southern faunas. The deposits have yielded a number of
specimens of Protorosaurus , a long-necked reptile related to Prolacerta (Lower Triassic, South
Africa and Antarctica), and of the glider Coelurosauravus ( = Weigeltisaurus = Gracilisaurus , Evans
and Haubold 1987) which has also been found in Madagascar (Carroll 1978). Haubold and Schaum-
berg (1985), reviewing the Kupferschiefer fauna, also note the presence of a pareiasaur, Parasaurus ,
and Nothosauravus which they tentatively link to the aquatic diapsid Claudiosaurus (Upper Permian,
Madagascar).

In 1870, Hancock and Howse described a small skeleton from the Marl Slate of Middridge,
Durham. They compared it with known examples of Protorosaurus speneri and concluded that the
new find was congeneric with Protorosaurus. The small size of the specimen, in addition to differences
in rib structure and limb proportions, led Hancock and Howse to erect a new species, P. huxleyi.
Watson (1914), however, noted differences between P. huxleyi and other specimens of Protorosaurus.
Most notable were the proportions of the cervical vertebrae— short in P. huxleyi and elongate in P.
speneri. On this basis, he created a new genus, Adelosaurus, for the P. huxleyi specimen, but left its
taxonomic position unresolved. Huene (1956) and Kuhn (19696) referred Adelosaurus to Broomi-
idae, and Romer (1966) to either Younginiformes or Protorosauridae; Vaughn (1955) left it incertae
sedis. Haubold and Schaumberg (1985) list P. huxleyi as a junior synonym of P. speneri and omit
any mention of Adelosaurus.

(Palaeontology, Vol. 31, Part 4, 1988, pp. 957-964J

© The Palaeontological Association

958

PALAEONTOLOGY, VOLUME 31

SYSTEMATIC PALAEONTOLOGY

Class REPTILIA
?Subclass DIAPSIDA
Genus adelosaurus Watson 1914

Type species. Adelosaurus huxleyi (Hancock and Howse 1870).

Holotype. G.26.49, The Hancock Museum, Newcastle upon Tyne.

Type locality. Railway cutting, 1 km south-south-west of Middridge, Durham, England (NZ 2455 2535).

Type horizon. Marl Slate (Upper Permian).

Diagnosis. A small, probably terrestrial, reptile showing the following combination of character
states: amphicoelous, notochordal vertebrae with broad neural arches and low spines; no develop-
ment of cervical or dorsal transverse processes; an estimated sixteen to eighteen dorsal vertebrae;
gastralia present; preserved ribs single-headed; scapula and coracoid fused; scapular blade low;
cleithrum probably retained; no trace of sternum; short rhomboid interclavicle with broad clavicular
facets; long, almost horizontal glenoid; humerus with broad shaft but little expansion of proximal
and distal ends; entepicondylar foramen present, but no trace of ectepicondylar foramen; radius
and ulna of equal length; radius 64 % of humeral length; ulna lacks olecranon and sigmoid notch;
ulnare and intermedium notched for perforating artery; medial and lateral centralia retained; medial
centrale fails to contact distal carpals 3 or 4; metacarpals and digits short; phalangeal formula
2:3 :4:(3 + ):3; ilium with triangular blade; long slender sigmoidal femur; tibia almost 90% of
femoral length; fibula very slender; metatarsals long.

DESCRIPTION

The reptile lies on its back (not on its belly, as described by Hancock and Howse 1870). The skull has been
lost. The skull fragment mentioned in the original description is part of the pectoral girdle. A mass of bone
fragments below the right arm may be part of the occiput and/or atlas-axis complex (text-fig. 1).

The axial skeleton. Hancock and Howse (1870) made a count of fourteen or fifteen dorsal rib pairs; there are
fifteen pairs preserved. One anterior vertebra has shorter ribs associated with it and is probably a cervical (see
below).

Each dorsal rib has a small single head. The proximal shaft is flattened and slightly expanded; distally it
becomes more circular in cross-section. The longest ribs are in the mid-dorsal region, but towards the rear of
the body they become shorter and the enclosed body cavity narrows. Between consecutive ribs, there are
slender gastralia, apparently three pairs per vertebral segment. These are clearest on the left side of the body
where they appear to begin between the sixth and seventh rib pair.

Because of the position of the animal at death, many of the vertebrae are seen in ventral or ventrolateral
view, with the neural spines obscured by ribs and gastralia. A total of nineteen presacrals and six fragmentary
caudals is preserved.

The vertebral centra are of roughly equal length. On vertebral morphology alone, it would be difficult to
distinguish dorsals from cervicals, but the ribs provide a key. Each of the fourteen vertebrae at or behind the
level of the proximal humeral heads is associated with a pair of long dorsal ribs. An additional five vertebrae
lie clustered around the most anterior (left) scapulocoracoid. Of these, at least one may be a dorsal (the most
anterior rib pair); the other four are probably cervicals. This confirms Watson’s (1914) conclusion that the
cervical vertebrae of Adelosaurus are short, in contrast to those of Protorosaurus. Unfortunately, these anterior
vertebrae are poorly preserved. They are similar to the dorsals except that the rib facet lies slightly further
back (text-fig. 2a). Hancock and Howse (1870) give a count of seven cervicals, but this was, presumably, an
estimate. There are at least fifteen dorsals, one for each rib pair. The femoral heads lie just behind the last
preserved presacral. The sacrum is missing, but from the diameter of the body at the end of the vertebral
series, it seems unlikely that there are many missing presacrals. If the first rib preserved is that of the first
dorsal, then an estimate of sixteen to eighteen dorsals seems reasonable.

The dorsal centra (text-fig. 2c, d) are relatively short (compared, for example, with those of the contemporary

EVANS: UPPER PERMIAN REPTILE

959

text-fig. I . Skeleton of Adelosaurus huxleyi, holo-
type, G. 26.49. Abbreviations used in figures: a.zy,
anterior zygapophysis; Cd.V, caudal vertebra;
Cla, clavicle; Cle, cleithrum; C.r, cervical rib; D.r,
dorsal rib; Fe, femur; Fi, fibula; H, Humerus; I,
intermedium; II, ilium; Int, interclavicle; lc, lateral
centrale; me, medial centrale; Mt, metatarsal; n.sp,
neural spine; P, pisiform; p.pt, posterior pit; R,
radius; rad, radiale; r.ft, rib facet; Sc.C, scapuloco-
racoid; Ti, tibia; U, ulna; ul, ulnare. Numbers 1-5
refer to distal carpals.

.4

glider Coelurosawavus (Evans 1982; Evans and Haubold 1987)). They are rounded, lack a ventral keel, and
are amphicoelous- probably notochordal (text-fig. 2c). The neural arch is much wider than the centrum, so
that, even allowing for some compression, the arch pedicels diverge upward in end view. There is a short low
neural spine (again in sharp contrast to Protorosaurus ), above the broad, flattened arch. The zygapophyses
are almost horizontal. The posterior zygapophyses are swollen; between them, at the base of the neural spine,
there is a deep pit— probably for the insertion of intervertebral ligaments. This does not, however, show the
pit and tubercle arrangement found in the intervertebral facets of younginiforms (Currie 1981). The anterior
zygapophyses are broad and flat. For the most part, they lie anterior to the neural spine. There are no
transverse processes and the rib facet lies at the anterior edge of the arch pedicel.

Only a few caudal vertebrae are preserved, separated from the last dorsal by a gap of about 50 mm (text-
fig. 1). They match the mid to posterior caudals of other genera in being cylindrical with small zygapophyses
and no neural spines (text-fig. 2b). Ventrally, there is a deep groove for the caudal blood vessels. A weak line
of discontinuity runs down the centrum at the mid-point of the vertebra passing on to the ventral surface and
obstructing the caudal groove. This may be a developmental feature rather than a functional autotomy plane.

960

PALAEONTOLOGY, VOLUME 31

text-fig. 2. Adelosaurus huxleyi , holotype, G. 26.49. a, cervical vertebra, left lateral view, b, caudal vertebra,
ventrolateral view, x marks the line of discontinuity (see text), c, d, associated dorsal vertebrae, e, Left hand,
dorsal view, f, interclavicle, clavicle, and possible cleithrum, ventral view. G, restoration of interclavicle, ventral
view. H, right scapulocoracoid, lateral view. I, right ilium, medial view.

Most of the vertebrae are disarticulated, but in a few places there are bone fragments between adjacent
centra. Watson (1914) interpreted these as tiny intercentra but they could also be fragments of ribs or gastralia.

The appendicular skeleton. The preserved parts of the pectoral girdle include the two scapulocoracoids, the
interclavicle, a clavicle, and a possible cleithrum.

The interclavicle is exposed in ventral view (text-fig. 2f). The left crus is almost complete, but the right crus
and the interclavicular stem are damaged leaving a few bone fragments and an incomplete impression. None
the less, the bone can be partially reconstructed (text-hg. 2g). The shape is that of a short rhomboid, almost

EVANS: UPPER PERMIAN REPTILE

961

T-shapcd, with wide clavicular facets that taper laterally. Anteriorly, the clavicles are separated by a narrow
spur of bone.

In association with the interclavicle, there are two slender bones (text-fig. 2f). The larger, probably the left
clavicle, has a long, narrow shaft expanding into a broad terminal plate. Adjacent to its shaft, there is a
fragment of a more slender bone which may be a cleithrum.

Both scapulocoracoids are preserved in lateral view (text-fig. 2h). The scapula and coracoid are fused
without trace of a suture. The two parts are of roughly equal size, with a low scapular blade and a relatively
short coracoid portion. This suggests that only one coracoid ossification was involved— a conclusion reached
by Watson (1914) and Kuhn (1969fi). The glenoid cavity lies at the junction of the scapula and coracoid. It is
long and almost horizontal in orientation, ending anteriorly in a well-developed boss. In front of this is the
coracoid foramen. There is no supraglenoid buttress.

The forelimbs are well preserved. The right arm described Hancock and Howse (1870), and figured by
Kuhn (1969A, p. 31, fig. 14.2) is, in fact, the left. The humerus is strong with a relatively thick shaft and little
proximal or distal expansion. A depressed area at the distal end of the left humerus may be a small entepicondy-
lar foramen but there is no visible ectepicondylar groove. The joint surfaces are unfinished. Taking the length
of an average dorsal vertebra as the standard, x (see Currie 1981), the length of the humerus is 51x.

The radius and ulna are strong and rather short (radius, 3-4x). The radius is 64 % of the humeral length. It
is slightly twisted and of similar width throughout. The ulna is expanded at both ends, with the greatest width
proximally but there is no sigmoid notch or olecranon (contra Huene 1956 and Kuhn I969A).

The left hand is preserved in dorsal (extensor) view; the right in plantar (flexor) position. This accounts for
small differences in detail between the two. The left hand is the more complete (text-fig. 2e). As in all primitive
reptiles, there are three rows of carpals— proximal, central, and distal. The proximal carpal row of Adelosaurus
contains four bones— radiale, intermedium, ulnare, and pisiform. Of these, the radiale is the smallest, with
the ulnare roughly twice its size. The pisiform is nearly as large as the intermedium. Both ulnare and
intermedium are notched for the passage of a perforating artery. The central carpal row contains medial and
lateral centralia of roughly equal size. The medial centrale contributes to the radial border of the carpus but
does not contact distal carpals (DC) 3 or 4 (contra Tangasauridae, see Currie 1981). There are slight differences
between the two hands with respect to the distal carpal row. Distal carpals 1 and 4 are clearly preserved but
rounded impressions mark the positions of 2 and 3. While it is conceivable that these carpals were simply lost,
their absence in both hands when the remaining carpals are relatively undisturbed, renders this improbable. It
is more likely that DCs 2 and 3 were incompletely ossified at the time of death. In the left hand, DC4 is
smaller relative to DC1 than on the right, but there is a small lateral bone which may be an unfused DC5.
The five metacarpals (MC) are short and stout (longest, l-3x), with expanded ends. MCI and 5 are of
roughly equal length, followed in increasing order of size by MCs 2, 3, and 4. The proximal phalanges are
even shorter. Ungual phalanges are poorly preserved on both hands, but they seem short and triangular. The
phalangeal formula is 2:3:4:(3 + ):3.

Our knowledge of the pelvic girdle is restricted to the ilium, although, surprisingly, Kuhn (1969 b) describes
the pubis and ischium as plate-like. The ilium is, unfortunately, preserved in medial view (text-fig. 21), its
ventral border angled by facets for the pubis and ischium. The blade is directed posterodorsally and is
triangular with a blunt tip. This may indicate incomplete ossification (Currie 1981). The surface is roughened
for the attachment of sacral ribs. Compared to those of contemporary genera, the ilium of Adelosaurus is
small (length 2-4x as compared to 3-5x in Younginiformes and Millerettidae).

Except for a fragment of the left femoral head, only the right hindlimb is preserved (text-fig. 1). The femur
is long and slender (6-3x), with a gently sigmoid shaft. Proximal and distal ends are of nearly equal width.
The femur is longer than the humerus, but is a more gracile bone. The tibia (5-4x) is nearly 90 % of the
femoral length. Its proximal end is wider than the distal end, but there is no crest. The fibula, by comparison,
is very slender. The foot is represented by isolated metatarsals and phalanges but there is no trace of the
tarsus. The longest metatarsals (2-4x) are almost twice the length of the longest metacarpal. As a whole,
the forelimb (humerus + radius) is 75 % of the length of the hindlimb (femur + tibia) but the proportions of
the pro- and epipodials are different, such that while the humerus is 86 % of the femoral length, the radius is
only 62 % of the tibial length.

Life stage and habit. Although the scapulocoracoid and vertebral centres are fully co-ossified, the specimen
shows signs of incomplete ossification: absence of joint surfaces on the long bones; non-ossification of DCs 2
and 3; the differences in the ossification of DCs 4 and 5 in the two hands; and the blunt-ended iliac blade.
This could be taken as evidence of either immaturity or an aquatic lifestyle.

In the terrestrial younginiform Thadeosaurus (Currie and Carroll 1984) the ossification centres of the carpals

962

PALAEONTOLOGY, VOLUME 31

appear before the scapulocoracoid suture closes. In this respect, Adelosaurus more closely resembles the
aquatic younginiform Hovasaurus (Currie 1981), where the scapula and coracoid fuse before some of the
carpal centres appear. However, in Hovasaurus , as is common in aquatic animals, the neurocentral sutures
remain open until late in life. In Adelosaurus and Tangasaurus (Currie 1981), they are closed.

On balance, it seems more likely that the skeleton of Adelosaurus described here is that of an immature,
rather than juvenile, animal in which the body proportions are unlikely to be significantly different from those
of the adult. The long, rather slender, hindlimbs suggest an agile terrestrial form. The Marl Slate and
Kupferschiefer are thought to have been laid down in the relatively shallow coastal waters of the Late Permian
Zechstein Sea (Smith 1970; Pettigrew 1980). In addition to fish, the deposits yield abundant plant remains
suggestive of coastal forest or woodland (Pettigrew 1980; Haubold and Schaumberg 1985) which would have
been home to the glider Coelurosauravus and some, at least, of the remaining reptiles, including Adelosaurus.

DISCUSSION

Adelosaurus differs from Protorosaurus , to which it was originally referred (Hancock and Howse
1870), in several respects, most notably the proportions of the humerus and cervical vertebrae, and
the length of the dorsal neural spines. None of the known specimens of Protorosaurus shows a clear
series of cervical and dorsal vertebrae and estimates of vertebral numbers vary. Huene (1926) and
Seeley (1888) count seven cervical vertebrae, but Huene’s reconstruction shows a long eighth
vertebra which may also be a cervical. Similarly, estimates of dorsal numbers vary from sixteen to
eighteen, although there seems to be a general agreement on sixteen dorsal ribs (Huene 1926; Seeley
1888; Haubold and Schaumberg 1985; pers. obs.). If Adelosaurus were a juvenile Protorosaurus ,
then we would expect elongated vertebrae in front of the first long rib. This is not the case.

The pareiasaur Parasaurus is known from three fragmentary specimens. It shares with Adelosau-
rus the primitive captorhinomorph condition of the vertebrae but the proportions of the two
animals are quite different, even allowing for the immaturity of Adelosaurus. Parasaurus is stoutly
built, with four to six sacral ribs meeting a broad iliac blade. The vertebrae are short and very wide,
and there are no gastralia (Kuhn 1969a). Coelurosauravus is a highly specialized glider (Carroll
1978; Evans 1982; Evans and Haubold 1987) with long ribs and elongated cervical and dorsal
vertebrae. Nothosauravus is represented by a single notochordal vertebra with either long transverse
processes or fused ribs. Neither genus bears any resemblance to Adelosaurus.

Adelosaurus therefore represents a fifth member of the Kupferschiefer/Marl Slate reptilian assem-
blage. In the absence of the skull and ankle, however, its phylogenetic position remains equivocal.
The general structure of the vertebrae, shoulder girdle, and carpus are primitive. The low neural
spines, broad neural arches, short rib pedicels, notochordal centra, and barely inclined zygapophyses
are primitive amniote character states (Heaton and Reisz 1986) but the slender sigmoidal femur,
triangular iliac blade, and unexpanded humerus are derived states.

The Upper Permian millerettids have been linked to captorhinomorph reptiles by Gow (1972)
and Heaton (1980), and it is generally agreed that they represent either modified or juvenile
(incompletely ossified) anapsids (Gauthier 1984; Benton 1985; Evans 1988), although their precise
relationships are still debated. Adelosaurus shares several character states with millerettids, including
a single coracoid, loss of the supraglenoid buttress and short rhomboid interclavicle, but these states
are found in other genera. Adelosaurus differs from millerettids in the shape of the iliac blade, the
sigmoid femur, the probable retention of a cleithrum, and the proportions of the radius and
humerus.

Huene (1956) and Kuhn (19696) link Adelosaurus with Broomia (Middle Permian, South Africa).
Broomia has recently been redescribed by Thommasen and Carroll (1981), who classify it as a
millerettid on the basis of the anterior position of the quadrate condyles and the structure of the
foot. Adelosaurus is more gracile than Broomia , and has broader clavicles. Both have a sigmoid
femur. In the carpus, the ulnare of Broomia is long and narrow while that of Adelosaurus is short
and broad. The perforating foramen in Adelosaurus passes between intermedium and ulnare, but in
Broomia the foramen is larger and includes the lateral centrale in its borders. In both carpal

EVANS: UPPER PERMIAN REPTILE

963

characters, Broomia shows the more primitive condition. There is little to support a relationship
between Adelosaurus and Broomia.

Broad neural arches with low neural spines are also found in pareiasaurs (discussed above) and
procolophonids. Procolophonids are known from Permian and Triassic deposits world-wide. They
combine primitive vertebrae with a dorsoventrally compressed body, short tail, and very short
epipodials. The iliac blade has an anterior process which meets an additional sacral rib. These
derived character states are not shared by Adelosaurus.

One feature of the Adelosaurus skeleton which differentiates it from the majority of primitive
reptiles, including those discussed above, is the short, triangular iliac blade. With the exception of
some pelycosaurs, such as Ophiacodon and Dimetrodon (in which the proportions of the humerus,
neural spines and scapula blade, and the structure of the rib facets preclude relationship), this type
of blade is usually found in diapsids. The Diapsida are diagnosed largely on the basis of cranial
characters, most notably the possession of an upper temporal fenestra. In the absence of a skull,
confirmation of diapsid status is difficult unless the specimen clearly shows the derived character
states of one of the diapsid subgroups. Recent reviews of the Diapsida (Gauthier 1984, 1986; Benton
1985; Evans 1988) recognize a primary dichotomy which produced an early radiation of essentially
primitive, but gracile, genera— the Araeoscelidia— on the one hand, and the majority of typical
diapsids (including archosaurs, rhynchosaurs, prolacertiforms, lepidosaurs, and younginiforms) on
the other. This second group has been alternatively named Sauria (Gauthier 1984, 1986) and
Neodiapsida (Benton 1985). The latter term is used here.

Adelosaurus lacks the majority of diagnostic araeoscelid character states for which it could be
coded: elongated cervical vertebrae; ventral keels on cervical and dorsal vertebrae; neural arches
with deep lateral excavations; elongated coracoid process for triceps; radius nearly equal in length
to the humerus (Reisz el al. 1984). Of fourteen neodiapsid character states (Evans 1988), Adelosaurus
can be coded for only four: single coracoid, loss of the supraglenoid buttress, slender sigmoidal
femur, and absence of an ossified olecranon and sigmoid notch, although the last could reflect
immaturity. Adelosaurus stands in much the same position as the contemporary South African
genera Galesphyrus and Heleosaurus whose diapsid status is equally tenuous. These genera are
provisionally classified as early offshoots from the diapsid stem (Benton 1985; Evans 1988) since
they lack the diagnostic character states of any major diapsid group. Placed with them is C/audio-
saurus from the Upper Permian of Madagascar. This genus has been described as a sauropterygian
ancestor allied to younginiforms (Carroll 1981). It is a diapsid, but it lacks the derived character
states of the Younginiformes, as diagnosed by Currie (1982). Claudiosaurus , like Adelosaurus ,
Galesphyrus , and Heleosaurus , has broad vertebrae with low neural spines— confirming that this
primitive condition can be found in early diapsids. It differs from Adelosaurus in the elongation of
the cervical and dorsal vertebrae, the less expanded clavicles, and the greater width of the distal
humeral head. Adelosaurus resembles tangasaurid younginiforms in the general proportions of the
scapulocoracoid and limbs, and in the possession of short cervical vertebrae, but it lacks young-
iniform character states (Currie 1982; Evans 1988) including the specialized intervertebral joints,
long radius, contact between medial centrale and DC4, and the presence of an ossified sternum.

Adelosaurus clearly lies at a similar evolutionary level to primitive diapsids, but it lacks the
diagnostic character states of any known genus or group and its inclusion within the Diapsida
remains provisional until further material is recovered.

Acknowledgements. I thank the Trustees of the Hancock Museum, Newcastle upon Tyne, for the invitation to
study this specimen. The Royal College of Surgeons, London; Museum fur Naturkunde, Berlin, DDR;
Geiseltal Museum, Halle, DDR; Geology Department, University of Freiburg, DDR; and the South African
Museum, Cape Town, provided access to comparative material, with funding from the British Council and
the Central Research Fund, University of London.

964

PALAEONTOLOGY, VOLUME 31
REFERENCES

benton, m. j. 1985. Classification and phylogeny of the diapsid reptiles. Zool. J. Linn. Soc. 84, 97-164.
Carroll, r. L. 1978. Permo-Triassic ‘lizards’ from the Karroo System. Part 2: A gliding reptile from the
Upper Permian of Madagascar. Pcilaeont. afr. 21, 143-159.

- 1981. Plesiosaur ancestors from the Upper Permian of Madagascar. Phil. Trans. R. Soc. B293, 315-383.
currie, p. j. 1981. Hovasaurus boulei , an aquatic eosuchian from the Upper Permian of Madagascar. Pcilaeont.

afr. 24,99-168.

- 1982. The osteology and relationships of Tangasaurus menelli Haughton (Reptilia: Eosuchia). Ann. S.
Afr. Mus. 86, 247-265.

- and carroll, r. l. 1984. Ontogenetic changes in the eosuchian reptile Thadeosaurus. J. vertebr. Paleont.
4, 68-84.

evans, s. E. 1982. The gliding reptiles of the Upper Permian. Zool. J. Linn. Soc. 76, 97-123.

1988. The early history and relationships of the Diapsida. In benton, m. j. (ed.). The Phylogeny and
Classification of the Tetrapods , vol. 1, Syst. Ass. Spec. Vol. 35A, 221 -253. Oxford University Press, Oxford.

and haubold, h. 1987. A review of the Upper Permian genera Coelurosauravus , Weigeltisaurus and
Gracilisaurus (Reptilia: Diapsida). Zool. J. Linn. Soc. 90, 275-303.

Gauthier, j. A. 1984. A Cladistic Analysis of the Higher Systematic Categories of the Diapsida. Ph.D. disser-
tation (unpublished), University of California, Berkeley.

1986. Saurischian monophyly and the origin of birds. In padian, k. (ed.). The Origin of Birds and the
Evolution of Flight. Mem. Calif. Acad. Sci. 8, 1 55.

gow, c. E. 1972. The osteology and relationships of the Millerettidae (Reptilia: Cotylosauria). J. Zool. Lond.
167, 219-264.

Hancock, a. and howse, r. 1870. On Protorosaurus speneri von Meyer, and a new species, Protorosaurus
huxleyi , from the Marl Slate of Middridge, Durham. Q. Jl geol. Soc. Lond. 26, 565-572.
haubold, h. and schaumberg, g. 1985. Die Fossilien des Kupferschiefers. Neue Brehm-Bucherei, Ziemsen
Verlag, Wittenberg Lutherstadt.

heaton, m. j. 1980. The Cotylosauria: A reconsideration of a group of archaic tetrapods. In panchen, a. l.
(ed.). The Terrestrial Environment and the Origin of Land Vertebrates. Syst. Ass. Spec. Vol. 15, 497-551.

and reisz, r. r. 1986. Phylogenetic relationships of captorhinomorph reptiles. Can. J. Earth Sci. 23,
402-418.

huene, f. von. 1926. Zur Beurteilung von Protorosaurus. Zentbl. Miner. Geol. Palaont. B. 1926,469-475.

1956. Pdlaontologie und Phylogenie der niederen Tetrcipoden. Gustav Fischer Verlag, Jena.
kuhn, o. 1969a. Cotylosauria. Handbuch der Paldoherpetologie, 6. Gustav Fischer Verlag, Stuttgart.

19696. Proganosauria — Protorosauria. Handbuch der Paldoherpetologie, 9. Gustav Fischer Verlag,
Stuttgart.

pettigrew, T. H. 1980. Geology. In DUNN, t. c. (ed.). The Magnesium Limestone of Durham County , 1-26.
Gilpin Press, Houghton-le-Spring.

reisz, r. r. 1981. A diapsid reptile from the Pennsylvanian of Kansas. Spec. Publ. Mus. nat. Hist. Univ. Kansas
7, 1-74.

berman, d. s. and scott, d. 1984. The anatomy and relationships of the Lower Permian reptile Araeo-
scelis. J. vertebr. Paleont. 4, 57-67 .

romer, a. s. 1966. Vertebrate Paleontology. 3rd edn. University of Chicago Press, Chicago.
seeley, h. g. 1888. Researches on the structure, organisation and classification of the fossil Reptilia. 1. On
Protorosaurus speneri (\ on Meyer). Phil. Trans. R. Soc. B178, 187-213.
smith, d. b. 1970. Permian and Trias. In hickling, g. (ed.). Geology of Durham County. Trans, nat. Hist. Soc.
Northumb. 41, 61 91.

thommasen, H. and Carroll, R. L. 1981. Broomia , the oldest known millerettid reptile. Palaeontology, 24,
379-390.

vaughn, p. p. 1955. The Permian reptile Araeoscelis restudied. Bull. Mus. comp. Zool. 1 13, 305-467.
watson, d. m. s. 1914. Broomia perplexa gen. et. sp. nov., a fossil reptile from South Africa. Proc. zool. Soc.
Lond. 1914,995-1010.

SUSAN E. EVANS

Department of Anatomy and Developmental Biology
University College and Middlesex School of Medicine
Typescript leceived 28 October 1987 Windeyer Building, Cleveland Street

Revised typescript 4 January 1 988 London, W 1 P 6BN

COMPARATIVE TAXONOMY OF THE BIVALVE
FAMILIES ISOGNOMONIDAE, INOCER AM I DAE,
AND RETROCERAMIDAE

by J. S. CRAMPTON

Abstract. Fossil Isognomonidae (Pteriacea) can be difficult to distinguish externally from the biostrati-
graphically important Mesozoic family Inoceramidae (Ambonychiacea?). Internal details of ligament area
morphology provide valuable taxonomic data at the family and species levels, as documented for many New
Zealand species. Definitive distinction between these two families is furnished by the shell structure underlying
the ligament area: in Isognomonidae the ligament attaches to the inner (presumed nacreous) shell layer,
whereas in Inoceramidae it attaches to the outer prismatic shell layer. Retroceramus , formerly included in
Inoceramidae, has the ligament attached to the inner shell layer, and should be placed in the Pteriacean
family Retroceramidae. These findings are consistent with a polyphyletic origin for the multivincular ligament
in Isognomonidae and Inoceramidae.

Two new species of Isognomon are described from New Zealand, I. wellmani (Palaeocene) and /. rekohuensis
(Late Cretaceous). They probably lived on soft or shelly substrates with otherwise similar life habits to Recent
forms.

Recognition of two new species of fossil Isognomonidae (Bivalvia), and the need to distinguish
these from biostratigraplncally important Cretaceous taxa, prompted this paper. The new species
are described from Late Cretaceous strata of the Chatham Islands and Palaeocene strata of Castle
Hill Basin, Canterbury (map area K34, see text-fig. 1 ). Rocks of these ages are currently undergoing
detailed study as part of the New Zealand Geological Survey’s Cretaceous-Cenozoic Programme.

Bivalves, while not being biostratigraplncally important in the New Zealand Palaeocene, are a
major tool in global and local Cretaceous biostratigraphy. Those of family Inoceramidae formed
the basis of Wellman’s (1959) pioneering subdivision of the New Zealand Cretaceous, and their
significance has since diminished little (for example, see Stevens and Speden 1978; Suggate el al.
1978). Both bivalves described herein closely resemble species of Inoceramidae, and have previously
been assigned to that family.

Late Cretaceous rocks occur extensively throughout New Zealand, and are characterized by
terrestrial coal measures and marine sandstone and siltstone sequences which are locally richly
fossiliferous. Often complexly faulted and folded, they achieve a thickness of 1000-2000 m
(Johnston 1980; Moore 1980). Comparatively unlithified and undeformed marine Cretaceous strata
were first recognized on the Chatham Islands by Boreham (1959), subsequently described by Hay
et al. (1970), and assigned to the Late Cretaceous by Speden ( 1976), Wilson ( 1976), Mildenhall (1977),
and Strong (1979). They comprise conglomerate, sandstone, and fossiliferous tuff and limestone.

Palaeocene rocks, on the other hand, are not well exposed in New Zealand, being restricted
mainly to thin sequences on the east coast of both islands. They generally consist of poorly
fossiliferous commonly glauconitic mudstone, sandstone, and limestone, which in many places
overlie Late Cretaceous strata. The Late Cretaceous-Palaeocene rocks of Castle Hill Basin,
Canterbury, have a long history of description, beginning with Hector (1881) and McKay (1881),
and were mapped most recently by Gage (1970). Lithologies at this locality include sandstone
(carbonaceous at the base and glauconitic above) with minor mudstone, limestone, and rare shell-
beds.

(Palaeontology, Vol. 31, Part 4, 1988, pp. 965-996, pis. 88-90.|

© The Palaeontological Association

966

PALAEONTOLOGY, VOLUME 31

text-fig. 1 . Map of New Zealand showing all fossil localities referred to in
the text in terms of their NZMS 260 1 : 50 000 map sheet areas.

Material described is housed in the Geology Department, University of Auckland, Auckland; Geology
Department, University of Otago, Dunedin; and the New Zealand Geological Survey, Lower Hutt. The
following prefixes indicate specimen repositories and localities:

L(AU) Specimen number. Geology Department, University of Auckland.

AU Collection number, Geology Department, University of Auckland.

OU Specimen number. Geology Department, University of Otago.

TM Type Mollusca specimen number, NZ Geological Survey.

WM World Mollusca specimen number, NZ Geological Survey.

GS Collection number, NZ Geological Survey.

L Palynology sample number, NZ Geological Survey.

K34/f48 Fossil locality number of the New Zealand Fossil Record File, based on metric NZMS 260 1 : 50 000
map sheets. K34 refers to the map sheet number, and f48 refers to a unique fossil locality within that
area. All New Zealand fossil localities mentioned in the text have their map sheets areas shown on
text-fig. I

CRAMPTON: ISOGNOMONIDAE, INOCERAM IDAE, AND RETROCERAMIDAE 967

Synonymy lists employ the annotations outlined by Matthews (1973) to indicate degrees of confidence for
references. Full bibliographic references for all bivalve taxa below superfamily level are given.

ISOGNOMONIDAE COMPARED TO INOCERAMIDAE AND RETROCERAMUS

Edentulous (in the adult stage), multivincular Pteriacea of variable form are included in
Isognomonidae Woodring, 1925. Members of this family are sometimes difficult to distinguish
from Inoceramidae Giebel, 1852, a problem addressed by a number of authors, notably Heinz
(1932), Cox (1940), and Hayami (1960): see Table 1. As discussed below, failures to recognize
some New Zealand fossil Isognomonidae, the uncertain taxonomic position of Retroceramus
Koschelkina (1959), and evidence for greater phylogenetic distance between Isognomonidae and
Inoceramidae than previously recognized, make it prudent to review differences between the two
groups.

Prior to Cox (1955) most authors (Heinz (1932) being one notable exception) included nearly
all inoceramids within the single genus Inoceramus Sowerby, 1814, which was grouped with the
isognomonids and bakevelliids in Isognomonidae. Cox (1954, p. 47) wrote. The removal of
Inoceramus and related genera from the Isognomonidae does not at present seem necessary . .
Indeed, he had earlier criticized Heinz (1932) for over-intensive subdivision of what was ‘ . . .
formerly regarded as a single genus . . and stated, Tt is possible that two or three distinct genera
and several subgenera may eventually prove to be distinguishable among the species hitherto
included in Inoceramus . . .’ (Cox, 1940, p. 125). Similarly, the genus Isognomon Solander in
Lightfoot (1786), as used by Cox (1940, 1954) and Hayami (1957, 1960: see Table 1) included most
species previously assigned to Perna Bruguiere, 1789 (not Perna Retzius, 1788 (Mytilidae)) and
now referable to several genera in Isognomonidae. (A number of other isognomonid genera had
been described before 1960, but apart from Crenatula Lamarck, 1803, they were little used.)

Subsequent to Cox (1955) not only have Inoceramus and related bivalves been placed in their
own family, Inoceramidae, but Kauffman and Runnegar (1975) tentatively suggested they should
be removed to a different superfamily, Ambonychiacea. This was based on evidence for their
evolution from the Permian Atomodesma Beyrich, 1864, as opposed to the widely accepted view
that most Inoceramidae evolved from Isognomonidae (for example, Hayami 1957, 1960). Separation
of the two families would not be remarkable given that many authors have postulated a polyphyletic
origin for the multivincular ligament, the single most distinctive character of both taxa (Heinz
1932; Cox 1940; Hayami 1960; Browne and Newell 1966; Kauffman and Runnegar 1975; Dickins
1983).

External characters have generally been used to distinguish Isognomonidae from Inoceramidae
(Table 1). Most importantly, Isognomonidae usually have terminal umbones which are little, if at
all, incurved and commonly project beyond the rest of the anterior shell margin; they are rarely
markedly prosocline; they may possess an anterior byssal gape; and they have a smooth,
commarginally lamellose, or in a few taxa radially sculptured surface lacking commarginal plicae.
Most Inoceramidae, on the other hand, generally possess a gibbous more or less incurved,
subterminal umbo; they may be acline to strongly prosocline; most do not possess a byssal gape
(recently some early forms with large byssal gapes have been referred to this family, for example
Permoceramus Waterhouse, 1970); and almost all have commarginal or (in fewer taxa) radial
plicae. These criteria hold true in most material examined, although in some cases differences may
be subtle, for example compare Isognomon rekohuensis (sp. nov., described herein) and Inoceramus
opetius Wellman, 1959. Isognomon rekohuensis has a weakly inflated, terminal, prosogyrate umbo
projecting beyond the anterior end of the hinge line, whereas Inoceramus opetius has a more
gibbous orthogyrate umbo close to, but not at, the anterior end of the hinge line (contrast PI. 89,
fig. \e and PI. 90, fig. 7). Both species are acline or nearly so, and Isognomon rekohuensis has only
a narrow byssal gape, if any gape at all. The latter does bear the lamellose ornament characteristic
of the family, but in addition it has weak commarginal plicae between the shell layers which
resemble the weak and irregular external ornament of Inoceramus opetius (PI. 89, fig. 16).

Descriptions of selected diagnostic and differential characters of Isognomomdae and Inoceramidae according to different authors.
Inoceramidae Isognomonidae

968

PALAEONTOLOGY, VOLUME 31

G

<D

a

*3

G

<D

fH O

.5 g-

O/ r) — <

(D ^

H " o
B dU

C O

*D

% ON

n , G

G in dj

| 5 S
u x

o

X

too

G

d

G

a

G

too

X

too On

c x

ON

5 x
12 o

G — H

G cN
m

a

G
O
X -
+-* ON

’> ^
> ON

^ G

. G ”
O toO,
rfr <D

. on G '

CN X
m O

«v C/5

On C

X o

ON -G

G

X

r9 X

X £
w G
o

X 5

5 £

X <D
X O

£ C
g o
G o

c/5 toO

<d (D c

3 6-|

S cl/

X

GOG

c/5 g

C/5

CD C/5 C/5

G
O

X ON
X

- G 2
<N X
— < X

• X °

agu

bO

CD

<D

.2 -S X

G

toO

G

s

CD

O X O G
•S U G £

£ d£

o' £

O

o y

G G
1 o

C/5

>NX
— G
G
CD

'G G

8 6

CD G

G G
O

o o

X 73
o ;3

O G

O

X *— <
£ 2

o o
o

£ G y

a S.f
£ $; o

x
o
U

03 g
-o Oh o

u . bo
== 3
O U • rj
3

O

C/5

3d.£ £
£ -O 2 -m

d u d
§ d a

C/5

CD

& d "3
o 3 £
Cl-O

S * c

2 G G
g 'I! toO
X c

o cd is

p ^ o
G G
(D O O
‘ too *-■
G </)

G

G

X X o

toO

G X
G

G

Qh.G 75

G ^ ‘
<D ^

6 *

S o

too X

CO 2

c y
.2 *
■e £

O G

g-d

2 ^
a p

Wife.

.2 03

>> x

V- OS

So.S
d d
^ 3
d >
G G

a, '

d

d

<u

add

. 03

a £

CD r~ .

SC ^ '
c sx -

C/5

CD r^.
took

Id

03 O
bD O

d

d d
d d
? 21

c d

O os
<u ~ <d

— X SO

d o a

.2 CJ ^

C ^ x
> c o

X (*J

<s-l

d

<u o

2- £
C/5 G

G

^ G
G ■-

G

S C >»
d op

_, '-t-H

.2 o

H o

toO

G

-o

G

G

G

C/5

X

id

G

£

O

O

G-

ON

X

o

u

G

X

’£

o

£

o

G

toO

o

O

G"

ON

X

o

U

G

X

G

G

G

£

5 ■«. g-

’3 "d d

W cS

d £
oj • 5
> a d

3 O d

y c- c
Sou

“ X g

dog
!- U £
o' — • a
<u y <u

b 3 i-i
O d O

s 2

ON

X

ON

. G

r~~ ^
oc

Od G
G

ci G

^ CD

° 2 ■
x .5 1

ON X5 c

^ G ’

s J ;

G O ,

.

G ■ — 1

x|

on

x

o

U

t/D
ON

G- ^
O

- £

ci >,

r-. G

G •

a,

G

ffi 2

v — ^ G"
^ ON

X

G

x y
—H X

m 00

. CD

a §

$1
os £

r“ 1 d

0 g

y §

C/5 T3

X CD

•r

^ r2

G G

1 s

G XJ
O

X) tS
x .£
S a

x ^
^ x

G g
£■£

• X £

CD

X ^ o

ON

X

ON

X

o

u

<*)

_ a
a 8
a y d

. £ c c

a y o <u

a o jo d

a < <

^ X
os o

C3
0) O

g a
-2 >.

a H

X £

o g
U c
w o
U

3

O

—

CD

£

G

G

X

£

G

G

X

x

£ x

o

G"

VON

2 g ><

G J O
d^U
2 x

d g «
S U d

d 3 §

^ o £
o

, X G

G

G

X

’£

o

o

G

too

O

G
O
CD

G
. G

/— d G

oo O
OO Oh

^ £
d2

. G

o o

SO CJ

2 G
G

*G X

toO ON

c/D 73

21 £
G

d 7

2 2
s; .2

8 g

6 "2

G
toO ’—

• C/5

G

O _g

O

G

G

z

C/5 CD

CD ^

1 1

^ G

G m

G V.O

e 2

(D

> G

(D G

■“I

o

3 2

O

3

o 2

X ^
G >*
G

k C/5
p-3

^2 o

1 a

G-" £
O G
^ G

d O
- o

d o 2

§§ 2 'a
(N ^ CD

d ^ 8

c c ^
o S £

ON

a
X o

CD X

> <D

G

g 3

> G
cd X

X c
^ <D

d >*•§

3 .2

d 03
03 DO

E 3

O &

w

o3 . ^ .

o c ^
a £ ^

>i G ff)

G-I

£2 o a

a

a

03

d

a

CD

C

O

d

£

P

c

<C

60

C

o

a

a

03

bO

X .

CO

£

£ 3
3 £

o £
U

CRAMPTON: ISOGNOMONI D AE, INOCERAMIDAE, AND RETROCER AM I D AE

969

o

ON

X

o

U

M

O

'O

c

'O

'O

o3

1 O X5

ON

NO

ON

X -

o r-

C/2 TN

Os
<D ^

W) _

j§ s

03

0) S'

S

jrin

.a

D.

O' •£

>1

c « I-

33 O

„ X >
<4 ° ^

s

03 X-

970

PALAEONTOLOGY, VOLUME 31

obliquity

dorsal

text-fig. 2. Shell dimensions referred to in this paper.

anterior

height

posterior

ventral

Details of internal morphology can provide useful diagnostic high- and low-level taxonomic
data on fossil Isognomonidae and Inoceramidae (a fact stressed by many authors, for example
Kauffman 1965, 1977; Troger 1976; Zonova and Yefremova 1976; Yonge 1978; Zonova 1980a;
Pokhialaynen 1985). Although such features are often difficult to observe and have in the past
been poorly documented, an increasing amount of information on ligament area morphology is
becoming available (Airaghi 1904; Kauffman 1965; Zonova and Yefremova 1976; Pokhialaynen
1969, 1977, 1985; Ivannikov 1979; Zonova 1980a, b , 1982). Use of such data must, however, be
tempered with caution since details of the ligament area can be unstable at the species level (for
example, Cox 1940, p. 122, this study Isognomon (I.) sp., PI. 89, figs. 2-5) and family level (for
example, Kauffman and Runnegar 1975, p. 36). In the present paper the ligament areas of several
New Zealand Jurassic to Palaeocene Isognomonidae, Inoceramidae, and Retroceramus are described
(Appendix, terminology explained in text-fig. 3) and illustrated for the first time. Table 1 summarizes
differences between the ligament areas of Isognomonidae and Inoceramidae as perceived by some
other workers. Of these characters the following appear to hold true in species described in the
literature or examined first hand (excepting Retroceramus , discussed below).

Isognomonidae have a multivincular ligament in all cases, whereas Inoceramidae may carry in
addition or exclusively an elongate longitudinal ligamental groove (Kauffman 1965; described in
detail in Speden 19706). Isognomonidae have monoserial resilifers on an area that is flat or only
slightly concave (for example, PI. 88, figs. 8 and 9; PI. 89, figs. 1<?, 2-7), whereas Inoceramidae
have relatively numerous monoserial, multilobate, or multiserial pits (refer to text-fig. 3) on a
weakly convex to strongly concave area (PI. 90, figs. 4-14). In Isognomonidae the resilifers are
nearly always approximately rectangular, breaching the ventral margin of the area, which may be
crenulated (though this can vary between individuals of a population, compare PI. 89, figs. 2 and
5). In Inoceramidae, on the other hand, resilifers may be rectangular to ovate and are commonly
elongate-ovate (for example PI. 90, figs. 9 and 13), they may or may not breach and do not
markedly crenulate the ventral margin of the area, and they apparently never have the broad flat
or concave interspaces common in Isognomonidae.

In all specimens examined by Dr N. J. Morris (British Museum, Natural History, written pers.
comm. 1986) the ligament in Isognomonidae attaches to the inner, presumed nacreous, aragonitic

CRAMPTON: ISOGNOMONIDAE, INOCER AM I DAE, AND RETROCERAMIDAE 971

resilifer to
margin of area

text-fig. 3. Schematic diagram of multivincular ligament area, explaining measurements and morphological
terms used in this paper. For other terms and morphologies see Pokhialaynen (1977).

shell layer (see text-fig. 4a, terminology after Taylor et al. (1969), shell structures discussed briefly
below), whereas in Inoceramidae it attaches to the outer, prismatic, calcitic layer (Koschelkina
1971; Pokhialaynen 1972; Morris, pers. comm. 1986; see text-fig. 4c). This distinction is maintained
in eleven New Zealand Cretaceous lnoceramus species, five New Zealand Jurassic to Palaeocene
Isognomon species (see Appendix), numerous younger New Zealand and world Isognomonidae,
and the type species of the type genera of both families. The type species are lnoceramus cuvieri
Sowerby, 1814 (two specimens seen, WM 14879 and 14880, these being casts of specimens B6683
and B20997, illustrated by Woods (1905, text-figs. 78 and 80 respectively) from the Sedgwick
Museum, Cambridge, England, and kindly made available by Dr D. Price), and Isognomon
isognomon (Linnaeus, 1758) (twelve specimens seen,WM 12064, from Stadbroke Island, Queensland,
Australia). However, six New Zealand Jurassic Retroceramus species (formerly included in
Inoceramidae) have the ligament attached to the inner shell layer (see text-fig. 4b and Appendix).
Attachment of the ligament to different shell layers in lnoceramus and Retroceramus was noted
previously by Koschelkina (1971, 1975, 1980) and Pokhialaynen (1972, p. 58), who referred to a
‘padding’ of prismatic ‘ligamentat’ beneath the ligament area of lnoceramus s.l.

The phylogenetic and taxonomic significance of the relationship between the ligament and
different shell layers is difficult to assess since few species descriptions include such detail and the
state of this character through ontogeny is poorly known. Larval bivalves show remarkably little
variation in ligament form, with differentiation occurring after settlement (Trueman 1969, p. 62).
In Recent Isognomon , the dentate larva gives rise to a juvenile with a single amphidetic resilium.
Subsequently the lamellar ligament extends posteriorly with (secondary) areas of fibrous ligament
(forming resilifers) appearing at intervals within it. Simultaneously the ligament area extends
ventrally, with the more dorsal parts becoming separated and obsolete (structure and development
of the ligament in Isognomon is discussed by Bernard 1898, Trueman 1954, Yonge 1968, and Siung
1980). It seems likely that the multivincular ligament of Inoceramidae formed in a similar manner
since different growth stages of the same species show expansion of the ligament area in the
posterior and ventral directions (for example, lnoceramus cuvieri , specimens WM 14879 and 14880).

972 PALAEONTOLOGY, VOLUME 31

text-fig. 4. Cross-sections through the ligament areas of representative Isognomonidae, Retroceramidae,
and Inoceramidae, drawn from photomicrographs. Sections perpendicular to ligament area and approximately
half-way between umbo and posterior end of area. All drawings oriented dorsal-up, shell interiors to the left,
and planes of commissure vertical. Inner, presumed nacreous, shell layer shown blank, prism shapes indicated
in outer prismatic shell layer, la = ligament area, a. Isognomon (Isognomon) sp. (see Appendix); TM 6790,
H47/f6494; coast opposite Bloody Jacks Island, Tuhawaiki, Southland; x 3.3. b, Retroceramus (Relroceramus)
haasti (Hochstetter, 1863); TM 6792, R 1 5/f 8564; north side of Kowhai Point, from western tip for
approximately 50-100 m east, Kawhia Harbour, south-west Auckland; x 10. c, Inoceramus opetius Wellman,
1959; TM 6791, P30/f6895; middle branch of Wharf Stream, approximately 1-2 km upstream from junction

with south-east branch, Marlborough; x 6.

Since, in Isognomonidae, the first-formed dorsal part of the ligament lies very close to or on the
boundary between the outer prismatic and inner nacreous shell (see, for example, Trueman 1954,
fig. 2), the different states seen in Inoceramidae and Isognomonidae could only arise by differential
thickening of the prismatic or nacreous shell layers respectively at the time of formation of the
first resilium. Once established in the individual, this pattern apparently does not change, and it
appears to be consistent over taxa that are geographically and temporally widely separated. Hence
it is postulated that the relationship of the ligament to the inner and outer shell layers is a
fundamental character fixed at the time of larval settlement, when other major adult growth
configurations are established. Detailed ontogenetic studies and more exhaustive surveys of fossil
Pterioida are required to test this hypothesis.

Based on the present data, however, it seems that Jurassic Retroceramus species should be
removed from Inoceramidae and accommodated in the family Retroceramidae Pergament in
Koschelkina, 1971. A diagnosis of Retroceramidae is given below. Attachment of the ligament to
the inner (rather than outer) shell layer separates Retroceramidae from Inoceramidae, and indicates
close relationship between Retroceramidae and Isognomonidae. Most species of Retroceramidae
are distinguished from Isognomonidae, externally, by their regular commarginal plicae, more
prosocline shape, greater inflation, and more prominent subterminal umbones (compare text-fig.
5a and d). Internally, some taxa in the two families have very similar ligament area morphologies
(compare PI. 89, fig. 2 and PI. 90, fig. 1). However, Retroceramidae have the ligament area inclined
to the plane of commissure, and they commonly have relatively broad, somewhat irregular, square
or ovate resilifers, separated by broad, concave interspaces. In some taxa the interspaces resemble
a second class of resilifer (see PI. 90, fig. 2). The morphology of the ligament area in Retroceramus
is described and illustrated by Crame (1982) and Koschelkina (1963, 1969, 1971). Some taxa,
apparently intermediate in form between Isognomonidae and Retroceramidae, can be difficult to
assign to either family, and they support an inferred common ancestry of the two families in the
Late Triassic or earliest Jurassic. For example, the Early Jurassic Isognomon ( Mytilopernal ) sp. B

CRAMPTON: ISOGNOMONIDAE, INOCERAMIDAE, AND RETROCERAMIDAE 973

text-fig. 5. External morphology of typical and atypical Isognomonidae and Retroceramidae (see text for
discussion). All specimens whitened with ammonium chloride sublimate, a. Isognomon ( Isognomon ) sp. (see
Appendix); TM 6793, H47/f6494; left valve; coast opposite Bloody Jacks Island, Tuhawaiki, Southland;
xO-7. b, I. (Mytiloperna!) sp. B (see Appendix); TM 4062, R15/f8006; left valve; 60-240 m north-east of
stack at Ururoa Point, south-west Auckland; xO-6. c, Retroceramus ( Fractoceramus ) inconditus (Marwick,
1953); TM 2373, F47/f7492; Quarry Hills, Waikawa district, Southland; x 1. d, R. (R.) galoi (Boehm, 1907);
TM 6719, R 1 5/f 8553; point west of Heteri Promontory, across Waikutakuta Inlet, Kawhia Harbour, south-
west Auckland; x 1, negative reversed.

(see Appendix) has the smooth to lamellose surface and weakly developed umbo diagnostic of
Isognomonidae, and the typically strongly prosocline shape of Retroceramidae (see text-fig. 5b).
Indeed, members of the subgenus R. ( Fractoceramus ) Koschelkina, 1959, which are distinguished
from most Retroceramidae by their weak and irregular ornament, differ little from I. ( Mytiloperna1. )
sp. B (compare text-fig. 5b and c). Internally, the ligament area morphology of I. (M.?) sp. B
resembles, in most respects, species in both families. However, the plane of the area is twisted, a
character state not observed in other members of either family. Hence, while separation of the two
families, Retroceramidae and Isognomonidae, is believed justified, it is acknowledged that some
morphologically intermediate taxa may be difficult to assign to either family.

Furthermore, the different relationships of ligament and shell layers in Isognomonidae,
Retroceramidae, and Inoceramidae are consistent with the postulated existence of two distinct
Late Palaeozoic lineages of multivincular Pterioida, included in superfamilies Pteriacea (containing
Isognomonidae and Retroceramidae) and Ambonychiacea (containing Inoceramidae), as suggested
by Kauffman and Runnegar (1975) and discussed above. The attachment of the ligament to the
prismatic shell in Atomodesma sp. (Ambonychiacea, and an early member of Inoceramidae
according to Kauffman and Runnegar, 1975) from the Early Permian of Queensland, Australia
(B. Runnegar, University of New England, Australia, pers. comm. 1986) lends additional support
to this phylogenetic scheme.

Other internal features, such as muscle scars and the presence and form of the umbonal septum
(for example, on Inoceramus rangatira, PI. 90, fig. 5), may prove useful in characterizing
Isognomonidae, Retroceramidae, and Inoceramidae (see Table 1). As yet little is known about
these characters in the latter two families, although Kauffman (1965) and Pokhialaynen (1985)
briefly described patterns of musculature in inoceramids.

Details of shell microstructure are also likely to be of value in high- and low-level taxonomy of
these groups. Again, little data is presently available, although at higher levels the taxonomic

974

PALAEONTOLOGY, VOLUME 31

distribution of major shell structures is known. While rarely preserved, the middle and inner shell
layers of fossil Pterioida are assumed to have been nacreous aragonite, based on studies of recent
species (Taylor et al. 1969). Waller (1978, p. 351) stated that all the Pterioida have a simple
prismatic, calcitic outer shell layer and a nacreous aragonitic inner layer on each valve. Exceptions
to, and variations within, this pattern were reviewed briefly by Pokhialaynen (1985). Shell
microstructures are documented for the two new species of Isognomon described below, and for
species of Inoceramus resembling them. More systematic studies are planned to discover variations
in shell structure across the individual (for example, the form of prisms in the prismatic layer may
vary immensely between the ligament region and the disc), between taxa, and as the result of
diagenesis.

SYSTEMATIC PALAEONTOLOGY

Superfamily pteriacea Gray, 1847; nom. transl. Dali, 1894 (ex Pteriidae; = Aviculidae Goldfuss,
1820. Pteriidae retained under Article 40 a of the International Zoological Code by Cox, 1969)

Family isognomonidae Woodring, 1925
Genus isognomon Solander in Lightfoot, 1786
Subgenus isognomon Solander in Lightfoot, 1786

Type species. Ostrea isognomon Linnaeus, 1764 (by tautonymy, = O. isognomum Linnaeus, 1758; see Rehder,
1967, p. 6).

Isognomon (Isognomon) wellmani sp. nov.

Plate 88; Plate 90, figs. 15-18

v. 1970 Inoceramus mato torus Wellman, 1959; Gage (p. 516).

Name. Named after Dr H. W. Wellman, formerly Professor of Geology at Victoria University, Wellington,
in recognition of his contribution to the understanding of Cretaceous biostratigraphy in New Zealand.

Material. Many specimens of both valves, several articulated, preserved as internal moulds and outer prismatic
shell layers with internal faces exposed.

Type locality. K34/f48, K34 097770 (imperial grid reference NZMS1, S66/236960), GS 14183: in the lower
part of the prominent 0-4 m thick oyster shell-bed at the top of Broken River Lormation (after Gage 1970,
modified by Andrews et al., in press, described by Browne and Lield 1985, and column from Field and
Browne, written pers. comm. 1986), in the south bank of Broken River, approximately 700 m downstream
from its confluence with Porter River, Castle Hill Basin, Canterbury (see text-fig. 1). Collected by J. S.
Crampton and G. H. Browne, 1986.

explanation of plate 88

Figs. 1-9. Isognomon (Isognomon) wellmani n. sp. All specimens from Broken River Formation, Broken
River, Canterbury, NZ (Palaeocene). 1, TM 6689, K34/f48, holotype; internal mould left valve, x0-7. 2,
TM 6690, K34/f48, paratype; internal mould right valve, x 0-7. 3, TM 6691, K34/f48, paratype; a,

steinkern viewed from right side, x0-7; b, anterior face of steinkern, dorsal to right, xO-7; c, dorsal view
of steinkern, anterior to left, x 0-7. 4, TM 6695, K34/f48, paratype; internal mould right valve, x 0-7.
5, TM 6692, K34/f9096, paratype; steinkern viewed from right side, x 0-7. 6, TM 6693, K34/f48, paratype;
steinkern viewed from right side, x 0-7. 7, TM 6694, K34/f48, paratype; steinkern viewed from right side,
x0-7. 8, TM 6690, K34/f48, paratype; latex mould of ligament area of right valve, x 1-3. 9, TM 6689,
K34/f48, holotype; latex mould of ligament area, x 1-3.

All specimens whitened with ammonium chloride sublimate.

PLATE 88

CRAMPTON, Isognomon (Isognomon) wellmani n. sp

976

PALAEONTOLOGY, VOLUME 31

Type specimens. Holotype: TM 6689, K34/f48, GS 14183 LY internal mould chosen for its reasonably
complete outline and ligament area. Paratypes: TM 6690, K34/f48, GS 14183, RV internal mould; TM 6691,
K34/f48, GS 14183, steinkern; TM 6692, K34/f9096, GS 67, steinkern; TM 6693, K34/f48, GS 14183,
steinkern; TM 6694, K34/f48, GS 14183, steinkern; TM 6695, K34/f48, GS 14183, RV internal mould; TM
6697, K34/f48, GS 14183, thin sections showing subradial and tangential cross-sections through the prismatic
shell layer on the discs of two valves; TM 6761, K34/f48, GS 14183, thin section showing longitudinal cross-
section through the prismatic shell layer of anterior margin of articulated specimen.

Diagnosis. Isognomon of large size, subrectangular to mytiliform and acline to weakly prosocline;
adult ligament area with approximately eleven deep rectangular resilifers which may be the same
width as or considerably wider than their interjacent ridges.

Description. Dimensions (in millimetres, refer to text-fig. 2, outlines reconstructed where feasible):

Specimen

Length

Height

Width

(both valves)

Obliquity

Internal moulds complete with ligament

area

TM 6689

- 80

~ 107

—

- 60°

TM 6690

> 77

- 90

—

- 45°

TM 6695 >60 -100

Steinkerns, excluding ligament area

—

~ 65°

TM 6691

> 63

- 88

31

~ 65°

TM 6692

> 74

> 82

31

- 50°

TM 6693

~ 64

~ 84

27

- 60°

TM 6694

- 75

- 75

21

- 45°

Shell large; variable in shape (PI. 88, figs. 1, 2, 3a, 4-7), subrectangular to mytiliform, acline to weakly
prosocline; prosogyrous with umbo at, or close to, anterior end of hinge line, which projects anteriorly
beyond rest of shell; roughly equivalve, moderately inflated. Dorsal outline incomplete in all specimens,
straight or gently convex. Anterior margin weakly to moderately concave dorsally, then becoming more or
less straight before curving backward to ventral margin. Posteroventral outline slightly concave to convex.
Posterior wing poorly defined, deep, outline not preserved on type material. No evidence for anterior auricle.
Shell wedge-shaped in longitudinal section (PI. 88, fig. 3c), maximum width close to anterior margin and
approximately midway dorsoventrally (PI. 88, fig. 3 b). Anterior face roughly perpendicular to commissure;
disc gently convex except posteriorly where it becomes slightly concave. Commissure flat, presence of byssal
gape not determined.

External sculpture, as determined from thin sections through the outer prismatic shell layer (TM 6697 and
6761; PI. 90, fig. 15), of irregular commarginal lamellae. These increase in density and prominence close to
shell margins, especially on the anterior where individual lamellae may protrude by many millimetres. Shell
layers interface with weak, irregular, commarginal plicae or lamellae.

Hinge edentulous, ligament multivincular (PI. 88, figs. 8 and 9). Ligament area flat or slightly concave;
parallel to or inclined a few degrees to commissural plane; scarcely undercut close to umbo. At least eleven
rectangular, concave-floored resilifers on adult shell, which breach and may weakly crenulate the gently
convex ventral margin of the area (refer to text-fig. 3). Inter-resilifer ridges more or less flat-topped, with
sharp edges and steep sides. Relative and absolute widths of resilifers: ridges vary anteriorly to posteriorly
from 2-7 mm: 1 mm to 4 mm: T5 mm on the holotype and from 2-5 mm: 1-5 mm to 3 mm: 3 mm on
specimen TM 6690. Ligament area at least 12-5 mm high, no growth lines observed although they might be
expected on better preserved material. Ligament attached to inner (nacreous) shell layer.

No adductor muscle scar visible on any of the type specimens. At least twelve discrete pallial muscle scars
form a line close to and parallel to the anterior margin of the shell, from the umbo to the anteroventral part
of the disc (PI. 88, figs. 2 and 3a, b).

The shell is only partially preserved on the type specimens. The inner two layers, which probably consisted
originally of nacreous aragonite (see earlier discussion) have been recrystallized and subsequently dissolved,
leaving only a layer of granular calcite (removed in specimens TM 6689 6691) coating internal moulds and
the internal faces of external shell layers. The external shell layer is preserved intact, and consists of polygonal
regular simple prismatic to rod-type fibrous prismatic calcite (PI. 90, figs. 15-18; terminology after Carter and
Clark 1985). The shell achieves a maximum thickness close to the anterior margin, where the prismatic layer
is at least 7 mm thick and the inner layers 10 mm or more thick. Total thickness towards the centre of the

CRAMPTON: ISOGNOMONI DAE, I NOCER AM I D AE, AND RETROCERAMIDAE 977

disc, however, is probably only of the order of a few millimetres. Within the prismatic layer, prisms are
reclined (dipping towards the shell boundaries), slightly sigmoid-shaped, generally smaller towards the outside
surface, and commonly bearing transverse discontinuities (typically off-set, see PI. 90, fig. 15) which rise to
the outside surface of the shell at an acute angle, resulting in the lamellae already described. Where the shell
is thick there may be many stacked lamellae. Where the shell is thin, in the centre of the disc, single prisms
traverse the whole thickness of this layer, and achieve a maximum size of approximately 2 mm long x
013 mm wide. Adjacent prisms show approximately coincident undulose, patchy, or relatively uniform
extinction. Fractured prisms, examined under SEM (PI. 90, fig. 17), display either a fine-grained granular
substructure (granules ~ 1 p across), or less commonly a smooth cleavage-like surface, while etching reveals
the presence of longitudinal and transverse blocks approximately 10 p across (PI. 90, fig. 18). The latter
probably result, in part, from closely spaced transverse tabulae 8 10 p apart, visible under transmitted light
(PI. 90, fig. 16), and interpreted as accretion lines.

Distribution. Thus far Isognomon wellmani is known with certainty from only the type locality. Specimen TM
6692, collected by McKay in 1886, is from the ‘Saurian Beds, Trelissic Basin’, which most probably
corresponds to the type locality, or very close by (G. H. Browne, pers. comm. 1986). It is very likely that
further sampling, and re-examination of earlier collections, will reveal the presence of this bivalve in other
Palaeocene and possibly Late Cretaceous faunas. Its distribution and biostratigraphic value, however, will
be difficult to assess because of previous confusion with Inoceramus matotorus (discussed below).

Age. The Broken River Formation in the Castle Hill Basin has hitherto been considered entirely Haumurian
(latest Campanian-Maastrichtian), based on the presence of I. matotorus Wellman and Conchothyra parasitica
(Hutton) (Gage 1970; Browne and Field 1985). However, dinoflagellates in the matrix of the shell-bed at the
type locality of Isognomon wellmani (K34/f48, sample L 12989) indicate a Teurian age (Danian-Landenian;
G. J. Wilson, written pers. comm. 1986). This is consistent with Teurian ages for two pollen samples
(K34/f9611 and 9612, samples L 4194 and 4195) from a short distance downstream and stratigraphically
below the shell-bed; and a pollen sample (K34/f9565, sample L 1706) from just above Torlesse basement in
Whitewater Creek, 5 km to the south-west (J. I. Raine, written pers. comm. 1986). Of the macrofossils from
the type locality, ‘ Inoceramus matotors' has here been referred entirely to Isognomon wellmani , and re-
examination of C. parasitica proved inconclusive: the specimens are poorly preserved but show traces of
ornament that may be remnants of the prominent spiral cords characteristic of the Teurian C. australis
(Marshall). The age of I. wellmani at its type locality is, therefore, considered Teurian (Danian-Landenian),
based on fossil dinoflagellates and pollen.

Discussion. Prismatic shell in the Broken River Formation has, until now, been assumed to
represent Inoceramus matotorus (Gage 1970). While little is known about the shape of I. matotorus ,
it may be distinguished from Isognomon wellmani by its huge adult size, juvenile ornament of
irregular commarginal plicae (which do affect the internal mould), adult ornament of weak relatively
regular frills (Wellman 1959, fig. 1), and nature of the ligament area (see earlier discussion, and
description in Appendix). At present the two species cannot be distinguished simply from details
of the prismatic shell structure, although Inoceramus matotorus appears to have more uniform and
regular hexagonal rod-type fibrous prismatic shell than Isognomon wellmani (terminology after
Carter and Clark 1985).

The presence of Isognomon at Broken River has been suggested previously by Sir Charles
Fleming (in unpublished faunal lists), based on hinges observed in situ at the type locality of /.
wellmani , and material collected in this region by McKay in 1880 (GS 6620 and 67 respectively);
and by Professor J. D. Campbell (written pers. comm. 1986). However, no Cretaceous or Palaeocene
Isognomon have hitherto been described from New Zealand, and furthermore, relatively few have
been documented overseas.

Comparisons with other species are hindered by the poor preservation of both the material being
described and much of that being compared. In addition, members of this genus can show extreme
morphological variability. For example, Fischer-Piette (1976, pi. 1, 2) illustrated the huge range
in form of Recent I. isognomum (Linnaeus, 1758) from a single population, and proposed a
remarkable synonymy list (containing seventy-eight species names) for that bivalve. Furthermore,
Duran-Gonzalez et al. (1984) documented considerable genetic variation between geographically
separated populations of Recent I. alatus (Gmelin, 1791). Even the small sample of individuals

978

PALAEONTOLOGY, VOLUME 31

being described here show a marked variation in morphology. Hence it is with caution that I.
wellmani is described as a new species, and discovery of more and better preserved material may
show that this form is indistinguishable from, and perhaps conspecific with, a number of other
species mentioned below.

I. wellmani differs from I. rekohuensis (described herein) by its smaller size, its possession of a
posterior wing, and its coarser ligament area structure. It resembles some Cretaceous and
Palaeogene species from Australia, North America, Europe, USSR, and Japan in the outline of
internal moulds, but is distinguished by its considerably larger dimensions.

/. wellmani is, however, very similar in size, shape, and ligament structure to I. ricordeana
(Orbigny, 1845) (pp. 494-495, pi. 399, figs. 1-3; illustrated also in Woods 1905, figs. 16-18) from
the Neocomian of Europe, and I. sanchuensis (Yabe and Nagao, 1926) (p. 57, pi. 12, figs. 1-4)
from the Aptian-Albian of Japan. The former possesses a more projecting umbo, while the latter
appears to be more strongly invaginated on the anterodorsal margin. In addition they are both
considerably older than the present record of I. wellmani.

Of the few Palaeogene Isognomon described, 7. bazini (Deshayes, 1860) (pi. 76, figs. 1-2; described
in Deshayes 1861, p. 57) from the Thanetian of the Paris Basin most resembles I. wellmani. I.
bazini is slightly smaller, has a straighter anterior margin with a less produced umbo, lacks a
posterior wing, and has more numerous resilifers than I. wellmani.

Isognomon ( Isognomon ) rekohuensis sp. nov.

Plate 89, fig. 1 a-e

vp. 1976 Inoceramus opetius Wellman, 1959; Speden (p. 385, fig. 1).

Name. Derived from the Maori name for the Chatham Islands: Rekolnt.

Material. A single articulated bivalved specimen with all shell material preserved (though partly recrystallized).

Type locality. CH/f213, 772202 (imperial grid reference NZMS 240/298673), GS 1 1521: from the Kahuitara
Tuff (Hay et al. 1970; Austin et al. 1973; Campbell et al., 1988) in the northern half of the bay immediately
south of Kahuitara Point, Pitt Island, Chatham Islands (see text-fig. I). Collected by H. R. Katz and P. Hill,
1975.

Type specimen. Holotype: TM 5453.

Diagnosis. Isognomon of large size, mytiliform and acline; ligament area smooth dorsally, and
carrying approximately twenty-six resilifers ventrally, resilifers of variable size and becoming
alternately wide-shallow and narrow-deep towards the posterior.

EXPLANATION OF PLATE 89

Fig. 1 a-e. Isognomon ( Isognomon ) rekohuensis n. sp. TM 5453, CH/f213, holotype; Kahuitara Tuff, Kahuitara
Point, Pitt Island, Chatham Islands (Late Cretaceous); a, external view of right valve, section of prismatic
shell missing revealing irregular ribs on surface of inner shell layer, x 0-7; b, internal view of left valve,
xO-7; c, dorsal view of articulated specimen, anterior to right, xO-7; d , anterior face of articulated
specimen, dorsal to right, x0-7; e, ligament area of left valve, x 1-3.

Figs. 2-7. Segments of ligament areas of some New Zealand Jurassic Isognomonidae, dorsal-up in all figures.
All figs, x 1 -3.

Figs. 2-5. Isognomon ( Isognomon ) sp. 2, OU 14399a, F46/f 7 1 ; latex mould right valve, Mataura, Southland.
3, L(AU) 3614, H47/f001 ; left valve, Tuhawaiki, Southland. 4, TM 6701, H47/f6494; right valve, Tuhawaiki,
Southland. 5, OU 143996, F46/f 7 1 ; latex mould right valve, Mataura, Southland.

Fig. 6. I. (Mytiloperna) sp. A. L(AU) 3413a, R16/H71; left valve, Kairimu Valley, south-west Auckland.
Fig. 7. Isognomon (M.l) sp. B. TM 4062, R15/f8006, paratype; left valve, Ururoa Point, south-west Auckland.
All specimens whitened with ammonium chloride sublimate.

PLATE 89

CRAMPTON, Isognomon ( Isognomon ), Isognomon ( Mytiloperna )

980

PALAEONTOLOGY, VOLUME 31

Description. Dimensions (in millimetres, refer to text-fig. 2):

Specimen Length Height Width Obliquity

(both valves)

TM 5453 113 142 59 55°

Shell large; mytiliform; acline; prosogyrous, umbo terminal and projecting beyond rest of anterior margin;
equivalve; moderately inflated. Dorsal and posterior outlines form unbroken curve with more convex ventral
margin. Anterior outline straight ventrally, concave dorsally. Anterior auricle and posterior wing absent.
Wedge-shaped in longitudinal profile, maximum width at anterior margin and approximately one third of
the way below the hinge line. Anterior face recurved from umbonal carina (although on the holotype this
face may have been depressed during preservation). Disc gently convex to planar. Commissure planar, byssal
gape narrow (although it is unclear whether this has resulted from the deformation suggested above). Small
ear-like projections on anterior margins either side of the commissure (these are not auricles since they do
not support and extend the hinge line).

External ornament of closely spaced, irregular, fine, commarginal lamellae. Interface between prismatic
and nacreous shell layers carries weak asymmetrical plicae and is lamellose in places (PI. 89, fig. 16).

Hinge edentulous, ligament multivincular (PI. 89, fig. le). Ligament area flat and inclined a few degrees to
plane of commissure; somewhat undercut close to umbo. At least twenty-six rectangular, shallow, concave-
floored resilifers which breach and weakly crenulate (judged from the shape of growth lines) the sigmoid-
shaped ventral margin of the area (refer to text-fig. 3). Resilifers, separated by very narrow angular ridges,
are differentiated posteriorly into alternate shallow wide (2-5 mm) and deeper narrow (1 mm) pits, this
differentiation decreasing close to the umbo. They carry an ornament of fine transverse growth lines which
vary between concave-up and convex-up on adjacent pits. The ligament area achieves a maximum height of
approximately 18-5 mm, although the resilifers extend over only the ventral half of this, being truncated
sharply, and leaving a smooth platform dorsally. Ligament attached to inner shell layer.

The posterior adductor muscle scar, shaped like an inverted comma, is situated midway between the dorsal
and ventral margins (PI. 89, fig. la). No pallial muscle scars are visible on the holotype.

The shell consists of an outer layer of polygonal rod-type fibrous prismatic calcite and an inner layer of
coarse granular calcite, presumed to be recrystallized nacreous aragonite originally forming the middle and
inner shell layers (discussed earlier). Over much of the disc the shell is somewhat less than 10 mm thick,
comprising a thicker prismatic layer towards the margins, and thicker inner layers close to the umbo. Near
the anterior margin the shell is approximately 20 mm thick. The structure of the prismatic shell layer is very
similar to that described for I. wellmanv. prisms achieve a maximum size of approximately 3-2 x 0-2 mm, they
are reclined, sigmoid-shaped, larger towards the inside surface than towards the outer, and formed into
discrete lamellae close to the outside surface. However, examination of I. rekotmensis prisms under SEM and
transmitted light revealed little substructure, and the uniform coincident extinction and apparent fracture
along cleavage planes may indicate diagenetic recrystallization.

Distribution. I. rekohuensis is known thus far from only the type locality. Fragments of prismatic shell from
elsewhere in the Kahuitara Tuff (CH/fl 1 and CH/fl la) may represent this species (although an undescribed
bakevelliid with thick prismatic shell also occurs in the Kahuitara Tuff). As with I. wellmani , the distribution
of I. rekohuensis may be difficult to gauge due to confusion with species of Inoceramus.

Age. Macrofossils in the Kahuitara Tuff were originally assigned to the lower or middle Cretaceous by
Boreham (1959). This unit was subsequently removed to the late Cretaceous based on Teratan-lowest
Haumurian (Senoman) dinoflagellates (Wilson 1976), probable Mata Series (Campanian-Maastrichtian)
palynomorphs (Mildenhall 1977), and poorly determinate Teratan to Haumurian foraminifera (Strong 1979).
With the referral of Inoceramus opetius to Isognomon rekohuensis , the only age-diagnostic macrofossil from
this formation is the Haumurian (latest Campanian-Maastrichtian) belemnite Dimitobelus hectori Stevens
(1965) from the localities on the north-west of Pitt Island (CH/f587, Rocky Side, and CH/f466, Flower Pot
Harbour).

A limestone filling cracks in the top of the Kahuitara Tuff at Flower Pot Harbour contains well-preserved
late Haumurian foraminifera (Strong and Edwards 1979). Radiometric analyses from the Southern Volcanics
and Whakepa Trachyte, which overlie and underlie the Kahuitara Tuff respectively, in the region of Kahuitara
Point, gave dates of 77-3 + 1 my and 79-0+1 my (Grindley et al. 1977). These dates correspond to mid to
late Piripauan (mid Campanian), according to the timescale of Stevens (1981).

Hence it seems likely that the Kahuitata Tuff is no older than Teratan (Coniacian), no younger than late
Haumurian (late Maastrichtian), and is Piripauan (Campanian) at the type locality of I. rekohuensis.

CRAMPTON: ISOGNOMONIDAE, INOCERAMIDAE, AND R ETROCER AM I D AE 981

Discussion. The holotype of I. rekohuensis was referred to Inoceramus opetius by Speden (1976).
The external features of the two species are contrasted in the family discussion above. In
addition, I. opetius has a variable ligament area with commonly multilobate or multiserial resilifers
(see PI. 90, figs. 6-8) which are constructed on the outer prismatic shell layer. The shell of
I. opetius is much thinner than that of Isognomon rekohuensis , the prismatic layer being only
about 1-5 mm thick on the disc of a large specimen, and having a polygonal simple prismatic
structure.

/. rekohuensis is readily separated from most other Late Cretaceous and Palaeogene Isognomon ,
including I. wellmani (described herein) by its large size and delicate resilifers. However, I. williardi
Stephenson (1923, pp. 125-126, pi. 23, figs. 1-3; pi. 24, figs. 1-2; pi. 25, fig. 3), from the Senonian
of eastern United States, is remarkably similar to the present species, but its resilifers are not
truncated dorsally and its ligament area is not undercut (although the latter is not a reliable
character, see for example Isognomon (/.) sp.; PI. 89, figs. 2 and 5).

The distinctive form of the ligament area seen in I. rekohuensis and I. williardi , as well as their
shape, resemble members of the subgenus I. (Hippochaeta) Sangiovanni, 1844, notably I. (H.)
sandhergeri (Deshayes in Sandberger 1863, p. 367, pi. 31, figs. 4-4a; and well illustrated in Ludwig
1864, pi. 13, fig. 1; pi. 14, figs. 1-3; pi. 15, figs. 1-lc) from the Middle Oligocene of France, and
I. (//.?) lamarcki (Deshayes, 1830) (p. 284; illustrated in Deshayes 1837, pi. 40, figs. 7 and 8) from
the Bartonian of France. However, in most species of I. ( Hippochaeta ) the area is very high, the
differentiation of broad shallow resilifers and deep narrow grooves is much more marked, these
grooves bifurcate dorsally in many species, and they appear to reach the dorsal margin of the area
in all species. (These features are not well developed in I. (//.?) lamarcki , and hence its referral to
this subgenus is queried.) Thus far, I. ( Hippochaeta ) is known only from Eocene to Pliocene rocks
(Cox 1969).

Family retroceramidae Pergament in Koschelkina, 1971

Type genus. Retroceramus Koschelkina, 1959. (The subgenus Inoceramus ( Retroceramus ) was first proposed
informally by Koschelkina (1957), and was validated by Koschelkina (1959) with the designation of a type
species (Crame 1982). The subgenus was elevated to generic level by Koschelkina (1962).)

Diagnosis. Variously shaped edentulous multivincular Pteriacea bearing commarginal plicae that
are large and regularly spaced in nearly all taxa, and in which the ligament is fixed to the nacreous
shell layer.

Discussion. Other characters typical, though not necessarily diagnostic, of this family include
marked obliquity of the valves (see text-fig. 5 d)\ subterminal, moderately to strongly gibbous
umbones; moderate to high angles between the planes of the ligament area and commissure; a flat
to weakly concave ligament area; relatively broad rectangular to sub-ovate resilifers which in
all (?) taxa breach and in most taxa crenulate the ventral margin of the area (refer to text-fig. 3;
PI. 90, figs. 1-3); and resilifer interspaces which are typically broad and concave. In some taxa
these interspaces are sufficiently deep to appear as a second class of resilifer, resulting in alternat-
ing broad-deep and narrow-shallow pits separated by narrow angular ridges.

The family-group name Retroceramidae was first published by Koschelkina (1971), who
attributed authorship to Pergament (1969, unpublished). Koschelkina’s description of the type
genus, Retroceramus , follows:

Shell equilateral or practically equilateral, with uneven sides, elongated along the axis of growth from the
beaks, which are near the anterior margin, but not terminal. Sculpture concentric, less often radial. Prismatic
and nacreous layers well developed. Ligament platform located upon nacreous layer. In adult forms it consists
of ligamental pits and ridges varying in outline. Posterior muscle— adductor large, anterior— strongly reduced.
Mantle line discontinuous

Lower Jurassic(?) Mainly in Middle Jurassic of Boreal province. Less numerous in Upper Jurassic. Lower
Cretaceous?

982

PALAEONTOLOGY, VOLUME 31

The name Retroceramidae has subsequently been used by Koschelkina (1980) and Pokhialaynen
(1985). Characters, facies relationships, and inferred life habits of Retroceramus, thus far the only
genus referred to Retroceramidae, are described by Koschelkina (1963, 1969, 1971) and Crame
(1982).

NOTES ON LIFE HABITS OF ISOGNOMON WELLMANI AND I. REKOHUENSIS

Recent Isognomon are physiologically tolerant filter-feeding byssate bivalves found in tropical and
subtropical littoral or inner shelf low- to high-energy marine and estuarine environments. They
typically live epifaunally in crowded beds attached by massive byssi to hard surfaces, and oriented
vertically (ventral up) or with their right valves against the substrate. Less commonly they are
found on or within soft substrates. (For accounts of the ecology and biology of Recent species of
Isognomon see Read 1964, Yonge 1968, Siung 1980, and Reid 1985.) Similar life habits for I.
wellmani and I. rekohuensis cannot be assumed since they are considerably larger, more inflated,
and thicker shelled than Recent species. Fiirsich (1976, 1980, 1981 ) and Fiirsich and Werner (1986)
inferred that a number of fossil species from the Jurassic of Europe lived close to shore, were setni-
endobyssate in generally fine-grained sediments, and were probably opportunistically euryhaline,
forming clusters and banks in hypersaline to mesohaline environments (i.e. hypersaline lagoons to
brackish bays).

While few data are available on fossil-lithofacies relationships of the new species, sediments and
faunas at both type localities suggest deposition in moderate- to high-energy shallow marine

EXPLANATION OF PLATE 90

Figs. 114. Segments of ligament areas and/or umbones of some New Zealand Jurassic Retroceramidae and
Cretaceous Inoceramidae. All figures dorsal up, x 1-3. All specimens whitened with ammonium chloride
sublimate.

Fig. 1 . Retroceramus (Retroceramus) galoi( Boehm, 1907). TM 6719, R 1 5/f 8546; right valve, Kawhia Harbour,
south-west Auckland.

Fig. 2. R. (R.) haasti (Hochstetter, 1863). TM 6720, R 1 5/f 8564; left valve, Kawhia Harbour, south-west
Auckland.

Fig. 3. R. (R.) cf. subhaasti (Wandel, 1936). TM 5774, R 1 5/f 80 1 2; latex mould right valve, Kawhia Harbour,
south-west Auckland.

Figs. 4 and 5. Inoceramus rangatira Wellman, 1959. Y19/f7494, Hapuku River, Marlborough. 4, TM 6712
umbo (umbonal septum directed into page) of right valve. 5, TM 6711, umbo and umbonal septum of left
valve.

Figs. 6-8. I. opetius Wellman, 1959. 6, TM 6708, W22/f8504, right valve, Waimarama, Hawke’s Bay. 7,
TM 6707, V23/fl6; left valve, Mangakuri River, Hawke’s Bay. 8, TM 6709, U25/f6462; valve unknown,
Akiteo River, Wairarapa.

Fig. 9. I. concentricus Parkinson, 1819. OU 4056, P30/f6551; left valve, Cover Creek, Marlborough.

Fig. 10. I.fyfei Wellman, 1959. TM 2114, X 1 6/f9539, holotype; latex mould right valve, Motu River, East
Cape.

Fig. 1 1. Inoceramus sp. A. TM 6716, W22/f8504; left valve, Waimarama, Hawke’s Bay.

Fig. 12. I. bicorrugatus Marwick, 1926. TM 6704, Y14/f7850; right valve, Waikura River, East Cape.

Fig. 13. I. australis Woods, 1917. TM 6703; plaster cast right valve, Gisborne district.

Fig. 14. Inoceramus sp. B. TM 6717, Z 1 4/f 1 06; valve unknown, Taurangakautuku Stream, East Cape.

Figs. 15 18. Prismatic shell layer of Isognomon ( Isognomon ) wellmani n. sp. Broken River Formation, Broken
River, Canterbury. 15 and 16, TM 6697, K34/f48; photomicrographs, plain polarized light, radial thin
section from disc of shell, outside to top, margin to left. 15, entire thickness of prismatic layer, x 35.
16, details of prisms showing transverse tabulae interpreted as accretion lines, x216. 17, SEM, oblique

to long axis of prisms, showing granular substructure of most prisms, x 546. 18, SEM, perpendicular to
long axis of prisms, etched sample (90 seconds, 0-5% HC1), showing block-like substructure of prisms,
x 762.

PLATE 90

CRAMPTON, Retroceramus , Inoceramus

984

PALAEONTOLOGY, VOLUME 31

environments. (The bivalved condition of specimens from both places indicates they were not
significantly transported prior to burial.) The Broken River Formation is non-marine at the base,
passing up into a medium to fine sandstone interpreted as an inner shelf deposit with thick shell-
beds (containing Ostreidae and Isognomon) developed on an offshore bar system (Browne and
Field 1985). Such an interpretation is consistent with thick reef-like accumulations of Ostreidae,
which are found today in estuaries and on shallow offshore shelves subject to moderate energy
conditions. Similarly, the Kahuitara Tuff, comprising coarse tuff, conglomerate, and breccia, may
be non-marine at the base (Huy et al. 1970), and contains a diverse marine fauna characteristic of
an epifaunal habit in a high energy inner shelf to subtidal environment (Speden 1976) and
foraminifera typical of near normal salinity and depths of 5-50 m (Strong 1979).

Sedimentary relationships suggest that I. wellmani and I. rekohuensis lived on sandy or shelly
substrates. Furthermore, both species are thick-shelled, particularly close to the dorsal and anterior
valve margins: a stabilizing strategy common in secondary soft-bottom dwellers (Seilacher 1984).
Stanley (1972) described morphologic adaptations of soft substrate byssate bivalves to epifaunal
and infaunal life habits. He concluded that endobyssate bivalves can be recognized by their elongate
prosocline shape, dorsoanteriorly lobate shell, broad byssal sinus, and absence of appreciable
anterior flattening (NB Stanley used the term ‘ventral flattening’, based on the orientation of the
shell with respect to the substrate). Neither species described here displays any of these characters,
and they may both therefore appear to have been epibyssate. However, Fiirsich (1980) documented
the apparent preserved life positions of three Jurassic Isognomon species, which, contrary to
theoretical predictions, must have been partly infaunal to maintain their vertical ‘mudsticking’
attitudes: umbo downwards, hinge line oblique to bedding, in a manner similar to Recent Pinna
(terminology of Seilacher 1984). Seilacher (1984), on the other hand, interpreted these preserved
positions as the result of ‘unnatural’ rotation on the byssus as the normally epifaunal animals
responded to burial.

Morphology, then, cannot necessarily be used to determine life positions of I. wellmani and /.
rekohuensis. The former occurs in a densely packed bed of large Ostreidae, and it most probably
lived epifaunally, attaching to, and providing attachment for, other bivalves. It may either have
rested on the right and left valves, using the posterior wing as a stabilizer (an ‘outriggered recliner’),
or on the flattened anterior face (an ‘edgewise recliner’, see Seilacher 1984, fig. 5). Spatial
competition in such a situation may account for the intraspecific morphological variation seen in
this species. The holotype of I. rekohuensis , on the other hand, was the only specimen found in
the outcrop. Its shape, and the presence on both valves of serpulids and possibly clionid sponges
(represented by abundant fine borings), are consistent with an edgewise reclining or semi-infaunal
mudsticking life position (see Fiirsich 1980, fig. 9).

In summary, I. wellmani and I. rekohuensis probably lived in marginal marine to inner shelf,
moderate- to high-energy marine environments which hosted faunas dominated by epifaunal
cemented and bysally attached suspension-feeding organisms. I. wellmani was probably an
epibyssate outrigger or edgewise recliner, attaching to other shells, while I. rekohuensis may have
been an epibyssate edgewise recliner or a semi-endobyssate mudsticker, attaching to sediment or
shell particles.

SUMMARY AND CONCLUSIONS

Fossil Isognomonidae can be difficult to distinguish externally from Inoceramidae, a problem
which has resulted in erroneous age determinations. Differences between these families are
summarized in Table 2. Internal details of ligament area morphology are characteristic at the
family level. Definitive distinction between these two families, however, is apparently furnished by
the shell structure underlying the ligament area (a character easily determined from whole shells
or thin sections). In Isognomonidae the ligament attaches to the inner (presumed nacreous) shell
layer, whereas in Inoceramidae it attaches to the outer prismatic shell layer. Retroceramus , formerly
included in Inoceramidae, has the ligament attached to the inner shell layer. The family

table 2. Summary of principal differences between Inoceramidae, Retroceramidae, and Isognomonidae.

CRAMPTON:

ISOGNOMONIDAE, INOCERAMIDAE, AND RETROCERAMIDAE

985

<D

P

JO

'5

o

P

d

T3

'5b

Tb

c

o

13

p

<L>

O

2 'p

P dJ

C O
0) cj

x> .6
§

t4 u 3) “

cj C _cj g
Cu 7j Ph o

H H

C3

.0 c

2 c

Ki 2

c

■o >■> — T OJ

y u .2 £

.5 > h

44

d

d

p *5 2
23 o 2
£ o

£ if

C/5 C/5

O O

22
g 3

a .S
^ >

£ 2
c£ p

f— 1 _j
dJ 03

o
c
o
■o £

c

§ £
£ <u

• C/5 C/5

c c/5 dj

o •£ *-

j-h § p

o £ *

o 03
13 o ^

C*_ c/)

2 0 2
d <u
Ph p
d

0

P P
-< Oh

P . „

P c3 P
u 0) P

- S £

b P-

2 ^ ^
p o ^
t>o _r

c c ^
5fl 2

P

2 S
£ o

P rrf

J- P
Dh C/5

cj P--P

> bfj

03 O '2
o -P -P

d p3
cj £

a p b

>7 <D P

-S' > c

00 "

C C

o •-

T3

p

, p
p 2

2 5
‘5b _

^H . — I

p p

s >,
£ ^
o 2

is o o

- y §

<al s

O- „

-S' o Id

o c

O 6

p

C* O

c b
cu ^p

2 "3

p f?

p 2

p- 3

>».£ C/5

w cj <5
p

j-T rp

p .r c/5

-p cj cj

O c3
P CJ • -

> g 2

• •— j O 3

s u «

22 |
P P
P ?J

I <H—

P o

0

CJ C/5

'•a £
§ <->

2 8
rs CJ
4J dj

^ ^ .

1 *
£> a*
0> ^

P cj

x 55

p S

a.s

w p
O <u

P tH
> P
O c^

O o

t; .s

2 ^

3 c3 ^ p g £

W) g ^ ^ -

p

CJ J-H

oj ti
V-( P

r dJ

-p >

E y :-=

OJ ^

OJ g

dj cj
P P
P

Ph

(DXJ

P

o

CJ

<

Ph

P

X

C/5

P

O

x>

£

P

Ph

P

W)

p

CQ c/5

P

60

dJ

P

P

o

3 -p P

2 pi CJ
P P cj

^ 2 S p

_ P p CJ
p dj to i:
cjpp p

C/5

986

PALAEONTOLOGY, VOLUME 31

Retroceramidae Pergament in Koschelkina, 1971, should therefore be used to accommodate those
multivincular Pteriacea which bear typically strong, regular, commarginal plicae and which have
the ligament attached to the inner shell layer. Thus far, only Retroceramus is referred to
Retroceramidae.

The present data are consistent with a polyphyletic origin for the multivincular ligament in
Inoceramidae and Isognomonidae; the evolution of Inoceramidae from Atomodesma; the removal
of Inoceramidae from Pteriacea to Ambonychiacea; and a close relationship between Retrocera-
midae and Isognomonidae.

At the species level, details of ligament area morphology are valuable, if not essential, for
discriminating between similar and, in many cases, morphologically highly variable species within
families Isognomonidae, Retroceramidae, and Inoceramidae.

Many Mesozoic and early Tertiary Isognomon , like Recent species, occupied inner shelf to
marginal marine environments. However, unlike Recent forms, it seems they were more typically
epibyssate on or semi-endobyssate in soft or shelly substrates. This difference in life habits accounts
for the stabilizing structures common in fossil Isognomon (for example, I. wellmani and I.
rekohuensis), such as a large anteriorly and dorsally thickened shell, elongate posterior wing, and
broad flat anterior face.

APPENDIX. LIGAMENT AREA MORPHOLOGIES OF SOME NEW ZEALAND
JURASSIC AND CRETACEOUS ISOGNOMONIDAE, RETROCERAMIDAE, AND

INOCERAMIDAE

These descriptions are based on few specimens of each taxon and are intended as introductory notes (to
facilitate the present discussion) pending more complete population-based taxonomic studies. Consequently,
names by which some specimens are identified may require future revision. Brief descriptions of ligament
area (refer to text-fig. 3 for an explanation of descriptive terms) and pertinent taxonomic comments are
followed by repository catalogue and Fossil Record File numbers (abbreviations explained in the Introduction),
relevant details of the whole specimens, locality information with grid references (where available), collector(s),
and ages (ages are bracketed if based solely on the species under discussion). Map sheet areas of New Zealand
fossil localities referred to are shown on text-fig. 1.

Superfamily pteriacea Gray, 1847; nom. transl. Dali, 1894 (ex Pteriidae; = Aviculidae Goldfuss, 1820 (see

earlier) )

Family isognomonidae Woodring, 1925
Genus isognomon Solander in Lightfoot, 1786
Subgenus isognomon Solander in Lightfoot, 1786

Type species. Ostrea isognomon Linnaeus, 1764 (see earlier).

Isognomon ( Isognomon ) sp.

Plate 89, figs. 2-5; text-figs. 4a and 5b

Ligament area morphology apparently very variable (compare PI. 89, figs. 2-5, specimens from the same
population). Area flat, at least six to seven broad concave rectangular resilifers (approximately 2-4 mm wide)
separated by narrower flat to moderately concave ridges (approximately 1 -7-2-5 mm wide), which on some
specimens resemble a second class of resilifer (PI. 89, fig. 4). Resilifers breach ventral margin of area on all
specimens, and strongly crenulate it on some. Height of area approximately 2-7 mm close to umbo, and
6 mm or more posteriorly. Ligament attached to inner shell layer.

Material. L(AU ) 3614, H47/H001, AU 11096. Plate 89, fig. 3. Shelly beak of LV. Jacks Bay, Tuhawaiki,
Southland; N. Hudson, 1986. Temaikan (Bajocian mid Callovian).

OU 14399a , F46/f071. Plate 89, fig. 2. Internal mould RV; length ~ 57 mm, height 72 mm. Stewart’s
Farm, near Mataura, Southland; M. C. Gudex?

OU 14399b, as for previous specimen. Plate 89, fig. 5. Internal mould RV; length > 56 mm, height 72 mm.

CRAMPTON: ISOGNOMONIDAE, INOCER AM IDAE, AND RETROCERAM IDAE 987

TM 6701, H47/f6494, GS 148. Plate 89, fig. 4. Partly shelly internal mould RV; length 54 mm, height
65 mm. Coast opposite Bloody Jacks Island, Tuhawaiki, Southland; A. McKay, 1873. Temaikan (Bajocian-
mid Callovian).

TM 6702 , H46/f6752, GS 7102. Not figured. Internal mould LV; length 43 mm, height 56 mm. Old coastal
face, south-west side of Jacobs Hill, Catlins River, Southland; H46 567104; I. G. Speden, 1957. Temaikan.

TM 6790, H47/f6494, GS 148. Text-fig. 4a. Two thin sections perpendicular to ligament area, approximately
half-way between umbo and posterior end of ligament area. Locality as for TM 6701 (above).

TM 6793, as for previous specimen. Text-fig. 5a. Partly shelly LV; length 48 mm, height 63 mm.

Subgenus mytiloperna Ihering, 1903
Type species. Perna americana Forbes in Darwin, 1 846.

Isognomon ( Mytiloperna ) sp. A
Plate 89, fig. 6.

This species is referred to Isognomon ( Mytiloperna ) on the basis of its prosocline shape, subterminal beak,
small size, lack of a distinct posterior wing, and small number of well-spaced resihfers (the first two criteria,
atypical of most Isognomonidae, characterize this subgenus). However, it does also resemble some forms of
Bakevilliidae King, 1850, notably Cuneigervillia Cox, 1954, and study of juvenile stages may reveal the
presence of hinge teeth characteristic of the latter genus.

Ligament area flat, nearly parallel to plane of commissure. Few (probably no more than five or six)
subrectangular resilifers, which narrow ventrally (from ~ 18 mm to ~ 1-2 mm), and breach but scarcely
crenulate the ventral margin of the area. Interspaces wider than resilifers (~ 2 mm to ~ 2-5 mm), flat or
weakly concave, most bounded by narrow upstanding rims. Area may become irregularly thickened and
extended ventrally, its height on two similar-sized individuals being ~ 1-5 mm and > 4 mm. Ligament
attached to inner shell layer.

Material. L(AU ) 3413, R16/H71, AU 4604. Plate 89, fig. 6. Shelly RV and ligament area of LV; length
(RV) > 30 mm, height ~ 25 mm. Paraohanga Stream, Kairimu Valley, Kawhia, south-west Auckland;
R16 662189; D. A. Francis. Heterian (Early Kimmeridgian).

L(AU) 3412, as for above. Not figured. Internal and external moulds of RV; length 35 mm, height 23 mm.

Isognomon ( Mytiloperna!) sp. B
Plate 89, fig. 7; text-fig. 5b

The specimen figured here, a paratype of Inoceramus ururoaensis Speden (1970n, pp. 836-842, figs. 12-20),
is tentatively referred to Isognomon ( Mytiloperna ) based on its prosocline shape, weakly developed subterminal
umbo, smooth to lamellose surface, and attachment of the ligament to the inner shell layer (see earlier
discussion of family characters). It is distinguished from Retroceramidae by the weakly developed umbo and
lack of commarginal plicae. This specimen, however, differs from typical I. (Mytiloperna) by being considerably
more obliquely elongate, lacking a distinct posterodorsal angle, having a convex anterior margin, and having
a strongly undercut ligament area. In addition the area (described below) of this specimen, while being similar
to I. (M.) ageroensis Hayami, 1957 (pp. 101-103, pi. 6, figs. 4-8), has relatively abundant and uniform
resilifers, which contrast with the well-spaced and somewhat irregular resilifers of most I. (Mytiloperna) and
strongly resemble those of Retroceramidae described herein. Hence it is with caution that this fossil is referred
to I. (Mytiloperna), although it is removed from Inoceramus with some confidence. It is not yet clear whether
Isognomon (M .1) sp. and the holotype of Inoceramus ururoaensis are conspecific.

Ligament area flat, slightly twisted so that it is subparallel to plane of commissure close to umbo, inclined
posteriorly. Probably no more than ten broad (2-3 mm) rectangular resilifers on figured specimen, which
breach and strongly crenulate ventral margin of area. Interspaces narrower (1-5-2 mm), weakly concave.
Area 5-5— 6-5 mm high. Ligament attached to inner shell layer.

Material. TM 4062, paratype, R15/f8006, ex Laws Collection. Incomplete LV; length > 75 mm, height
> 41 mm. 60 -240 m north-east of stack at Ururoa Point, south-west Auckland; RI5 648431. Ururoan
(Pliensbachian Aalenian).

988

PALAEONTOLOGY, VOLUME 31

Family retroceramidae Pergament in Koschelkina, 1971
Genus retroceramus Koschelkina, 1959
Subgenus retroceramus Koschelkina, 1959

Type species. Inoceramus retrorsus Keyserling, 1848.

Retroceramus ( Retroceramus ) aff. everesti (Oppel, 1865, p. 298)

Not figured here

Referred to Retroceramus by Crame (1982). Ligament area diminutive, poorly known. Ligament attached to
inner shell layer.

Material. L{AU) 3597 , R13/f6969, AU 4410. Internal mould, RV. Cliff in northern bank of Huriwai Stream
just east of confluence with south-flowing tributary, Port Waikato, Auckland; R13 645184; A. B. Challinor,
1969. Puaroan (Tithonian).

Retroceramus (Retroceramus) galoi (Boehm, 1907, p. 68, pi. 9, figs. 10-14; pi. 10, figs. 1 and 2)

Plate 90, fig. 1; text-fig. 5d

Referred to Retroceramus by Crame (1982). Ligament area steeply inclined (~ 45°) to plane of commissure,
very weakly concave. Broad rectangular or elongate-ovate resilifers (probably no more than nine or ten on
figured specimen) ~ 2 mm wide separated by narrower concave interspaces (~ I mm wide, these concave
interspaces constituting a second class of resilifer, according to Koschelkina 1969). Resilifers scarcely breach
crenulated ventral margin of 4-7 mm high area. Ligament attached to inner shell layer.

Material. TM 6719 , R15/f8553 (considered the same as R15/f8546), GS 5944. Partly shelly internal mould
RV; length 43 mm, height 41 mm. Point west of Heteri Promontory, across Waikutakuta Inlet, Kawhia
Harbour, south-west Auckland; R15 659401; K. J. McNaught, 1953. Heterian (Early Kimmeridgian).

Retroceramus ( Retroceramus ) haasti (Hochstetter, 1863, p. 190)

Plate 90, fig. 2; text-fig. 4b

Referred to Retroceramus by Crame (1982). Ligament area moderately inclined (~ 20°) to plane of
commissure, very weakly concave. Broad (3-5 mm) square to ovate resilifers breach crenulated ventral margin
of area. Resilifer interspaces broad (~ 17 mm) and markedly concave dorsally, narrow (< 1 mm) ventrally.
Area 4 mm high. Ligament attached to inner shell layer.

Material. TM 6720 , R 1 5/FB564, GS 5955. Plate 90, fig. 2. Partly shelly internal mould LV; length 71 mm,
height 84 mm. North side of Kowhai Point, from west of tip for ~ 50-100 m east, Kawhia Harbour, south-
west Auckland; R15 67434076; K. J. McNaught, 1953. Lower Ohauan (mid Kimmeridgian).

TM 6792, as for previous specimen. Text-fig. 4b. Thin section perpendicular to ligament area, approximately
half-way between umbo and posterior end of ligament area.

Retroceramus (Retroceramus) marwicki (Speden, 1970a, pp. 842-850, figs. 22-34)

Not figured here

Here referred to Retroceramus. Ligament area diminutive. Resilifers shallow, approximately square, 1-5 mm
wide, breach and crenulate ventral margin of area. Interspaces narrower (~ 1 mm wide), flat or slightly
concave. Area ~ 1-5 mm high. Ligament attached to inner shell layer.

Material. TM 4052 , holotype, R 17/F8636, GS 7886. Partly shelly internal and external moulds RV; length
~ 35 mm, height 43 mm. West side of Rauroa Stream, at the back of the flood plain, 480 m upstream of
ford on Tuamatamairie Road, south-west Auckland; R17 65068961; I. G. Speden, G. R. Stevens, 1961.
Upper Temaikan (late Bathonian-Callovian).

CRAMPTON: ISOGNOMONIDAE, INOCERAMIDAE, AND RETROCERAMIDAE 989
Retroceramus ( Retroceramus ) cf. subhaasti (Wandel, 1936, p. 469, pi. 15, fig. 2; pi. 16, fig. 5a, b)

Plate 90, fig. 3

Referred to Retroceramus by Crame (1982). Ligament area low and almost perpendicular to plane of
commissure. Resilifers shallow, square, breach but only weakly crenulate ventral margin of area, and only
slightly wider and deeper than concave interspaces (i.e. the ‘interspaces’ seem to form the second class of
resilifer described by Koschelkina 1969). Width of resilifers ~ 2-4 mm, width of interspaces ~ 1-7 mm, height
of area 2 mm. Ligament attached to inner shell layer.

Material. TM 5774, R 1 5/f80 1 2, GS 9937. Partly shelly internal mould LV; length > 68 mm, height ~ 65 mm.
Old Kihi Road, Hauturu, Kawhia, south-west Auckland; R15 864420; G. R. Stevens, I. W. Keyes, 1968.
Heterian (Early Kimmeridgian).

Subgenus fractoceramus Koschelkina, 1959
Type species. Inoceramus formosulus Voronetz, 1937.

Retroceramus ( Fractoceramus ) inconditus (Marwick, 1953, p. 93, pi 13, fig. 13)

Text-fig. 5c

Here referred to Retroceramus. This species is referred to subgenus Fractoceramus on the basis of its weak
and irregular ornament which is atypical of the genus as a whole. A single specimen with poorly-preserved
ligament area shows the ligament attached to the inner shell layer.

Material. L(AU) 3598, NC/T435, AU 7266. Not figured. Internal mould LV. South end of west coast of Uitoe
Peninsula, New Caledonia; La Tontouta 0615775556; J. A. Grant-Mackie, 1975. Temaikan (Bajocian-
Callovian).

TM 2373, holotype, F47/f7492, GS 2998. Text-fig. 5c. Internal mould LV; length ~ 45 mm, height 32 mm.
Quarry Hills, Waikawa district. Southland; F47 060001; R. A. S. Browne, 1944. Temaikan (Bajocian-
Callovian).

Superfamily ambonychiacea? Miller, 1877; nom. transl. Newell, 1965 (ex Ambonychiidae)

Family INOCERAMIDAE Giebel, 1852; nom. transl. Steinmann, 1903 (ex Inoceraminae)

Genus inoceramus Sowerby, 1814
Subgenus indeterminate

Inoceramus australis Woods (1917, pp. 27-28, pi. 13, figs. 1-3)

Plate 90, fig. 13

Ligament area strongly concave with more than twenty-six (possibly twice this number) deep, sigmoid-
shaped, elongate-ovate to ovate resilifers which are slightly oblique to the hinge line and become deeper and
less elongate towards the posterior. Interjacent ridges upstanding, well defined, angular to rounded, higher
and broader ventrally, ~ 2-5 mm between crests. Resilifers shallow steeply close to and breach but do not
crenulate ventral margin of area. Area at least 7 mm high. Ligament attached to outer shell layer.

Material. TM 6703, GS 8385. Articulated specimen; length ~ 140 mm, height ~ 170 mm. Gisborne district,
locality not known. (Piripauan (Campanian).)

Inoceramus bicorrugatus Marwick (1926, pp. 380 -381, fig. 1)

Plate 90, fig. 12

Ligament area known, so far, from a single large poorly preserved specimen. Many ovate to square resilifers
of moderate depth separated by narrow upstanding ridges, ~ 2-5 mm between crests. Depressed smooth
platform ventral to resilifers. Total height of area > 1 4- 5 mm, resilifers occupying dorsal ~ 4-5 mm. Ligament
attached to outer shell layer.

Material. TM 6704, Y14/f7850, GS 11601. Articulated specimen; length ~ 280 mm, height > 370 mm.
South-western tributary of Waikura River, East Cape; Y14 582736; R. T. Farmer, G. W. Grindley, 1975.
(Mangaotanean (Turonian).)

990

PALAEONTOLOGY, VOLUME 31
Inoceramus concentricus Parkinson (1819, pp. 58 59, pi. 1, fig. 5)

Plate 90, fig. 9

Moderately concave ligament area with > 12 shallow elongate-ovate resilifers separated by low rounded
ridges. 1-5 mm between ridge crests close to umbo, > 2-5 mm posteriorly. Resilifers barely breach and do
not crenulate ventral margin of area. Height of area ~ 5-3 mm. Growth lines may be strongly formed to
give stepped appearance. Area smooth beneath umbo. Ligament attached to outer shell layer.

Material. OU 4056 , P30/f6551, GS 5815. Plate 90, fig. 9. Incomplete articulated specimen; length ~ 70 mm,
height ~ 80 mm. Cover Stream 180 m upstream from junction with Wharf Stream, Marlborough; P30
826173; R. A. Cooper, 1953. Ngaterian (late Albian-early Cenomanian).

TM 6706 , P30/f 1 93, GS 14017. Not figured. Partly shelly LV; length > 88 mm, height > 120 mm.
Wharekiri Stream, Marlborough; P30 748912; I. G. Speden, M. G. Laird, 1981. (Ngaterian (late Albian-
early Cenomanian).)

Inoceramus fyfei Wellman (1959, p. 157, pi. 11, fig. 5)

Plate 90, fig. 10

On the holotype, the weakly concave and longitudinally undulose ligament area is steeply inclined to the
plane of commissure and underlain by a smooth platform lying approximately parallel to the commissure.
At least thirteen ovate, moderately deep resilifers separated by narrow upstanding ridges which are peaked
at their ventral ends. Approximately 1-5 mm between ridges, area ~ 4 mm high. Ligament attached to outer
shell layer.

Material. TM 2114 , holotype, X16/f9539, GS 6277. Distorted internal mould RV; length ~ 55 mm. Mill
Road, end branch road from No. I Quarry, Motu River, East Cape; XI 6 14171736; G. J. Lensen, 1956.
(Ngaterian (late Albian-early Cenomanian).)

Inoceramus matotorus Wellman (1959, p. 155, pi. 10, fig. 1)

Not figured here

The holotype of this specimen displays a small part of the concave ligament area which carries shallow
elongate-ovate to rectangular resilifers similar to those of Inoceramus sp. A illustrated in Plate 90, fig. 10.
Approximately 2 mm between ridges, area 7 mm high. Ligament attached to outer shell layer.

Material. TM 2110, holotype, Y16/f7489, GS 1604. Distorted bivalved specimen; height ~ 220 mm. Lower
part of Ihungia Stream, East Cape; M. Ongley, 1922. (Haumurian (Maastrichtian).)

Inoceramus opetius Wellman (1959, pp. 155-156, pi. 10, fig. 3)

Plate 90, figs. 6-8; text-fig. 4c

Weakly convex to weakly concave ligament area of variable form, parallel or inclined to plane of commissure.
Resilifers monoserial, multilobate, or multiserial, probably no more than three rows of pits. Many columns
of shallow resilifers, individual pits ovate or scooped where multiserial, otherwise formed into extended
narrow troughs which are more or less lobate and separated by fine ridges which pinch and swell.
Approximately 1 mm between ridge crests, area at least 8-5 mm high. Ligament attached to outer shell layer.

Material. TM 6707, V23/f 1 6, GS 13069. Plate 90, fig. 7. Partly shelly LV; length ~ 80 mm, height ~ 107 mm.
Castle Hill Station, Mangakuri River, southern Hawke’s Bay; V23 419290; R. D. Black, 1981 . Mangaotanean
Teratan (Turonian-Santonian).

TM 6708 , W22/f8504, GS 3225. Plate 90, fig. 6. Partly shelly incomplete RV. Approximately 3 km south
of Waimarama, southern Hawke’s Bay; J. D. H. Buchanan, 1983. (Teratan (Coniacian-Santonian).)

TM 6709, U25/f6462, GS 118. Plate 90, fig. 8. Isolated section of ligament area. Akitio River, eastern
Wairarapa; J. Hector, J. D. Enys, A. McKay, 1873-1875. (Teratan (Coniacian Santonian).)

TM 6791, P30/f6895, GS 9047. Text-fig. 4c. Thin section perpendicular to ligament area approximately
half-way between umbo and posterior end of ligament area. Middle branch of Wharf Stream, approximately
1-2 km upstream from junction with south-east branch, Marlborough; P30 862170; W. D. M. Hall, 1962.
(Teratan (Coniacian-Santonian).)

CRAMPTON: ISOGNOMONIDAE, INOCERAMIDAE, AND RETROCE R AM I D AE 991

Inoceramus pacificus Woods (1917, p. 28, pi. 14, figs. 1-2)

Not figured here

A single specimen with a poorly preserved ligament area shows that the ligament attached to the outer shell
layer.

Material. TM 6710, 029/f9863, GS 9355. Partly shelly internal mould RV. Ribble Stream, Awatere Valley,
Marlborough; 029 612244; G. J. Lensen, 1964. Piripauan (Campanian).

Inoceramus rangatira Wellman (1959, p. 156, pi. 10, fig. 4)

Plate 90, figs. 4-5

This species differs from all those described thus far by possessing a thick umbonal septum. On the moderately
inflated left valve the septum is slightly concave and parallel to, though depressed from, the plane of the
ligament area, extending outwards > 27 mm. On the weakly inflated right valve, on the other hand, it is
perpendicular to the area. Details of the resilifers are barely preserved on the specimens illustrated, and are,
as yet, poorly known. Ligament attached to outer shell layer.

Material. TM 6711 , 031 /f95 1 4, GS 6051. Plate 90, fig. 5. Umbo of LV. Long Creek, Hapuku River,
Marlborough; 031 667782; H. E. Fyfe, 1935. Arowhanan (late Cenomanian).

TM 6712, as for above. Plate 90, fig. 4. Umbo of RV.

Inoceramus sp. A
Plate 90, fig. I I

Identified previously as Inoceramus bicorrugatus on the basis of its juvenile ornament and marked growth
stop, this specimen has adult ornament and ligament area morphology very similar to I. matotorus. At present
it cannot be referred to the latter with confidence.

Ligament area non-cylindrically concave, split into three longitudinal bands at ~ 40° to each other. Many
relatively broad shallow elongate-ovate to rectangular resilifers which do not breach the ventral margin of
the area, separated by low ridges which increase in height and width on the dorsalmost band of the area.
Area pinches out close to umbo, and achieves a height of ~ 10 mm posteriorly, where the ridges are 2-5 mm
apart. Ligament attached to outer shell layer.

Material. TM 6716, W22/f8504, GS 3225. Incomplete portion of LV. Location as for TM 6708 (I. opetius).

Inoceramus ? sp. B
Plate 90, fig. 14

Known from isolated beaks with very distinctive ligament area, this taxon cannot at present be identified
with any described species. Resilifers consist of ovate pits enclosed by raised box-like structures with sharp
upstanding ridges on three or four sides. Areas between ‘boxes’ depressed and of varying widths. Resili-
fers carry transverse sculpture and decrease in size in one direction, from a width of 2-5 mm and a height of
> 2-5 mm. Ligament attached to outer shell layer.

Material. TM 6717, Z 1 4/f 1 06, GS 13400. Distorted beaks and ligament areas. 400 m up south flowing
tributary of Taurangakautuku Stream, East Cape; Z14 722725; I. G. Speden, 1979. Haumurian (Maastrichtian).

Inoceramus tawhanus Wellman (1959, pp. 156-157. Figured in Woods, 1917, pi. 4, fig. la, b)

Not figured here

From a single specimen with a poorly preserved section of the ligament area it would appear that the resilifers
may alternate in size and strongly breach the ventral margin of the area. Ligament attached to outer shell
layer.

Material. TM 6718, O29/f9630, GS 6534. Incomplete partly shelly LV. Near mouth of Limestone Creek,
Awatere Valley, Marlborough; H. E. Fyfe, 1956. Ngaterian (late Albian-early Cenomanian).

992

PALAEONTOLOGY, VOLUME 31

Acknowledgements. This study has benefited greatly from discussion with and critical appraisals of Dr A. G.
Beu, Dr H. J. Campbell, Dr J. A. Crame, Dr P. Maxwell, Dr I. G. Speden, and an anonymous referee. I
gratefully acknowledge permission to refer to unpublished data gathered by Dr N. J. Morris, Dr J. I. Raine,
and Dr J. G. Wilson; and the loan of fossils by Professor J. D. Campbell, Professor J. A. Grant-Mackie,
and N. Hudson. G. H. Browne assisted in the field, while invaluable technical assistance was provided by I.
Beu, E. McGregor, I. Galuszka, I. Keyes, A. Lee, J. Simes, W. St George, and library staff of the New
Zealand Geological Survey.

REFERENCES

airaghi, c. 1904. Inocerami del Veneto. Boll. Soc. geol. ital. 23, 178-199.

Andrews, p. b., field, b. d., browne, G. h. and Mclennan, j. m. 1987. Lithostratigraphic nomenclature for
the upper Cretaceous and Tertiary sequence of central Canterbury, South Island. Rec. geol. Surv. NZ, 24,
1 40.

Austin, p. m., sprigg, r. c. and braithwaite, J. c. 1973. Structural and petroleum potential of the eastern
Chatham Rise, New Zealand. Bull. Am. Ass. Petrol. Geol. 57, 477 497.

Bernard, f. 1898. Recherches ontogeniques et morphologiques sur la coquille des lamellibranchs. Annls Sci.
nat. Zoologie , (8) 8, 1-208.

beyrich, E. 1864. Uber eine kohlenkalk-Fauna von Timor. Abh. dt. Akad. Wiss. Berl. 61-98.
boehm, G. 1907. Beitrage zur Geologie von Niederlandisch-Indien. 1. Abt. die Sudkiisteb der Sula-Inseln
Taliabu und Mangoli. 3. Abs. Oxford des Wai Galo. Palaeontographica , 4 (supplement), 59-120. [Not
seen.]

boreham, a. u. E. 1959. Cretaceous fossils from the Chatham Islands. Trans. R. Soc. NZ , 86, 1 19-125.
browne, g. h. and field, b. d. 1985. The lithostratigraphy of Late Cretaceous to Early Pleistocene rocks of
northern Canterbury, South Island. Rec. geol. Surv. NZ , 6, 1 63.
browne, i. a. and newell, n. d. 1966. The genus Aphanaia Koninck, 1877, Permian representative of the
Inoceramidae. Am. Mus. Novit. 2252, 1 -10.

bruguiere, j. g. 1789. [In 1789-1792], Encyclopedic methodique (dictionaire encyclopedique methodique), ou
par ordre de matieres; par une societe de gens de lettres , de savans et d' artistes . . . precede d'un vocabulaire
universel. Histoire naturelle des vers. Vol. 1. Liege, Paris. [Not seen.]

CAMPBELL, H. J., ANDREWS, P. B., BEU, A. G., EDWARDS, A. R., HORNIBROOK, N. DE B., LAIRD, M. G., MAXWELL,

p. a. and watters, w. a. 1988. Cretaceous-Cenozoic lithostratigraphy of the Chatham Islands. Jl R. Soc.
NZ, 18, 285-308

carter, j. g. and Clark, g. r. 1985. Classification and phylogenetic significance of molluscan shell
microstructure. In broadhead, t. w. (ed.). Mollusks, notes for a short course, 50-71. University of Tennessee
Department of Geological Sciences, Studies in Geology 13, Knoxville, Tennessee,
cox, l. r. 1940. The Jurassic lamellibranch fauna of Kuchh (Cutch). Mem. geol. Surv. India, Palaeont. indica.
9, 3 (3), 1 157.

— 1954. Taxonomic notes on Isognomonidae and Bakevellidae. Proc. malac. Soc. Lond. 31, 46-49.

— 1955. Proposed determination of the nominal species to be accepted as the type species of the genus
‘Inoceramus ’ Sowerby (J), 1814 (Class Pelecypoda) and proposed addition of that name to the ‘Official list
of generic names in zoology’. Bull. zool. Norn. 11, 239-245.

— 1969. Families Bakevelliidae to Isognomonidae. In moore r. c. (ed.). Treatise on invertebrate paleontology.
Part N. Mollusca, 6 (1), N306-N326. University of Kansas and Geological Society of America, Lawrence,
Kansas and Boulder, Colorado.

crame, j. a. 1982. Late Jurassic inoceramid bivalves from the Antarctic Peninsula and their stratigraphic use.
Palaeontology, 25, 555-603.

darwin, c. R. 1846. Geological observations on South America. Being the third part of the geology of the voyage
of the Beagle during the years 1832 to 1836, 286 pp. Smith, Elder and Co., London.
deshayes, p. G. 1830. [In 1824-1832], Description des coquilles fossiles des environs de Paris, 1, 392 pp. Chez
I’auteur Paris. [1 80, 1824; 81-170, 1825; 171-238, 1829; 239 322, 1830; 323-392, 1832.]

— 1837. Description des coquilles fossiles des environs de Paris, atlas, 28 pp. + 106 pi. Chez F.-G. Levrault,
Paris and Strasbourg.

1860. Description des animaux sans vertebres decouverts dans le bassin de Paris, atlas, 1, 88 pp. + 87 pis.
J.-B. Bailliere et fils, Paris.

— 1861. [In 1861 1864], Description des animaux sans vertebres decouverts dans le bassin de Paris, 2,
968 pp. J.-B. Bailliere et fils, Paris. [1 432, 1861; 433-640, 1862; 641 -920, 1863; 921-968, 1864.]

CRAMPTON: ISOGNOMONI D AE, INOCERAMIDAE, AND RETROCERAMIDAE 993

dickins, J. m. 1983. Posidoniella , Atomodesma, the origin of the Eurydesmidae, and the development of
the pelecypod ligament. Palaeontological Papers 1983. Bull. Bur. Miner. Resour. Geol. Geophvs. Aust. 217,
59-65.

duran-gonzalez, a., rodriguez-romero, F. and laguarda-figueras, a. 1984. Polymorphisme chromoso-
mique et nombre diploide dans une population D’ Isognomon alatus (Bivalvia: Isognomonidae). Malac.
Rev. 17, 85-92.

fischer-piette, e. 1976. Revision des Aviculidees. 1. Crenatula, Pedalion , Foramelina. J. Conch. Paris , 113,
3-42.

fursich, f. t. 1976. Fauna-substrate relationships in the Corallian of England and Normandy. Lethaia , 9,
343-356.

1980. Preserved life positions of some Jurassic bivalves. Palaont. Z. 54, 289-300.

1981. Salinity-controlled benthic associations from the Upper Jurassic of Portugal. Lethaia , 14, 203-

223.

and werner, w. 1986. Benthic associations and their environmental significance in the Lusitanian Basin

(Upper Jurassic, Portugal). Neues Jb. Miner. Geol. Palaont. Abh. 172, 271-329.
gage, M. 1970. Late Cretaceous and Tertiary rocks of Broken River, Canterbury. NZ Jl Geol. Geophys. 13,
507-559.

giebel, c. g. 1852. Allgemeine Palaeontologie: Entwurf einer systematischen Darstellung der Fauna und Flora
der Vorwelt, viii + 413 pp. Ambrosius Abel, Leipzig.

GMELIN, J. F. 1791. In linnaeus, c. (ed.). Systema naturae, sive regna tria naturae systematice proposita per
classes ordines, genera, and species. 13th edn., 1 (6) Regnum Animale, 3021 3910. Impensis Direct. Laurentii
Salvii, Holmiae.

grindley, G. w., adams, c. j. D., lumb, j. T. and watters, w. a. 1977. Paleomagnetism, K-Ar dating and
tectonic interpretation of upper Cretaceous and Cenozoic volcanic rocks of the Chatham Islands, New
Zealand. NZ J! Geol. Geophys. 20, 425-467.

hay, R. F., mutch, a. R. and watters, w. a. 1970. Geology of the Chatham Islands. Bull. geol. Surv. NZ,
83, 1-86.

hayami, I. 1957. Liassic Gervillia and Isognomon in Japan. Jap. J. Geol. Geogr. 28, 95-106.

1960. Jurassic inoceramids in Japan. J. Fac. Sci. Tokyo Univ. Section 2, 12, 277-328.

hector, J. 1881. Progress report 1879-80. Rep. geol. Explor. geol. Surv. NZ, (13), 2-30.
heinz, R. 1932. Aus der neuen systematik der inoceramen. Mitt. geol. St Inst. Hamb. 13, 1 26.
hochstetter, f. 1863. Neu-Seeland, 556 pp. J. G. Cotta, Stuttgart.

ihering, h. von. 1903. Notes sur quelques mollusques fossiles du Chile. Revta chil. Hist. nat. 7,
120-127.

Ivannikov, A. v. 1979. Inocerami of the upper Cretaceous sediments of the southwest Eastern European
Platform, 102 pp. Akademiya Nauk Ukrainskoi CCP, Kiev. [In Russian ]
johnston, m. r. 1980. Geology of the Tinui-Awatoitoi district. Bull. geol. Surv. NZ, 94, 1-62.
kauffman, e. G. 1965. Taxonomic, ecologic, and evolutionary significance of interior shell morphology in
the Inoceramidae (Mesozoic Bivalvia) (abstract). Spec. Pap. geol. Soc. Am. 87, 86.

1977. Systematic, biostratigraphic, and biogeographic relationships between middle Cretaceous Euamer-

ican and North Pacific Inoceramidae. Spec. Pap. palaeont. Soc. Japan, 21, 169-212.

and runnegar, B. 1975. Atomodesma (Bivalvia), and Permian species of the United States. J. Paleont.

49, 23-41.

keyserling, a. f. m. l. a. von. 1848. In middendorff, a. t. von. (ed.). Reise in den dussersten Norden und
Osten Sibiriens wahrend . . . 1843 und 1844 . . . auf Veranstaltung der kaiserlichen Akademie der Wissenschaften
zu St. Petersburg ausgefuhrt und . . . herausgegeben von. I (1), lvi + 274 pp. St Petersburg. [Not seen ]
king, w. 1850. A monograph of the Permian fossils of England, xxxvii + 258 pp. Palaeontographical Society,
London.

koschelkina, z. v. 1957. Palaeontological basis of the distribution of strata in the marine Jurassic deposits
of the Vilyusk Syncline and Verkhoyansk depression. In Transactions of the interdepartmental conference
on the development of the unified stratigraphical schemes in Siberia 1956. Reports on the stratigraphy of
Mesozoic and Cainozoic deposits, 27-31. State scientific-technical publishing house of the oil and mining
industry, Leningrad section, Leningrad. [In Russian; not seen.]

1959. Stratigraphy of the Jurassic deposits of the Vilyusk Syncline and Verkhoyansk depression. Trudy
mosk. geol.-razv. Inst. 33, 89-100. [In Russian; not seen.]

— 1962. Field atlas of the fauna of the Jurassic deposits of the Vilyusk Syncline and Verkhoyansk depression.
Magadan. [In Russian; not seen.]

994

PALAEONTOLOGY, VOLUME 31

1963. Jurassic stratigraphy and Bivalvia of the Vilyusk syncline ad Verkhoyansk depression. Tr. Sev.-
Vost. Kompleksn. nauchno-issled. Inst. 5, 1-220. [In Russian.]

— 1969. General characteristics and morphology of the genus Retroceramus. In shilo, n. a. (ed.). Jurassic
and Cretaceous inocerams of the northeastern USSR, 5-13. Ibid. 32. [In Russian.]

— 1971. Problems with the nomenclature and systematics of Jurassic retroceramids. Kolyma, 1971 (5), 43-
45. [In Russian.]

1975. Microstructural features of shells of some families. . . . In koschelkina, z. v., bychkov, yu. m.
and pokhialaynen, v. p. (eds.). Mezozoy Severo-Vostoka SSSR; tezisy dokladov mezhvedom-stvennogo
stratigraficheskogo soveshchaniya, 23-24. Akad. Nauk SSSR, Geol. Inst., Magadan. [In Russian; not seen.]
1980. The family Retroceramidae and the zonal stratigraphy of the Middle Jurassic of the northeastern
USSR. In krassilov, v. a. and blokhina, n. i. (eds.). Ekosistemi v Stratigrafii (Materiali Vsesoyuznogo
Soveshchaniya Vladivostok , Oktyabr 1978), 131-135. Vladivostok. [In Russian.]
lamarck, j. b. de. 1803. Sur la Crenatule, nouveau genre de coquillage. Annales du Museum National
D'Histoire Naturelle, par les Professeurs de cet Etablissement. 3 (13), 25-31 .
lightfoot, j. 1786. A Catalogue of the Portland Museum, lately the property of the Duchess Dowager of
Portland, deceased, which will be sold at auction, by Mr Skinner and co . . . London, viii + 194 pp. [Not
seen.]

Linnaeus, c. 1758. Systema naturae, per regna tria naturae, secundum classes, ordines, genera, species, cum
characteribus, differentiis, synonymis, locis. 10th edn., reformata, 1 (6) Regnum Animale, iv + 824 pp.
Laurentii Salvii, Holmiae.

1764. Museum S.R.M. Ludovicaae Ulricae Reginae . . . in quo Animalia rariora, exotica, inprimis Insecta
and Conchilia describuntur and determinantur . . ., vi + 720 pp., 2 vols. Holmiae.
ludwig, r. 1864. Fossile conchylien aus den Tertiaren susswasser-und meerwasser-ablagerungen in Kurhessen,
Grossherzogthum Hessen und der Bayer’schen Rhon. Palaeontographica, 14 (2), 40 -97.
mckay, a. 1881. On the Trelissick Basin, Selwyn County. Rep. geol. Explor. geol. surv. NZ, (13),
49-53.

marwick, j. 1926. Cretaceous fossils from Waiapu Subdivision. NZ Jl Sci. Teclmol. 8, 379-382.

1953. Divisions and faunas of the Hokonui System (Triassic and Jurassic). Paleont. Bull., Wellington,
21, 1 141.

Matthews, s. c. 1973. Notes on open nomenclature and on synonymy lists. Palaeontology, 16, 713-719.
mildenhall, d. c. 1977. Cretaceous palynomorphs from the Waihere Bay Group and Kahuitara Tuff,
Chatham Islands, New Zealand. NZ Jl Geol. Geopliys. 20, 655-672.
moore, p. r. 1980. Late Cretaceous-Tertiary stratigraphy, structure, and tectonic history of the area between
Whareama and Ngahape, eastern Wairarapa, New Zealand. Ibid. 23, 167-177.
oppel, a. 1865. Ueber ostindische Fossilreste aus den secondaren Ablagerungen von Spiti und Gnari-Khorsum
in Tibet. Beschreibung der von der Herren Adolf Herrman und Robert v. Schlagintweit wahrend der Jahre
1854 1857 gesammelten Arten. In oppel, a. (ed.). Palaentologische Mittheilungen aus dem Museum des
Koenigl. buyer. Staates, 293-304. Ebner und Seubert, Stuttgart.
orbigny, A. d\ 1845. [In 1843-1847], Paleontologie Francaise. Description zoologique et geologique de tous les
animaux mollusques et rayonnes fossiles de France. Terrains Cretaces. 3 (Lamellibranchiata). 807 pp. Chez
Arthus Bertrand, Paris. [Pagination dates unknown.]

Parkinson, j. 1819. Remarks on the fossils collected by Mr W. Phillips near Dover and Folkstone. Trans,
geol. Soc. Fond. 5(1), 52-59.

pokhialaynen, v. p. 1969. On characters of the articulation folds of Neocomian inoceramids. In shilo,
n. a. (ed.). Jurassic and Cretaceous inocerams of the northeastern U.S.S.R. Tr. Sev.-Vost. Kompleksn.
nauchno-issled. Inst. 32, 118-123. [In Russian.]

1972. Systematic location of Inoceramidae of the Neocomian. Trudy Vsesoyuznogo Kollokviuma po
Inotseramam , 1, 57-65. Moskow. [In Russian.]

— 1977. The particular construction of Cretaceous inoceramids. Mater. Geol. polez. Iskop. Sev.-Vost.,
SSSR, 23, 52-62. [In Russian.]

1985. The basis for an above-species taxonomy of Cretaceous inoceramid bivalves, 37 pp. Cevero-
Vostochnogo Kompleksnogo Nauchno-Issledovatelskogo Instituta DVNTS Akademia Nauk SSSR,
Magadan. [In Russian.]

read, K. R. h. 1964. Ecology and environmental physiology of some Puerto Rican bivalve molluscs and a
comparison with Boreal forms. Caribb. J. Sci. 4, 459-465.
rehder, h. a. 1967. Valid zoological names of the Portland Catalogue. Proc. US Natn. Mus. 121 (3579),

1 51.

CRAMPTON: ISOGNOMONIDAE, INOCERAM IDAE, AND RETROCER A M I D AE 995

reid, r. g. b. 1985. Isognomon: Life in two dimensions. In morton, b. and dudgeon, d. (eds.). The malacofauna
of Hong Kong and Southern China. 2. Proceedings of the Second International Workshop on the Malacofauna
of Hong Kong and Southern China , 6~24 April 1983 , 311 319. Hong Kong University Press, Hong Kong.
sandberger, c. L. F. von. 1863. [In 1858-1863]. Die Conchylien des Mainzer Tertidrbeckens, v + 468 pp.
C. W. Kreidel’s Verlag, Wiesbaden. [I 72, 1858; 73-112, 1859; 1 13-152, 1860; 153-232, 1861; 233-272,
1862; 273 468, 1863.]

sangiovanni, g. 1844. In philippi, r. a. 1836 1844. Enumeratio Molluscorum Siciliae , cum viventium turn in
tellure tertiaria fossilium quae in itinere sus observant. 2 vols. Halis Saxonum, Berolini. [Not seen.]
seilacher, a. 1984. Constructional morphology of bivalves: evolutionary pathways in primary versus
secondary soft-bottom dwellers. Palaeontology , 27, 207-237 .
siung, a. m. 1980. Studies on the biology of Isognomon alatus Gmelin (Bivalvia: Isognomonidae) with notes
on its potential as a commercial species. Bull. Mar. Sci. 30, 90-101.
sowerby, j. 1814. In Article 6; Proceedings of philosophical societies — Linnaean Society. Ann. Phil. 4, 448.
speden, i. G. 1970a. Three new inoceramid species from the Jurassic of New Zealand. NZ Jl Geol. Geophys.
13, 825-851.

— 19706. Generic status of the Inoceramus! tegulatus species group (Bivalvia) of the latest Cretaceous of
North America and Europe. Postilla , 145, 1-45.

- — 1976. Inoceramus opetius in the Kahuitara Tuff, Chatham Islands, New Zealand (note). NZ Jl Geol.
Geophys. 19, 385-387.

Stanley, s. m. 1972. Functional morphology and evolution of byssally attached bivalve molluscs. J. Paleont.
46, 165-212.

Stephenson, l. w. 1923. The Cretaceous formations of North Carolina, part 1: invertebrate fossils of the
upper Cretaceous formations. Rep. N. Carol, geol. econ. surv. 5(1), i-xi, 1-604.

Stevens, G. R. 1965. The Jurassic and Cretaceous belemnites of New Zealand and a review of the Jurassic
and Cretaceous belemnites of the Indo-Pacific region. Paleont. Bull., Wellington, 36, 1-233.

— 1981. Geological time scale. Geological Society of New Zealand, miscellaneous publication 28.

— and speden, i. g. 1978. Chapter 8: New Zealand. In moullade, m. and nairn, a. e. m. (eds.). The Mesozoic,
A. The Phanerozoic geology of the world II, 251 328. Elsevier Scientific Publishing Co., Amsterdam.

strong, c. p. 1979. Late Cretaceous foraminifera from Kahuitara Tuff, Pitt Island, New Zealand. NZ Jl
Geol. Geophys. 22, 593-61 1.

— and edwards, a. r. 1979. Late Haumurian (Maastrichtian) microfossils from Chatham Islands, New
Zealand. Ibid. 22, 613-619.

suggate, r. p., stevens, G. R. and te punga, m. T. (eds.). 1978. The geology of New Zealand , 2 vols., xx +
820 pp. Government Printer, Wellington.

taylor, J. D., Kennedy, w. J. and hall, a. 1969. The shell structure and mineralogy of the Bivalvia.

Introduction. Nuculacea-Trigonacea. Bull. Br. Mus. nat. Hist. Zoology, Supplement 3, 1-125.
troger, k. a. 1976. Evolutionary trends of Upper Cretaceous inocerames. Evolut. Biol., Praha, 1976, 193-203.
trueman, e. r. 1954. The structure of the ligament of Pedalion ( Perna ). Proc. malac. Soc. Lond. 30, 160- 166.

— 1969. Ligament. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part N. Mollusca, 6(1),
N58-N64. University of Kansas and Geological Society of America, Lawrence, Kansas and Boulder,
Colorado.

voronetz, n. k. 1937. Representatives of the genera Trigonia and Inoceramus from the Jurassic of South
Ussuri-Land. Records of the Geology and of the Mineral Records of the Far East, 67 (for 1936), 1-36.
waller, t. r. 1978. Morphology, morphoclines, and a new classification of the Pteriomorpha (Mollusca:
Bivalvia). Phil. Trans. R. Soc. B284, 345-365.

wandel, G. 1936. Beitrage zur Palaontologie des Ostindischen Archipels. XIII. Beitrage zur Kenntnis der
Jurassischen Molluskenfauna von Misol, Ost-Celebes, Buton, Seran und Jamdena. Neues Jb. Miner. Geol.
Paldont. BielBd. 75, 447 -526. [Not seen.]

Waterhouse, J. b. 1970. Permoceramus, a new inoceramid bivalve from the Permian of Eastern Australia.
NZ Jl Geol. Geophys. 13, 760-766.

wellman, h. w. 1959. Divisions of the New Zealand Cretaceous. Trans. R. Soc. NZ, 87, 99-163.
wilson, g. J. 1976. Late Cretaceous (Senonian) dinoflagellate cysts from the Kahuitara Tuff, Chatham
Islands. In Notes from the New Zealand Geological Survey (9). NZ Jl Geol. Geophys. 19, 127 130.
woodring, w. p. 1925. Miocene molluscs from Bowden Jamaica, pelecypods and scaphopods. Pubis Carnegie
Instn. 366, 1 -222.

woods, H. 1905. A monograph of the Cretaceous lamellibranchia of England, 2 (2), 57-96. Palaeontographical
Society, London.

996

PALAEONTOLOGY, VOLUME 31

— 1917. The Cretaceous faunas of the north-eastern part of the South Island of New Zealand. Paleont.
Bull., Wellington, 4, 1-41.

yabe, h. and nagao, t. 1926. In yabe, h., nagao, t. and shimizu, s. 1926. Cretaceous Mollusca from the
Sanchu Graben in the Kwanto mountainland, Japan. Sci. Rep. Tohoku Univ. (Ser. 2), 9, 33-76.
yonge, c. m. 1968. Form and habit in species of Malleus (including the ‘Hammer Oysters’) with comparative
observations on Isognomon isognomon. Biol. Bull. mar. biol. Lab., Woods Hole , 135, 378-405.

— 1978. Significance of the ligament in the classification of the Bivalvia. Proc. R. Soc. B202, 231-248.
zonova, t. d. 1980a. Ligamental striae of a type new to inoceramids of Central Asia. Ezheg. vsgs. paleont.

Obshch , 23, 50-56. [In Russian.]

— 1980A Representatives of Albian inoceramids in the Soviet Far East and descriptions of their ligament
bands. In ablaev, a. g., poyarkov, b. v. and poyarkova, z. n. (eds.). Fossil mollusks of the Far East and
their stratigraphical significance, 10-18. Dalnevostochnyi Geologicheskii Institut, Vladivostok. [In Russian.]

1982. The ligament apparatus of shells of new inoceramid species of the Penchina Series of northeastern
USSR. Ezheg. vses. paleont. Obshch , 25, 244-252. [In Russian ]

and yefremova, v. i. 1976. A new type of ligamental band in Late Cretaceous inoceramids. Paleont. J.
10, 108-110.

JAMES S. CRAMPTON

New Zealand Geological Survey
PO Box 30368

Typescript received 9 June 1987 Lower Hutt

Revised typescript received 24 November 1987 New Zealand

ALLOMETRY AND HETEROCHRONY IN THE
GROWTH OF THE NECK OF TRIASSIC
PROLACERTIFORM REPTILES

by KARL TSCHANZ

Abstract. The functional morphology of the elongated neck of Tanystropheus longobardicus (Bassani) has
long been controversial. It is suggested here, that the neck was not very flexible because the elongated cervical
ribs are bundled along the ventrolateral margin of the vertebrae. The result, a stiffened neck, is advantageous
in an aquatic environment. The ontogenetic development of the neck in T. longobardicus and Macrocnemus
bassanii Nopcsa, both included within the Prolacertiformes, is another point of interest. During ontogeny,
the neck exhibits constant, positive allometric growth with differing growth parameters for the two taxa.
This difference most likely resulted from heterochronic processes. The marked elongation of the neck in T.
longobardicus was primarily caused by hypermorphic growth. Additional factors, modifying the growth
pattern, include predisplacement of growth and an increased number of cervical vertebrae.

The monophyly of the Prolacertiformes is corroborated by a number of synapomorphies, such as
an incomplete lower temporal bar, elongated cervical vertebrae, low neural spines on the cervical
vertebrae, and a short ischium (Benton 1985). It includes the taxa Prolacerta , Macrocnemus ,
Tanystropheus , Tanytrachelos (Olsen 1979; Wild 19806; Benton 1985), and possibly Protorosaurus
(Carroll 1981; Benton 1985). Other taxa like Cosesaurus (Olsen 1979) and Malerisaurus (Chatterjee
1980) have been included in the Prolacertiformes, but their relationship is not firmly established.
A characteristic feature of the Prolacertiformes is their elongated neck. The elongation results
mainly from lengthening of the cervical vertebrae. An increase of their number occurs in some
taxa, adding to the elongation of the neck. The shortest relative length of the neck is observed in
Prolacerta (8 cervical vertebrae); it increases slightly in Macrocnemus (8 cervical vertebrae) and
markedly in the most advanced species of Tanystropheus (9-12 cervical vertebrae). In adult T.
longobardicus the cervical vertebral column equals more than half of the total body length.

The earliest interpretation of the elongated neck vertebrae of Tanystropheus , from the Triassic
Muschelkalk Beds of Germany, was by Munster (1834) who believed them to represent elements
of dinosaur extremities. In 1855, H. von Meyer interpreted the same bones as caudal vertebrae of
a dinosaur which he named T. conspicuus. The most unconventional hypothesis has been proposed
by Nopcsa (1923), who considered the elongated bones as wing elements (phalanges) of a
pterosaur (Tribelesodon longobardicus Bassani). Apparently only poorly preserved material from the
Grenzbitumenzone Beds of Besano (northern Italy) was available to Nopcsa. The discovery of a
complete skeleton of a reptile with an elongated neck in the Grenzbitumenzone Beds from Monte
San Giorgio (Switzerland) by Peyer in 1929 finally revealed the true identity of the elongated bones
as cervical vertebrae of Tanystropheus longobardicus (Bassani). Systematic excavations in the
Middle Triassic Grenzbitumenzone of Monte San Giorgio (Switzerland) yielded about fifteen fairly
complete skeletons of T. longobardicus (Wild 1973).

FUNCTIONAL MORPHOLOGY

Ever since the first discovery of a complete skeleton, the life style of T. longobardicus has remained
enigmatic. According to Peyer (1931) Tanystropheus was a terrestrial animal. This view was based
on morphological characters such as the form of the pelvic girdle, the presence of claw-like terminal

| Palaeontology, Vol. 31, Part 4, 1988, pp.j)97-1011.| © The Palaeontological Association

998

PALAEONTOLOGY, VOLUME 31

phalanges, and the proportions of metatarsals and metacarpals. Consequently, his reconstruction
shows Tanystropheus in a terrestrial environment. Locomotion was supposed to have been clumsy,
not more than a slow crawling. The short limbs may occasionally have supported locomotion
which was effected mainly by lateral undulations of the vertebral column. Normally the body was
thought to have lain directly on the ground, and the neck was oriented more or less horizontally.

A neck with a degree of flexibility comparable to that observed in birds (Boas 1929) was assumed
by Peyer (1931). He therefore subdivided the neck of T. longobardicus into parts of different
mobility. But the compartmentalization was not considered to be as advanced as in birds.
Nevertheless, the elongated neck could have been used as a perfect instrument to grasp highly
mobile prey. Sitting safely near the shoreline, T. longobardicus was believed to be able to snap at
fishes (text-fig. la). One problem, the phylogenetic development of the elongated neck of
Tanystropheus , remained enigmatic to Peyer (1931). He postulated that this development would
not have been possible if Tanystropheus had always lived in a terrestrial environment. This is why
the hypothetical ancestor was thought to have been at least partly aquatic.

Wild’s (1973) description of T. longobardicus was based on a sample of twenty-seven nearly
complete specimens from the Swiss part of the Grenzbitumenzone Beds, and on some isolated,
cervical vertebrae of T. conspicuus recovered from the German Upper Muschelkalk. Interpretation
of the mobility of the neck was based on a detailed analysis of the position of the zygapophyses.
Some of Peyer’s (1931) hypotheses were confirmed, e.g. the subdivision of the neck into three parts
of different mobility. The cervical ribs, even more elongated than the cervical vertebrae, were
supposed to be elastic and to protect the blood vessels, the trachea, and the oesophagus. In
addition, the ribs supported the elongated neck at the intervertebral joints. Wild (1973) postulated
that the neck was very flexible. When on land, the neck of T. longobardicus would have been
relatively elevated, and an S-shaped posture would have resulted (text-fig. 1 b). Adult specimens
show some adaptations to an aquatic life, as indicated by the proportions of fore and hind limbs.
Together with the characteristic tooth replacement, this would be evidence for ecological changes
during the life of T. longobardicus. According to Wild (1973, 1980a, b ), the juveniles of Tanystropheus
lived as terrestrial insectivores (tricuspid teeth), whilst the adults lived as aquatic carnivores
(recurved, conical teeth). Stomach contents of adult T. longobardicus have yielded unquestionable
hooks from cephalopod arms (phragmoteuthids) (Wild 1973).

Kummer (1975) reconstructed the position of the neck of T. longobardicus, according to
fundamental static constraints. He concluded that the neck could not have been held horizontally
without tilting of the animal. Consequently, his reconstruction shows T. longobardicus with the
neck strongly recurved in a swan-like position (text-fig. lc). This position appears advantageous
if static constraints are considered in isolation. The shear stress on the cervical column resulting
from this position would be minimal (Preuschoft 1976).

In this study (see also Tschanz 1986) the anatomy of the cervical vertebrae of T. longobardicus
was compared with that of recent lacertilians. The following structural differences were recorded:
reduced neural spines result in reduced attachment areas for important parts of the cervical
musculature. Only the short, intervertebral muscles had extensive insertional areas. The musculature
was too weak to lift the neck beyond the horizontal to the curved position postulated by Wild
(1973) or Kummer (1975). The muscles would not only have had to counteract gravitational forces,
but also to bend the cervical ribs dorsally. The ribs, slender and longer than the cervical vertebrae,
are assumed to have been bundled (text-fig. 2). In this way they acted as two rods, lateroventral
to each side of the cervical column. The cervical ribs are thickened at the intervertebral joints. The
stiffened rods supported the vertebral column and would have reduced gravitational shearing
stresses. This construction restricted dorsal bending of the neck of T. longobardicus (Tschanz 1985,
1986). Therefore, the reconstructions of Tanystropheus, given by Wild (1973) and Kummer (1975),
with S-shaped or swan-like curved necks, have to be rejected. If T. longobardicus was capable of
a terrestrial life, its neck would have been held out horizontally (text-fig. 1 d).

In an aquatic environment the same neck construction would appear more advantageous. Relief
for the musculature would have resulted from the buoyancy of the surrounding medium. Therefore,

TSCHANZ: ALLOMETRY IN TRIASSIC PROLACERTIFORM REPTILES

999

d

text-fig. i . Different reconstructions of Tanystropheus longobardicus. a, as a
mainly terrestrial reptile (redrawn after Peyer 1931). b, as a terrestrial reptile,
with its neck in a ‘normal’, elevated position (redrawn after Wild 1973). c, with
the neck in a swan-like position. If the head is positioned more forward the
animal is supposed to tilt (redrawn after Kummer 1975). d , with the neck in
horizontal position. This represents the most advantageous position for terrestrial
and aquatic life (Tschanz 1985, 1986).

1000

PALAEONTOLOGY, VOLUME 31

1cm

a

b

text-fig. 2. a, tenth cervical and last dorsal vertebra compared (specimen T2791). b,
reconstruction of the rib arrangement in the anterior part of the neck in Tanystropheus
longobardicus. The neck is supported by the elongated ribs.

the musculature counteracting gravitation would not have to be as extensive as for an animal with
a terrestrial mode of life. In addition, a stiffened neck would have been advantageous for aquatic
locomotion. Propulsion in Tanystropheus most likely resulted from lateral undulations of trunk
and tail. The extension of these undulations forward beyond the trunk was restricted by the
stiffened neck. Additionally, lateral bending in the region of the neck-trunk transition must have
been prevented by the musculature of the shoulder girdle region. This enabled T. longobardicus to
hold its skull straight in the direction of locomotion. At any rate, T. longobardicus with its reduced
cervical musculature and its stiffened neck, was adapted to an aquatic environment. A more
interesting question, however, addresses the growth parameters which would have created the
elongation.

DEVELOPMENTAL PROCESSES

According to Wild, the elongation of the neck results from pronounced positive allometric growth
relative to absolute body size (as represented by the length of the precaudal vertebral column). A
linear regression line could be fitted approximately to the point cluster of the logarithmically
transformed length measurements of the cervical vertebral columns. This regression line is supposed
to show two sharp breaks in its slope (text-fig. 3). For the first increase in slope no explanation
was found. The second increase was correlated by Wild with sexual maturity in T. longobardicus,
at an overall body length of about 2 m. This hypothesis seems reasonable because the regression
lines of other elements (radius, humerus, skull) show a similar pattern of slope change at the same
body size (Wild 1973, p. 138) (text-fig. 3). In addition, a characteristic pattern of tooth replacement,
from tricuspid to conical, takes place at this time.

Another hypothesis explaining slope changes is that the sample contains specimens of two species
with different body size. Different slopes of the regression lines then would reflect differential
growth rates of the neck in these two species. This paper will concentrate on the analysis of the
ontogenetic growth of the neck in T. longobardicus and the closely related, contemporary
Macrocnemus bassanii. A basic premiss is that the studied specimens of Tanystropheus belong to
a single species, T. longobardicus. The following hypotheses are tested: 1, ontogenetic growth of
the cervical vertebrae of T. longobardicus and M. bassanii is positively allometric relative to absolute
body size and remains constant during growth; 2, linear regression lines fit the data best and the
allometric parameter b (slope of regression line) is equal for all the cervical vertebrae of one taxon;

TSCHANZ: ALLOMETRY IN TRIASSIC PROLACERTIFORM REPTILES

1001

In

text-fig. 3. Approximately fitted regression lines of the neck (n), humerus (h),
and radius (r) versus precaudal vertebral column (pcvc) in Tanystropheus longo-
bardicus. The regression lines show sharp breaks in slope (redrawn after Wild

1973).

3, the allometric growth of the neck in Tanystropheus and Macrocnemus is comparable. The pattern
in M. bassanii represents the pattern of a hypothetical ancestor of T. longobardicus.

The last hypothesis is supported by the fact that some characters of M. bassanii are not as
advanced as in T. longobardicus , especially in the cervical vertebral column. The cervical column
of M. bassanii consists of only eight cervical vertebrae with relatively high neural spines, and the
cervical ribs are relatively short. A similar pattern is shown by T. antiquus, stratigraphically the
oldest representative of the genus Tanystropheus. Its neck consists of nine cervical vertebrae, and
they are more elongated. It is possible to establish the polarity of evolution within the Prolacerti-
formes, based exclusively on morphological and growth characters of the cervical vertebrae.

1002

PALAEONTOLOGY, VOLUME 31

Prolacerta shows the most primitive condition in neck elongation. The eight cervical vertebrae are
only moderately elongated and possess high neural spines. M. bassanii retains the primitive number
of cervical vertebrae but they are more elongate than in Prolacerta. The neural spines are relatively
high. T. antiquus shows similarly built cervical vertebrae, but additionally a ninth vertebra is
included in the series (Wild 1987). In the most advanced forms, T. longobardicus and T. conspicuus ,
the cervical vertebrae are more elongated and their number is increased to twelve. The two species
possibly have to be unified within a single species pending the discovery of cranial material (Wild
1980/?). Tanytrachelos, a small prolacertiform from the Upper Triassic of North America, is allied
to Tanystropheus as it shares the same number of cervical vertebrae, although these are not as
elongate. If M. bassanii is the hypothetical ancestor of T. longobardicus it would be possible to
analyse ontogenetic growth of the latter in terms of heterochronic processes. Phylogenetic and/or
ecological implications of the elongated neck of the Prolacertiformes may hence be inferred. At
any rate, the effects of ontogenetic change on growth pattern will be better understood. In
particular, structures with no recognizable adaptive value may be more reasonably explained as
results of allometric growth.

MATERIAL AND METHODS

Most specimens of T. longobardicus and M. bassanii on which this study is based come from the Middle
Triassic Grenzbitumenzone (Anisian/Ladinian) and one specimen of M. bassanii comes from the Lower
Meridekalk (Ladinian). The specimens were found at several localities on the Monte San Giorgio (Switzerland).
They are housed at the ’Palaontologisches Institut und Museum der Universitat Zurich’ (Table 1).

The allometric analysis is exclusively based on specimens with a partially preserved trunk region. These
were eight specimens of T. longobardicus (two with complete cervical column) and five specimens of M.
bassanii. Most specimens lack one or more cervical vertebra, or they are incompletely preserved. Measurements
were taken of the lengths of the centra of all cervicals, a middle dorsal vertebra, and the last presacral
vertebra. A slide caliper with mm scaling was used to a degree of accuracy of ±0-5 mm (Tables 2 and 3). If
one end of a vertebra was incomplete, the total length of the centrum was extrapolated.

table 1 . List of the analysed specimens.

Specimen Stratigraphy Status of preservation

Tanystropheus longobardicus (Bassani)

T1277

Grenzbitumenzone

T2482

Grenzbitumenzone

T2485

Grenzbitumenzone

T2787

Grenzbit umenzone

T2791

Grenzbitumenzone

T2795

Grenzbi l umenzone

T2817

Grenzbi tumenzone

T2818

Grenzbitumenzone

Macrocnemus bassanii Nopcsa

T2472

Grenzbitumenzone

T2476

Grenzbitumenzone

T2815

Grenzbitumenzone

Cava Tre Lontane (CTL)

Grenzbitumenzone

Alla Cascina (AC)

untere Meridekalke

Disarticulated skeleton, only posterior cervicals pre-
served

Disarticulated skull and anterior cervicals, last presacral
vertebra not preserved

Disarticulated skeleton, only posterior cervicals pre-
served

Disarticulated but nearly complete skeleton
Complete skeleton, anterior cervicals disarticulated
Disarticulated skeleton, cervicals partly in sequence
Skull and anterior cervicals missing, posterior cervicals
poorly preserved
Complete skeleton

Articulated, nearly complete skeleton
Cast of the specimen Besano 2 (Peyer), disarticulated,
incomplete skeleton
Disarticulated, incomplete skeleton
Articulated, nearly complete skeleton
Articulated skeleton, anterior part of the trunk and
posterior cervicals missing

table 2. Length (in mm) of the vertebrae of Tanystropheus longobardicus (Bassani).

TSCHANZ: ALLOMETRY IN TRIASSIC PROLACERTIFORM REPTILES

1003

0000*0*0*00
OS so Tt <b\ <N oo O'!
nooYiTtriL-
— —
***——• (N

0>0*000000

0000000*0

00*000*0*00

> — ' m ^ rf P-

0*00000*00

000000*00

00*000000
or^t^r^or;'(N*iT
•oiosor^ooTfso^t
* — < — ' <n

oooooooo

oopooooo

in ri oo oo r i rh 6 oo

r*-! \f ’t 'Ti > — i *0 (N

CX CL* * ~ rl

o3 o3

0*0100*0000

Ohi-OodsLob

rh r, rf T) \0 O ^ M
* * — < -H <N

*000*00000

•oooooooo

dssbr^*o^obrl.ir)

<N<o*o«osoo<o —

* * - - M

00*00*00*00

000*0100*00

Oh

o

z

•£ °
c —

OJ V

0 0 0 0*0

0 0*00*0
m co x

Cl Cl

to to to o O
*o r- oo ds 6s

o o o o o

ds rn *o sb

ioo*o»oo

p o o o o
— I ds O rb rn
— * <N <N <N

0*0 0*0*0

*0*0*0 0*0

0*0*000

0*0*0 0*0

r- o ' — ' rn rN

u

cd

o3 ^

UpHOmr?
. t'- * — i Mh

JS rj- oo ,u

-3 M D M i— <

< H H H H

•a ^
o ^

S 2P

cd

<u

— i*oeN*or^r-r^oo
OiONXxr'X-n- <
r^r^-rj-Tf(Nr^oooo
(N(N(N(N— h<N(N(N

«

* 3

.g-s

s 2

D.’S

S “>
^ §

1004

PALAEONTOLOGY, VOLUME 31

The lengths were transformed logarithmically. Regression lines (reduced major axis) were then fitted to
the point clusters produced by plotting the lengths for the cervical vertebrae versus the lengths of the last
presacral vertebra (text-figs. 4-6). Reduced major axis was given preference over least squares because it
operates symmetrically on the two variables (Imbrie 1956). Isometry for the relation length of the last
presacral vertebra versus absolute body size (e.g. body weight) is required. The resulting allometric parameters
b (slope of the regression lines) were subjected to statistical testing. Possibly undetectable distortions did
occur because of the small sample size. Therefore, despite statistical significance, the value of confidence may
be reduced. The total length of the cervical vertebral column was calculated as the sum of the lengths of the
cervical vertebrae. The lengths of missing vertebrae were calculated, based on the particular regression line.

Analysis of longitudinal growth makes the definition of a standard measure for absolute body size necessary.
The standard measure chosen by Wild (1973) was the total length of the precaudal vertebral column. This
is inaccurate, however. First, the axial skeleton is usually incompletely preserved, and secondly, the vertebrae
to be analysed are part of the standard length. Because it is usually well preserved, the last presacral vertebra
was chosen for this analysis. Also this vertebra is easily identified. According to Currie and Carroll (1984),
the length of the centrum of any other dorsal vertebra could serve as a standard as well. Indeed, it could be
shown that growth of the last presacral vertebra proceeds isometrically relative to any other dorsal vertebra.

The functions in general use for quantifying allometric growth are power functions (y = axb). Logarithmic
transformation therefore will result in regression lines with the function Y = A + bX. The parameter A (log
a) corresponds with the intercept of the y-axis by the regression line. The parameter b (allometric coefficient)
is the slope of the regression line. Growth is positively allometric with b > TO. The significance of the positive
allometric growth was tested (z-test; H0 : bcerv = b)ast dors) (Imbrie 1956). The regression lines were
additionally tested to substantiate the hypothesis that they are members of the same cluster (H0 : bn = bn_i),
and therefore have to be treated as parallel lines.

RESULTS

Tanystropheus longobardicus

The regression analysis indicates that ontogenetic growth of the cervical vertebrae is strongly
positively allometric. Correlation is high with coefficients (r) close to TOO (Table 4). Therefore, the
point clusters are best represented by linear regressions (text-fig. 4). The values for the slopes of

table 4. Reduced major axis slopes (b), standard deviations Sb, y-
intercept (A), correlation coefficient (r) from the regression of log
length of cervical vertebra on log length of last presacral vertebra for
Tanystropheus longobardicus. If the z value of the test for equality of
the slopes of the cervical vertebrae and the last presacral vertebra is
> 1-96 the probability is > 0 05.

Vertebrae

b (slope)

Sb

A

r

z

m.dors.

104

006

008

0-983

2

1 30

003

-014

0-999

3-71

3

1-28

007

0-29

0-992

2-67

4

1-26

006

0-43

0-995

2-75

5

1-27

012

0-43

0-975

1-77

6

1-33

012

0-34

0-975

2-23

7

1-27

002

0-44

0-999

3-83

8

1-22

005

0-56

0-995

2-25

9

1 05

002

0-80

0-999

0-17

10

109

004

0-75

0-997

0-71

11

118

013

0-53

0-964

TOO

12

119

015

008

0-961

0-64

neck

1-22

006

1-49

0-989

2-25

neck (1 8)

1-27

002

1-23

0-998

3-83

TSCHANZ: ALLOMETRY IN TRIASSIC PROLACERTIFORM REPTILES

1005

(0

1.0

1,5

dC8

— aC 7
»C 5

— °C4

tCIO
VC 9

aC 6

— +C11

— *C3

— OC 12

— IC2

— «D

*•

log dors

text-fig. 4. Reduced major axis for the cervical vertebrae (C2-C12) and a
middle dorsal vertebra (D) of Tanystropheus langobardicus. log cerv =
logarithmically transformed length values of the cervical vertebrae, log
dors = logarithmically transformed length values of the last presacral vertebra.

1006

PALAEONTOLOGY, VOLUME 31

the regression lines (parameter b) vary between 1 05 ±0 02 (9th cervical vertebra) and 1-33 + 0- 12
(6th cervical vertebra) (Table 4). The hypothesis that growth is isometric (H0 : bcerv = biast dors)
has to be rejected for most cervical vertebrae (Table 4). Therefore, they grow in a significantly
positively allometric fashion, with the exception of the cervicals 9 to 12. The cervical vertebrae 9
and 10 are the most elongated of T. longobardicus. Therefore, it is surprising that growth is not
significantly positively allometric for these vertebrae. But it is possible that this is only an artifact
of the small sample size (only five length values in each case).

The hypothesis of parallel regression lines (H0 : bn = bn_i) cannot be rejected in most cases.
The regression lines have thus to be treated as a bundle of parallel lines. The slope of the regression
for the total neck length of T. longobardicus has a value of 1-22 + 0 06 (Table 4; text-fig. 6). As
expected, the ontogenetic growth is also positively allometric. To calculate this regression, the
standard errors of the calculated lengths of missing cervical vertebrae have not been taken into
account. Therefore, the standard error of the allometric parameter b would be higher than
calculated ( + 0 06).

Macrocnemus bassanii

The values of the allometric parameter b vary between 1-12 + 0-14 (3rd cervical vertebra) and
1-47 + 0-08 (6th cervical vertebra) (Table 5). The variability is greater than in T. longobardicus.
Growth of all cervical vertebrae, except for the third, is significantly positively allometric (Table
5). The regression lines have to be treated as parallel lines, but the significance is not as strong as
for the regression lines of T. longobardicus (Table 5; text-fig. 7). The growth of the cervical vertebrae
of M. bassanii seems to have been accelerated as compared to T. longobardicus since the regression
lines are steeper. The regression line of the total neck length has a slope (parameter b) of 1-27 + 0 07
(Table 5; text-fig. 6). The difference from the slope of the regression line of the neck in T.
longobardicus (b = 1-22 + 0 06) is not very spectacular, the ontogenetic growth of the neck of M.
bassanii is slightly increased. The difference was not found to be statistically significant (zm/t =
0-56). It is possible that this is again due to the small sample size. In addition, the size range
of the five specimens of M. bassanii is not as great as the size range of the eight specimens of
T. longobardicus.

table 5. Reduced major axis slopes (b), standard deviation Sb, y-
intercept (A), correlation coefficient (r) from the regression of log
length of cervical vertebra on log length of last presacral vertebra
for Macrocnemus bassanii. If the z value of the test for equality
of the slopes of the cervical vertebrae and the last presacral
vertebra is > 1 -96 the probability is > 0-05.

Vertebrae

b (slope)

Sb

A

r

z

m.dors

1-00

0-03

0-02

0-997

2

1 20

0 10

0 01

0-987

1-92

3

112

0 14

0-30

0-968

0-86

4

1-29

0 1 1

0-21

0-986

2-64

5

1 -37

0-09

0 13

0-989

4 1 1

6

1-47

0-08

0-02

0-993

5-22

7

1 26

0-05

0 14

0-997

4-33

8

1-07

—

0-21

—

—

neck

1-27

0-07

0-99

0-993

3-38

TSCHANZ: ALLOMETRY IN TRIASSIC PROLACERTIFORM REPTILES

1007

text-fig. 5. Reduced major axis of the cervical vertebrae
(C2-C8) and a middle dorsal vertebra (D) of Macro-
cnemus bassanii. log cerv = logarithmically transformed
length values of the cervical vertebrae, log dors = log-
arithmically transformed length values of the last pre-
sacral vertebra.

DISCUSSION

Evidence for positive allometric growth of the cervical vertebral column has been found in both
genera analysed. Growth was constant during life (linear regression line). The two accelerations
of the growth rate, as postulated by Wild (1973), one caused by unknown effects and the other by
sexual maturity, could not be substantiated. No slowing of the growth rate (deceleration) could
be observed for the largest specimens of T. longobardicus , in which the neck is relatively most
elongated. Wild (1973) also postulated ontogenetic growth changes of the humerus and the femur
(text-fig. 3). His data has been reanalysed too, and again linear regression lines resulted. The
hypotheses formulated in the introduction, postulating unchanged positive allometry, are thus
confirmed. The possibility remains that a larger sample size would result in modifications, but the
fundamental trends are obvious.

Different allometric parameters b for the cervical vertebrae of T. longobardicus and M. bassanii
indicate decelerated growth of the neck in the former taxon. Although this pattern is statistically
unsubstantiated, it merits closer scrutiny. Decelerated growth of the neck of T. longobardicus can
be explained by its body growth. It is possible that a structure with strong positive allometric
growth will become functionally inappropriate or inadaptive if the same allometric growth
parameter is maintained into a new size range (Gould 1966). Two strategies can be invoked to
avoid loss of adaptation in structures generated by positive allometric growth:

«, decrease of the allometric parameter b (slope of the regression line),

b , decrease of the allometric parameter A (y-intercept).

Adult specimens of T. longobardicus are obviously much larger than adult specimens of M.
bassanii ; in other words the two taxa belong to different size classes. T. longobardicus can grow to

1008

PALAEONTOLOGY, VOLUME 31

up to 6 m in length, while M. bassanii does not exceed a length of 1 m. Consequently, the
decelerated growth of the cervical vertebrae of T. longobardicus can be correlated with increased
body size. Decelerated growth indicates that the neck of Tanystropheus had reached an adaptive
limit. Half of the total body length of a 4-5 m long animal was taken up by its neck. As explained
above, the cervical musculature was not well developed in T. longobardicus. This would have
caused functional restrictions of the neck if Tanystropheus had given rise to larger forms. The only
way of bypassing the adaptive limit of neck growth with increasing body size would have been
decelerated allometric growth of the neck.

Decelerated growth of the neck is caused by a decreased rate of morphological development.
Decreased morphological development in the ontogeny of a hypothetical descendant indicates
neoteny (McKinney and Schoch 1985). In other words, if the neck growth of M. bassanii
corresponds with the neck growth of a hypothetical ancestor of T. longobardicus, the latter would
show neoteny in relative neck length. Unfortunately the differences of slope could not be verified
statistically.

Consequently, it is postulated that the extreme elongation of the neck in T. longobardicus results
from accelerated body growth. This pattern of growth to a new size range is called hypermorphosis.
Therefore, T. longobardicus could be no more than a hypermorphic M. bassanii. In a hypermorphic
taxon the onset of sexual maturity has been retarded in relation to the hypothetical ancestor
(McNamara 1986; McKinney 1986), in this case represented by M. bassanii. Because the cervical
vertebrae grew over a longer period, an extremely elongated neck resulted from its positive
allometry.

The hypothesis that T. longobardicus is no more than a hypermorphic M. bassanii neglects some
observations, e.g. the different values for the parameter A (y-intercept) (text-fig. 6) and the different
number of cervical vertebrae. Differences of the parameter A can be explained as a means of
avoiding inadaptive elongation of the neck. If the resulting neck is relatively shorter in the
descendant its functionality is retained. There are two possible ways of shortening the neck; either
its development starts out from shorter primordia, or the onset of its development is delayed. The
latter mechanism of paedomorphosis is called postdisplacement (McNamara 1986). Both of these
explanations are not applicable to T. longobardicus. In comparison to M. bassanii the cervical
vertebrae of T. longobardicus are not shorter but relatively longer, as is indicated by the higher
values for the parameter A for the latter. The onset of morphological development starts earlier.
This pattern is called predisplacement. The result is a prolonged period of growth of the cervical
vertebrae and resulting in a longer neck.

The cervical vertebral column of T. longobardicus comprises twelve vertebrae, four more than
in M. bassanii, and three more than in T. antiquus (Wild 1980a, b: Benton 1985; Wild 1987). T.
antiquus is closely related to the other two taxa, but it comes from older sediments than T.
longobardicus. The number of presacral vertebrae seems to be the same (24 to 25; Peyer 1937) in
all three taxa. Wild (1973) advanced the hypothesis that the 1st dorsal vertebrae of an ancestral
form had been transformed to cervical vertebrae in T. longobardicus. In other words, a backward
shift of the shoulder girdle with simultaneous transformation of the vertebrae would have occurred
during phylogeny. Assuming the hypothesis to be correct, this transformation would provide an
additional explanation for the extremely elongated neck of T. longobardicus. It might seem possible
that the addition of cervical vertebrae is more important for the elongation of the neck than are
the other parameters, such as hypermorphosis and predisplacement. At any rate, the effect of the
addition of vertebral elements should be detectable. If the regression analysis for the total neck
length is performed including only the anterior eight cervical vertebrae, identity of the resulting
regression line with that of M. bassanii is to be expected. However, the two regression lines only
approach each other, but are not superimposed (text-fig. 6). Therefore, the extreme elongation
of the neck of T. longobardicus can only partially be explained by the addition of dorsal
vertebrae. Hypermorphosis and predisplacement as parameters of heterochronic change are more
important.

TSCHANZ: ALLOMETRY IN TRIASSIC PROLACERTIFORM REPTILES

1009

log neck

text-fig. 6. Reduced major axis of the neck of Tanystropheus longobardicus
(traced line includes all the cervical vertebrae, dotted line includes only the
anterior cervical vertebrae 2 to 8) and Macrocnemus bassanii. log neck = logarith-
mically transformed length values of the cervical vertebral columns, log dors =
logarithmically transformed length values of the last presacral vertebra.

1010

PALAEONTOLOGY, VOLUME 31
CONCLUSIONS

Reinvestigation of the growth of the cervical vertebrae of T. longobardicus and M. bassanii has
shown that the elongation of the neck within the Prolacertiformes is caused by changes during
early ontogenetic development and differences of adult body size. The hypothesis that all the
studied specimens of T. longobardicus belong to a single species was confirmed because no evidence
for ontogenetic changes of the allometric parameters has been found. The second hypothesis,
dealing with constant positive allometric growth, has also been verified for both taxa analysed.

It is postulated that the elongation of the cervical vertebral column of T. longobardicus is caused
by several processes of heterochronic change, characterized as peramorphic growth (McNamara
1986). All assumptions have been made relative to a hypothetical ancestor of T. longobardicus ,
with a morphology exemplified by M. bassanii. The most important cause for the elongation of
the neck is hypermorphic growth, an evolutionary trend that occurs in the phylogenetic line of T.
longobardicus. Other causes such as predisplacement and an increased number of cervical vertebrae
have only a modifying character. It is supposed that in T. longobardicus the elongation of the neck
had reached a point where further elongation would have produced a functionally impossible
structure. Support for this hypothesis is given by the trend to reduce the allometric parameter b
in T. longobardicus.

The hypothesis, that the evolution of the Prolacertiformes can be deduced from the development
of the neck elongation, remains unresolved. More data would be needed about the ontogenetic
growth in other taxa, such as T. antiquus Hiihne, T. meridensis Wild, T. fossai Wild, Prolacerta ,
and Tanytrachelos. One form, T. ahynis Olsen, would be the most interesting to study. Approximately
100 specimens have been collected, which should form the basis for a successful statistical analysis
of ontogenetic growth. On the other hand, Tanytrachelos is geologically the youngest prolacertiform
known so far. In addition, this form remains very small. Therefore, it is possible that analysis of
its ontogenetic growth would reveal a stronger positive allometry than for Tanystropheus
longobardicus.

I propose that T. antiquus had a body size intermediate between M. bassanii and T. longobardicus.
Its cervical vertebral column comprises nine cervical vertebrae, one more than M. bassanii , and
these vertebrae are supposed to be more elongated. Recently, Wild (1987) reported a great number
of complete skeletons of juvenile T. antiquus from the Black Forest. Until the description of this
material, the intermediate position of T. antiquus remains open to question.

Acknowledgements. I would like to thank Dr Olivier Rieppel and Martin P. Sander from the Palaontologisches
Institut, as well as Professor Robert D. Martin from the Anthropologisches Institut, for their review of the
manuscript and editorial comments.

REFERENCES

bassani, f. 1886. Sui fossili e sull’eta degli schisti bituminosi triasici di Besano in Lombardia. Atti Soc. ital.
Sci. nat. 29, 15-72.

benton, m. j. 1985. Classification and phylogeny of the diapsid reptiles. Zool. J. Linn. Soc. 52, 575-596.
boas, J. E. v. 1929. Biologisch-Anatomische Studien fiber den Hals der Vogel. Mem. Acad. r. Sci. Lett.
Danemark, Sect. Sci. 9, 105-222.

carroll, R. L. 1981. Plesiosaur ancestor from the Upper Permian of Madagascar. Phil. Trans. R. Soc. B,
293, 315-383.

chatterjee, s. k. 1980. Malerisaurus, a new eosuchian reptile from the late Triassic of India. Ibid. 267, 209-
261.

currie, p. j. and carroll, r. l. 1984. Ontogenetic changes in the eosuchian reptile Thadeosaurus. Vertebr.
Paleont. 4, 68-84.

could, s. J. 1966. Allometry and size in ontogeny and phylogeny. Biol. Rev. 41, 587-640.
imbrie, j. 1956. Biometrical methods in the study of invertebrate fossils. Bull. Am. Mus. Nat. Hist. 108, 217-
252.

TSCHANZ: ALLOMETRY IN TRIASSIC PROLACERTIFORM REPTILES

101 1

kummer, b. 1975. Biomechanik fossiler und rezenter Wirbeltiere. Natur Mus. 105, 156 167.
mckinney. m. l. 1986. Ecological causation of heterochrony: a test and implications for evolutionary theory.
Paleobiology, 12, 282-289.

- and schoch, r. m. 1985. Titanothere allometry, heterochrony, and biomechanics: revising an evolutionary
classic. Evolution , 39, 1352-1363.

mcnamara, k. j. 1986. A guide to the nomenclature of heterochrony. J. Paleont. 60, 4-13.
meyer, h. von. 1855. Die Saurier des Muschelkalkes mit Riicksicht auf die Saurier aus Buntem Sandstein
und Keuper. In: Zur Fauna der Vorwelt , zweite Abtheilung, Frankfurt a.M.
munster, G. von. 1834. Vorlaufige Nachricht liber einige neue Reptilien im Muschelkalk von Baiern. Neues
Jb. Miner. Geol. Palaont. Abh. (1834), 521 526.

nopcsa, F. von. 1923. Neubeschreibung des Trias-Pterosauriers Tribelesodon. Palaont. Z. 5, 161-181.
olsen, p. e. 1979. A new aquatic Eosuchian from the Newark Supergroup (Late Triassic-Early Jurassic) of
North Carolina and Virginia. Postilla , 176, 1-14.

peyer, b. 1931 . Die Triasfauna der Tessiner Kalkalpen. II. Tanystropheus longobardicus Bassani. Abh. schweiz.
palaont. Ges. 50, 9 110.

1937. Die Triasfauna der Tessiner Kalkalpen. XII. Macrocnemus bassanii Nopcsa. Ibid. 54, 3-140.

preuschoft, h. 1976. Funktionelle Anpassungen evoluierender Systeme. In Evoluierende Systeme I und II;

Aufsatze u. Reden. Senckenberg. naturf. Ges. 28, 98-117.
tschanz, k. 1985. Tanystropheus— an unusual reptilian construction. In Konstruktionsprinzipien lebender
und ausgestorbener Reptilien. Konzepte SFB 230, Heft 4 (1985), 169 178.

1986. Funktionelle Anatomie der Halswirbelsaule von Tanystropheus longobardicus (Bassani) aus
der Trias (Anis/Ladin) des Monte San Giorgio (Tessin) auf der Basis vergleichend morphologischer
Untersuchungen an der Halsmuskulatur rezenter Echsen. Unpublished Dissertation, University of Zurich.
wild, r. 1973. Die Triasfauna der Tessiner Kalkalpen. XXIII. Tanystropheus longobardicus (Bassani), Neue
Ergebnisse. Schweiz, palaont. Abh. 95, 1-160.

- 1985a. Tanystropheus (Reptilia: Squamata) and its importance for stratigraphy. Mem. Soc. geol. Fr. ns
139, 201-206.’

19806. Die Triasfauna der Tessiner Kalkalpen. XXIV. Neue Funde von Tanystropheus. Schweiz, palaont.

Abh. 102, 1-43.

— 1987. An example of biological reasons for extinction: Tanystropheus (Reptilia: Squamata). Mem. Soc.
geol. Fr. ns 150, 37 44.

Typescript received 11 September 1987
Revised typescript 25 February 1988

KARL TSCHANZ

Palaontologisches Institut und Museum
der Universitat Zurich
Kiinstlergasse 16
CH-8006 Zurich

vi

VARIATION OF RECENT AND FOSSIL
CRASSOSTREA IN JAMAICA

by D. TIMOTHY J. LITTLEWOOD and STEPHEN K. DONOVAN

Abstract. Biological studies have indicated that the oysters Crassostrea virginica (Gmelin) and C. rhizophorae
(Guilding) may be a single species. This is surprising as they are morphologically dissimilar, C. virginica
being far larger and thicker than C. rhizophorae. We postulate that this variation may be ecophenotypic in
origin, a cause of gross variation in form in other oysters. To test our hypothesis, we have compared the
palaeoecology and ecology of Plio-Pleistocene C. virginica and Recent C. rhizophorae from Jamaica. A
spectacular Plio-Pleistocene deposit is dominated by C. virginica , other organisms being almost absent. One
exceptional bed, over 3 m thick, is dominantly composed of oysters. This sequence appears to have been
near-shore marine, or possibly estuarine, but, somehow, the environment was obviously highly favourable
for C. virginica. Conversely, modern C. rhizophorae mainly attach to mangrove rhizophores and may compete
with a very broad variety of organisms. Physical factors, such as salinity, can vary rapidly within this
environment. In consequence, C. rhizophorae seems to grow fast, reproduce early and die early, whereas Plio-
Pleistocene C. virginica grew to a large size which probably indicates considerable maturity. Environmental
stress necessitates a rapid life cycle for C. rhizophorae. Therefore, ecophenotypic variation may indeed be the
cause of morphological variation between C. virginica and C. rhizophorae. However, detailed studies on living
populations of both species are considered essential to test this hypothesis further.

In the fossil record of many organisms, such as oysters and scallops, we can only attempt to
differentiate between evolutionary and ecophenotypic variation if we have tight stratigraphic
control and large samples for statistical analysis (for example, Bayer et al. 1985; Johnson 1981).
However, different methodologies are used to determine such variation more precisely in modern
organisms. Gunter (1954, p. 134) stated that ‘within certain limits, defined by the fact that the
shells consist of two hinged valves, oysters are among the most plastic organisms known’. This
plasticity in shell form has caused much confusion in oyster taxonomy as many morphological
variants of one species are similar to those of others. Indeed, the distinction of an oyster genus
upon shell morphology alone has been questioned by a number of authors (for example, Ranson
1942; Gunter 1950), as macroform is strongly influenced by substrate (Galtsoflf 1964; Palmer and
Carriker 1979).

The two oysters Crassostrea virginica (Gmelin) and C. rhizophorae (Guilding), which are on first
sight morphologically distinct, may be two end members of a single, highly variable taxon (a
review of the literature comparing C. rhizophorae with C. virginica is given in Newball and Carriker
1983).

C. virginica and C. rhizophorae each have a diploid number (2n) of 20, hybridize readily (Menzel
1972, 1973), have morphologically similar karyotypes (Rodriguez-Romero et al. 1979), and, by
means of electrophoretic studies, it has been shown they share approximately 72% of the same
genes (Buroker et al. 1979). Menzel (1972, 1973) suggested C. rhizophorae may be a subspecies of
C. virginica , but although these ‘species’ hybridize readily in the laboratory, such a phenomenon
would not necessarily occur under natural conditions (Menzel 1971). Survival of hybrids between
these ‘species’ was 34% after one year and compares favourably with survival rates of 25% and
72% of pure bred C. rhizophorae and C. virginica over the same period (Menzel 1971). Detailed
ultrastructural examinations of young individuals of each species have led Newball and Carriker
(1983) to suggest that C. rhizophorae is an ecotype of C. virginica.

(Palaeontology, Vol. 31, Part 4, 1988, pp. 1013-1028, pi. 91. |

© The Palaeontological Association

1014

PALAEONTOLOGY, VOLUME 31

C. rhizophorae and C. virginica are not the only species within the genus Crassostrea Sacco, 1897
to show close affinities. For instance, Singarajah (1980) believed C. rhizophorae to be synonymous
with both Ostrea arhorea and C. ( Ostrea ) brasiliana Lamarck, and Durve (1986) has likened C.
madrasensis (Preston) to C. virginica. On the other hand, physiological variation within the species
C. virginica has also been demonstrated (Stauber 1950; Loosanoff 1958), where morphologically
indistinguishable groups within this species are considered to be physiological races that are
functionally different from one another. Palmer and Carriker (1979) review factors suspected to
affect shell morphology in C. virginica and other ostreids. The list includes substrate, culture
technique (bottom and off-bottom), temperature, current velocity, turbidity, salinity, and exposure
to direct sunlight. However, C. virginica is never seen to vary so much that it appears to approach
C. rhizophorae closely in morphology.

text-fig. 1 . Small specimens of the attached valves in
Crassostrea virginica (Gmelin). A, fossil specimen from the
Plio-Pleistocene Round Hill Beds of Jamaica, b. Recent
specimen from Prince Edward Island, Canada. Both x 0-45.

The features which differentiate the two ‘species’, C. rhizophorae and C. virginica, include heavier
muscle scar pigmentation and greater lower left valve plication in C. virginica (Gunter 1951;
Galtsoff 1964). Additionally, maximum height of C. virginica approaches 400 mm, whereas that
of C. rhizophorae rarely exceeds 100 mm. It has already been mentioned that substrate affects
macroform. Even though the habitats of C. rhizophorae and C. virginica, mangrove prop roots
and soft sediment or hard, shelly substrates, respectively, may not explain the difference in plication,
other factors, such as different growth rates, may be at least indicative of cause and effect. Although
Mattox (1949) failed to find evidence of alternational hermaphroditism in C. rhizophorae, a feature
common to the genus, Angell (1986) suggested protandrous hermaphroditism may occur in this
species. The evidence includes the predominance of females in populations of C. rhizophorae
(Angell 1973), the presence of hermaphroditic gonads, and the observation that males tend to be
smaller than females (Angell 1986).

If, indeed, C. virginica and C. rhizophorae are members of a single, highly variable species, then
nobody has yet explained why they are so different. In the case of such variation within a species
of fossil oyster, difference of environment is usually cited as the probable principal reason for
variation in form. Herein, we examine the environments of C. virginica and C. rhizophorae in
Jamaica. C. rhizophorae, the mangrove oyster, is a common element of the modern fauna, but C.
virginica is extinct in Jamaica and is only known from the Plio-Pleistocene. One exceptional fauna,
dominated by the latter taxon, sheds light on the environment of C. virginica and enables us to
make at least some comparisons with modern C. rhizophorae. The fossil C. virginica are
morphologically indistinguishable from Recent members of the same ‘species’ (text-fig. 1).

LITTLEWOOD AND DONOVAN: CRASSOSTREA IN JAMAICA

1015

CRASSOSTREA VIRGINICA IN THE PLIO-PLEISTOCENE OF JAMAICA

The highly fossiliferous succession in the August Town Formation of the Coastal Group (late
Miocene to Pleistocene) at Round Hill, Clarendon, Jamaica (text-fig. 1) is a sequence of more or
less sandy limestones, with a fauna dominated by benthic molluscs, foraminiferans, and corals,
with rare clypeasteroid echinoids. Dips are steep to the south or vertical, and the outcrop is
cut by occasional faults. This coastal section was first described by Duncan and Wall (1865, p. 6,
fig. 4), who considered the succession to be comprised of Miocene sediments overlain by a white
limestone. Robinson (1968) correctly reinterpreted the structure as a possibly conformable contact
between the underlying Newport Formation of the White Limestone Group and the younger
Round Hill Beds which, however, are in turn unconformably overlain by cemented limestone screes
of late Pleistocene age derived from Round Hill itself. Robinson (1968, p. 46) noted \ . . Several
remarkable beds of oysters occur near the base of the sequence, with the oysters in an original
position of growth, and with many individual shells reaching 15 inches or more in length’. Prescott
and Versey (1958, p. 39) considered that these oysters resembled O. haitiensis Sowerby. The age
of the Round Hill Beds is probably Pliocene, perhaps extending into the early Pleistocene
(E. Robinson, written comm.).

The Round Hill Beds have yet to be described in detail. Herein we only wish to discuss a small
part of the sequence that includes a remarkable bed, over 3 m thick and dominated by C. virginica
(Gmelin) (PI. 91, figs. 1 and 2; text-fig. 3), which outcrops on Farquhars Beach at Jamaica grid
reference H415345 (text-fig. 2). A measured section from this locality is illustrated in text-fig. 3.
Eight beds are recognized in this part of the sequence. Bed 1 (the lowest in text-fig. 3) is a sandy
limestone with limestone pebbles, some of which are bored. The fauna consists solely of
dissociated valves of C. virginica , which are only present towards the top of the bed. This is
succeeded by a unit with an abrupt, planar, and apparently erosive base. The top is uneven
and thickness is variable. This bed is dominated by C. virginica , most shells being dissociated and
often apparently broken. No other faunal elements are present at this horizon. This unit may
represent a channel fill or shell bank, with all valves recumbent, unlike the vertically orientated
concentrations of dead, dissociated C. virginica valves found off the Florida coast (Grinnell 1974).
The overlying bed 3 is a sandy, nodular, white to orange banded limestone. This has been cut into
by bed 4, which has the geometry of a channel. As with bed 2, bed 4 is principally composed of
mainly dissociated, recumbent, and possibly broken valves of C. virginica. Although some valves
retain encrusting basal plates of Balanus spp., no complete barnacles are preserved and no other

text-fig. 2. Locality map showing the
position of the principal outcrop of
the Round Hill oyster bed, Claren-
don, south-central Jamaica, WI. Fossil
locality on Farquhars Beach marked by a
star; summit of Round Hill by a triangle.
Inset map shows position of Round Hill
(RH) and Bowden (B). North towards
top of page in both maps.

1016

PALAEONTOLOGY, VOLUME 31

EROSIVE TOP

Laminations about 1cm thick- Some nodules-
Sphaerogypsina very common- Rare vertical
burrows cf- Skolithos-

Less well cemented towards top- No C- virginica
but Sphaerogypsina common-

Well-cemented sandy limestone with pebbles,
C- virginica arid Sphaerogypsina-

C- virginica Bed- Base gradational- Shells close
packed- Associated valves common throughout
section - Shells mainly in life position near top-

KEY

PEBBLES

ms

BORED PEBBLES

<C> <5*

CALCAREOUS NODULES

SANDY LIMESTONE

u

VERTICAL BURROWS

• * # t

SPHAEROGYPSINA

RECUMBENT C- VIRGINICA

/

C- VIRGINICA IN LIFE

_y

POSITION-

Channel filled by C- virginica-

Sandy, nodular limestone-

Dissociated C- virginica valves-

Sandy, pebbly limestone with dissociated
C- virginica-

text-fig. 3. Graphic, annotated log of
the Round Hill Beds at the fossil locality
marked in text-fig. 1. Widths of units
indicate how beds have weathered rela-
tive to each other at this locality.

EXPLANATION OF PLATE 91

Figs. 1 11. Crassostrea virginica (Gmelin) at Farquhars Beach, Clarendon, Jamaica. I, general view of north-
west end of sequence illustrated in text-fig. 3. Top and bottom of bed 6 (about 3-3 nr thick) indicated. 2,
detail of beds 1 (bottom) to base of 6, shown towards the left of text-fig. 1. Hammer (280 mm long) resting
against bed 5. 3, bored oyster in bed 7, x0-40. 4, curved, adult shell in upright, life position and

encrusted by numerous, juvenile oysters, xO-17. 5, large valve with single boring, xO-46. 6, large, upright
valve showing a triangular ligament area about 45 mm in length, x 0-38. 7, particularly thick shell, x 0-25.
8, very large, recumbent oyster, x0-18. 9, paired, upright valves showing external evidence of boring,
xO-34 10, large recumbent shell encrusted by a pair of younger oysters which are almost as large, and

in the same orientation, as the adult, xO-22. 11, upright valve encrusted by Balanus sp., x0-18.

Specimens in figs. 4 11 all from bed 6. All figures are of uncoated specimens taken in the field.

PLATE 91

LITTLEWOOD and DONOVAN, Crassostrea virginica

1018

PALAEONTOLOGY, VOLUME 31

fauna noted. Both beds 2 and 4 have sharp contacts with their underlying and overlying units. Bed
5 is similar to bed 3 but occasional dissociated valves of C. virginica occur near the top. This unit
grades into the overlying main oyster horizon, bed 6 (PI. 91, fig. 1), which is 3-3 m thick and
dominantly composed of C. virginica , preserved variously as broken shell fragments, dissociated
valves (PI. 91, fig. 6), recumbent, associated valves (PI. 91, figs. 7, 8, 10), and upright, associated
valves (PI. 91, figs. 4 and 9). Balanus spp. (PI. 91, fig. 11) and juvenile C. virginica (PI. 91, figs. 4
and 10) encrust valves, on both the inner and outer surfaces. Young oysters are particularly
prominent on some of the largest, upright, mature specimens of C. virginica near the top of the
bed (PI. 91, fig. 4). Additionally, some shells are encrusted on their lower valve by juvenile C.
virginica. The only other body fossils are rare, thick-walled calcareous tubes of uncertain affinity
(possibly annelids?) and a single gastropod. Some shells of C. virginica have been bored (PI. 91,
figs. 5 and 9), probably post-mortem, by bivalves and clionid sponges (an exposure of a further
C. virginica horizon to the north-west includes common calcareous tubes, plus valves bored by
polydorid polychaetes). The matrix is an orange limestone, with larger, sand-sized grains probably
being derived from fragmented oyster shells. The matrix is more muddy towards the bottom of
the bed and more gritty towards the top. Valves in life position occur throughout this unit but are
concentrated at particular horizons, especially towards the top, where shells reach 400 mm in
height. Such shells are amongst the largest C. virginica known.

Bed 7, in contrast, contains only rare, mainly disarticulated and occasionally bored, valves of
C. virginica in its lower half (PI. 91, fig. 3), with occasional pebbles and the spherical benthic
foraminifera Sphaerogypsina, in a well cemented, sandy, orange limestone. C. virginica shows little
or no encrustation at this level. In the upper half of this bed C. virginica is absent but Sphaerogypsina
is very common, often being preserved as clusters of tens or hundreds of individuals. The overlying
bed 8 consists of finely laminated limestone horizons, each about 10 to 40 mm thick and
differentiated by being alternately more or less well cemented. Some of these horizons appear to
be nodular. Sphaerogypsina is very common and dominates some units. Occasional moulds of
bivalves and simple vertical burrows, cf. Skolithos , are present. The sequence is truncated by an
angular unconformity with the overlying limestone screes derived from Round Hill.

The presence of common, upright, articulated shells in bed 6, many of which retain a well-
preserved epifauna (PI. 91, figs. 4, 10, 11), indicates that some, if not all, of these oysters are
preserved in situ, with minimal or no transport. The origin of oyster beds 2 and 4 is more
problematic. Bed 4 has the geometry of a channel fill; bed 2 is either a channel fill or a shell bank.
It is difficult to envisage large C. virginica valves being transported very far, except under very
high energy conditions, perhaps due to storm action, and abrasion is minimal. There are several
indications that this sequence was shallow water in origin (see discussion below) and, therefore,
well within the lower limits of storm wave base. Nevertheless, it is possible that these are in situ
shell deposits which have been little altered in geometry by gross physical processes. Although
Ager (1963, p. 200) concluded that \ . . epibiontic communities will almost invariably be moved
and dispersed before fossilization . . .’, some studies indicate that dead shells often accumulate
with little or no post-mortem transport. Reineck and Singh (1973, pp. 134-136) recognized that
shell concentrations are produced both by post-mortem transport and dumping or in situ
accumulation. Warme (1969) concluded that, even within a high energy sand channel environment,
transportation of shells away from their life habitat was minimal within a coastal lagoon. Holme
(1961, pp. 433, 443) and, in a much more detailed study, Carthew and Bosence (1986), noted that
live and dead shell-gravel assemblages on the shallow shelf off Plymouth, UK, had essentially
similar molluscan faunal compositions and agreed that post-mortem transport was negligible.
These are important conclusions when we recognize the great size of C. virginica compared with
most other bivalves. Intuitively, we must conclude that only particularly high energy conditions
would be capable of transporting even an uncemented C. virginica. Seilacher (1984, pp. 215-217)
considered Crassostrea (possibly thinking more of the common European species C. angulata
(Lamarck), the Portuguese oyster) to be well adapted as a ‘boulder-shaped recliner’ on soft sediment
and noted that storm tells of this taxon sometimes reach 20 m thick, soft sediment presumably

LITTLEWOOD AND DONOVAN: CRASSOSTREA IN JAMAICA

1019

A B

text-fig. 4. A suggested sequence for the passive formation of a Crassostrea
virginica channel fill, a, soft, calcareous sediment stabilized by sea grass, b,
channel formed by storm action, c, invasion by C. virginica. D, eventual burial

of channel.

being removed by winnowing. This is a potential explanation of all Crassostrea beds at Round
Hill, particularly bed 6.

An alternative scenario for development of an in situ C. virginica channel fill deposit is illustrated
in text-fig. 4. It is possible that the sea-floor sediment was stabilized by vegetation, possibly sea
grass, at least in the lower part of the section (Brasier 1975; Eva 1980; text-fig. 4a). Modern sea
grass communities of Jamaica are not favourable habitats for C. rhizophorae and we might speculate
that they would also have been unsuitable for C. virginica at Round Hill; certainly, in those
limestone units apart from the three shell beds (= substrates that may have been stabilized by sea
grass), C. virginica is uncommon and almost always disarticulated. However, removal of the sea
grass might have encouraged successful spatfalls of oysters. One event that would remove sea grass
would be the formation of a channel (text-fig. 4b), possibly during a storm. The substrate, cleared
of vegetation, would now be more suitable for colonization by C. virginica (text-fig. 4c), although
the oyster would not be able to spread out of the channel. We could thus develop a passive channel
fill, with disarticulation and abrasion being produced by relatively low energy post-mortem
processes with some slight transport. It is unlikely that breakage of valves would be produced by
weight of overburden (Rettger 1935). Final burial (text-fig. 4d) could result from a number of
causes.

The main oyster bed, 6, is much thicker than either beds 2 or 4. It is visible over about 90 m
of coastal exposure and may represent a very large channel deposit, appearing to thin to the south-
east (north-west end obscured by slipped material), or is perhaps even a laterally extensive bed.
Many of the oysters are in life position (PI. 91, figs. 4-11). The only other mollusc found was a
single gastropod near the top of the bed. Conditions thus appear to have been extraordinarily
favourable for C. virginica , to the virtually complete exclusion of all potential molluscan competitors.
What might those conditions have been? Certainly evidence from various parts of the Round Hill
section indicate that this sequence was deposited in a shallow water environment. In the overlying
bed 8 there are occasional vertical burrows, suggestive of Seilacher’s (1967) Skolitlios ichnofacies
and indicative of littoral deposition. Channelling, possibly due to storm action in shallow water,
is found in bed 2 and possibly 4. Elsewhere in the section molluscan assemblages appear similar
to those found within snorkelling depth today. The presence of two species of clypeasteroid, Encope
aff. sverdrupi Durham and Clypeaster cf. rosaceus (Linnaeus), is possibly also indicative of shallow
water conditions. In particular, we have never seen the large, heavy tests of modern Clypeaster
washed up on beaches; it is always found subtidally, even after death, forming a hard substrate
for encrusting and cryptic organisms. The two species of acorn barnacle found in bed 6 suggest
restricted marine to brackish conditions. Balanus improvisus assimilis Darwin is common in modern

1020

PALAEONTOLOGY, VOLUME 31

inshore, near-marine habits, whereas B. eburneus Gould is characteristically estuarine (Dr P. R.
Bacon, written comm.). There is no indication that Crassostrea virginica was a mangrove oyster,
unlike C. rhizophorae.

If we accept this environmental assessment, then it is apparent that C. virginica was living in
shallow, well oxygenated and highly energetic water. Plankton would probably have been in ample
supply, but the substrate would have been unsuitable for the growth of sea grass, being composed
primarily of oyster valves. Other organisms were obviously largely excluded, although we cannot
speculate whether this was due to the oysters influencing the environmental conditions or to a
prevalent condition that encouraged C. virginica initially. Certainly, once established, a substrate
dominated by oyster valves would have been unsuitable for burrowing molluscs to colonize. A
third possibility, perhaps less probable, is that other mollusc shells have been winnowed away.
Nevertheless, large valves of Strombus sp., found elsewhere at Round Hill, were probably as heavy
as the shell of C. virginica , yet are absent from the measured section.

Salinity and dissolved calcium carbonate content were probably normal or possibly brackish.
The above wave base, high energy conditions would have kept sediment mobile and prevented
inorganic fouling of the valves. Indeed, energy conditions appear to have been so high that sediment
within bed 6 was largely winnowed away. Oyster spatfalls could settle on both soft and, more
probably, hard substrates. Experiments by D.T.J.L. have shown that growth in young C. rhizophorae
is most vigorous on the underside of attachment surfaces. Well-preserved shells of young C.
virginica seen growing on the lower valves of adult oysters are thus possibly indicative of similar
settlement rather than of reworking.

CRASSOSTREA RHIZOPHORAE IN THE RECENT OF JAMAICA

C. rhizophorae lives in many of the mangrove stands found around the coast of Jamaica. The
largest population of C. rhizophorae is found at Bowden, St Thomas (text-fig. 2; GR N788362),
where the red mangrove, Rhizophora mangle , fringing the smaller inner bay, supports most of the
population. Collection and culture of young spat for commercial purposes takes place in the larger
outer bay (Wade et al. 1981).

The bottom of each bay is covered in thick layers of fine, muddy sand with occasional outcrops
of the turtle grass, Thalassia testudinum. The inner bay is less than 1 m deep and is fed by two
small rivers. The salinity and temperature vary between 5-35 %0 and 25-28 °C, respectively,
throughout the year. Although salinity in the outer bay rarely falls below 35 %0 (unpublished data,

D. T.J.L. and Oyster Culture Jamaica Project, Ministry of Agriculture), the oyster thrives in these
marine conditions. The tidal range is approximately 350 mm (Meteorological Service, Kingston)
but occasionally varies with heavy rainfall and winds.

Hubbard (unpublished data) studied the distribution of C. rhizophorae in the swamps at Bowden
and found the greatest number to occur 6-9 m behind the mangrove fringe. Characteristically the
oyster cements itself to any substrate relatively free from other organisms. Although this settlement
is usually on young rhizophores, the shells of the bivalve Isognomon alatus Gmelin and mature C.
rhizophorae often serve as a substrate for the oyster. Siung ( 1976) showed that 70-7% of mangrove
oyster spat settle in the intertidal zone and that competition for food and space from other
organisms prevents successful recruitment in the subtidal zone.

Table 1 is a list of fauna and flora found in association with C. rhizophorae in Bowden. Many
of these species were collected from subtidally hung oyster substrate and are therefore not
necessarily present in the intertidal zone of the mangrove swamp where C. rhizophorae is naturally
dominant. The listing largely reflects the interests of those collectors who are responsible for
identifying the species. None the less, similar fouling communities have been described for mangrove
swamps in Puerto Rico (Glynn 1964; Cerame-Vivas 1974), the Bahamas (Riitzler 1969), Martinique
(Saint-Felix 1972), Venezuela (Sutherland 1980), and Port Royal, Jamaica (Goodbody 1961; Bruce
1968; Siung 1976). More extensive lists of fauna associated with R. mangle, in Cuba, may be found
in Rueda and Moreno (1985) and Rueda et al. (1985). Although continuous breeding and settlement

LITTLEWOOD AND DONOVAN: CRASSOSTREA IN JAMAICA

1021

table 1. A list of species collected from mangrove stems and artificial
substrates in the inner and outer bays at Bowden. This is not a complete list
and is largely biased by the collecting specialities of those who identified the
organisms. However, such a compilation is more complete than similar faunal
lists presented for fossil oyster beds. Life habitat and trophic group are as
follows:

ec = epifaunal cemented; b = epifaunal byssate; f = epifaunal free-
living; s = suspension feeding; h = herbivorous; d = deposit feeding;
c = carnivorous.

Identifications (ID); £ = D.T.J.L.; * = I. Goodbody (Zoology Dept.,
UWI, Jamaica); § = P. T. Hatfield (Biology Dept., Dalhousie University,
Canada); f = K. E. Conlan and E. L. Bousfield (National Museum of
Natural Sciences, Canada); A = R. H. Hubbard (Institute of Marine Affairs,
Trinidad); A = S. Prudhoe (retired, British Museum (Natural History),
UK).

Phylum

Life Trophic ID
habitat group

FAUNA

PORIFERA

Various unidentified groups ec s

COELENTERATA

Hydroids ec s

Aiptasia tagetes ec s

BRYOZOA

Bugula sp. ec s

Caulibugula sp EC s

Membranipora tenuis EC S

MOLLUSCA
GASTROPODA

Murex recurvirostris rubidus F. C. Baker f c

Littorina angulifera Lamarck f h

Melongena melongena Linnaeus F c

Caecum nebulosum (Rehder) f

Cymatium pileare Linnaeus f c?

C. muricinum Roding f c?

Vermicularia knorri Deshayes ec s

BIVALVIA

Ostrea frons Linnaeus EC s

O. equestris Say ec s

Isognomon alatus Gmelin b s

Anomia simplex Orbigny b s

Brachidontes recurvus Rafinesque B s

Modiolus americanus Lamarck b s

PLATYHELMINTHES

Stylochus ( Stylochus ) frontalis Verrill f c?

ANNELIDA

Sabellastarte magnifica (Shaw) EC s

Poly dor a sp. ec s

Spirorbis sp. EC s

Serpulidae

ARTHROPODA

CRUSTACEA

Balanus eburneus (Gould) ec s

B. amphitrite Darwin ec s

£

A

A

A

£

s

§

£

£

£

£

£

£

£

£

£

£

£

£

£

Table 1 continued overleaf ]

1022

PALAEONTOLOGY, VOLUME 3

TABLE 1 ( COUt .)

B. improvisus assimilis Darwin EC s

Ch thalamus angustitergum (Pilsbry) EC s

C. proteus Dando & Southward ec s

AMPHIPODA

Ericthonius brasiliensis Dana F D

Dulichiella appendiculata Say f d

Corophium bonellata Milne-Edwards F d

Ampithoe ramondi Audouin f d

Elasmopus sp. F D

Grandidierella sp. f d

Caprellids F d

DECAPODA

Panopeus herbstii H. Milne-Edwards f c

Aratus pisoni (Milne-Edwards) f

Goniopsis cruentata F

Mithrax mithrax spinosissimus (Lamarck) F H

Callinectes sapidus Rathbun f c

Alpheids f c

CHORDATA

Ascidians

Botrylloides nigrum Herdman EC s

Symplegma brackenhielma Michaelsen ec s

Diplosoma listeranum Milne-Edwards ec s

D. glandulosum Minniot EC S

Lissoclinum abdominale Minniot EC s

Didemnum psammathodes Sluiter ec s

Didemnum sp. ec s

Polyclinum constellation Savigny EC S

Perophora viridis Verrill ec s

Ecteinascidia styeloides Transtedt EC S

Ascidia nigra Savigny EC S

Styella canopus Savigny EC s

Fish

Bathygobius sopor at or (Valencinnes) f c

Hypleurochilus aequipinuis (Gunther) f c

FLORA

ALGAE

Enteromorpha sp.

Ulva sp.

Caulerpa racemosa (Forsk)

Dictyota sp.

§

§

§

t

t

t

t

t

t

#

#

t

*

*

*

*

*

*

*

*

*

*

*

*

A

A

#

#

#

of marine invertebrates tends to occur in the tropics (Goodbody 1962, 1965), Sutherland (1980),
studying the dynamics of the epibenthic mangrove root community in Venezuela, noted that there
was little recruitment of species or change in specific composition during an 18 month period. He
also showed that the low rate of recruitment on mangrove prop roots could be correlated with a
low rate of supply of new roots (an increase of ~ 8% yr_1 in Venezuela).

The ecological role of individual members of temperate littoral communities is better understood
than that of their tropical counterparts. Organisms sharing similar biologies are most likely to
compete with one another for food and space, but there appears to be little experimental evidence
in the literature demonstrating this with tropical species or for those groups of organisms

LITTLEWOOD AND DONOVAN: CRASSOSTREA IN JAMAICA

1023

responsible for biofouling. For example, ascidians, of which there are twelve species at Bowden
(see Table 1 ), are frequently referred to as possible competitors with oysters (for example, Loosanoff
1962; Arakawa 1980), but there is little evidence to confirm this. Overgrowth of one species by
another is a frequent method of competition for space between sessile feeders and occurs amongst
bryozoans, ascidians, sponges, bivalves, gastropods, tube-forming polychaetes, barnacles, hydroids,
and corals (see reviews by Jackson 1977; Branch 1984). Didemnum psammathodes certainly appears
to affect subtidally cultivated C. rhizophorae in this way, although the ascidian is not found in the
intertidal zone where natural mangrove oyster populations dominate. Indeed, only bivalve species
have been shown to compete with other bivalves for food and thereby reduce the growth and
condition of their competitors (for example, Engle and Chapman 1952).

Growth rates of C. rhizophorae vary between 0-25-0-35 mm day"1 in the current culture system
used in Jamaica (Littlewood 1987), 0-42-0-50 mm day"1 when cultivated on mangrove sticks in
Cuba (Saenz 1965), and 0T-0-2 mm day"1 on natural mangrove roots in Puerto Rico (Mattox
1949). Warmke and Abbott (1961) report that this species varies in 'length’ between 50 mm
and 150 mm but Nikolic and Alfonso (1971) have recorded maximum heights of approximately
100 mm after 9 months in Cuba. There is little information on the mortality of C. rhizophorae in
its natural habitat, but some data are available on its performance in culture systems. Mortality
values varying between 15-59% during the dry season and 1-20% during the rainy season have
been recorded (Bosch and Nikolic 1975). Mortalities as high as 91-2% have been recorded before
animals had reached 50 mm in shell height (Nikolic and Alfonso 1971; Nikolic et al. 1976) and
97% within 6 months of settlement (Bosch and Nikolic 1975), although little information is
available on what causes these high mortalities (Nikolic 1969). Studies in progress suggest the
flatworm Stylochus ( Stylochus ) frontalis Verrill, the hairy triton Cymatium pileare Linnaeus, the
porcupine fish Diodon hystrix Linnaeus, and the blue crab Callinectes sapidus Rathbun all contribute
to heavy mortality through predation, although post-spawning stress, disease, and the effects of
silt load in the water column have yet to be investigated.

Although Crassostrea rhizophorae tends to breed and settle all year round in the Caribbean (e.g.
Mattox 1950; Nascimento et al. 1980), there are generally two distinct spatfalls in Jamaica which
coincide with the rainy seasons, beginning in May and October of each year. This contrasts with
the single spatfall of extant C. virginica which extends from July to October depending on the
locality (see, for example, Andrews 1955; Beaven 1955).

DISCUSSION

There are obvious difficulties in attempting to compare the life habits of a group of fossil animals
with an extant form (see Hallam 1 965). For example, we are largely unable to discuss the importance
of predation or competition from associated fauna when much of this may have been either soft-
bodied or too brittle to be represented in the fossil beds. None the less, certain broad comparisons
and inferences are possible. First, the Round Hill Crassostrea species probably lived longer than
the Bowden species does today. Galtsoff (1964, p. 20) noted that '. . . as a rule, oysters do not
stop growing after reaching certain proportions but continue to increase in all directions and,
consequently, may attain considerable size’. The fossil C. virginica are considerably larger than C.
rhizophorae. Indeed, the largest specimens from Round Hill seem to be some of the largest shells
of C. virginica ever to be found. According to Galtsoff (1964), the largest, living specimen of C.
virginica to be documented was that found by Ingersoll (1881, pi. 30, p. 32). The shell measured
355 mm in height and 1 10 mm in length. We observed a fossil oyster shell from Round Hill that
measured approximately 390 mm in height and 125 mm in length (PI. 91, fig. 8).

Secondly, the Plio-Pleistocene group of Crassostrea appear to have been essentially shallow
water or estuarine, either intertidal or subtidal, and benthic in their habitat, whereas the Recent
group are predominantly intertidal, cemented directly or indirectly to narrow mangrove rhizophores.
The features of these habitats may suggest how the life styles of each ‘species’ differed following
settlement.

1024

PALAEONTOLOGY, VOLUME 31

Mangrove swamps are typically muddy environments with high concentrations of suspended
matter resulting from a continuous rain of leaf litter and detritus. Oyster spat settling on the
muddy bottom would be quickly buried in an organic, and occasionally silty, downpour and would
have to grow at a tremendous rate to facilitate water exchange for respiration, feeding, and
excretion. Stenzel (1971, pp. N1044-N1045) noted that \ . . oyster larvae avoid settling on mud-
covered substrata . . . [and due to heavy siltation] tend to colonise the undersurfaces of inclined
mangrove stems rather than their top surfaces’. By attaching to a rhizophore, or to the shell of an
animal already attached, the oyster ensures that it is above the detritus settling zone, although it
is restricted to a vertical range limited by the position of the prop root in the mangrove swamp.
Seaward fringing mangrove trunks and rhizophores are in shallow but relatively deeper water than
those further landward. With the low rate of root generation (Sutherland 1980), available substrate
is scarce for all epibionts settling in the subtidal and intertidal zone. Consequently, inter- and
intraspecific competition for space may limit recruitment and high population densities may limit
growth through competition for space and food.

The oyster’s ability to withstand aerial exposure by adducting its valves predisposes it to an
intertidal existence where tolerance to respiratory, thermal, and desiccation stress is required. In
the intertidal zone the substrate is subject to movement relative to mean tidal levels. If oysters
settle on leaf bearing stems or trunks, they may also be carried out of the tidal range and left
permanently exposed as the mangrove tree grows. Given these features of the mangrove
environment, one can see why the mangrove oyster may be restricted within the intertidal zone.

Hallam (1965) reviewed environmental features which may cause stunting in living and
fossil marine benthonic invertebrates. Following his guidelines for determining whether or not
environmental features may be responsible for a relatively smaller, ‘stunted’ animal (in this case
C. rhizophorae versus C. virginica ), we can investigate further the possibility that these two bivalves
are ecophenotypes of one species and that their environments have caused the observed differences.
Flallam (1965) considered the following to be principal factors: food supply, salinity, oxygen
content, turbidity, agitation, and temperature of the sea water, together with population density.
In view of our lack of associated fossil evidence or technical ability to describe more clearly the
Round Hill Beds, we are restricted to considering only a few of these features.

Hallam (1965, p. 134) noted that ‘. . . the actual consumption of food is more important than
its availability and is obviously the prime factor controlling size variations’. As an essentially
intertidal bivalve C. rhizophorae may only feed during periods of tidal immersion, but the high
growth rates and high organic content of the water do not suggest a food shortage. Furthermore,
Littlewood (in press) has shown that aerial exposure may enhance growth in the mangrove oyster.

The inner bay at Bowden is fed by two small rivers and salinity may fall markedly and rapidly
during periods of high rainfall. Although Crassostrea species are known for their euryhalinity,
rapid drops in salinity caused by heavy rainfall are known to result in mass mortalities of tropical
marine fauna (Goodbody 1961). By closing their shells the oysters can withstand limited periods
of physiological stress (cf. aerial exposure). However, if exposure to fresh water is prolonged,
oysters are unable to feed or respire aerobically and eventually die (see Andrews 1982). The sudden,
low salinities brought on by heavy rain may therefore limit the life span of organisms in tropical
environments such as mangrove swamps which, if not actually fed by rivers, would certainly
experience coastal run off. The relatively calm waters in mangrove swamps may retain the fresh
water for long periods. Goodbody (1961) noted that C. rhizophorae , and many other species in
the Kingston Harbour mangrove swamps, were adversely affected by heavy rainfall during the
rainy seasons. Only those species well below the less dense hyposaline waters were capable of
survival. None the less, C. rhizophorae was one of the first organisms to recolonize the swamps
following return to more marine salinites, although it was unable to re-establish as quickly as the
ascidians. This may further explain the exclusion of the mangrove oyster from the subtidal zone,
which is often dominated by the soft-bodied ascidians. Goodbody (1961, p. 155) suggested that
‘. . . the mangrove root communities of the lagoons in Kingston Harbor may seldom reach a
climax condition due to repeated destruction of the developing communities’.

LITTLEWOOD AND DONOVAN: CRASSOSTREA IN JAMAICA

1025

At Round Hill, C. virginica was coastal and near shore, either intertidal or shallow subtidal,
thus experiencing at least moderate wave action which thereby supplied sufficient oxygen. Mangrove
swamps are generally well oxygenated (see, for example. Bacon 1970). As mentioned above, the
effects of heavy siltation, observed at Bowden, have not been investigated and remain a possible
cause for the ‘stunting’ of mangrove oysters. During heavy rainfalls water agitation and the large
volume of silt in the water column would result in a more turbid environment. Although no
evidence of heavy siltation at Round Hill exists, and despite a proposed moderate level of water
agitation, the fossil oysters cannot be considered as ‘stunted’. Seilacher (1984, p. 214) noted that
cemented bivalves ‘eventually lift-off the substrate in order to facilitate water circulation, to widen
the shell cavity, and to defend against overgrowth’. Some of the oysters from Round Hill appear
to be cupping at a tremendous rate (PI. 91, fig. 4). However, the majority of oysters are flat and
lay relatively horizontal within the fossil beds, suggesting little overgrowth or silt load.

Many authors have noted that substrate topography can affect growth and shape of cemented
shells and that new substrates may induce novel growth patterns (e.g. Stenzel 1971; Carreon 1973;
Seilacher 1984). Although shell morphology differs little between C. virginica and C. rhizophorae,
settling on mangrove roots may have induced ‘ecotypic “derailment”’ (Seilacher 1984, p. 214) in
an ancestral Crassostrea stock. However, in view of the similarities between these two ‘species’,
any genetic differences may have been strongly influenced by ecologic factors. Shell form and
growth may also be affected by population density but to what extent this has played a part in
the observed differences between C. virginica and C. rhizophorae cannot be investigated.

The scarcity of the substrate, the possibility of being smothered by other organisms (including
other oysters), the limited amount of time available for feeding when the oyster is covered by the
tide, the possibility of being killed by tropical rains, and the threat of high turbidity during such
rains would suggest that the Recent C. rhizophorae must reproduce soon after it has settled.
Gonadal development may proceed at the cost of shell and somatic growth with the result that
oysters would be small when sexually mature. Indeed, Urpi et at. (1984) found sexually mature
specimens of C. rhizophorae with a shell height of only 1 3 mm with an approximate age of between
15 and 22 days, although spawning was not observed in individuals smaller than 21 mm. Although
spermatids may form when C. virginica is about 4 months old, it does not appear to reach sexual
maturity until it is approximately 7 months old (Galtsoff 1964). The difference in sexual development
between these two oysters is no doubt, in part, due to the difference in their distribution. Mangrove
oysters, C. rhizophorae , are found in coastal regions of the West Indies from Cuba and Puerto
Rico at the tip of their northern limit extending southward into regions of Brazil (Ahmed 1975)
and C. virginica occurs along the east coast of North America from the Gulf of St Lawrence in
Canada, in the north to Florida, the Gulf of Mexico, Central America, the West Indies (Stenzel
1971), and Brazil (Gunter 1951). Colder winters slow down the development of the gonads in C.
virginica (Galtsoff 1964) and require that the oyster stores sufficient food reserves to be able to
survive the winter. In contrast, less energy needs to be stored by oysters in warmer waters and the
continuous breeding of C. rhizophorae which peaks during each of the two rainy seasons enables
the ‘species’ to survive in a relatively unstable environment. Perhaps the rainy season spawning
periods are an adaptation to maximize the chances of survival during times of environmental
stress.

The Plio-Pleistocene C. virginica at Round Hill were obviously not living in a mangrove
environment. Although the Round Hill Beds are obviously faulted, it is unlikely that their position
relative to Round Hill has altered much since the Plio-Pleistocene, apart from tilting and uplift
relative to present sea level. This, together with observations discussed above, suggests that these
fossil Crassostrea may have been near shore, either intertidal or subtidal, or possibly estuarine (cf.
Frey et al. 1987). Modern C. virginica is certainly known from both the intertidal and the subtidal
zone (Galtsoff 1964). The advantage of living subtidally is that feeding may be continuous and the
oyster may be better protected from floating layers of fresh water during and after rainfall.
Furthermore, in an open coastal environment C. virginica would not have been subjected to severe
or prolonged exposures to low salinity. Perhaps in this way C. virginica has been able to grow

1026

PALAEONTOLOGY, VOLUME 31

much larger and live much longer than C. rhizophorae, which is largely restricted to the more
unstable intertidal zone when settling on mangrove. However, although the above observations
and arguments all suggest ecophenotypic variation to be a plausible explanation of observed
morphological differences between C. virginica and C. rhizophorae , further tests of this suggestion
are desirable, particularly on extant populations of the two species living in geographic and/or
ecologic close association.

Acknowledgements. We thank Professor Edward Robinson for allowing us access to his unpublished field
data concerning the Round Hill section and the staff of the Oyster Culture (Jamaica) Project for access to
salinity data. We are grateful to the Oysterseed Cooperative Project (UWI/Dalhousie) for allowing us to use
the LMT vehicle. Dr Peter R. Bacon (UWI) kindly identified our barnacles and provided useful discussion
on their ecology.

REFERENCES

ager, d. v. 1963. Principles of paleoecology, xi + 371 pp. McGraw-Hill, New York.
ahmed, M. 1975. Speciation in living oysters. Adv. mar. Biol. 13, 357-397.

Andrews, J. D. 1955. Setting of oysters in Virginia. Proc. natn. Shellfish, Ass. 45, 38-46.

— 1982. Anaerobic mortalities of oysters in Virginia caused by low salinities. J. Shellfish Res. 2, 127-132.
angell, c. l. 1973. Maduracion gonadica y fijacion de Crassostrea rhizophorae en una laguna hipersalina
del nororiente de Venezuela. Mem. Soc. Cienc. nat. ‘ La Salle', 33, 216-240.

1986. The biology and culture of tropical oysters, vi + 42 pp. ICLARM Stud. Rev. 13, Seattle,
Washington.

arakawa, k. y. 1980. Prevention and removal of fouling on cultured oyster. A handbook for growers.
(Translated from the Japanese by R. Gillmor.) Mar. Sea Grant Tech. Rep. 56, 37 pp. University of Maine,
Orono.

bacon, p. R. 1970. The ecology of Caroni Swamp , Trinidad , 68 pp. Special Publication, Central Statistical
Office, Trinidad.

bayer, u., Johnson, a. L. a. and brannan, j. 1985. Ecological patterns in Middle Jurassic Gryphaea: the
relationship between form and environment. In bayer, u. and seilacher, a. (eds.). Sedimentary and
evolutionary cycles, 436-463. Springer-Verlag, Berlin.
beaven, G. F. 1955. Various aspects of oyster setting in Maryland. Proc. natn. Shellfish. Ass. 45, 29-37.
bosch, c. a. and nikolic, m. 1975. Algunas observaciones sobre el recrutamiento, el crecimiento y la
mortalidad de ostiones ( Crassostrea rhizophorae, Guilding 1828) experimentalmente cultivados. Resum.
Invest. Inst. nac. Pesca Cent. Invest. Pesq., Cuba, 2 (B/30), 96 100.
branch, G. m. 1984. Competition between marine organisms: ecological and evolutionary implications.
Oceanogr mar. Biol. ann. Rev. 22, 429-593.

brasier, M. d. 1975. An outline history of seagrass communities. Palaeontology, 18, 681-702.
bruce, j. a. 1968. Oceanographic cruise summary, marine biofouling studies in Montego and Oyster Bays,
Jamaica, January 1967 to January 1968. US Nav. oceanogr. Off. internal Rep. 68-116, 21 pp.
buroker, n. e., Hershberger, w. k. and chew, K. K. 1979. Population genetics of the Family Ostreidae. 2.

Interspecific studies of the genera Crassostrea and Saccos trea. Mar. Biol. 54, 171-184.
carreon, j. a. 1973. Ecomorphism and soft animal growth of Crassostrea iredalei (Faustino). Proc. natn.
Shellfish. Ass. 63, 12-19.

cartfeew, r. and bosence, d. 1986. Community preservation in Recent shell-gravels, English Channel.
Palaeontology, 29, 243-268.

cerame-vivas, m. j. 1974. Mangroves of Puerto Rico , 64 pp. San German, Puerto Rico: P.F.Z. Properties
Inc., and M. J. Cerame-Vivas, Inc. Environmental Consultants.
duncan, p. m. and wall, G. p. 1865. A notice of the geology of Jamaica, especially with reference to the
district of Clarendon; with descriptions of the Cretaceous, Eocene, and Miocene corals of the islands. Q.
Jl geol. Soc. Lond. 21, I 15.

durve, v. s. 1986. On the ancestry and distribution pathways of three species of Indian oysters. Indian J.
mar. Sci. 15, 56-58.

engle, j. b. and chapman, c. r. 1952. Oyster condition affected by attached mussels. Sth. Fisherman, August,
28-30.

eva, a. n. 1980. Pre-Miocene seagrass communities in the Caribbean. Palaeontology, 23, 231 -236.

LITTLEWOOD AND DONOVAN: CRASSOSTREA IN JAMAICA

1027

frey, r. w., basan, p. b. and smith, j. m. 1987. Rheotaxis and distribution of oysters and mussels, Georgia
tidal creeks and salt marshes, U.S. Palaeogeog., Palaeoclimat ., Palaeoecol. 61, 116.
galtsoff, p. s. 1964. The American oyster, Crassostrea virginica Gmelin. Fishery Bull. Fish Wildl. Serv. US,
64, 1-480.

glynn, p. w. 1964. Common marine invertebrate animals of the shallow waters of Puerto Rico. Historia
natural de Puerto Rico, 1-53

goodbody, i. 1961. Mass mortality of a marine fauna following tropical rain. Ecology , 42, 150 155.

— 1962. Breeding seasons in tropical marine invertebrates. Proc. 4th Meeting Ass. Mar. Labs., Curasao,
N.A., Nov. 1962.

1965. Continuous breeding in a population of two tropical crustaceans, Mysidium columbine (Zimmer)
and Emerita portocricensis Schmidt. Ecology, 46, 195- 197.
grinnel, r. s., jr. 1974. Vertical orientation of shells on some Florida oyster reefs. J. sedim. Petrol. 44, 1 lb-
122.

gunter, G. 1950. The generic status of living oysters and the scientific name of the common American species.
Am. Midi. Nat. 43, 438 439.

— 1951. The species of oysters of the Gulf, Caribbean and West Indian region. Bull. mar. Sci. Gulf Caribb.
1, 40-45.

1954. The problem in oyster taxonomy. Syst. Zool. 3, 134-137.

hallam, a. 1965. Environmental causes of stunting in living and fossil marine benthonic invertebrates.
Palaeontology, 8, 132-155.

holme, n. a. 1961. The bottom fauna of the English Channel. J. mar. biol. Ass. UK, 41, 397-461.
ingersoll, e. 1881. The oyster industry. In The history and present condition of the fishery industries, 251 pp.

Tenth Census of the United States, Department of the Interior, Washington, DC.
jackson, j. b. c. 1977. Competition on marine hard substrata: the adaptive significance of solitary and
colonial strategies. Am. Nat. Ill, 743-767.

Johnson, a. l. a. 1981. Detection of ecophenotypic variation in fossils and its application to a Jurassic
scallop. Lethaia, 14, 277-285.

littlewood, d. t. j. 1987. Biological consequences of aerial exposure of the mangrove oyster Crassostrea
rhizophorae (Guilding, 1828) (Mollusca: Bivalvia). Ph.D. thesis (unpublished). University of the West Indies
(Mona).

— In press. Subtidal versus intertidal cultivation of Crassostrea rhizophorae. Aquaculture.
loosanoff, v. l. 1958. Challenging problems in shellfish biology. In buzzati-tra verso, a. a. (ed.). Perspectives
in marine biology, 483-495. University of California Press, Berkeley and Los Angeles.

— 1962. Recent advances in the control of shellfish predators and competitors. Proc. Gulf Caribb. Fish.
Inst. 13, 113-128.

mattox, n. t. 1949. Studies on the biology of the edible oyster, Ostrea rhizophorae Guilding, in Puerto Rico.
Ecol. Monogr. 19, 339-356.

— 1950. Studies on the edible oyster, Ostrea rhizophorae Guilding in Puerto Rico. Proc. Gulf Caribb. Fish.
Inst. 2, 12-14.

menzel, r. w. 1971. Selective breeding in oysters. In Proceedings of the Conference on Artificial Propagation
of Commercially Valuable Shellfish — Oysters. Oct. 22-23, 1969, 8 1 -92. College of Marine Studies, University
of Delaware, Delaware.

— 1972. Selection and hybridisation in the mariculture of oysters and clams. Proc. 3rd aim. Workshop
Wld. Mariculture Soc. 309-317.

— 1973. Some species affinities in the oyster genus Crassostrea. Bull. Am. malac. Un, March (1973), p. 38.
nascimento, i. a., pereira, s. a. and souza, r. c. 1980. Determination of the optimum commercial size for

the mangrove oyster (Crassostrea rhizophorae) in Todos os Santos Bay, Brazil. Aquaculture, 20, 1-8.
newball, s. and carriker, m. r. 1983. Systematic relationship of the oysters Crassostrea rhizophorae and C.

virginica : a comparative ultrastructural study of the valves. Am. malac. Bull. 1, 35-42.
nikolic, m. 1969. Informe provisional sobre las actividadas desarrolladas durante el periodo comprendido
entro marzo 1963 y mayo 1969. Doc. interned Cent. Invest. Pesq. Inst. nac. Pesca, Cuba, Pt. 2 (4.1), 1 0 52.

— and alfonso, s. j. 1971. El ostion del mangle Crassostrea rhizophorae Guilding 1828 (exploitacion del
recurso y posibilidades para el cultivo). FAO Fish. Rep. 71.2, 201-208.

— bosch, a. and vazquez, y. b. 1976. Las experiencias en el cultivo de ostiones del mangle (Crassostrea
rhizophorae). FAO Technical Conference on Aquaculture, Kyoto, Japan 1976. FIR: AQ/Conf/76/E.52.

palmer, r. e. and carriker, m. r. 1979. Effects of cultural conditions on morphology of the shell of the
oyster Crassostrea virginica. Proc. natn. Shellfish. Ass. 69, 58-72.

1028 PALAEONTOLOGY, VOLUME 31

prescott, G. c. and versey, h. r. 1958. Field meeting at Hayes Common and Round Hill, Jamaica. Proc.
Geol. Tss. 69, 38-39.

ranson, G. 1942. Note sur la classification des Ostreides. Bull. Soc. geol. Fr, 5th ser. 12, 161-164.

reineck, H. E. and singh, I. B. 1973. Deposit ional sedimentary environments with reference to terrigenous
elastics, xvi + 439 pp. Springer-Verlag, Berlin.

rettger, r. e. 1935. Experiments on soft-rock deformation. Bull. Am. Ass. petrol. Geol. 19, 271-292.

robinson, e. 1968 (for 1967). The geology of Round Hill, Clarendon. J. geol. Soc. Jamaica , 9, 46-47.

RODRIGUEZ-ROMERO, F., URIBE-ALCOCER, M., LAGUARDA-FIGUERAS, A. and DIUPOTEX-CHENG, M. E. 1979. The
caryotype of Crassostrea rhizophorae (Guilding, 1828). Venus, Kyoto, 38, 135-140.

rueda, R. L. and moreno, m. p. 1985. Estudio cualitivo y cuantitativo de la fauna asociada a las raices de
Rhizophora mangle en la cayeria este de la Isla de la Juventud. Revta Investnes. mar. 6, 45-57.

— conesa, m. a., ortiz, m., moreno, M. p. and veledo, T. 1985. Organismos asociados a las raices de
mangle, Rhizophora mangle, en lagunas costeras y de cayos. Ibid. 59-71.

rutzler, k. 1969. The mangrove community, aspects of its structure, faunistics and ecology. In castanares,
a. a. and phleger, f. b. (eds.). Lagunas Costeras, Un Simposio, Nov. 28-30, 1967, Mexico, Mem. Simp,
int. Laguna Costeras, 515-536.

SACCO, f. 1897. I molluschi dei terreni terziarii del Piemoonte e della Liguria. Boll. Musei Zoo!. Anat. comp.
R. Univ. Torino, 12, 99-102.

saenz, B. A. 1965. El ostion antillano: Crassostrea rhizophorae Guilding y sum cultivo experimetal en Cuba.
Nota sobre Investigaciones Pesqueras, 6, 34 pp.

saint-felix, c. 1972. Les gisements huitres de Crassostrea rhizophorae en Martinique. Sci. Peche, 214, 23 pp.

seilacher, a. 1967. Bathymetry of trace fossils. Mar. Geol. 5, 413-428.

1984. Constructional morphology of bivalves: evolutionary pathways in primary versus secondary soft-
bottom dwellers. Palaeontology, 27, 207-237.

singarajah, k. v. 1980. On the taxonomy, ecology and physiology of a giant oyster, Crassostrea paraibanensis,
new species. Bull. mar. Sci. 30, 833-847.

siung, a. 1976. Studies on the biology of three species of mangrove 'oysters’ (Isognomon alatus Gmelin,
Crassostrea rhizophorae Guilding and Ostrea equestris Say) in Jamaica. Ph.D. thesis (unpublished),
University of the West Indies (Mona).

stauber, l. a. 1950. The problem of physiological species with special references to oysters and oyster drills.
Ecology, 31, 109-118.

STENZEL, H. B. 1971.. Oysters. In MOORE, r. c. (ed.). Treatise on invertebrate paleontology. Part N, Bivalvia 3,
N953-N1224. Geological Society of America and University of Kansas Press, Boulder, Colorado and
Lawrence, Kansas.

Sutherland, j. p. 1980. Dynamics of the epibenthic community on roots of the mangrove Rhizophora mangle,
at Bahia de Buche, Venezuela. Mar. Biol. 58, 75-84.

urpi, o. p., pena, G. c. and madriz, E. z. 1984. Crecimiento y madurez sexual de Crassostrea rhizophorae
(Guilding, 1828) cultivado en sistema suspendido en Estero Viscoya, Limon, Costa Rica. Revta Biol. trop.
31, 177-281.

WADE, B. A., BROWN, R. A., HANSON, C., HUBBARD, R., ALEXANDER, L. and LOPEZ, B. 1981. The development of
a low-technology oysterculture industry in Jamaica. Proc. 33rd Meet. Gulf Caribb. Fish. Inst. Nov. 1980,
6-18.

warme, j. E. 1969. Live and dead molluscs in a coastal lagoon. J. Paleont. 43, 141-150.

warmke, G. L. and Abbott, r. t. 1961. Caribbean Seashel/s, x + 348 pp. Livingston Publishing Company,
Pennsylvania.

D. TIMOTHY J. LITTLE WOOD

Present address: Shellfish Research Laboratory
New Jersey Agricultural Experiment Station
Cook College, Rutgers University
Port Morris, New Jersey, 08349, USA

STEPHEN K. DONOVAN

Typescript received 5 November 1987
Revised typescript received 8 March 1988

Department of Geology,
University of the West Indies
Mona, Kingston 7
Jamaica, WI

THE HOLOTYPE OF THE WEALDEN CONIFER
BRACHYPHYLLUM PUNCTATUM MICHAEL

by JOAN WATSON, HELEN L. FISHER and NICOLA A. HALL

Abstract. The missing holotype of the conifer Brachyphyllum punctatum Michael originally described from
the Wealden of Germany has been rediscovered. B. castatum Watson, Fisher and Hall from the English
Wealden has proved to be synonymous with B. punctatum. Tarphyderma glabra Archangelsky and Taylor
from the Lower Cretaceous of Argentina is probably also specifically identical.

The single well-preserved conifer shoot used by Michael (1936) to establish the species Brachyphyl-
lum punctatum was part of a small collection originally housed in the Geological Survey of Berlin.
Since 1961 enquiries and searches made in Berlin and elsewhere had failed to locate any of her
hand specimens or preparations. It was thus thought likely that they were lost along with many
of the nineteenth-century Wealden type and figured specimens, and several conifer species in the
English Wealden flora have subsequently been described without the benefit of comparison with
similar German material. Watson, Fisher and Hall (1987) discussed the unusual and uncertain
nature of B. punctatum and its possible synonymy with their new English species but from Michael's
figures alone were unable to draw any satisfactory conclusions. However, the holotype of B.
punctatum has now quite unexpectedly been found lying unrecognized amongst a collection of
unfigured material in the Geologisch-Palaontologisches Institut and Museum of the Georg-August-
Universitat, Gottingen. Though it had no registration number it was easily recognizable as
Michael’s original. Study of its cuticle shows it not only to be identical to the English material of
Watson et al. (1987) but probably also to a newly erected species from the Lower Cretaceous of
Argentina (Archangelsky and Taylor 1986).

SYSTEMATIC PALAEONTOLOGY

Brachyphyllum punctatum Michael, 1936
Plate 92, figs. 1 -6

1936 Brachyphyllum punctatum Michael, p. 60, pi. 3, figs. 7 and 8; pi. 4, figs. 2 and 3.

1976 34 CONIF BrA; Oldham, p. 466, pi. 75, figs. 1-8 (code number used in place of Linnean name).

1987 Brachyphyllum castatum Watson, Fisher and Hall, p. 169, pi. 1, figs. 1 5; pi. 2, figs. I 8; pi. 3,

figs. 1-6; pi. 4, figs. 1-8; pi. 5, figs. 1-7; pi. 6, figs. 1-6; text-fig. 1a-d; text-fig. 2a-d.

The following is probably also synonymous:

1986 Tarphyderma glabra Archangelsky and Taylor, p. 1578, figs. 1-30.

Material and age. Specimen 53.1.4 from Egestorf, Deister: Berriasian.

Description. The holotype, shown at natural size in Plate 92, fig. 1, is of similar dimensions and morphology
to the shoot figured by Watson et al. (1987) in their plate 1, fig. 3. The formula devised for us by Dr Alan
Charlton (see appendix in Watson et al. 1987) for determining phyllotaxis has given parastichy numbers of
5 + 8 which agrees with British Museum (Natural History) specimen V.2321. The cuticle is of the type having
moderately long stomatal tubes (PI. 92, fig. 2) and has the enigmatic ‘thick cells’ which permit the instant
recognition of this species in the light microscope. Unfortunately we have yet again been unable to demonstrate
these cells satisfactorily in the SEM. Plate 92, fig. 6 is the inner surface of the adaxial cuticle showing the
typical elongated cells with strongly cutinized, pitted inner periclinal walls. The convoluted cuticle lining the

| Palaeontology, Vol. 31, Part 4, 1988, pp. 1029-1031, pi. 92. |

© The Palaeontological Association

1030

PALAEONTOLOGY, VOLUME 31

stomatal tubes in several English specimens has not been seen in the holotype. However, this is a variable
feature by no means always present. It is not present in the English specimens with the longest tubes but is
seen in the Argentinian material which has equally long tubes.

DISCUSSION

Watson et al. (1987) have discussed Michael’s description of the cuticle of B. punctatum which
they eventually concluded must be different from the English material in having the outer surface
‘covered by a thick, densely arranged hair-like tomentum’ (translation of Michael 1936 by Dr H.
Jahnichen). Michael sectioned a leaf and her photograph of this (Michael 1936, pi. 3, fig. 8) shows
these protuberances quite clearly with no question of the cuticle having been inadvertently reversed.
We are now able to demonstrate that her description was indeed a misinterpretation, caused by
unusual preservation of the holotype cuticle. Plate 92, fig. 3 shows the outer surface of the abaxial
cuticle, intact on the right-hand side but with all the cutinized outer periclinal walls missing on
the left-hand side. Plate 92, fig. 4 shows a close-up of the junction between these two areas, it now
seems clear that the leaf sectioned by Michael must have had the outer walls of the epidermal cells
missing. Plate 92, fig. 5 shows the vertically cut edge of such a piece of cuticle, at high tilt in the
SEM with the outer surface uppermost. Sections of this would certainly give the appearance of
strong surface protuberances.

The Argentinian material is so far known only as large leaves with the longest stomatal tubes.
There seems to us no doubt about it being B. punctatum but there is a puzzling difference in the
form of the cells of the adaxial surface of the leaf. The adaxial cuticle of the English material
shows elongate cells of a very distinctive and consistent form, indistinguishable from Michael’s
plate 4, fig. 2. Archangelsky and Taylor (1986, fig. 2) figure polygonal adaxial cells with thick
walls. We have seen nothing like them in the European specimens although the holotype does have
much shorter cells in places. Archangelsky and Taylor in their diagnosis mention ‘sometimes
elongate cells with straight walls’ but they are not figured. This discrepancy should be studied
further before the diagnosis for the species is emended.

REFERENCES

archangelsky, s. and taylor, t. n. 1986. Ultrastructural studies of fossil plant cuticles. II. Tarphy derma
gen. n., a Cretaceous conifer from Argentina. Am. Jl Bot. 73, 1577-1587.
michael, f. 1936. Palaobotanische und kohlenpetrographische Studien in der nordwestdeutschen Wealden-
Formation. Abh. preuss. geol. Landesanst. 166, 1-79.
oldham, t. c. b. 1976. Flora of the Wealden plant debris beds of England. Palaeontology, 19, 437-502.
watson, j., fisher, h. l. and hall, n. a. 1987. A new species of Brachyphyllum from the English Wealden
and its probable female cone. Rev. Palaeobot. Palynol. 51, 169-187.

Typescript received 23 September 1987
Revised typescript received 21 January 1988

JOAN WATSON, HELEN L. FISHER,
NICOLA A. HALL

Departments of Botany and Geology
The University
Manchester Ml 3 9PL

EXPLANATION OF PLATE 92

Figs. I 6. Brachyphyllum punctatum Michael. 2-6 are scanning electron micrographs. All 53.1.4 the
holotype. 1, leafy shoot, x 1. 2, inside of abaxial cuticle showing stomatal tubes, x 150. 3, outside of
abaxial cuticle; surface intact on right-hand side, outer periclinal walls missing from all cells on left-hand
side, x 150. 4, close up of junction between two areas in fig. 3, x 400. 5, cut edge of abaxial cuticle at
high tilt showing anticlinal walls, outer surface, lacking periclinal walls, uppermost, x400. 6, inside of
adaxial cuticle, x 400.

PLATE 92

WATSON, FISHER and HALL, Brachyphyllum punctatum

HETEROCHRONIC TRENDS IN NAMURI AN
AMMONOID EVOLUTION

by ANDREW R. H. SWAN

Abstract. Theoretical models of heterochronic processes are based on the comparison of ontogenetic age-
shape curves of ancestor and descendant. An existing principal components analysis of an exhaustive body
of Namurian ammonoid morphological data is a suitable source of information for assessment of heterochrony
in this context. Using size as an indicator of age and principal component score as a shape index, heterochronic
analyses of two evolutionary radiations of the Gastriocerataceae demonstrate that one was strongly influenced
by neoteny, the other by acceleration. From an ancestor which undertook a change in habitat during ontogeny
from benthic to nektonic, the occupation of the benthic habitat was increased in the neotenous trend and
decreased in the lineage showing acceleration. Widespread changes in marine benthic conditions are suggested
as the cause of these trends.

This paper compares the ontogeny with the evolutionary trends of Namurian (mid-Carboniferous)
ammonoids in order to evaluate the contribution of heterochronic processes. There is a long history
of ammonoid studies of this sort, due largely to the good record of ontogeny exhibited by many
specimens. Ammonites were cited as evidence of recapitulation by, for example, Hyatt (1889), and
of proterogenesis by Schindewolf (1936). Carboniferous ammonoids do not have such a history of
analysis in this context, although Newell (1949) documented phyletic size increase in an Upper
Palaeozoic ammonoid lineage which included Pennsylvanian forms. The restriction of the present
study to Namurian ammonoids is due to the existence of an extensive and appropriate data base
with accompanying analyses compiled by Saunders and Swan (1984).

Heterochrony has been defined by Gould (1977) after de Beer (1930) as: ‘phyletic change in the
onset or timing of development, so that appearance or rate of development of a feature in a
descendant ontogeny is either accelerated or retarded relative to the appearance or rate of
development of the same feature in an ancestor’s ontogeny.’ Following the publications of Alberch
et al. (1979) and McNamara (1986), the terminology of heterochronic processes has been clarified
(text-fig. 1). Where the mature morphology of a descendant is similar to the immature morphology
of its ancestor, the result is termed paedomorphic; where the opposite is true, the descendant is
peramorphic. A paedomorphic descendant can arise by a decrease in the rate of change of
morphology in ontogeny (neoteny), by early attainment of maturity (progenesis), or by delayed
onset of morphological change. Paedomorphosis results in the loss of some morphologies from
the ontogeny. Peramorphosis occurs by increase in the rate of change of morphology (acceleration),
by late maturity (hypermorphosis) or by early onset of morphological change, and involves
transcendance into morphologies absent from the ancestral ontogeny.

In this terminology, recapitulation results from peramorphosis and phyletic size increase may
be due to hypermorphosis. Proterogenesis, however, is not a purely heterochronic process, but
requires a morphological innovation specifically in early ontogeny: an event known as cenogenesis.
This novelty then spreads to the adult stage, perhaps by neoteny (text-fig. 2).

The evolutionary importance of the heterochronic processes results from the potential of
achieving great changes in mature morphology by isolated mutations in the regulatory genes.
Through such mutations, major morphological changes are not accompanied by the high risk of
low viability which is invoked by large-scale structural mutations resulting in ‘hopeful monsters’.
In the case of paedomorphosis, the viability of the resulting adult is likely to be high, due to the
previous viability of the same morphology in the juvenile. In addition to his arguments for the

(Palaeontology, Vol. 31, Pari 4, 1988, pp. 1033-1051.]

© The Palaeontological Association

1034

PALAEONTOLOGY, VOLUME 31

PAEDOMORPHOSIS

PERAMORPHOSIS

text-fig. 1 . Definitions of heterochronic modes on the basis of ontogenetic trajectories of ancestor (solid
line) and descendant (dashed line). The onset of morphological change is indicated by a solid circle; cessation
is indicated by an open circle for the ancestor, a square for the descendant.

In paedomorphosis, the terminal shape of the descendant is the same as the shape of a younger ancestor;
in peramorphosis, the terminal shape of the ancestor is the same as the shape of a younger descendant. In
neoteny and acceleration there is change in the ontogenetic gradient; in progenesis and hypermorphosis there
is change in the length of time during which shape change occurs.

Redrawn from Alberch el al. (1979, figs. 15 and 16).

evolutionary importance of heterochrony, Gould (1977) contended that certain heterochronic
modes are compatible with specific ecological strategies: progenesis with r-type strategy (rapid
reproduction allowing opportunistic colonization), and neoteny with K-strategy (‘fine tuning’ to a
stable environment with high investment in few offspring).

Analysis of heterochrony in Namurian ammonoid evolution is therefore significant both in the
continuing global assessment of modes of evolution and in the interpretation of specific Namurian
evolutionary trends and environments. The recently rationalized terminology of heterochrony has
given the procedure of heterochronic analysis a well-defined suite of requisites and criteria which
form the basis of the methodology in this study.

DATA USED

The approach used here is strongly dependent on the principal components analysis of an extensive
data base presented by Saunders and Swan (1984). These data include measurements of size and
external morphology (expressed by 20 shape variables which incorporate shell geometry, aperture
form and ornament, see Table 1) of 371 Namurian ammonoid specimens (281 species, 81 genera),
compiled largely from published illustrations. An attempt was made to include all published species
from North-West Europe, the South Urals (USSR), and North America, though many species

SWAN: HETEROCHRONY IN NAMURI AN AMMONOIDS

1035

text-fig. 2. Theoretical sequence of ontogenetic trajectories resulting
in proterogenesis. The first descendant (2) of the ancestor ( 1 ) ditfers
from it due to an evolutionary innovation affecting only early ontogeny
(cenogenesis). This innovation then spreads to later ontogeny in sub-
sequent descendants (increments 3 to 5) by neotenous decrease of
ontogenetic gradient.

CD

table 1. Representative values of characters for the Namurian ammonoid morphotypes relevant here, with
the contribution of each character to the first three principal components of variation.

Character abbreviations: D, diameter of umbilicus; AH, aperture height; S, whorl shape; VW, ventral
acuity; W, whorl expansion rate; OW, areal expansion rate; T, spacing of transverse ornament; TVS, spiral
versus transverse ornament; LT, plication or tubercle length; HT, plication or tubercle elevation; RIB, ribbing
strength; ARC, arching of aperture; HS, depth of hyponomic sinus; OS, depth of ocular sinus; UP, umbilical
aperture projection; VG, ventral structure; VLG, ventrolateral structure; UR, umbilical ridge; CON, number
of constrictions; BIF rib bifurcation.

Most of these are expressed as ratios; see Saunders and Swan (1984) for definitions and additional details
of each character.

Character

Morphotype

P.C. loadings

III

V

VII

VIII

P.C.l

PC. 2

P.C. 3

D

0-307

0-133

0-433

0-45

0-885

0-069

-0-017

AH

0-586

0-55

0-565

0-624

0-237

0-502

0-373

S

1-49

1-2

1-75

1-58

0-757

-0-257

-0-236

VW

0-476

0-437

0-446

0-527

-0-102

0-083

-0-05

w

2-12

2-18

1 -51

1 -62

-0-529

0-295

0-328

OW

0-77

1-18

0-611

0-595

-0-785

0-254

0-153

T

10-0

6-0

16 0

18-0

-0-226

0-274

-0-553

TVS

0-8

1-0

1-0

0-8

0-026

— 0-154

0-423

LT

0-3

0-0

0-0

0-65

0-659

0-521

0-093

HT

0-025

0-0

0-0

0-02

0-65

0-411

0-158

RIB

0-2

0-9

0-0

0-2

0-304

-0-134

0-734

ARC

9-0

— 12-3

-0-7

13-0

0-683

0-193

-0-239

HS

21-0

10 1

1-7

3-5

-0-555

0-618

0-074

OS

24-5

0-5

0-0

11-0

0-099

0-807

-0-194

UP

0-0

1-5

2-0

0-0

-0-579

0-067

0-109

VG

1-1

1-0

1-0

1-0

-0-283

-0-067

0-107

VLG

0-8

1-0

1-0

TO

0-094

-0-554

0-174

UR

0-0

0-0

0-0

0-0

-0-005

-0-149

0-041

CON

3-0

0-0

0-0

3-0

0-31

0-047

-0-281

BIF

1-0

2-0

1-0

3-0

0-507

0-125

0-553

Total % 24-4

12-32

9-08

1036

PALAEONTOLOGY, VOLUME 31

had to be eliminated from analysis due to incomplete morphological data. Other faunas, for
example from North Africa and China, are at present only partially documented. For each specimen
the general location and stratigraphic level were recorded. The data were originally compiled with
the objective of including the whole range of morphologies present, regardless of size. Consequently,
where possible, species were assessed at two different sizes (10-25 mm and > 25 mm) to allow for
ontogenetic change, though the exact sizes used were dictated by the available documentation.
Specimens smaller than 10 mm diameter were not assessed due to paucity of information and
difficulty of measurement.

Within the realms of logistical feasibility, it would be difficult to improve on this data base as
a source of information on Namurian ammonoid morphology with respect to ontogeny, time, and
space. The data have been deposited with the British Library, Boston Spa, Yorkshire, UK, as
supplementary publication no. SUP 14032 (41 pages).

T AX A CONSIDERED

Saunders and Swan (1984, figs. 13-20) documented the changes in the diversity of external
morphology of ammonoids through the Namurian of North-West Europe, North America, and the
South Urals. Some taxa were shown to be morphologically conservative (e.g. Dimorphocerataceae,
Prolecanitina) while others declined (e.g. Neoglyphiocerataceae) or became extinct (e.g. Muenstero-
cerataceae). The Gastriocerataceae, in contrast, arose within the Namurian and evolved rapidly,
showing radiation and innovation into new morphologies. The most striking trends in the evolution
of this superfamily were the development of two distinctive morphotypes: 1, evolute, depressed,
coarsely ornamented forms with fairly simple apertures, e.g. Cancelloceras ; 2, involute, compressed
forms with strong hyponomic and ocular sinuses and prominent ventrolateral lingua, e.g. Bilinguites.
In Saunders and Swan’s (1984) principal components analysis of twenty external morphologic
characters in 281 Namurian ammonoid species, these two morphologies are resolved as positive
P.C.l, low P.C.2 scores (designated morphotype VIII), and as positive P.C.2, low P.C.l scores
(designated morphotype III), respectively (text-fig. 3; Table 1). The development of these two
morphotypes in the Gastriocerataceae is the focus of the heterochronic analysis in this paper. The
constituent families of the Gastriocerataceae are: Homoceratidae, Decoritidae, Reticuloceratidae,
Gastrioceratidae, and Bisatoceratidae.

CONSTRAINTS ON THE DATA

Unambiguous recognition of heterochrony according to the scheme of Alberch et al. (1979) requires
the following: 1, knowledge of ancestor-descendant relationship; 2, recognition of ontogenetic
stages corresponding to the onset and cessation of morphologic change; 3, age at these stages; 4,
shape at these stages; and 5, size at these stages (only needed to resolve the special cases of
proportioned gigantism and dwarfism). The problems associated with satisfying each requisite for
the chosen group of Namurian ammonoids must be carefully considered.

Ancestor-descendant relationship

The Namurian marine record is punctuated by strong eustatic regressive events in all the important
stratigraphic sections (Ramsbottom 1977; Saunders et al. 1979); consequently it is not possible to
trace individual lineages through the succession with any confidence. This limits the analysis to
trends, rather than details, in evolution. Hence the comparison is of successive faunas rather than
species, and only fairly major morphological shifts can be resolved. For this situation, the
methodology only demands that the stem groups of the analysed species in each fauna are included
within the data for the chronologically previous fauna. The more recent of the relevant phyletic
hypotheses are generally supportive (Ruzhencev and Bogoslovskaya 1978; Swan 1984); specific
problems will be assessed where appropriate. It should be noted that faunas from North-West
Europe, the South Urals, and North America (the main sources of data) are not phyletically

SWAN: HETEROCHRONY IN NAMURIAN A M MONOIDS

1037

jQ

-t-* a

o °
x. a

d> d>

3 6

^ c
e °
■2 2
2 .2
’C .ts

ctf c/d
> O

f , , Oh

O d)

c "•

d> Q->

o 'S

O o

Oh (J

£ .5

cs *5

O

-S ^

^ r-1

g o E

r? Q. r\

a.

Kj

a-

c ^
E c

O d>

^ "S

d> <3
^ 1/5 ’

^ d)

- S-<

O c3

^ 3

a"

O C/D

d>

Oh r 5

C/D W
0) - •

5-h Oh

JO

7f

^

d)

> 3
d) o»

C/D rn

3 o

o n

a

— c o

C/D rH

qo S

a <d e a is <~

•£ c

c ^
E d>

o J-H

73 ^
d> d)

c3 "5

c

§ s

’I E

X&

o

Oh

J-H

3

Oh

O d>

>

C \G

O ctf

O ^

J-> C/D

O 71-

oo

a o\

C r-H

)-2 c

£ f

'S ™

C/D 73

'■3 S

a c«

-O J-H

■*“* d)

r ”3
-C> G
_ 3
O <3
00

o %

w c2

d>

-O ^

EZ w
G
• <-G

H-g

d> u
G c/D

• S 3
O

d> M

Oh^

A Ӥ

C/D Oh

;g <i>
^ Pi
d)

JO ^

^ 7f

o

bfl O
« <*
K4-

oo

Os ^ .

- d)

£ «

^ d) W

: "2 <u

i § E

: cs -s
1 ^

ifl °

i g 1

; « §
5 6
! £ 2
i 0*3 O
i - g
) g —

- 2 .2

> £ «

1 7- 43

i . 60

p C 3

i 2 2
! -2 ^
i 00 ~

a

1038

PALAEONTOLOGY, VOLUME 31

discrete; they share many genera and some species throughout the Namurian, due presumably to
migration during transgressive maxima. This validates the analysis of the disparate regions together.

Ontogenetic stages

The onset of morphological change can readily be regarded in ammonoids as the earliest secreted
structure— the protoconch. The cessation of morphological change is also definable in ammonoids
because, like recent Nautilus (Saunders 1983) there is decline and cessation of growth at maturity.
Recognized symptoms are: approximation of septa, development of apertural modifications, change
in aperture size or shape (for example by constriction), decline in ornament, and change in tightness
of coiling (Kennedy and Cobban 1976). The recognition of cessation of morphological change is
critical in documenting progenesis and hypermorphosis because, for example, a hypermorphic
descendant identical to the terminal morphology of its ancestor only differs from it in that its
ontogeny continues.

However, problems exist in the consistent identification of maturity in Namurian ammonoids.
Documentation of symptoms of maturity is not as thorough as for Mesozoic forms, so the
application of criteria is open to doubt. Septa are not readily observable in most specimens,
ornament frequently declines long before maturity, changes in coiling are never more than subtle,
apertures do not show the extreme modifications associated with sexual dimorphism in the
Mesozoic, and in any case they are often destroyed by various taphonomic processes along with
the rest of the body-chamber. Detailed knowledge of the terminal stage of ontogeny is therefore
unavailable for most Namurian species. In addition, the use of symptoms of maturity in these
analyses may be inadvisable in that they may be intimately linked with gonadal development.
Hence the apertural modifications associated with sexual maturity of an ancestor may be expected
to occur at maturity in a neotenous descendant, even though up to that point the descendant
morphology had been that of the juvenile ancestor.

At the expense of precision, size is here adopted as an indicator of ontogenetic stage, as it is the
only remaining parameter which is at all correctable with development. The effect of this
imprecision on the results is discussed later.

Age

Age provides the measure of ontogeny used on the x-axis of the theoretical ontogenetic trajectories
of Alberch et al. (1979) (text-fig. 1). It is, of course, notoriously difficult to assess in fossil material.
Estimates for age at maturity of ammonoids are all contestable and vary from 4 to 30 years; it is
clearly not feasible to ordinate large numbers of specimens against an age axis. Once again, size
is the only available parameter conceivably related to age. The tentative equation of size and age
precludes the recognition of proportioned gigantism and dwarfism, and renders the result of
heterochronic analysis indefinite to a degree which will be discussed later.

Shape

The theoretical ontogenetic trajectories established for heterochronic analysis (text-fig. 1) use a
single parameter on they-axis to characterize shape. Although various authors have used univariate
data to discern heterochrony (e.g. Newell 1949), this procedure is logically unsound. If an ancestral
ontogeny involves the change in a single character value from p to q, then a mature value in the
descendant of between p and q could be regarded as due to a type of paedomorphosis; a value
beyond q could be regarded as due to peramorphosis, even though the change might be the result
of any evolutionary mode. Therefore, in some cases, any conceivable character value in the mature
descendant could be explained by a heterochronic process and the hypothesis of heterochrony
would not be falsifiable.

With reference to the theoretical age-shape trajectories (text-fig. 1), this situation can be stated
in terms of vectors. The various heterochronic modes are transformations of the ancestral ontogeny
which, in combination, could produce any result in the plane defined by the age and shape axes.

SWAN: HETEROCHRONY IN NAMURI AN AMMONOIDS

1039

providing the direction ( + or — ) of the gradient is conserved. However, the test of the heterochronic
hypothesis improves if more characters are used, increasing the dimensionality of shape-space. As
a result, the descendant ontogenetic trajectory is not constrained within the plane defined by the
age axis and the ancestral trajectory. If the descendant trajectory is within this plane, then a
heterochronic hypothesis is supportable, and becomes more so with larger numbers of dimensions
of shape-space. Heterochronic studies should therefore consider as many morphological characters
as possible.

Whole morphology has been assessed in heterochronic studies by many authors (e.g. McNamara
1982) but rarely with any numerical reinforcement. Gould (1968), however, used factor analysis
of seven shape measurements to demonstrate the similarity between adult snail paedomorphs and
juvenile non-paedomorphs on a plot with two varimax axes. The eigenvectors which form the basis
of this type of multivariate analysis are directions in multi-dimensional shape-space; consequently,
if a plane defined by an eigenvector and the age axis contains both ancestral and descendant
ontogenetic trajectories, then heterochrony is a likely hypothesis. Eigenvectors are therefore an
appropriate means of resolving shape as one parameter which can be used for constructing
ontogenetic trajectories for comparison with the theoretical heterochronic modes of Alberch et al.
(1979).

In this context, the principal components analysis of Saunders and Swan (1984) is a suitable
source of data. Much of the information contained in the 20-character data set for each specimen
is conveyed by co-ordinates in three-dimensional principal component ( = eigenvector) space.
Although one principal component describes no more than 25 % of the total variation, the two
evolutionary radiations chosen for the present work, namely the gastrioceratacean excursions into
morphotypes III and VIII, are roughly linear in at least the first two principal components. The
score against one of the principal component axes for these morphotypes consequently gives a
good estimate of total morphology. For morphotype VIII, the score on the P.C. 1 axis is appropriate;
for morphotype III, the score on the P.C. 2 axis.

Size

Size is the least problematic of the required parameters. Diameter of the conch is a standard
measure of ammonoids, and is adopted here. The possible objection that this does not necessarily
correlate with body volume is not critical because body-chamber length is fairly consistent within
the Gastriocerataceae, and whorl height tends to be inversely correlated with whorl width, giving
little variety in whorl cross-sectional area (Swan and Saunders, 1987). For each species, Saunders
and Swan (1984) assessed morphology in two different size ranges: at 10-25 mm diameter and
> 25 mm diameter, wherever possible. The lack of data on smaller sizes was imposed by the
available published information, and there is seldom any definite knowledge of the maximum size
attained by species, which is likely to be in excess of the largest documented specimen. Consequently,
the data available from this study are of segments of ontogeny of variable length, usually without
knowledge of morphology at onset or cessation of growth.

In summary, the information available is in some respects not ideal for the stringent assessment
of heterochrony; it corresponds to the third restricted model of Gould (1977, p. 260): standardization
by size when neither age nor developmental stage are known. Nevertheless, as noted by Gould, it
is typical of the quality of data available to palaeontologists. It is likely that most palaeontological
heterochronic analyses will have to proceed using some of the assumptions adopted here, and care
must be taken to interpret the results with due consideration to the assumptions, and to use the
results to review their validity.

ARRANGEMENT OF THE DATA AND GRAPHICAL ANALYSIS

Ontogenetic trajectories, expressed as curves of shape (principal component score) against size
(diameter) are required for potential ancestors and descendants amongst the Namurian

1040 PALAEONTOLOGY, VOLUME 31

Gastriocerataceae which adopted morphotypes VIII and III. Data are selected and arranged as
follows:

Morphotype VIII

The radiation into this morphotype occurred apparently abruptly at level 6 of Saunders and Swan
(1984) (text-fig. 36). However, distinct stratigraphic horizons within level 6, though not confidently
correlatable between continents, are recognized locally and can be used to improve the resolution.
For this purpose the stratigraphic detail used here is at the zonal level; zones used are: in North-
West Europe, H2e, Rlal, Rla2, Rlb, Ric> ^-2a’ ^2c’ Gia, Gib; in the USSR, Nm2bx, Nm2b2,

Nm2b3, Nm2Cj, Nm2c2. Data for this morphotype from North America are sparse (five specimens)
and are not considered here due to problems in correlation. The isolated incursion into morphotype
VIII in level 4 (text-fig. 3b) is ignored as this species, Homoceras alveatum Ruzhencev and
Bogoslovskaya (1978), is only known from one apparently pathological specimen. With these
exceptions, the complete data for all specimens recorded by Saunders and Swan (1984) in each of
the zones listed above which have at least part of their ontogeny in morphotype VIII (as defined
on text-fig. 3b) have been plotted. The result is shown in text-fig. 4, and representative morphologies
are illustrated in text-fig. 5.

Morphotype III

The adoption of this morphology by the Gastriocerataceae was not abrupt, but shows increasing
strength through zones Rla to R2c (Nm2b2-Nm2c2 in USSR). For this reason, and the lesser
quantity of information available, the stratigraphic resolution into the four levels 6-9 of Saunders
and Swan (1984) provides adequate account of the evolution of this morphology. Note that,
although this study only concerns the Gastriocerataceae, morphotype III was extensively represented
by girtyoceratids earlier in the Namurian (which declined in level 3, E2c zone), and by the
conservative and rare nomismoceratids, Hudsonoceras and Baschkirites , in the later Namurian (levels
5-9, zone H.2a onwards). The gastrioceratacean genus Bilinguites is to an extent homoeomorphic with
these genera.

As with morphotype VIII, the complete data for all Gastriocerataceae in each of the stratigraphic
levels which have at least part of their ontogeny in morphotype III (as defined in text-fig. 3b) have
been plotted. The result is shown in text-fig. 6, with representative morphologies illustrated in text-
fig. 7.

INTERPRETATION OF RESULTS

Morphotype VIII

Two important features are apparent on examination of the graphical results (text-figs. 4 and 5)
and the interpretative, schematic summary (text-fig. 8 a). First, the innovation into this morphotype
initially occurs only in the smaller ontogenetic stages: in level 6 there are no specimens larger than
30 mm diameter in this quadrant of the principal components plot, and the ontogenetic gradients
are steep. The innovation of this morphotype, therefore, can be described as cenogenesis, and this
is not a heterochronic process. Secondly, whilst the morphology at small sizes is retained or
accentuated, the ontogenetic gradients decline through time (text-fig. 9), with the result that later
stages in ontogeny become more similar to the early stages, and similarity tends to be between
early ontogeny of ancestors and late ontogeny of descendants. If these graphs are compared directly
with the theoretical models (text-fig. 1), it is clear that this decline in ontogenetic gradient is
compatible with neoteny.

Before the hypothesis of neoteny is confirmed, the effect of the assumptions needs to be discussed.
The assumption of descent between successive faunas is not in question for much of the data; for
example, Ruzhencev and Bogoslovskaya (1978, pp. 59-60) and Swan (1984, p. 319) both traced a
simple lineage through the morphotype VIII reticuloceratids (levels 6-7). However, the relationships
with higher and lower faunas are uncertain. The earliest European species in the present analysis

SWAN: HETEROCHRONY IN NAMURIAN AMMONOIDS

1041

10 20 30 40 50 60 70 80 Diam. (mm)

G, a

text-fig. 4. Size versus first principal component score for all analysed species with at least part of ontogeny
in morphotype VIII. Data points from the same species are connected by lines.

The data for each zone are plotted separately; zonal schemes of North-West Europe (left) and southern
Urals (right) are approximately correlated and in stratigraphic order (youngest at top) with the stratigraphic
levels (6-9) of Saunders and Swan (1984) indicated on the far right; international correlation within each
level is not definite. Labelling on all axes as for Glb zone.

The graphs show a general evolutionary decline in ontogenetic gradient. The faunas from zones Rlc, R2a,
and R.2c are not part of the main morphotype VIII phylogenetic lineage.

1042 PALAEONTOLOGY, VOLUME 31

text-fig. 5. The development of morpholype VIII, illustrated by sketch profiles (with samples of ornament)
and aperture shapes for representative specimens ordinated against diameter and stratigraphic level. Similarity
is generally between smaller stratigraphically lower specimens and larger, higher specimens. This is particularly
true with respect to the characters which comprise the first principal component (see Table 1): whorl width
(S), whorl expansion rate ( W ), diameter of umbilicus ( D ), coarseness of ornament (LT, HT, T), depth of
hypnomic sinus ( HS ), bifurcation of striae ( BIF ). The acute venter shown by the two larger specimens is
apparently associated with approach of maturity in some reticuloceratids.

For each zone, morphology is shown for two or three ontogenetic stages of a species which is, where
possible, representative and average for the fauna, as ascertained from the graphs shown in text-fig. 4. The
species illustrated are, in ascending order: level 6 —Homoceratoides prereticulatus, H2c zone; Phillipsoceras
inconstans , R]al zone; P. alpharhipaeum , Nm2b2 zone; level 7 — Tectiretites posterns , Nm2b3 zone; level 9 —
Cancelloceras rurae, Nm2c2 zone; C. martini , Gla zone. A sliding scale is used — 5 mm scale bars are shown
for each sketch.

SWAN: HETEROCHRONY IN NAMURI AN AMMONOIDS

1043

text-fig. 6. Size versus second principal com-
ponent score for all analysed species with at least
part of ontogeny in morphotype III. Data from
same species connected by lines. Data for each of
the four stratigraphic levels (6 9) plotted separ-
ately; labelling on all axes as for level 9.

Through levels 6 to 8 the ancestral P.C. score
at larger sizes is shown by descendants at smaller
sizes.

text-fig. 7. The development of morphotype III, illustrated by sketch profiles (with samples of ornament)
and aperture shapes for representative specimens ordinated against diameter and stratigraphic level. Similarity
is generally between larger, stratigraphically lower specimens and smaller, higher specimens. This is particularly
true with respect to the characters which are important in the second principal component (see Table 1):
whorl width (S'), depth of hyponomic and ocular sinuses ( HS , OS), presence of groove in the ventro-lateral
region ( VLG ).

For each level, two ontogenetic stages are shown of a species which is, where possible, representative and
average for the fauna, as ascertained from text-fig. 6. Species illustrated are: level 6 — Phillipsoceras inconstans,
Rlal zone; level 7 — Reticuloceras reticulatum, Rlc zone; level 8 — Bilinguites gracile , R2a zone; level 9 B.
superbilingue, R2(, zone. Scale bars 5 mm are shown for each sketch.

1044

PALAEONTOLOGY, VOLUME 31

text-fig. 8. Interpretations of evolutionary trends in ontogenetic trajectories through time. Dashed lines
indicate extrapolation beyond the available analysed data, a, Morphotype VIII. The ancestor in level 5, in
common with other early gastriocerataceans, probably showed ontogenetic transition between morphotypes
VII and V (see text-fig. 3). As P.C.l is used here as an index of the morphotype VIII direction, the level 5
trajectory can be regarded as flat. The innovation of morphotype VIII in level 6 is cenogenetic, involving
just the smaller stages, and further development is by decrease in the gradient of the trajectory, suggesting
neoteny. Compare with text-fig. 2. fi, morphotype III. The interpretation of low P.C.2 scores for small stages
(< 10 mm) in levels 8 and 9 is based on unanalysed evidence (see text). The trajectories show an evolutionary
increase in gradient, suggesting acceleration, though this is constrained by an upper limit to the P.C.2 score.

is Homoceratoides prereticulatus from the H2c zone, which is not regarded as ancestral to the
reticuloceratids either by Bisat (1933), who derived this species and the reticuloceratids separately
from Homoceras , nor by Ruzhencev and Bogoslovskaya (1978, p. 59), who placed the genus in a
different superfamily, the Thalassocerataceae, and derived the reticuloceratids from Surenites.
Nevertheless, Homoceratoides prereticulatus does show similarities with Russian Surenites and
Brevikites; this suite of taxa is in need of systematic revision. On balance, the H.2c zone fauna
should not be regarded as an integral part of the documented evolutionary trends. The earliest
contribution from the Urals is a species of Surenites in Nm2b1 zone, which can be included with
greater confidence.

The derivation of the Gastrioceratidae from the Reticuloceratidae is also debatable. Bisat (1933)
postulated an origin for the family in Homoceras , but this is not supportable in the light of more
recently available information. Ruzhencev and Bogoslovskaya (1978, p. 60) preferred Surenites

SWAN: HETEROCHRONY IN NAMURIAN AM MONOIDS

1045

text-fig. 9. Decline in gradient of morphotype VIII onto-
genetic trajectories through time. Faunal averages are plot-
ted for each of the zones shown in text-fig. 4. Triangles
North-West European faunas; circles. South Urals faunas.
Open symbols denote faunas which are phylogenetically
distinct from the main morphotype VIII lineage, or are
doubtful.

.18

.16

.14

.12

C

a>

■5 • 1

A

.06

.04

▲

A

.02

A

o

h2c R1ai R1aa R1 b R1 c

(Nm,b2) via Monitoceras (Nm2c1) as an ancestor, but without any record of the lineage in Nm2b3
zone. I (Swan 1984, p. 196) have emphasized the close similarity between Cancelloceras (Nm2c2)
of the Gastrioceratidae with Alurites (Nm2bo_3) of the Reticuloceratidae. An origin for the
Gastrioceratidae amongst the morphotype VIII reticuloceratids of level 7 is most likely, and the
validity of the evolutionary comparison is sustained for these groups. Note, however, that this
lineage does not include the European specimens of Bilinguites from zones R.2a and R2c, plus
Reticuloceras in Rlc, which stray into morphotype VIII. These can be eliminated from further
considerations regarding the morphotype VIII lineage.

The other major assumption, the standardization of ontogeny by size, needs careful appraisal.
If the size-age relationship and the size at maturity were constant through the Namurian, then
the evolutionary trend in morphotype VIII is certainly neotenous, as the size axis of text-figs. 4
and 8r/ would be directly comparable to the age axis in text-fig. 1. An error in the assumption of
constant size at maturity would not directly affect this result, as this is only critical in discerning
hypermorphosis and progenesis, and does not affect the gradient of the ontogenetic trajectory
(text-fig. 1). An error in the assumption of constant size-age relationship, however, could affect
the hypothesis of neoteny. For example, more rapid growth in a descendant (resulting in
proportioned gigantism if life-span is retained) would not affect the shape-age curve, so could not
be termed neoteny, but would decrease the gradient on a shape-size curve. The distinctions between
neoteny and proportioned gigantism on a shape-size curve are the length of the trajectory
(proportioned gigantism results in attainment of larger size) and the consequent attainment
by proportioned giants of all morphologies present in the ancestor. Thus the descendant
gastriocerataceans, for example in level 9, would be derived by proportioned gigantism (rather
than neoteny) from their ancestors, for example in level 6, if the complete ontogenies of the former
continue beyond the apparent maximum size to cover the morphologic range of the latter. The
observed ontogenetic gradients (text-fig. 9) indicate that this would require an increase in maximum
size of the descendant over the ancestor by a factor of more than 5. Published and other data
suggest that the phyletic size trend involves less than a twofold increase (text-fig. 1 Or/), and this
may be overestimated due to the bias imposed by frequent fragmentation of larger specimens in
the earlier European faunas. The data are then only compatible with proportioned gigantism if
the ontogeny of the ‘giants’ were shortened by progenesis. This hypothesis can be regarded as less
likely than neoteny by the simple application of Occam’s razor.

In conclusion, although Gould (1977) argued that the ‘restricted model’ used here of ‘standardiza-
tion by size’ cannot yield definitive heterochronic results, it is clear that, following a cenogenetic
innovation, neoteny is the most parsimonious hypothesis for the evolutionary development of

1046

PALAEONTOLOGY, VOLUME 31

a b

text-fig. 10. Species size through time for a, morphotype VIII, b , morphotype III. The maximum recorded
diameters for all species allocatable to the respective morphotypes are plotted. Triangles, North-West
European species; circles. South Urals species. Solid symbols are used where specimens show some sign of
approaching maturity (e.g. loss of ornament, change of aperture shape), open symbols denote specimens
without such indication (though this may be due to lack of morphological change at maturity).

Overall, the size trends through time are not sufficient to account for the evolutionary trends as gigantism
or dwarfism.

morphotype VIII by the Gastriocerataceae. This combination of modes of evolution is identical
to proterogenesis of Schindewolf (1936).

Morphotype III

The important trends apparent from the graphical results (text-figs. 6 and 7) and the interpretation
(text-fig. 8 b) are in contrast to those of the previous morphotype. First, the innovation of the
morphology is in the larger ontogenetic stages, with positive slopes on ontogenetic trajectories in
levels 6 and 7. Secondly, the gradient of the trajectories at smaller sizes is interpreted as becoming
steeper higher in the Namurian, though the trajectory levels off at larger sizes. Unfortunately, this
second contention requires knowledge of specimens from levels 8 and 9 of smaller sizes than those
for which data are available for analysis. However, the interpretation of a low or negative P.C.2
score for early ontogenies in levels 8 and 9 is supported by Bisat (1924, p. 116), who states that
the lateral lingua (a distinctive characteristic of morphotype III) are not developed until 5 mm
diameter.

The increase in slope of the ontogenies through time is compatible with the hypothesis that
heterochronic acceleration has occurred. In this mode, descendants recapitulate ancestral ontogeny,
and may transcend it in late ontogeny by simple extrapolation of ontogenetic trends. The negative
slope on the trajectory in level 9 (text-fig. 6), if representative, suggests a minor cenogenetic event,
but is insufficient to warrant further consideration.

The hypothesis of acceleration for morphotype III is subject to the same constraints as was
neoteny for morphotype VIII. First, the assumption of descent needs to be justified, but here this
involves fewer uncertainties. There is complete agreement amongst all workers that the progressive
accentuation of the morphology of Bilinguites (levels 8 and 9) from Reticuloceras (level 7) represents

SWAN: HETEROCHRONY IN NAMURIAN AM MONOIDS

1047

a monophyletic lineage. Indeed, Bisat (1924) described successive species, now allocated to
Bilinguites, as ‘mutations’ of R. reticulatum. The derivation of this lineage in level 6 is less
certain, but is clearly within the reticuloceratids, and the ancestral ontogeny must necessarily
have shown a low gradient of morphological change in the direction of the morphotype III
vector.

The problem of the standardization by size can be assessed using the same logic as for morphotype
VIII. Acceleration would not be a supportable hypothesis if the change of gradient of the shape-
age size curve was due to change of rate of growth rather than change in the gradient of the shape-
age curve. This would require that descendants grew slower but with the same timing of shape
changes, which would result in proportioned dwarfism. The change in size of species necessary to
explain the slope changes is approximately a 10-fold decrease. The hypothesis of dwarfism is not
supported by data on species size through time (text-fig. 10/?), which shows little if any trend.
Dwarfism is only tenable as an explanation if accompanied by a delay in timing of maturation
(hypermorphosis) so that larger sizes were attained. This combination must be deemed less likely
than simple acceleration.

DISCUSSION

Neoteny and acceleration, then, are the most likely processes to have dominated the evolutionary
trends of the two ammonoid groups studied. Although few species-to-species lineages are known
with confidence, the systematic trends, involving large numbers of species through a substantial
period of time, are sufficient indication of the operation of heterochrony. It remains to assess the
importance of heterochrony relative to other modes of evolution, and to infer its significance with
regard to the organisms and their environment.

How much heterochrony?

The lack of evolutionary lineages forbids an estimate of the number of actual species-to-species
transitions which were affected by heterochrony. The percentage of gastrioceratacean genera
affected by the documented heterochronic trends, however, is 60-70 %, and there may be other,
more subtle heterochronic events in the residual genera. Other Namurian superfamilies, for example
the Prolecanitaceae, Medlicottiaceae, Dimorphocerataceae, and Goniatitaceae, do not exhibit
extreme ontogenetic changes in external morphology, and evolve comparatively little in the
Namurian, so for these groups heterochrony is less likely and would be difficult to detect. Amongst
the Neoglyphiocerataceae, however, the derivation of the genus Eumorphoceras from the Dinantian
Girtyoceras closely parallels the origin of Bilinguites recorded here, and probably involved the
same process. A minimum estimate for Namurian genera affected strongly by heterochrony is
25 % (approximately 20 % neoteny, 5 % acceleration). Interestingly, these genera supply about one-
half of the zonal ammonoid species in the USSR and about two-thirds in the USA and North-
West Europe. Either heterochrony is important in the rapid evolution necessary for zonal division,
or the rapidity enables the mode of evolution to be deciphered.

Within the heterochronic trends, the effect of heterochrony relative to other, less systematic
evolutionary modes can be estimated. The first three principal components of variation, in which
we know the distribution of specimens of morphotypes VIII and III, comprise 45-8 % of the total
variation: 24-4% in P.C.l; 12-3% in P.C.2; 9-1 % in P.C.3 (Table 1). For morphotype VIII,
removing the cenogenetic effect and the randomizing, non-heterochronic scatter, we can estimate
that about half of the P.C.l and P.C.2 directions of variation are due to neoteny; that is 18 % of
the total or about 40 % of the first three principal components. For morphotype III there is less
scatter in P.C.s 2 and 3 but more in P.C.l; the estimated percentages are 23 % of the total, 50 %
of the first three principal components. The degree of scatter within the morphotypes with respect
to the remaining components of variation can be assumed to be similar, so that the latter figures
(40 % and 50 %) are realistic estimates for the contribution of heterochrony.

1048

PALAEONTOLOGY, VOLUME 31

Why heterochrony?

In interpreting evolution by heterochrony, mention must first be made of Gould’s (1977) attempt
to link progenesis with r-type and neoteny with K-type ecological strategies. Progenesis, involving
early maturity at the expense of morphological specialization, is plausibly claimed by Gould to be
advantageous for rapid turnover of generations in order to exploit ephemeral resources. His
contention that neoteny is a good K-strategy is less logical: morphological ‘fine-tuning’ to a stable
environment is not an automatic result of neotenous juvenilization. Hypermorphosis is the more
apparent converse of progenesis. Neoteny is clearly only advantageous if the ancestral immature
morphology is more successful tha’n the ancestral mature morphology in the particular environmen-
tal situation of the mature descendant. The review by Stearns (1976) of the complexity of life
history strategies and the problems of assessing competing hypotheses is an adequate critique of
such simplistic models, and Alberch et a/. (1979, p. 314) conceded that much depends on the
properties of the specific environment and organism being studied.

Spectacular anisometric ontogenies are well known amongst heteromorph ammonites, and have
invited speculation about changes in mode of life, for example by Klinger (1981). Changes in
Palaeozoic ammonoid ontogenies are more subtle, but have been analysed by Kullman and Scheuch
(1970) and by Kant and Kullman (1978), who consistently detected abrupt changes in allometric
growth constants, but did not attempt a functional interpretation. There is no previous literature
on Carboniferous ammonoid life history.

Swan and Saunders (1987) presented a detailed analysis of the functional morphology of
Namurian ammonoids. Using evidence from hydrostatics, hydrodynamics, apertural morphology,
ornament, and facies associations, modes of life were postulated for each of Saunders and Swan’s
(1984) morphotypes. Results relevant here are as follows: morphotype VIII shows a suite of
characteristics (high drag coefficient, low aperture orientation, potentially cryptic ornament, etc.)
all compatible with a benthic mode of life; morphotype III (with streamlined shell and strong, high
hyponome) was probably nektonic and pelagic; morphotype V, which was adopted by the mature
ancestors of the innovators of both morphotypes VIII and III, is interpreted as versatile, nekto-
benthic; and morphotype VII, which may have been the immature morphology of the morphotype
VIII ancestors, was probably a less sophisticated benthic adaptation. This functional morphological
analysis was based on data from ammonoids at various sizes, and the functional interpretations
are largely independent of size (the exception being a small component of hydrodynamic behavour,
Chamberlain 1981). Consequently, these results can be used in interpreting not only differences
between species but also changes in morphology during ontogeny.

The ontogeny of the immediate ancestor of the morphotype VIII innovator probably included
a transition from morphotypes VII to V: this is almost ubiquitous for the early Gastriocerataceae.
This may be interpreted, following Swan and Saunders (1987) as a transition from a benthic
adolescence towards greater versatility by improved swimming ability at maturity. The cenogenetic
evolution of morphotype VIII did not change this basic scenario: morphotype VIII differs from
VII only in ornament. The new distinctive ornamentation may have been cryptic in effect and
developed in response to predation of juveniles. The subsequent neotenous advance of this
morphology to later ontogeny suggests that benthic conditions became suitable for the entire life
history of the individual; the previously advantageous versatility at maturity became redundant.
High mobility may not be necessary before maturity where there are strongly localized resources
which can be intensively exploited, but in these circumstances mobility is important at maturity
for genetic variability in mating and for appropriate siting of eggs (as in caterpillar and imago
stages of butterflies). The neotenous progression of morphotype VIII may reflect an improvement
in the quality and lateral extent of benthic habitats, so that these habitats could support the entire
ontogeny of ammonoids, and allow sufficient lateral migration without the requirement of strong
swimming ability.

The ancestor of the morphotype III Reticuloceras- Bilinguites lineage would, in common with
other early reticuloceratids, have shown an ontogenetic transition from morphotype VIII to V.
According to Swan and Saunders’ (1987) work, this corresponds to a change from benthic to

SWAN: HETEROCHRONY IN N AMURIAN AMMONOIDS

1049

nekto-benthic, with improvement in swimming ability. The initial foray into morphotype III
occurred in late ontogeny by exaggerated development of a suite of characters: compression,
involuteness, smoothness, depth of hyponomic, and ocular sinuses. These characters favour
hydrodynamic efficiency, and the evolutionary development of the morphotype indicates further
improvement of swimming ability and less dependence on the benthic environment. The subsequent
acceleration of these characters in Bilinguites had the effect of pushing this morphology into earlier
ontogeny. In this way, less of the ontogeny remains suitable for a benthic existence until, in level
8, all but the first 5 mm is well adapted to a nektonic, pelagic lifestyle. It is notable, however, that
even in the terminal extreme of this lineage (represented by B. superbilingue ), the available material
showing early ontogeny, though usually poorly preserved, appears to retain vestiges of morphotypes
V and VIII. The ontogeny, therefore, is condensed by pure acceleration and not by deletion (Gould
1977, p. 75). It seems that the ancestral morphologies have been regressed into early ontogeny as
much as the process of acceleration allowed. In contrast to morphotype VIII, there is a strong
trend in this lineage to reduce the dependence on the benthic environment as much as possible.

The evidence here, then, does not support the concept of a profound relationship between
ontogeny and phylogeny envisaged by Haeckel (1866) and other nineteenth-century philosophers,
or the importance of the cryptogenic juvenile innovations of Schindewolf s proterogenesis; nor
does it support Gould’s (1977) hypothesis of ecological stragegies. Rather, the heterochronic mode
was determined by specific features of the organism’s ontogeny and specific aspects of the changing
environment. Thus, it appears that if an ammonoid ancestor was successful by being morphologically
adapted to exploitation of habitat X in early ontogeny and habitat Y in late ontogeny, then if
habitat X disappeared, a peramorphic descendant was ‘naturally selected’, if habitat X improved,
then a descendant was likely to be paedomorphic.

Namur ian en vironmen ts

The possibility emerges from the preceding discussion that there is evidence for two contrasting
environmental trends in the Namurian. The morphotype VIII development suggests improving
benthic conditions through the latter half of the series; morphotype III, in apparent contradiction,
may reflect a phase of deteriorating benthic conditions, perhaps due to anoxia. (The possible
importance of benthic anoxia in ammonoid evolution was proposed by House 1985.) Saunders
and Swan (1984) contended that the changes in morphologic diversity in the Namurian were, in
general, synchronous world-wide; the possibility of global environmental changes demands more
detailed investigation.

In terms of abundance and rate of neotenous evolution, morphotype VIII is strongest in zones
Rj and G (Russian Nm2b, Nm2c2); in the higher Rt the development is considerably stronger in
the carbonate shelf of the South Urals than in the basinal shales of North-West Europe (text-fig.
4). In the intervening zones (R2a,b,c’ Nm2c1), however, the morphotype is rare everywhere,
represented only by Bilinguites derivatives and early gastrioceratids, all the typical morphotype
VIII reticuloceratid genera having become extinct. This period coincides with the rise of morphotype
III, the maximum rate of acceleration for which was in zones Rlc to R2c (Nm2b3 to Nm2c2), with
greater abundance in North-West Europe. Following the resurgence of morphotype VIII, Bilinguites
declines in abundance markedly. In the European G zone, B. superbilingue occurs occasionally in
thin layers within Cancelloceras- dominated horizons and declines upwards; in the South Urals,
Bilinguites is only common in association with the earlier species of Cancelloceras , C. rurae
(Ruzhencev and Bogoslovskaya 1978, pp. 6-26). Investigation of these trends amongst other
superfamilies is not without difficulties: as might be expected, morphologies interpreted as strongly
pelagic (e.g. Dimorphoceras , Anthracoceratites) are unaffected by the inferred benthic changes, and
the remaining examples of benthic morphotype VII are difficult to interpret (e.g. Syngastrioceras ,
see Swan and Saunders 1987). The benthic morphotype VI, however, does decline in parallel
with VIII.

This evidence, then, is generally compatible with the following sequence of events: 1 , development
of morphotype VIII during radiation into diverse benthic habitats resulting from eustatic

1050

PALAEONTOLOGY, VOLUME 31

transgression in late H zone (Nm2b1); 2, neotenous progression of morphotype VIII in the
Gastriocerataceae particularly in shelf carbonate environments of the Urals in Nm2b2 and Nm2b3
zones, whilst morphotype III begins to develop by acceleration in reticuloceratids due to
deteriorating benthic oxygenation in Europe; 3, poor benthic conditions widespread in R2 zone
(Nm2c1), morphotype III proliferates whilst morphotype VIII survives in few remaining favourable
niches; 4, benthic conditions improve in G zone (Nm2c2), surviving examples of morphotype VIII
re-radiate and morphotype III gradually becomes obsolescent. It must be emphasized, though,
that the quality of the evidence ensures that this scenario is no more than a tentative hypothesis.

SUMMARY

A comprehensive body of morphological data for nearly all known Namurian ammonoid species
at various sizes and localized in space and time provides a suitable data base for the comparison
of ontogenetic and phylogenetic trends. Scores on axes representing principal components of
variation give a good estimate of shape, and ontogenetic trajectories can be constructed by
ordination against size. Differences between ancestral and descendant ontogenies can then be
compared with models of heterochronic results.

Despite the crude estimation of ontogenetic stage by size and the poor resolution of evolutionary
lineages, careful appraisal of the data leads to the parsimonious conclusion that the development
of two gastrioceratacean morphological radiations was strongly affected by heterochrony. A
depressed, evolute, coarsely ornamented morphotype (VIII) evolved by proterogenesis, comprising
an initial cenogenetic event followed by neoteny; the compressed, involute, smooth morphotype
with deep hyponomic and ocular sinuses (III) apparently developed by acceleration. Evolution by
heterochrony is estimated as having affected a minimum of 25 % of Namurian ammonoid genera,
which includes the majority of biostratigraphic index species.

Functional morphological analysis of Namurian morphotypes suggests that heterochrony is not
in itself an ecological strategy for this group. From ancestral ontogenies involving adaptation to
a change in mode of life from benthic to nekto-benthic, neoteny allowed specialization to benthic
habitats throughout ontogeny, and acceleration diminished the benthic stage in favour of nektonic
ability. For this type of evolution, heterochrony is appropriate in that whole morphology can be
transferred to different positions in ontogeny, by single changes in regulatory genes.

Cosmopolitan trends in Namurian ammonoid evolution lead to the suggestion that the
development of morphotype VIII occurred in response to improving benthic conditions in zones
R, and G1? whereas the success of morphotype III could be a symptom of widespread reduced
benthic oxygenation in R2 zone.

Acknowledgements. This paper is a product of a continuing programme of research in collaboration with
Bruce Saunders of Bryn Mawr College, Pennsylvania. I am grateful to Dave Evans for many fruitful
discussions during the past 3 years at Swansea. Thanks also to Wendy Johnson, University College Swansea,
and Paula Thorne-Jones, Kingston Polytechnic, who typed the manuscript with commendable enthusiasm,
and to my wife Isobel for help with some of the figures.

REFERENCES

alberch, p., gould, s. j., oster, G. F. and wake, D. b. 1979. Size and shape in ontogeny and phylogeny.
Paleobiology , 5, 296-317.

bisat, w. s. 1924. The Carboniferous goniatites of the North of England and their zones. Proc. Yorks, geol.
Soc. 20, 40-124.

— 1933. The phylogeny of the North of England goniatites. Proc. Geol. Ass. 44, 255-260.
chamberlain, j. a. 1981. Hydromechanical design of fossil cephalopods. In house, m. r. and senior, j. r.

(eds.). The Ammonoidea. Syst. Ass. Spec. Vol. 18, 289-336. Academic Press, London.
de beer, G. R. 1930. Embryology and evolution, 1 16 pp. Clarendon Press, Oxford.

SWAN: HETEROCHRONY IN NAMURIAN AMMONOIDS

1051

gould, s. j. 1968. Ontogeny and the explanation of form: an allometric analysis. In macurda, d. b. (ed.).
Paleobiological aspects of growth and development, a symposium. Paleont. Soc. Mem. 2, 81 98.

1977. Ontogeny and phytogeny , 501 pp. Harvard University Press, Cambridge, Massachusetts.

Haeckel, E. 1866. Genere/le Morphologie der Organismen, 2 vols., 574 pp., 462 pp. Georg Reimer, Berlin.
house, m. r. 1985. Correlation of mid-Palaeozoic ammonoid evolutionary events with global sedimentary
perturbations. Nature , Lond. 313, 17-22.

hyatt, a. 1889. Genesis of the Arietidae. Bull. Mus. comp. Zool. Harv. 16 (3), 1 238.
kant, R. and kullman, j. 1978. Gehause-Allometrie bei Cephalopoden. Neues Jb. Geol. Palaont. Abh. 157,
98-103.

Kennedy, w. J. and cobban, w. a. 1976. Aspects of ammonite biology, biogeography and biostratigraphy.
Spec. Pap. Palaeont. 17, 1-94.

klinger, h. c. 1981. Speculations on buoyancy control and ecology in some heteromorph ammonites. In
house, m. r. and senior, j. r. (eds.). The Ammonoidea. Syst. Ass. Spec. Vol. 18, 337-335. Academic Press,
London.

kullman, j. and scheuch, J. 1970. Wachstums— Anderungen in der Ontogenese Palaozoischer ammonoideen.
Lethaia, 3, 397-412.

mcnamara, K. J. 1982. Heterochrony and phylogenetic trends. Paleobiology, 8, 130 142.

1986. A guide to the nomenclature of heterochrony. J. Paleont. 60, 4 13.

Newell, n. d. 1949. Phyletic size increase, an important trend illustrated by fossil invertebrates. Evolution ,
3, 103-124.

ramsbottom, w. h. c. 1977. Major cycles of transgression and regression (mesothems) in the Namurian.
Proc. Yorks, geol. Soc. 41, 261-291.

ruzhencev, v. e. and Bogoslovskaya, m. f. 1 978. Namyurski etap y evolyutsii ammonoidei. Pozdenamyurskiye
ammonoidei. Trudy paleont. Inst. 167, 1-336.

saunders, w. b. 1983. Natural rates of growth and longevity of Nautilus belauensis. Paleobiology, 9, 280
288.

— ramsbottom, w. h. c. and manger, w. l. 1979. Mesothemic cyclicity in the mid-Carboniferous of the
Ozark shelf region? Geology, 7, 293-296.

and swan, a. r. h. 1984. Morphology and morphologic diversity of mid-Carboniferous (Namurian)
ammonoids in time and space. Paleobiology, 10, 195-228.
schindewolf, o. h. 1936. Palaontologie , Entwicklungslehre und Genetik, 108 pp. Berlin.

Stearns, s. c. 1976. Life-history tactics: a review of the ideas. Q. Rev. Biol. 51 (3), 3 47.
swan, a. r. h. 1984. A revision of some Silesian goniatites using cluster analysis. Ph.D. thesis (unpublished).
University of Leeds.

— and saunders, w. b. 1987. Function and shape in late Paleozoic (mid-Carboniferous) ammonoids.
Paleobiology , 13, 297-311.

ANDREW R. H. SWAN
Department of Geology
University College
Singleton Park
Swansea SA2 8PP

Present address:

School of Geological Sciences
Kingston Polytechnic
Penrhyn Road
Kingston upon Thames
Surrey KT1 2EE

Typescript received 8 October 1986
Revised typescript received 8 April 1988

■:

A NEW SPECIES OF STEM-GROUP CHORDATE
FROM THE UPPER ORDOVICIAN OF NORTHERN

IRELAND

by a. p. cripps

Abstract. A new scotiaecystid, Scotiaecystis collapsa sp. nov. is described from the Killey Bridge Beds, lower
Cautleyan Stage, Ashgill Series, near Pomeroy, Co. Tyrone, Northern Ireland. It is most closely related to
S. curvata Bather. The interrelations of cornutes are studied through a cladistic analysis using PAUP
(Phylogenetic Analysis Using Parsimony) involving twenty-one species and thirty-nine characters. Three
equally parsimonious trees are obtained and their information content summarized in the form of a consensus
tree. By the addition of the mitrates (primitive crown-group chordates) this consensus tree is resolved. As in
previous studies, the genus Cothurnocystis forms an uncharacterizable group with some species more
crownward than others. The genus Thoralicystis is also paraphyletic. The Scotiaecystidae are an extinct
monophyletic group more crownward than C. elizae Bather and less crownward than the Phyllocystidae,
Amygdalotheca , and Reticulocarpos. The scotiaecystids exhibit a departure from a primitively deposit-
feeding mode of life towards suspension feeding. The hind tail of S. collapsa is unusual for it is not
truncated at the end as is often the case in other cornutes and is flexible in both the horizontal and vertical
planes. The family Phyllocystidae is redefined to contain Phyllocystis and Chauvelicystis , and the family
Cothurnocystidae redefined to include only Cothurnocystis elizae Bather, C. cowtessolei Ubaghs, and C.
primaeva Thoral.

The aims of this paper are to describe a new species of stem-group chordate belonging to the
plesion (family) Scotiaecystidae (Caster and Ubaghs, in Ubaghs 1967) and to provide a cladistic
analysis of the cornutes. The cornutes and mitrates are controversial. More precisely, two groups
are currently proposed as the living models for these fossil forms: the phylum Echinodermata
(Ubaghs 1967; Philip 1979) and the phylum Chordata (Jefferies 1967). Whilst Ubaghs and Philip
agree that cornutes and mitrates are echinoderms, their interpretations of the structure believed
here to be the tail differ. Philip believed this organ to be a crinoid-type stem and called it a stele
(although he accepted its locomotory function), but Ubaghs argued that it was an aulacophore or
feeding arm. The interpretation of Jefferies, that these organisms were chordates, is adopted here.
The detailed arguments for this view are found in Jefferies (1986).

PHYLOGENETIC METHODOLOGY AND CLASSIFICATION

The stem-group concept of Hennig (1965, 1969, 1981) is used here in order to place fossils in
relation to extant groups when reconstructing phylogeny. Except for the case of a genuine polytomy,
every fossil organism must be more closely related to one living group than to any other. Two
living groups relevant in the case of the cornutes are the Echinodermata and Chordata. The view
held here is that all of the cornutes are more closely related to living chordates than to living
echinoderms.

Affinity with a particular living group can only be established on the basis of shared derived
characters. Thus, all cornutes share with living chordates a locomotory tail, muscle blocks, and
a notochord amongst other features (Jefferies 1986). But the features characterizing extant
chordates did not all arise at once, that is, they did not all suddenly appear in the most recent

IPalaeontology, Vol. 31, Part 4, 1988, pp. 1053-1077, pi. 93.|

© The Palaeontological Association

1054

PALAEONTOLOGY, VOLUME 31

common ancestor of extant chordates. Rather, they were acquired gradually by the cornutes and
inherited by the hypothetical common ancestor of all living chordates. This latest common
ancestor of the living chordates, together with all of its descendants, constitute the crown-group
Chordata.

There then remain those forms which do not belong to the crown-group Chordata, but which
nevertheless are more closely related to this group than they are to echinoderms. These are the
stem-group chordates (text-fig. 1). The total group of Jefferies (1986), or Hennig’s (1969)
‘Gesammtgruppe’, in this case the phylum Chordata, consists of the stem-group chordates plus
the crown-group chordates. All of the cornutes are stem-group chordates. The mitrates, not
discussed in this paper, are primitive crown-group chordates.

text-fig. 1 . The stem-group concept as applied to chordates.

The problems arising in classifying fossils have been considered at length by Patterson and
Rosen (1977), Wiley (1979, 1981), and Jefferies (1979, 1986). All these authors agree that the
addition of a fossil species or group to an existing classification should be possible without
disrupting that classification. Patterson and Rosen (1977) suggested that fossil groups be designated
plesions and that ‘it should no longer be necessary to rank fossils formally, except within extinct
monophyletic groups’ (p. 160). All of the taxa designated plesions in this work are stem-group
chordates. This means that each plesion possesses at least one derived character which links it to
the chordate crown-group.

In erecting a cladistic classification of the cornutes, Wiley’s convention four (1979) has been
adopted in placing certain groups ‘sedis mutabilis ’ at the hierarchical level where their interrelations
are known. Such groups form part of an unresolved trichotomy or polytomy. Jefferies (1986) has
recently discussed the plesion concept and its usage. Under his definition, a plesion always includes
a segment of the stem-lineage to which it is attached and consequently must always be paraphyletic.
Moreover, if there is a trichotomy or polytomy in the stem lineage, all the groups involved are
deemed to constitute a single plesion, since no member of this plesion can be shown to be more
closely related to the crown-group than any other. The different approach of Jefferies from that
of Wiley will not be considered further here and in fact would make no difference to the classification
resulting from this work.

CRIPPS: NEW ORDOVICIAN CORNUTE

1055

SYSTEMATIC PALAEONTOLOGY

Superphylum deuterostomia Grobben 1908
Subsuperphylum dexiothetica Jefferies 1979
Phylum chordata Bateson 1886

Plesion (Family) scotiaecystidae Caster and Ubaghs, in Ubaghs 1967
Genus scotiaecystis Caster, in Ubaghs 1967
Species Scotiaecystis collapsa sp. nov.

The trivial name collapsa refers to the fact that even the best specimens have collapsed upon burial.

Material, horizon, and localities. All the known specimens of Scotiaecystis collapsa sp. nov. are from the
lower Cautleyan part of the Killey Bridge Formation, Ashgill Series (Upper Ordovician), near Pomeroy, Co.
Tyrone, Northern Ireland. For an account of the stratigraphy see Mitchell (1977). According to R. P. Tripp
(pers. comm. Oct. 1987), the upper part of the Killey Bridge Formation is of Rawtheyan age, but the evidence
for this statement has not yet been published. The lower part of the Killey Bridge Formation remains of
Cautleyan age. About 140 specimens of S. collapsa are known, all preserved as empty moulds and most of
them disarticulated. Among the better specimens, E63072 carries part and counterpart of two almost complete
individuals lying on top of each other, and of these, the individual nearer the camera in Plate 93, fig. 3 is
chosen as holotype.

All the specimens are from Mitchell’s localities 2 and 3 (1977, text-fig. 2), mostly from the latter. The
details are as follows:

Locality 2. Warren Wood River. Grey shales on the banks and stream bed of the Warren Wood River,
2 km east-south-east from Pomeroy Square, 3-2 km south-south-west of Craigbardahessiagh, 160 m upstream
of the junction with Bonn River (Irish grid reference H 7130 7128).

Locality 3. Little River. In situ in grey shales in a river cliff on the south bank of Little River and also
from a small, old quarry tip, now largely removed, on the north bank of the river. 3-6 km east of Pomeroy
Square, 1-6 km south-south-east of Craigbardahessiagh, 160 m east of Slate Quarry Bridge (H 7297 7268).

The material was collected in three small expeditions from the British Museum (Natural History). These
took place in May 1977 (R. P. Tripp and S. F. Morris), in May 1978 (R. P. Tripp, S. F. Morris, and R. P. S.
Jefferies), and in June 1984 (R. P. S. Jefferies and E. H. Westergaard). The first specimens of this species
were noticed by Mr R. P. Tripp in August 1977 when breaking up material collected during May 1977. The
specimens are all conserved in the Department of Palaeontology, British Museum (Natural History) with the
following registration numbers:

Locality 2. E29662-E29681, E63088, E63089.

Locality 3. E29682-E29742, E63065- E63087 (including holotype on E63072), E63271 -E63284.

METHODS OF STUDY

In order to reconstruct the three-dimensional shape of S. collapsa on the drawing-board, a life-
size model was made using casts taken of the individual plates. To avoid any difficulties associated
with scale, all plate impressions for the model, save one, were taken from the same specimen
E29709u, b (part and counterpart). Plate 1, which was missing on this specimen and which occupies
a peripheral position, was taken from E29724a, b. The model was complete enough to give a
precise idea of the general shape, but lacked plates g, j, r, h, and i. For these plates, impressions
and latex casts taken from several other specimens were studied and they were reconstructed as
accurately as possible.

Casts were made using Reprosil, a low-viscosity, high-precision silicone impression material. In
making the model, the casts of individual plates were fixed to each other by a more viscous variant
of the same substance in a different colour, so enabling the cement between the plates to be
distinguished from casts of the plates themselves. This was applied around the plate junctions so
that the integrity of the sutures was preserved. Any silicone left in the natural mould was easily
removed with an organic solvent such as dichloromethane.

1056

PALAEONTOLOGY, VOLUME 31

Many of the specimens were cleaned in thioglycollic acid (10 % in water) before latex casts were
made using a red latex solution.

ANATOMICAL DESCRIPTION

Plate nomenclature follows that of Jefferies and Prokop (1972), as recently elaborated in Jefferies,
Lewis, and Donovan (1987). A simple alphabetical notation is used. Plates given the same letter,
initially on the basis of the crownward cornute Reticulocarpos hanusi Jefferies and Prokop, are
believed to be homologous. Ubaghs, in his work on these animals, has used a different system for
naming the plates. His notation is reproduced in Table 1 for comparison. The plates present in S.
collapsa are labelled in text-fig. 2.

S. collapsa , like all other cornutes, consists of a distinct head and tail (PI. 93, fig. 3; text-figs. 2,
3, 4, 5a). On the holotype the head is 13 mm across at its widest point. The whole animal, with
straightened tail and including the length of the b-appendage, is about 30 mm long, though this
length cannot be accurately determined. The head is asymmetrical and rather boot-shaped. It is
bordered by a marginal frame of thirteen calcite plates, two of which (1 and b) project from the
frame, are greatly elongate, and can be called appendages. The tail is attached to the posterior
part of the frame and is situated midway between the left and right sides of the animal.

Comparisons will be made, in particular, with S. curvata Bather (text-fig. 6) in the following
description and character analysis. 5. curvata is believed to be (for reasons discussed later) the

table 1. Comparison of plate nomenclature and organ terminology according to
Jefferies (1986) and Ubaghs (1970).

Jefferies

Ubaghs

1. Plates

a

m5

b-appendage

glossale

c-appendage

digitale

d

M'4

e

M'3

f

M'2

g

M\

h, i

adorales

J

Mi

k

m2

1-appendage

m4, m3

m

zygale

?

s

m4

t

m3

u, ii

?

V

M's

w

m6

x

not distinguished
from M,

y

central adoral

2. Organs

head

theca

tail

aulacophore

mouth

periproct

anus

right adoral opening

gills

cothurnopores, lamellipores

notochord

water vascular canal

CRIPPS: NEW ORDOVICIAN CORNUTE

1057

text-fig. 2. Scotiaecystis collapsa. a , dorsal aspect; b, ventral aspect; c, right lateral aspect; d , left lateral

aspect; e, anterior aspect; /, posterior aspect.

1058

PALAEONTOLOGY, VOLUME 31

text-fig. 3. Outline drawing of Plate 93, fig. 3. Labels other than
plate notation are; fac, facet for the attachment of the dorsal
integument; ft, fore tail; hto, hind tail ossicles; ie, interbranchial
elements; ksp, k-spike; pli, plates belonging to lower animal; sty,
stylocone; tt, terminal tail ossicle of lower animal.

closest relative of the animal described here. The cornute Cothurnocystis melchiori Ubaghs, 1983,
is, in the following description and discussion, referred to as Thoralicystis melchiori , for it is shown
later that this species is a scotiaecystid and not closely related to other species of Cothurnocystis.

There is only one oral appendage in S. Collapsa as in some other scotiaecystids. This is thought
to be plate b and equivalent to the left oral appendage of C. elizae Bather. However the fa-
appendage of S. collapsa is unique in two ways (text-fig. 7a): 1, running down each side of the
plate are numerous serrations which appear slightly larger towards the base of the appendage, and
2, no less obvious are two kinks one about half-way along the plate, the other two-thirds of the
way towards the rounded apex. These kinks are present on almost all specimens where there is a
mould of the b-appendage. The left appendage or plate 1 (spinale of Ubaghs) is similarly serrated
and both b and 1 are dorsoventrally compressed, presumably helping these appendages to slice
through the substrate during locomotion, like a pair of bread knives. Their action would create a
cloud of suspended mud particles which could be sucked in through the mouth for feeding. Plate
1, unlike b, terminates in a point (text-fig. Id).

Plate s, another scotiaecystid feature, is also found in S. collapsa. This plate is markedly curved
and together with plate a is responsible for the dorsally convex shape of the anterior frame (PI. 93,
fig. 2; text-fig. 2e, f). This feature differs only slightly from that of S. curvata , in that the summit
of the convexity is not at the s-a junction but on plate s. The functional significance of such an
arcuate shape is hard to explain, particularly as the posterior frame is not nearly as convex.
Interestingly, in one specimen (E29681) plate s is unusually long relative to the other plates and
very slender. This specimen was presumably juvenile in view of its small size. In growing to its
adult condition, s would have lengthened relatively less than the other plates.

Three protuberances are found attached to the frame on the ventral surface of the head which

EXPLANATION OF PLATE 93

Figs. 1 5. Scotiaecystis collapsa sp. nov. 1, 2, stereo-pairs of model, based upon silicone rubber casts of
plates, dorsal and anterior views, respectively. The posterior frame and strut are incomplete, x2-l. 3, 4,
latexes of the holotype BMNH E63072u (text-fig. 3). 3, two individuals, in ventral aspect, of which the more
complete upper individual was chosen as the holotype. The lower individual exhibits two important features
of the hind tail; its lateral flexibility and a rounded terminal ossicle, x5T. 4, the tail of the holotype, in
ventral aspect, showing the paired plates of the fore tail, the stylocone, and the ossicles of the hind tail,
x 15-3. 5, latex of BMNH E29715a showing stylocone (text-fig. 15). The median groove and two pairs of
oblique grooves can be seen, as can the laterally directed processes on the sides of the stylocone, x 7 0.

PLATE 93

CRIPPS, Scotiaecystis collapsa sp. nov.

1060

PALAEONTOLOGY, VOLUME 31

are extremely variable in shape and size. These spike-like processes are found on plates f, k, and
r, and are named after the plates that bear them. The f- and k-spikes also have significant laterally
directed projections which are not found in S. curvata. On plate k the lateral and ventral processes
are connected by a flange, whereas the lateral process on f terminates very abruptly with a pointed
anterior edge. The r-spike is the most variable of all, but commonly is a finger-like projection,
ovoid in cross-section and directed away from plate f. These spikes probably served to anchor the
head in the mud, preventing it from slipping forwards during locomotion (Jefferies et al. 1987).
Plate k, like b and 1, is serrated along part of its left edge.

Although plate a is slightly shorter relative to s, it is of the same general appearance as its
homologue in 5. curvata. It sends a ventral process, a part of the strut, which may have possessed
a slight kink, to meet plate g and support the ventral integument. Plate e, like k, b, and 1, is
serrated. The serrations are, as elsewhere, arranged in a line, but are found only upon the outer
edge of the plate. They are fewer and larger than those of the other plates. Four serrations were
counted. Neither e nor k is serrated in S. curvata.

The small plates h and i, near the tail-base, show important differences from those in S. curvata.
This region, where the fore tail joins the head, is the most complicated part of the animal (text-
figs. lb, 8, 9). Both h and i have a strongly convex anterior face and a concave posterior face, thus
distinguishing them from their homologues in S. curvata which are almost square in dorsal aspect.
Furthermore, in S. collapsa , unlike S. curvata, h and i do not send out ventral processes to meet
in the mid-line anterior to the posterior coelom (text-fig. 9c), and i is not in contact with the
interbranchial elements.

Plates g and j meet ventrally at the tail-base (text-fig. lb). Dorsally, these two plates have curved
facets for the attachment of h and i (text-fig. 9b). On the posterior surface of g and j in the region
of the mid-line is a symmetrical basin (text-fig. If). This excavation is believed to have carried the
brain. Also observed here, and appearing to leave the brain, are two canals for the median-line
nerves. Lateral to these canals are the so-called pyriform bodies which, like the nerve canals, are
preserved as natural casts (text-fig. 8). These bodies have been interpreted by Jefferies (1968) as
the trigeminal ganglia and seem to overlap the reception groove for the fore tail anteriorly (text-
figs. 8 and 9).

Plates g, j, and the small plate d exhibit no important differences from the same plates in S.
curvata. All the marginal plates are approximately triangular in cross-section (text-fig. 2a), except
for the appendages, and the latexes show that they were made of labyrinthic stereom (Smith 1984).
Appendage 1, unlike all the other plates (including b), was constructed from a fasciculate type of
stereom.

CRIPPS: NEW ORDOVICIAN CORNUTE

1061

text-fig. 5. Scotiaecystis collapsa sp. nov. A, latex cast of BMNH E63072fi showing two individuals, one
lying on top of the other, of which the lower individual is the holotype in dorsal aspect (text-fig. 4). Most of
the head of the upper individual is missing but the posterior frame is present, x 5 I . b, natural mould, BMNH
E63072Z), SEM close-up of fore tail and cerebral basin, x26-6 (text-fig. 8). c, latex of BMNH E29683n lit
from the bottom right in order to show more clearly the interbranchial elements as well as plates 1, k and s,

x 7-3.

1062

PALAEONTOLOGY, VOLUME 31

text-fig. 6. Scotiaecystis curvata. a, dorsal aspect; b. Scotiaecystis curvata; ventral aspect. Redrawn after
Jefferies 1986, text-figs. 7.12a and 7.12c, respectively.

HEAD OPENINGS AND INTERBRANCH I AL ELEMENTS

The interbranchial skeletal elements of the gill openings, which number about forty-live (an
accurate count is not possible), are arranged in a curved series and situated in the left part of the
dorsal integument. They are placed more posteriorly than in S. curvata. If the shape of the animal
is compared with a boot, then the gills occupy the posterior half of the ‘toe’ (text-fig. 10). In other
words, the series of gills does not neatly bisect this area (the left part of the dorsal integument),
unlike the gill series in S. curvata and T. zagoraensis (Chauvel). This relatively posterior position
of the gills is probably the primitive condition within the cornutes (see below). Unlike S. curvata ,
the interbranchial plates are not chevron-shaped and in fact are slightly concave dorsally (text-fig.
11). They are altogether much simpler structures than those of S. curvata , possessing none of the
grooves or more complicated processes of the latter and lying more or less parallel to each other
(text-fig. 5c). Ventrally, the elements bear relatively simple terminal processes (text-fig. 11/?).
Another interesting feature of the interbranchial elements is that, on one specimen (E29683), a
suture is visible running across some of the plates and found about one-quarter of the way down
each element from the front edge. If, as Jefferies believes, these plates represent the fusion of
adjacent halves of anterior and posterior u-plates, then these sutures are well placed to be the
vestiges of such a transformation.

The only evidence of the mouth is found on specimen E63072a in the form of a few pointed,
elongate plates which are largely hidden beneath a displaced e-plate. Nevertheless, these plates are
enough to show that the mouth is dorsally placed as in all known scotiaecystids, except T.
zagoraensis as described and figured by Chauvel (1971). In both S. collapsa and S. curvata there
is no sign of an external gonopore-anus and the gonorectal canal opens into the gills. This canal
is preserved as a natural cast on E630726 (text-figs. 5b and 8). It is a much larger structure than
in S. curvata and runs through the i-j suture on its way to the gill slits (text-fig. 9c).

CHAMBERS AND SOFT ANATOMY OF THE HEAD

The asymmetrical shape of most cornutes, including S. collapsa , has been explained by Jefferies
(1979) as a consequence of descent from a bilaterally symmetrical ancestor that lay down on its
right side. The arguments supporting this view will not be repeated here, but Jefferies (1986)

CRIPPS: NEW ORDOVICIAN CORNUTE

1063

text-fig. 7. Scotiaecystis collapsa sp. nov. a, latex of BMNH E29682 b showing plates f, e, d, and b in ventral
aspect. The serrations on the b-appendage can be clearly seen, x 6-5. b, latex of BMNH E29682u showing
the g-j junction, plate k and, faintly, some of the interbranchial elements, x 8-8 (text-fig. 9c). c, latex of
BMNH E29706 viewing the inner surface of plate a and showing the lines of attachment of the buccal cavity
and pharynx, x 13-8 (text-hg. 12). d, latex of BMNH E29715u showing the 1-appendage (top), plate k beneath

it, and plate f left of centre, x 4-8.

1064

PALAEONTOLOGY, VOLUME 31

cast of left pyriform body

cast of right pyriform

text-fig. 8. Scotiaecystis collapsa. Tail-base
region of natural mould (E630726).

text-fig. 10. Scotiae-
cystis collapsa. Bisector
of ‘toe’ region.

depression for left pyriform body

depression for right pyriform body

text-fig. 9. Scotiaecystis collapsa. a , reconstruction of tail-
base region; b , tail-base region with h and i removed; c, anterior
view of tail-base region.

anterior ventral process

text-fig. 1 1 . Scotiae-
cystis collapsa. Inter-
branchial elements: a,

external; b, lateral; c,
internal views.

provides a synthesis of the evidence. This hypothetical ancestor would most likely have been similar
to the living hemichordate Cephalodiscus resting upon its right side — an orientation called
dexiothetism. Such an ancestor, Jefferies believes, also gave rise to the echinoderms. Therefore the
two phyla Echinodermata and Chordata have been united by him to form the superphylum

CRIPPS: NEW ORDOVICIAN CORNUTE

1065

text-fig. 12. Scotiaecystis collapsa. Inner face of plate a,
BMNH E29706, text-fig. 7c.

Dexiothetica. Jefferies’ theory is uniquely able to account for the peculiar asymmetries observed
in the development of echinoderms, tunicates, and cephalochordates. It also explains the boot-like
shape so characteristic of many primitive cornutes.

In the cornutes there is evidence for the existence of four chambers in the head; the buccal cavity
just behind (in some cases below) the mouth, the pharynx situated in the ‘toe’ part of the ’boot’,
the posterior coelom just anterior to the tail base, and the right anterior coelom underlying the
pharynx. A fifth chamber is postulated to exist based upon comparative evidence alone. This is
the left anterior coelom and is reasoned to be present through a comparison with Cephalodiscus.
Such a comparison suggests that the right metacoel of Cephalodiscus is the homologous chamber
of the right anterior coelom of cornutes.

The evidence for the existence of the head chambers in S. collapsa is described below and shown
in text-figs. 12 and 13. The buccal cavity, pharynx, and posterior coelom have all left evidence of

text-fig. 13. Scotiaecystis collapsa. a , Reconstruction of internal cavities of head: i, dorsal; ii, left lateral; iii,
posterior aspects, b , reconstruction of internal cavities of head (for key see text-fig. 13a): i, anterior; ii, right

lateral; iii, ventral aspects.

1066

PALAEONTOLOGY, VOLUME 31

their presence in the form of a series of ridges and grooves which can be seen on the inner faces
of the marginal plates (text-figs. 1c and 12). In addition to these clues, the gross morphology of a
given region of the head can indicate the position of a chamber. The only good example of this is
in the case of the buccal cavity, with the shape and position of plates a and e enabling recognition
of its posterior boundary. This evidence is backed up by other clues which come from studying
the inner surfaces of the plates concerned (text-figs. 12 and 13).

On the inner faces of the marginal plates are found an upper, a middle, and a lower zone. The
upper and lower zones are concave excavations whilst the middle zone is a distinct ridge. The
height of this ridge, and the degree of separation of the upper and lower zones varies from place
to place. The upper and lower zones are concave facets for the attachment of dorsal and ventral
integuments, respectively. The middle zone represents the attachment of various head chambers
to the marginal frame. A clearly defined facet, also concave, can be found on both lateral surfaces
of the ventral strut which is formed by plates a and g. These facets are for the attachment of the
ventral integument alone.

The pharynx would have been the largest chamber in the head, occupying about two-thirds of
the space available. Its existence is confirmed by the presence of grooves on the inner surfaces of
the head plates. There is no observable boundary between the pharynx and the right anterior
coelom in S. collapsa , but grossly they probably had the same relative positions as in all other
cornutes for the following reasons: 1, the gill slits are in the left part of the dorsal integument and
would have opened out of the pharynx; 2, the gonorectal groove emerges from the presumed
position of the right anterior coelom and indicates that the gonad and most of the non-pharyngeal
gut were in the ‘heel’ part of the head; and 3, the height of the frame is greater to the right of the
tail than to the left of it, so this part of the animal was capacious enough to hold the right anterior
coelom and its contents.

The anterior border of the posterior coelom is marked by a groove in the natural mould of
E630726 corresponding to a ridge on plates g and j. This chamber, which is roofed over by plates
h and i, probably extended posteriorly as far as the fore tail. The left anterior coelom, in all
cornutes and mitrates, is purely hypothetical and virtual.

text-fig. 14. Scotiaecystis collapsa. Integument plates.

The integument plates are bobbin-shaped as in S. curvata. However, the axis of the bobbin is
more elongate and sometimes lacks one of the two heads— usually the outer one (text-fig. 14).
Plate density, i.e. number of plates per unit area, is greater on the ventral than dorsal surface, and,
on the dorsal surface, between the area surrounding the gills. The density variations are thus
similar to those in S. curvata.

THE TAIL

The tail of S. collapsa is represented on about twelve specimens but is most complete on E63072u
and b (PI. 93, figs. 3 and 4). On E63072# the tail is seen in ventral aspect on two individuals. It is
clearly divided into fore tail, mid tail, and hind tail. Although an accurate count of the number
of hind tail segments is not possible, it is fairly certain that there are at least twenty-one segments,
which is more than have so far been discovered in S. curvata (Jefferies 1968, p. 275, states that the
hind tail of S. curvata must have had about sixteen segments).

The fore tail is composed of eight segments. The skeleton of each segment consists of a pair of
small dorsal plates and a pair of much larger ventral plates. The ventral plates are somewhat wider

CRIPPS: NEW ORDOVICIAN CORNUTE

1067

text-fig. 15. Scotiaecystis collapsa. Stylocone; internal
structure in dorsal aspect.

than the dorsal plates (meaning width in a direction transverse to the long axis of the tail) and
hence are visible in dorsal as well as ventral aspect. The plates of each segment imbricate beneath
those of the segment immediately in front. This is probably an adaptation allowing for dorsoventral
flexibility of the fore tail, since when the tail is stretched during flexion the plates are able to slide
over each other without stretching the intervening soft tissues. The most anterior pair of ventral
plates overlap plates g and j.

The fore tail segments terminate posteriorly at the stylocone. This is a funnel-shaped structure,
overlain dorsally by two pairs of plates which are very different in shape to those found anywhere
else in the tail. In form they are much like the corresponding plates in S. curvata. The stylocone
itself is more ventrally situated than in most species and on many specimens is seen to bear a pair
of laterally directed processes which are like those observed by Ubaghs (1983) in T. melchiori ,
although in the latter they are directed upwards.

Several specimens also show the internal structure of the stylocone in dorsal aspect (PI. 93, fig.
5; text-fig. 15). There is a median groove, believed to have carried the notochord, which is always
preserved as a natural cast. The soft structure which it housed must have extended into the fore
and hind tails. The median groove sends out two pairs of lateral grooves which are directed slightly
rearwards. These grooves disappear underneath a shelf on either side. Some detail can be seen
upon this shelf too in the form of a pair of depressions. The posterior depression borders the bases
of the above-mentioned lateral processes, whilst the anterior one runs towards the front edge of
the stylocone. Similar depressed areas have been described by Ubaghs (1970) in the stylocones of
Cothurnocystis primaeva Thoral and T. griffei (Ubaghs).

The hind tail, as stated earlier, consists of at least twenty-one segments. As in the mid and fore
tail, the dorsal plates are paired and meet at the mid-line. There is only one ventral ossicle per
segment, however, as is true of most cornutes. The ventral ossicles approximate to hemicylinders
but gradually become less deep distally. Concomitant with this distalward flattening are two further
changes: 1, the sutures between the ossicles change from being planar and straight in ventral aspect
to being curved and convex anteriorly in ventral aspect; 2, each ossicle bears upon its dorsal face
a pair of transverse buttresses. As the ossicles flatten so these buttresses become more laterally
orientated (text-fig. 2c, d) and hence visible in ventral view.

Small, rounded protuberances appear on the ventral surfaces of the ossicles towards the tail end
(text-fig. 2b). There is one to each ossicle and they are situated in the mid-line, close to the posterior
suture. These were most likely used for gaining a good purchase on the substrate when the animal
used its tail like a hook during the locomotory cycle (see discussion of locomotion in Jefferies et
al. 1987). These knobs are clearly seen on the last seven ossicles in E63072(/ and may well be
homologous with the ventral spikes seen in more crownward cornutes such Reticulocarpos
hanusi. The terminal tail ossicle of the lower individual on specimen E63072^/ has a rounded tip.
A similar condition has been described in T. melchiori (Ubaghs 1983, p. 35, text-fig. 7c and pi.
VIII, fig. 1). The situation in these two species therefore contradicts Jefferies's assertion that
autotomy at the end of the cornute tail was ‘normal in cornute ontogeny’ (1986, p. 230). It appears,
however, to have been habitual in mitrates and in the most crownward cornutes such as R. hanusi ,
for this species is sometimes observed to have only one ossicle in the hind tail and has never been
shown to have more than four (Jefferies and Prokop 1972). The tail impression of the lower
individual on specimen E63072o also demonstrates that the hind tail was flexible in the horizontal
plane (PI. 93, fig. 3; text-fig. 3).

1068

PALAEONTOLOGY, VOLUME 31

The dorsal surfaces of the hind tail ossicles are not well shown in any specimen and nothing
can be said concerning the facets for articulation of the dorsal plates. A longitudinal median groove
can be recognized, as in the stylocone, with a pair of oblique, rearwardly directed lateral grooves
in each ossicle. The dorsal plates of the hind tail are hemicrescentic in outline (text-fig. 2a) when
viewed from above. Apart from the first pair of plates, which bear small, forward-facing processes,
these plates are uniform except that (like the ventral hind tail ossicles) they narrow gradually
towards the end of the tail. The oddest aspect of the whole hind tail is that each dorsal plate makes
contact with three ventral ossicles, i.e. it extends along the whole length of one ossicle and projects
anteriorly on to the more proximal neighbouring ossicle and posteriorly on to the more distal
neighbouring ossicle. Posteriorly, each plate overlaps the plate behind it. Each plate bears a large
anterior process which lies just in front of the transverse buttress belonging to one of the ossicles.
The plate extends posteriorly, overlapping the following plate, traversing an entire ossicle (text-
fig. 2c, d). In S. curvata, by contrast, each plate is in contact with only a single ossicle.

INTERRELATIONS OF STEM-GROUP CHORDATES

Twenty-one species of cornutes were coded for thirty-nine characters and subjected to a cladistic
analysis using PAUP (version 2.4.1), a computer program devised by Dr David Swofford of the
Illinois Natural History Survey. PAUP (Phylogenetic Analysis Using Parsimony) has a number of
option settings. No weights were applied to any of the characters and Ceratocystis perneri Jaekel
was used to root the tree. C. perneri is the only cornute to retain a hydropore and is considered
by Jefferies (1969, 1979, 1986), on the basis of this and other characters, to be the most primitive
known chordate. PAUP is able to deal with missing data, represented in the data matrix (Table
2) by a question mark, which is treated as either 0 or 1, and reversals of character-state are
permitted.

The program produced an initial tree and then undertook global branch swapping until a shorter
tree was found. PAUP discovered three trees with a minimum length of sixty-nine steps and with
a consistency index of 0-565. From these trees an Adams’s consensus tree (Adams 1972) was
constructed (text-fig. 16) which provides a summary of the different most parsimonious solutions.
The trichotomy consisting of Amygdalotheca griffei Ubaghs, R. hanusi and R. pissolensis Chauvel
can be resolved through the addition of primitive crown-chordates (mitrates) to the tree (text-fig.
17). R. pissotensis is the most crownward cornute species, sharing with the mitrates a convex
ventral surface (in other cornutes it is the dorsal surface which tends to be convex). Galliaecystis
lignieresi Ubaghs was placed by Jefferies (1986) in a more crownward position than shown in text-
figs. 16 and 17, due its possession of a dorsal bar. But set against this, Galliaecystis retains an
1-appendage, an asymmetrical head, and a large number of gill openings. It seems more parsimonious
to assume that Galliaecystis acquired its dorsal bar independently of Reticulocarpos.

Phyllocystis and Chauvelicystis form a clade on the basis of five synapomorphies. However, of
these five only one — tuberculated posterior marginals — is uniquely derived. In Chauvelicystis the
tubercles articulate with the more posterior spines characteristic of this genus. The two species of
Phyllocystis — P. blayaci Thoral and P. crassimarginata Thoral— appear to form a clade, both
having a heart-shaped marginal frame.

The monophyly of the Scotiaecystidae (text-fig. 18) is supported by two uniquely derived
characters. These are the possession of plate s and of interbranchial elements. The relations within
this group are more problematical. The relative positions of T. griffei and T. melchiori are uncertain.
On the one hand the two species may be sister taxa as shown in the consensus tree. This hypothesis
is based upon the possible reappearance in these species of one of the two small plates (v or w) at
the anterior edge of the head. But, as discussed below, it is not possible to know whether one,
both, or neither of these plates was present in the most primitive scotiaecystid, T. zagoraensis. If
both were absent then the reappearance of one of them (v or w?) in T. griffei and T. melchiori is
a possible synapomorphy. T. griffei also shares loss of the e-spike with more derived scotiaecystids,
but once again the condition in T. zagoraensis is unknown.

table 2. Character data matrix for twenty-one species of stem-group chordates. Character-states described represent the derived conditions.

CRIPPS: NEW ORDOVICIAN CORNUTE

1069

9SBpu3dde-[ jo 99U9sqy
SIBUI^JBUI

jouajsod p3jE[nDJ3qnjL
pB9q padeqs-yeajj
ssuids

JO 9<?UUJ SJB9q pB3jq
9§UBy jBJ9qdu9(j

auqpiui jBSJop ui
ajnjns S3)B[d |iB}-9Joq

S9SS990id |BJJ
-U9A SuuB9q jibi puijq

PBJUOO >J-l JO 93U9Sqy
n puB n jo 30U3sqy

S9A[Bq jBnb9 oa\j

ojui t90j, spiAip sqir)

X9AU09

X[[BSJOp 9UJBJJ JOIJ9}Uy

XllXjBqd J9A09 S9}B[d
}U9uing9iui pgpuno^j

S§UIU9do
[|i§ jo J3quinu 93iBq

XjpuiuiXs
S9jB|d-n JOU9juy
JBq [BSJOQ
}TU}S

UIOJJ p9pn[9X9 B 9}B[CJ

XyBSJOp
X9AUOO S9JBjd [BiqDUBig

SJU9UI9J9 [Biq3UBJqJ9}U]

j 9^ia

9X3 UBip9UI JO 99U9Sqy
uoi}B|n9qjB p-9

S[|I§ 0]UI
SU9do [BUB9 [BP9JOUOQ

m j° m

o] su3do snuB-3JodouoQ
qinoui [bsjoq
jrujs [Biju9y\

p 9JB|d OJ UO XjIAB9
[Boonq jo ]jiqs pjBAVioj

9 JO 99U9Sqy
A\ pUB A JO 93U3Sqy
A\ JO A JO 99U9Sqy
9>Jlds-9 JO 93U3Sqy
X jo 93U9sqy
x jo 93U9sqy
S 91B1J

1 3^ld
S9^lds

^ PUB J JO 99U3Sqy

jooy pB9q 3iqix9|j
JOOJ pB3q 3|qiX9IJ
9JodojpXq jo 93U9sqy

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o-

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

o

0000 000 o 00 O — — — — — — o

OOOO OOO O OO O O OO — — — o-
OOOO OOO o OO o o OO — — OO
OOOO OOO O OO O O OO OO — —
OOOO OO— O OO O — — o- OO OO

OOOO OOO

0 0 0-0 — — —

o — — —

o- o- o- O

— OO o* o-

— — OO — c-

00 00 00

o

o

OOOO o- o* O"

00—— OOO

o- o* o- — — —

OOOO OOO

o — —

— O OO OO o-o-

00 00

o

o

o

o

OOOO OOO

OOOO OOO

OOOO 00 —

OOOO — — —

OOOO OOO

O- O O- — —I — ^

— O — 1 — o- o- O

OOOO 00 —

— — o- o 00 00 00

00 o — — — — — 00

00 — o — — 00 00

O — —

00 00

o o

o o

o

o

0 0-00 — o — —

o o-

00 00 00

00 00 00

00 00 00

— — OO OO-
OO 00 00

00 00 00

00 — —

00 —

o- o-

OOOO 0 0-0 — — — o O OO o-o- OO

OOOO Oo-O — — — o- ^ o-o- — — <

OOOO — o- — — — ■ — — o- — — — — — —

OO— • — 00--H T— I O — — — ^ —H —

_H O — — — — — — — — — — — OO OO

OOO— O o- — — — — — — . — OO

OOOO — — — ^ — — < o o 00 00 00

O — — — 000 o 00 o o 00 — — — —

OOOO 0C--0 — 00

o — —

g

£ -d a*
S 5 s
a. t «

'•2 ^

Q <>j En

^ c ^

<G> 'G

act:
s = s
is,^> ^
G U

-2 a

■G -S ^ ^ .a
~S -a 5

- f" ~ V ^

r • ^

a

■N
| 2

.5 .2

5U o g M

. -5;

(Jk 2-

co

. a
CO <3

13 igt-

b« 1. ■

V ; t«

? « a

a s c .s

2 £?

R « & a
a? a, GO G

table 2. Character data matrix for twei

melchiori
T. zagoraens
T griffei

Bohemiaecystis

bouceki

? 7 7

0 00010000
0 10000000
0 1 0 0 I I 0 0 0

0 I 0 1 I 1 0 0 0

0 10 1 0

? I ? 1 0

110 1 0
0 10 1 0

1 ? 0 0
0 0 0 0
1 ? 0 0
110 0

0 ? 1 0 0 10

0 ? 1 0 0 10

0 ? I ? 0 10

0 01010100 1 11
7 0 1 ? 1 ? ? 7 7 7 10

0 0 1 1 1 I I 0 0 1 II

I 0
I 0

7 0 110 10

7 1110 10

7 0 110 II

Scoliaecyslis
cur vat a
S. collapsa

Reliculocarpos
hanusi I 1

R. pissotensis 1 1

Phyllocyslis

blayaci I 1

P crassimarginala I 1

Chauvclicystis
spinosu I I

C. ubaghsi 1 l

0 0 111111

0 0 1 I 1 I 1 I

0 0 0 110 11

1 0 0 I 1 1 7 7

1 0 0 1 I 1 I 1

I 0 0 11111

1 I 0 I 0 I 1 7

1 1 0 I 0 I 1 7

1 10 0 0 111

1 10 0 0 111

0 I 1
0 1 1

1 1 0
l 1 0

1 1 0
I I 0

7 1 1

7 1 1

1 1 7

1 1 7

I 1110
1 0 1 I I

17 111

0

0 7

1 7 7 0 1 0 1 1

7 7 7 0 1 0 7 0

I

1070

PALAEONTOLOGY, VOLUME 31

O

Cothurnocystis primaeva possesses a confusing array of plesiomorphic and apomorphic traits,
but a clue to its systematic position may lie with the contact that the c-appendage makes with
plate d. In C. elizae and C. courtessolei Ubaghs this contact is in the form of a mobile articulation
(albeit more obviously so in C. elizae ). The extent of its development in C. primaeva is unclear
from published photographs (Ubaghs 1970). However, in the same paper Ubaghs states that the
c-appendage (his digitale) is attached to plate d (his M'4) ‘par une articulation qui parait peu
differenciee’. It is therefore possible that this character, the c-d articulation, defines a group
consisting of C. elizae , C. courtessolei , and C. primaeva. Within this group C. elizae and C.
courtessolei share a plate t (acquired in parallel with the Phyllocystis-Chauvelicystis clade) and a
covering of circular integument plates for the pharynx.

The cladogram presented in text-fig. 17 is the most parsimonious solution to the distribution of
the thirty-nine derived character states used in this study if the assumption is made that all
characters carry equal weight. Some justification of certain of the character polarities as entered
in the data matrix will now be given. Since I had no access to specimens of most of the species
used in this analysis, it was necessary to rely upon published descriptions of these organisms.

First, the loss of plate x is here believed to be a derived feature despite the absence of this plate
in Ceratocystis perneri. This is because x is present in Protocystites menevensis Hicks and probably
also in Nevadaecystis americana Ubaghs, considered on the basis of other characters to be the
most anti-crownward known cornutes apart from C. perneri. The loss of x is in fact the only
character used in this analysis which I believe to be derived in C. perneri , yet there is also a

CRIPPS: NEW ORDOVICIAN CORNUTE

1071

text-fig. 17. Character-state tree for stem-group chordates with primitive crown-group chordates (initiates)
added. Synapomorphy scheme: 1, notochord; 2, locomotory tail; 3, loss of hydropore; 4, flexible head roof;
5, gonopore-anus opens to left of tail; 6, plate x; 7, ventral strut; 8, anterior u-plates; 9, fore tail ossicles meet
in the mid-line; 10, flexible head floor; 11, loss of u and u\ 12, loss of y; 13, loss of median eye; 14, c-d
articulation; 15, plate t; 16, pharynx covered by rounded integument plates; 17, loss of x; 18, loss of v and w;
19, forward shift of buccal cavity on to d; 20, loss of i-k contact; 21, loss of e-spike; 22, head symmetry; 23,
reduced number of gill openings ; 24, loss of 1-appendage; 25, plate t; 26, plate y; 27, dorsal mouth: 28, median
eye, 29, tuberculated posterior marginals; 30, heart-shaped head; 31, i-k contact', 32, loss of symmetry, 33,
plate x; 34, head with fringe of spines; 35, hind tail with ventral processes; 36, peripheral flange; 37, plate a
excluded from strut; 38, dorsal bar; 39, strut is not in contact with marginal plates; 40, convex ventral surface.

Characters in bold are parallelisms, italicized characters are reversals.

possibility that this species possesses a plate wax (Jefferies et at. 1987) which broke up in more
derived cornutes into the three plates w, a, and x.

At the anterior edge of the marginal frame, between plates a and d, there are commonly found
one or two smaller plates which have been called by Jefferies and Prokop (1972) v and w. In C.
perneri , P. menevensis, and throughout the genus Cothurnocystis (excluding T. melchiori ) both v

Sc o t c y til c o 1 1 a p i a

Thorallcyatla lagotaonali

PALAEONTOLOGY, VOLUME

1072

PALAEONTOLOGY, VOLUME 31

<0

Q.

o

o

(0

O

4>

«0

O

o

V)

o

o

co

o

05

«

N

O

■c

text-fig. 18. Character-state tree for the Scotiaecystidae. Synapomorphy scheme: 1, plate s; 2, interbranchial
elements; 3, dorsal mouth; 4, loss of c; 5, rearward shift of buccal cavity, 6, gonorectal canal opens into gills;
7, plate r; 8, anterior frame convex dorsally; 9, pharynx covered with rounded integument plates. Characters in
bold are parallelisms, italicized characters are reversals.

and w are present. This is probably the primitive condition. In T. melchiori and T. grijfei (the
situation in T. zagoraensis is unknown) one of the two plates has evidently been lost, though
whether v or w cannot be determined. In all the more derived scotiaecystids, Galliaecystis ,
Amygdalotheca, Reticulocarpos, and in Chauvelicystis both plates have disappeared. The condition
in Phyllocystis is uncertain. Because of the incompleteness of the only known specimen of T.
zagoraensis (Chauvel 1971), it is not possible to say whether the loss of just v or w is a parallelism
with reference to the cladogram given here. It could be assumed that T. zagoraensis, like all other
scotiaecystids, is at least without one of the two plates. The most parsimonious solution (text-fig.
17) is that v and w were lost in the common ancestor of G. lignieresi plus all more crownward
cornutes. Yet, as mentioned earlier, this implies the reappearance of one of these plates in T. griff ei
and T. melchiori.

CRIPPS: NEW ORDOVICIAN CORNUTE

1073

The exact situation of the posterior right boundary of the buccal cavity is difficult to determine,
especially when, as is sometimes the case, the mouth is not preserved. In the primitive condition,
seen in Ceratocystis perneri , the posterior boundary of the cavity was attached to plate e on the
right side of the marginal frame. This is also where it was attached in most other cornutes. In
Cothurnocystis courtessolei, Galliaecystis , Amygdalotheca, Reticulocarpos, and Chauvelicystis the
posterior boundary of the buccal cavity is attached, on the right, to plate d rather than to e. In
other words it has shifted forwards relative to the plates of the frame. The situation in the more
primitive scotiaecystids is less clear; T. zagoraensis is too incomplete to make any statement at all
regarding the position of the posterior right boundary of the buccal cavity. In T. melchiori and
T. griffei it appears that the posterior right boundary of this cavity was probably attached to d.
If true then this feature, the forward shift of the posterior border of the buccal cavity, characterizes
a group consisting of Galliaecystis plus all more crownward cornutes. The derived state is most
parsimoniously interpreted as having been independently acquired in Cothurnocystis courtessolei.
The dorsal situation of the mouth is derived within the cornutes and is used here to characterize
a group within the scotiaecystidae, excluding only T. zagoraensis (text-fig. 18). From Chauvel’s
(1971) description of T. zagoraensis, it appears that the mouth is almost terminally placed — the
primitive condition— and was most likely anterior to plates v and/or w. In this feature, T.
zagoraensis probably resembled C. elizae and there is also no sign of a frame anterior to the mouth
in Chauvefs photograph. From this terminal position of the mouth in T. zagoraensis it is possible
to establish an evolutionary trend based upon the scheme of interrelations proposed here (text-
fig. 18).

In T. melchiori the mouth is dorsally placed (Ubaghs 1984) and just posterior to v or w (Ubaghs’s
plate M6). In T. griffei, which also has a plate v or w, the mouth is clearly dorsal and lies some
distance inwards from this plate (Ubaghs 1970). In Bohemiaecystis houceki Caster the mouth is
obscured, but in S. curvata and S. collapsa it is dorsal and now found well away from the anterior
frame. This is presumably a more favourable position for suspension feeding and therefore marks
a change from the more primitive type of deposit feeding hypothesized for C. elizae and T.
zagoraensis. The significance of the varying distance between the dorsally placed mouth and
anterior buccal frame is unknown, as is the reason why a whole group of cornutes took to
suspension feeding. A dorsally situated mouth is also found in the two species of Phyllocystis and
most likely in Chauvelicystis too, having been acquired independently of the scotiaecystids.

In the majority of cornutes the opening of the gonorectal canal is clearly external and in all
except Ceratocystis perneri to the left of the tail. In S. curvata , S. collapsa, and T. griffei the
gonopore-anus opens into the gills. B. houceki shares with S. curvata and S. collapsa loss of plate
c and rearward shift of the buccal cavity on to plate e, yet B. houceki has an external gonopore-
anus. Either this species has reverted to the primitive state or T. griffei evolved the derived condition
independently (the view adopted here).

One of the three gill-associated characters used in this study is the possession of interbranchial
elements in the form of rigid skeletal units separating the gill slits. As stated earlier these are
believed to have evolved from adjacent halves of the u-plates found in more primitive cornutes.
These elements are found in all scotiaecystids including, it is asserted here, T. melchiori. From
published photographs in Ubaghs (1983) it seems unclear as to whether T. melchiori possesses such
elements or the more primitive anterior and posterior u-plates, but in the light of its systematic
position among the scotiaecystids, based upon other characters, the interpretation of these plates
as interbranchial elements seems reasonable, despite Ubaghs’s contention that they surround
cothurnopores. The interbranchial elements are convex dorsally in T. griffei and S. curvata. In the
latter species they are also chevron-shaped. In B. houceki they are vertically sloping lamellae,
convex ventrally.

Another character is the presence of anterior u-plates bordering the gill openings. Only C. perneri
and Protocystites menevensis among known cornutes primitively lack these plates (Jefferies et al.
1987), whereas the principle of parsimony dictates that in other cornutes, such as Reticulocarpos,
they have been secondarily lost.

1074

PALAEONTOLOGY, VOLUME 31

The number of gill openings has also been used as a character in this analysis, for despite the
fact that the number is highly variable and that the gill count is unknown in four species, there
are some valid distinctions to be made here. In C. perneri and N. americana there is a maximum
of seven branchial openings which I take to be the primitive number (in P. menevensis the count
is uncertain). In Cothurnocystis elizae this figure of seven has roughly doubled and in T. melchiori
the number has increased to twenty-five. T. zagoraensis and T. griffei both have about thirty-two
branchial openings, whilst in B. bouceki, S. curvata , and S. collapsa there is yet another increase
to around forty-five. If the scheme of interrelations depicted in text-figs. 17 and 18 is accepted,
there is a clear trend, at least among the scotiaecystids, to increase the number of branchial
openings, although there is one reversal in this tendency in T. melchiori. In the Phyllocystis-
Chauvelicystis clade and in all more crownward cornutes there is a secondary reduction in the
number of gill openings correlated with the attainment of symmetry.

The position of the gills relative to the anterior and posterior parts of the frame may be
functionally significant. The primitive and most widespread condition is typified by Cothurnocystis
elizae and T. melchiori in which the gills are positioned quite close to the posterior frame, as in
Ceratocystis perneri. In T. zagoraensis and S. curvata the gill-slit series bisects the ‘toe’ region. In
this position the respiratory current would leave the animal perpendicular to the integument
(Jefferies 1968) rather than parallel to the integument as in other forms in which the gills are
situated to one side of the bisector. If the bisector (text-fig. 10) corresponded to the line of greatest
stretching as the pharynx became swollen with water, which seems likely, then the water already
utilized for respiratory purposes and, in the case of S. curvata , the waste from the gonopore-anus,
would be ejected clear of the animal with maximum force. This character would therefore seem to
be advanced both on the basis of outgroup comparison and functional adaptation. However, it
has a limited taxonomic significance in the sense that it does not appear to characterize a natural
group. Unfortunately, branchial openings have not been described in four of the twenty-one species
included in this study.

The loss of i-k contact due to the contraction of plate i occurred independently in Protocystites
menevensis and in a more crownward group including Galliaecystis , the scotiaecystids, Phyllocystis -
Chauvelicystis, Amygdalotheca, and Reticulocarpos.

Derstler (1979) has put forward a rather different scheme of cornute interrelations to that shown
in text-fig. 17. As in this study, and others by Jefferies, he recognizes that Ceratocystis perneri is
the most primitive cornute so far described and also that some cornutes are more closely related
to mitrates than others— his Amygdalothecidae. This family consists of the three genera Galliaecystis ,
Amygdalotheca , and Reticulocarpos. All of the other cornutes, except for C. perneri and Phyllocystis ,
are contained within his suborder Cothurnocystida. Derstler’s Cothurnocystida and Amygdalothe-
cidae are shown here and elsewhere to be paraphyletic groups and as such uncharacterizable. The
results of this study indicate that if the taxon Cothurnocystidae is to be retained then it is perhaps
better restricted to Cothurnocystis elizae , C. courtessolei, and C. primaeva which may indeed form
a clade. ‘C.’ fellinensis Ubaghs is more anticrownward than any of these three and T. melchiori is
a scotiaecystid.

CLASSIFICATION OF STEM-GROUP CHORDATES

A classification of the cornutes, based upon text-figs. 17 and 18 can now be given:

Superphylum Deuterostomia
Subsuperphylum Dexiothetica
Phylum Chordata

plesion Ceratocystis perneri Jaekel
plesion Protocystites menevensis Hicks
plesion Nevadaecystis americana Ubaghs
plesion Cothurnocystis fellinensis Ubaghs
plesion (family) Cothurnocystidae

CR1PPS: NEW ORDOVICIAN CORNUTE

1075

Unnamed subfamily

Cothumocystis primaeva Thoral
Subfamily Cothurnocystinae

Cothumocystis courtessolei Ubaghs
Cothumocystis elizae Bather
plesion Galliaecystis lignieresi Ubaghs
plesion (family) Scotiaecystidae
Subfamily Thoralicystinae (new)

Thoralicystis zagoraensis (Chauvel)

Subfamily Scotiaecystinae (new)

Thoralicystis melchiori (Ubaghs), sedis mutabilis
Thoralicystis griffei (Ubaghs), sedis mutabilis
Tribe Scotiaecystini (new), sedis mutabilis
Bohemiaecystis bouceki Caster
Scotiaecystis curvata Bather
Scotiaecystis collapsa sp. nov.
plesion (family) Phyllocystidae
Genus Phyllocystis

Phyllocystis blayaci Thoral
Phyllocystis crassimarginata Thoral
Genus Chauvelicystis

Chauvelicystis spinosa (Ubaghs)

Chauvelicystis ubaghsi Chauvel
plesion Amygdalotheca griffei (Ubaghs)
plesion Reticulocarpos hanusi Jefferies and Prokop
plesion Reticulocarpos pissotensis Chauvel
Subphylum Cephalochordata
Subphylum Urochordata
Subphylum Craniata

CONCLUSIONS

The cornutes are a paraphyletic assemblage of stem-group chordates. Within this assemblage are
three recognizable monophyletic groups. The first such group is the Scotiaecystidae, characterized
by having an s-plate and rigid interbranchial elements. The second monophyletic group I have
called the Cothurnocystidae and is composed of Cothumocystis courtessolei , C. elizae , and C.
primaeva. These three species seem all to have an articulation between plates c and d. Because
some members of the genus Cothumocystis are more crownward than others it is clearly an artificial
grouping, as is the genus Thoralicystis and the genus Reticulocarpos. A third monophyletic group
I have called the Phyllocystidae which includes Phyllocystis and Chauvelicystis ; it is characterized
by a plate t, dorsal mouth, and tuberculated posterior marginals. The new species described here,
Scotiaecystis collapsa , is a member of the Scotiaecystidae and is most closely related to S. curvata.
It shares with this species a plate r, a dorsally convex frame, and a gonorectal canal that opens
into the gills. S. collapsa may be distinguished from S. curvata by its possession of the following
features:

a. Appendage b has two obvious kinks and is serrated along both edges, b , Appendage 1 is
serrated and terminates in a point, c. Plates e and k also bear serrations, though not along their
entire length, d. The interbranchial elements are approximately parallel to one another and are
relatively posterior in the left dorsal integument, c. The interbranchials are not chevron-shaped
but slightly concave dorsally and simpler internally./, Plates f and k both bear significant laterally
directed processes as well as ventral spikes, g, The summit of the dorsally convex anterior frame

1076

PALAEONTOLOGY, VOLUME 31

is formed by plate s. Relative to S. curvata , plate a has given ground to s, sending a relatively
shorter dorsal process to meet it. h, The integument plates are generally fewer per unit area and
larger, and, although bobbin-shaped as in S. curvata , the central process of each bobbin is more
attenuated, i, The r-spike is commonly a finger-like projection, sometimes bearing a terminal flange,
and is directed away from the f-plate. j, Plates h and i have convex anterior surfaces and do not
send out ventral processes which meet in the mid-line. Plate i does not contact the interbranchial
elements, k. The gonorectal canal passes through the i-j suture on its way to the gills. 1, The fore-
tail skeleton consists of eight plates and ossicles as opposed to six in S. curvata. m. The stylocone
bears two laterally directed processes, n, The dorsal plates of the hind tail are not lobate as in
S. curvata but semi-crescentic and each dorsal plate contacts three successive ventral ossicles, o.
There are ventral protuberances on the hind tail which have not been discovered in S. curvata. p.
The median line nerves leave the brain through separate notches in the natural mould, not through
a single tunnel-like canal as in S. curvata.

Acknowledgements. I would especially like to thank Dr R. P. S. Jefferies for many hours of stimulating
discussion and much invaluable advice throughout the course of this project. Drs P. L. Forey and C. J.
Humphries devoted a considerable amount of their time in steering me through the computing. I thank them
for their patience. Mrs Frances Mussett read an earlier draft of the manuscript and made several helpful
suggestions. I would also like to thank Mr Ron Tripp who discovered this species and whose enthusiasm
ensured that the material was at last described. He, Mr S. F. Morris, Mr E. H. Westergaard, and Dr
R. P. S. Jefferies collected all the known specimens of S. collapsa. Their efforts made this paper possible.

REFERENCES

adams, E. n. 1972. Consensus techniques and the comparison of taxonomic trees. Syst. Zool. 21, 380 387.
bateson, w. 1886. The ancestry of the Chordata. Q. J. microsc. Sci. 26, 535-571.

chauvel, J. 1971. Les Echinodermes Carpoides du Paleozoique inferieur marocain. Notes Serv. geol. Maroc.
31, no. 237, 49-60.

— and nion, J. 1977. Echinodermes (Homalozoa: Cornuta et Mitrata) nouveaux pour 1’Ordovicien du
Massif Armoricain et consequences paleogeographiques. Geobios , 10, 35-49.

derstler, k. 1979. Biogeography of the stylophoran carpoids (Echinodermata). In gray, j. and boucot, a.
(eds.). Historical Biogeography, Plate Tectonics and the Changing Environment, 91-104. Oregon State
University Press.

grobben, K. 1908. Die systematische Einteilung des Tierreiches. Verb. Zool.-bot. Ges. Wien, 58, 491 -51 1.
hennig, w. 1965. Die Acalyptratae des baltischen Bernsteins. Stuttg. Beitr. Naturk. 145, 1-215.

1969. Die Stammesgeschichte der Insekten , 436 pp. Kramer, Frankfurt-am-Main.

1981. Insect Phytogeny, 514 pp. Wiley, Chichester.

jefferies, r. p. s. 1967. Some fossil chordates with echinoderm affinities. Symp. zool. soc. Land. 20,
163-208.

— 1968. The subphylum Calcichordata (Jefferies, 1967), primitive fossil chordates with echinoderm affinities.
Bull. Br. Mus. nat. Hist. (Geol.), 16, 243-339.

— 1969. Ceratocystis perneri Jaekel— a Middle Cambrian chordate with echinoderm affinities, Palaeontology,
12, 494-535.

— 1979. The origin of the chordates— a methodological essay. In house, m. r. (ed.). The origin of major
invertebrate groups, 443-477. Systematics Association Special Volume 12, 1-515.

— 1986. The ancestry of the vertebrates, 376 pp. British Museum (Natural History), London.

— lewis, m. and donovan, s. k. 1987. Protocystites menevensis Hicks 1872— a stem-group chordate
(cornuta) from the Middle Cambrian of South Wales. Palaeontology, 30, 429-484.

— and prokop, R. J. 1972. A new calcichordate from the Ordovician of Bohemia and its anatomy,
adaptations and relationships. Biol. J. Linn. Soc. 4, 69- 1 15.

mitchell, w. i. 1977. The Ordovician Brachiopoda from Pomeroy, Co. Tyrone. Palaeontogr. Soc. [Monogr.],
138 pp.

patterson, c. and rosen, d. e. 1977. Review of ichthyodectiform and other Mesozoic teleost fishes and the
theory and practice of classifying fossils. Bull. Am. Mus. nat. Hist. 158, 85-172.

CRIPPS: NEW ORDOVICIAN CORNUTE

1077

philip, G. M. 1979. Carpoids— echinodenns or chordates? Biol. Rev. 54, 439 471.
smith, a. B. 1984. Echinoid Palaeobiology, 190 pp. Allen & Unwin.

ubaghs, G. 1967. Stylophora. In moore, r. c. (ed. ). Treatise on Invertebrate Paleontology. Part S. Echinodermata
1 (2), S495-S565. Geological Society of America and University of Kansas Press.

— 1970. Les echinodermes carpoides de l’Ordovicien inferieur de la Montagne Noire, Cab. Paleont. 1 12 pp.

— 1983. Echinodermata. Notes sur les echinodermes de l’Ordovicien inferieur de la Montagne Noire
(France). In courtessole, r., marek, l., pillet, j., ubaghs, g. and vizcaino, d. (eds.). Calymenina,
Echinodermata et Hyolitha de T Ordovicien de la Montagne Noire ( France Meridionale). Mem. Soc. Etud.
sci. TAude, 33-55, pis. 8

wiley, e. o. 1979. An annotated Linnaean hierarchy, with comments on natural taxa and competing systems.
Syst. Zool. 28, 308-337.

— 1981. Phylogenetics— the theory and practice of phylogenetic systematics , 439 pp. New York, Wiley.

ANTHONY P. CRIPPS

Department of Palaeontology
British Museum (Natural History)

Typescript received 9 November 1987 Cromwell Road

Revised typescript received 17 March 1988 London SW7 5BD

TAPHONOMY OF THE EOCENE LONDON CLAY

BIOTA

by PETER A. ALLISON

Abstract. The London Clay of Sheppey, Kent, is a grey plastic clay which was deposited in an offshore
marine environment. It contains a diverse assemblage of well-preserved plant and animal fossils in concretions
of either pyrite, apatite, or calcite. A diagenetic and geochemical study of the London Clay biota shows that
apatite was the first preservational mineral to form, followed by calcite and pyrite. Mineralogy is strongly
related to original biological composition. Only those organisms with an original skeletal phosphate content
(i.e. vertebrates and arthropods) have been phosphatized. Thus a geochemical bias accounts for the
preservation of the greatest detail in fossils of these groups. Early diagenetic mineralization is the only process
which can halt the information loss occurring during decay. For this reason organisms preserved during the
earliest phases of mineralization retain the most detail.

The organic precursors of fossils can be thought of as chemically exotic sedimentary particles
which achieve equilibrium with the surrounding sediment by decay and mineralization. Details of
the nature and mineralogy of preservation can therefore yield information on the diagenetic and
geochemical history of the enclosing strata. In addition, a fuller understanding of the processes
responsible for early diagenesis and exceptional preservation will pin-point possible locations for
new exceptionally preserved biotas.

Allison (1988a) has shown that anoxia is ineffective as a long-term preservational medium and
that only mineralization can halt decay-induced information loss in the fossil record. Further,
preservational mineralogy of a biota is strongly related to original organic composition. This
results in a geochemical taphonomic bias whereby those fossils associated with the earliest
phases of mineralization exhibit a higher level of preservation than those formed by later
events.

The London Clay biota presents a variety of exceptionally preserved plant and animal remains
and can be defined according to Seilacher (1970; see also Seilacher et al. 1985) as a Konservat or
conservation Lagerstatten. This study examines the mineralogy and preservation of the London
Clay biota and discusses the palaeoenvironment and diagenetic sequence of the host rock. The
London Clay flora represents one of the world’s most diverse fossil fruit and seed assemblages,
containing over 500 plant types including 300 named species (Collinson 1983). It is regarded as
one of the best preserved and most diverse assemblages of fossil plant material in Europe. For a
taxonomic review of the flora, see Chandler (1961, 1964) and Collinson (1983).

The animals of the London Clay biota are as well preserved as the plants. The hard parts of
mammals, birds, reptiles, fish, arthropods, and molluscs almost always occur in three dimensions
within pyrite or calcium phosphate concretions. Soft-part preservation is very scarce, but includes
a pyritized maggot (Rundle and Cooper 1970) and the pedicle of a terebratulid brachiopod (Rowell
and Rundle 1967). The diversity of the fauna has attracted considerable scientific attention from
invertebrate and vertebrate specialists alike, e.g. Murray and Wright (1974) on the Foraminifera,
King and King (1976) on the brachiopods, Davis and Elliott (1957) and Curry (1965) on the
molluscs. Keen (1978) on the ostracodes, Quayle and Collins (1981) on the crabs and lobsters,
Casier (1966) and Ward (1979) on some of the fish, and Hooker et al. (1980) on some of the
mammals and reptiles.

| Palaeontology, Vol. 31, Part 4, 1988, pp. 1079-1100, pi. 94.|

© The Palaeontological Association

1080

PALAEONTOLOGY, VOLUME 31

text-fig. 1. Map showing palaeogeography of
southern Britain during Eocene times. Principal
occurrences of London Clay are arrowed.

GEOLOGICAL SETTING

Outcrop of the London Clay in the British Isles is limited to the London and Hampshire basins (text-fig. 1).
By far the best and most complete sections occur along the coast, although brick pits and motorway cuttings
have created additional exposure (see Collinson 1983 for details). Although these basins are currently
separated by the Chalk ridges of the Downs, they were originally deposited in a single trough along the
northern flank of the Anglo-Paris Basin (Curry 1965; Davis and Elliot 1957; Wills 1951). Repeated subsidence
and sedimentation has resulted in a cyclic sequence of marine, brackish, and estuarine deposits. The clay
mineral suite occurring in Tertiary sediments from the western part of the Hampshire Basin is dominated by
a kaolin illite assemblage derived from the West Country granites (Gilkes 1967). However, the sediments
occurring in the east are rich in montmorillonite and may be derived from either locally eroded Chalk (Gilkes
1967) or the decomposition of pyroclastic ash deposits (Knox and Harland 1979). During marine transgressions
of the basin, such as that responsible for the deposition of the London Clay, the illite/montmorillonite suite
extended from the eastern province to include both the London and Hampshire basins. The shore line during
London Clay times is thought to have run south-west roughly from the Wash to a few miles west of the Isle
of Wight (text-fig. 1; Wills 1951). Sand and silt horizons occurring near what was the Eocene shoreline thin
eastwards (King 1981) into the stiff blue-grey muds which are so characteristic of the London Clay.

The London Clay within the London Basin attains its maximum thickness of 165 m on the Isle of Sheppey
where it crops out as a series of monotonous stiff blue-grey clays with numerous nodule bands (Davis 1936).
King (1981) proposed a lithostratigraphical classification of the London Clay and the associated beds, which
together he referred to as the Thames Group. In his classification this group is divided into the Oldhaven
and London Clay formations. The latter is further sub-divided into divisions A E of which the upper two
(D and E) crop out along the Sheppey coast.

By far the best exposed section at Sheppey occurs at Warden Point where almost 50 m of sediment are
exposed. Much of the material upon which this study is based was collected from either the pyrite and nodule
concentrates on the foreshore or in situ from the cliff.

PRESERVATIONAL STYLE

Vertebrates. At Sheppey fish teeth and vertebrae are the most common vertebrates, although
mammalian (Hooker et al. 1980) and avian remains have been recorded (Harrison and Walker
1977). The fossils are most commonly preserved in concretions of either phosphate or pyrite and
seldom occur in carbonate concretions. Phosphatie fossils show a greater degree of articulation
and in some cases may be completely intact. Preservation of some of the fish includes articulated
hard parts (skull, vertebrae, etc.) enclosed by a cylindrical bag of scales in what appears to be life
position. Soft parts (muscles and viscera, etc.) are absent from such fossils and the scales are

ALLISON: TAPHONOMY OF AN EOCENE BIOTA

1081

text-fig. 2. a, phosphatic crab-bearing concretions, x 0-8. b. X-ray radiographic print of concretion showing
partially disarticulated crab; c and d refer to areas of X-ray enlarged in text-fig. 2c, d, x 0-8. c, d, photographic
enlargements of B. c shows crab pincer with apodemes (a), x 7. d, crab leg with spine (s) and possible skirt
of sensory hair (h), x 10. X-ray photographs were taken with a Phillips mg 161-160 kV constant potential
X-ray unit. The X-ray tube had a focus of 0-4 x 0-4 mm and a film focus distance of 600 mm. Inherent
filtration of the X-ray tube was equivalent to 1 mm of beryllium.

separated from the skeleton by phosphatized sediment. Thus it is clear that decay had destroyed
soft tissues prior to phosphatization. Decay of soft parts commonly leads to tissue collapse and
flattening of carcasses (Zangerl and Richardson 1963; Zangerl 1971; Conway Morris 1979; Briggs
and Williams 1981). The preservation of scales in an uncompacted life position therefore implies
that sediment infill of the body cavity occurred during the decay of soft parts. Such a mode of
preservation could only be achieved with extremely rapid rates of sedimentation.

Vertebrate hard parts within these concretions are preserved in brown-black calcium phosphate
in contrast to the buff colour of the enclosing concretion.

Arthropods. The arthropod fauna is dominated by crabs and lobsters although barnacles and
insects also occur. Insects are pyritized and three-dimensional, and include both adult and larval
forms (Venables and Taylor 1963). Rundle and Cooper (1970) suggested that the insects at Sheppey
were wood-boring forms which may have been transported in floating timber.

Eumalacostracans are always preserved in dark francolite within buff-coloured concretions of
calcium phosphate (text-fig. 2a). They are invariably fragmented and disarticulated but are rarely
flattened and show fine morphological detail including cuticle lamination and pore canals (text-
fig. 6i). In some specimens it is possible to differentiate between endo- and exocuticle (text-fig. 6i).
X-ray radiographic methods have demonstrated the preservation of delicate structures such as
spines, sensory hairs (text-fig. 2b, d), the antennae of a crab (text-fig. 3a, b), and even the apodemes
or muscle attachment sites on the inside of a crab pincer (text-fig. 2c).

X-ray methods have usually been used in the past to detect pyritized structures. The phosphate
in which the London Clay arthropods are preserved is chemically similar to that of the enclosing
nodule. However, the phosphate of the crustacean cuticle is more densely crystalline than the
enclosing phosphatized sediment. Thus, the fossils are only rendered visible to X-rays by a slight
difference in density/porosity between the calcium phosphate of the concretion and that of the
arthropod cuticle. Flattened concretions are more amenable to X-rays than rounded forms. This
is because variations in thickness in the concretion affect the absorbance of X-rays and thereby
control the exposure of the radiograph. Thus flattened concretions with an even thickness have a
uniform exposure (text-fig. 3a). Rounded concretions present a technical problem in that very few

1082

PALAEONTOLOGY, VOLUME 31

text-fig. 3. a. X-ray radiographic print of crab in phosphate concretion, xl-6. b, drawing of X-rayed
specimen, L = appendages, S = carapace, A = antennae, C = pincers.

X-rays pass through the thicker centre of the concretion compared with the thinner periphery.
Prints taken from such radiographs of the periphery of the concretion require considerable
photographic ‘dodging’. In the case of the print showing the crab chelae (text-fig. 2c) the upper
margin of the print received 360 times as much light as the lower margin of the print.

Shelly fossils. Molluscs and brachiopods are commonly found as isolated elements amongst pyrite
concentrates on the beach at Sheppey. They are commonly infilled with pyritized sediment and/or
euhedral pyrite. Pyrite often replaces original shell, although even where this is the case, it is
possible to pick out original laminar skeletal structure. In some cases it is possible to identify
vertical rods of pyrite normal to the shell surface which appear to be pseudomorphing original
skeletal fabric (text-fig. 6d). Calcareous relics of the original shell are rare within these pyritized
remains but occasionally the more heavily calcified parts, such as the spire of gastropods survive
(text-fig. 6e, g). Original shell material of both gastropods and bivalves is found in concretions of
calcium carbonate (text-fig. 4) and calcium phosphate. In this instance even delicate shells are
uncracked and show no indication of sediment compaction. Concretion formation was therefore
pre-compactional.

Borings of the bivalve Teredina squamosa are common in calcified (text-fig. 4) and pyritized
(text-fig. 9e) wood within concretions. Modern representatives of Teredina have a reduced shell
which is rocked back and forth by the animal as it bores its way into the wood. Since most of the
soft parts lie outside the shell, the animal secretes a thin layer of calcite to line the boring. Within
calcified wood from Sheppey this lining calcite is ubiquitous, and the articulated shells are found
in the boring. Tangential sectioning of mineralized borings reveals a series of semi-vertical striations
running along the edge (text-fig. 5d) which were made by the rocking motion of the shell. Some
of the borings contain irregular spherical bodies of dark brown calcite up to 6 mm in diameter
(text-fig. 5a). These may be the calcified gastric contents left in their original location following
decomposition of the soft parts.

Plants. The plants demonstrate a greater preservational diversity than the animals (see Table 1 ).

Coalified plant matter. The cuticles of pyritized fruits and seeds may be preserved as coalified
remnants. In addition, isolated woody fragments occur throughout the clay which are part

ALLISON: TAPHONOMY OF AN EOCENE BIOTA

1083

text-fig. 4. Transverse section of wood-bearing concretion, x 0-8. Note extensive boring by bivalve Teredina :
s = gastropod shell, m = pyrite meniscus, 1 = shell lag, t = burrow, g = geopetal infill with pellets/sediment,

z = polyzonal-calcite lining Teredina boring.

table 1. Summary of mineralogical facies associated with fossilization in the London Clay.

Pyrite

Calcium carbonate

Calcium phosphate

Framboids

a, Isolated

b , Conjoined

Fossils

a , Plants: fruits and wood

a. Original skeletal material

a ,

Arthropods and verte-

b , Shelly material

b , Permineralized wood

brates

c. Vertebrates

b ,

Permineralized seed cases

Internal

a , Vessel infills in calcified

a. Cavity infills with pore

a ,

Mineralized sediment in-

moulds

wood

lining habit in Teredina

fill of cavities in seeds

b. Vesicle infills in bone

c. Cavity infills in gastro-

borings

pods and Teredina bor-
ings. Includes: pyritized
sediment, euhedral pore
lining pyrite, and pyrite
stalactites

Overgrowths

a, Burrow infills

b , Bi-pyramidal forming:

light ‘dustings’ and cauli-
form growths

c. Radiating cauliform

growths

Concretions

a. Sub-spherical to flat and

a. Sub-spherical to ovate

a ,

Sub-spherical to ovate

cauliform

concretions

nodules

Septarian

a. In phosphatic nodules

a, In calcareous nodules

infills

b. In pyritic nodules

1084

PALAEONTOLOGY, VOLUME 31

text-fig. 5. a, dark-brown calcareous concretions, supposed to be gastric contents of Teredina , deposited
in situ through decay of animal, x 4. b, geopetal pelletal infill of boring, x 5. c, bore-lining polyzonal
calcite, x 5. d, tangential longitudinal section of boring showing striations upon wood surface made during

excavation, x 3.

pyritized and part coalified (text-fig. 6a). Unmineralized coalified material is always compacted
and cellular detail is usually obscured.

Pyritization. Pyrite has preserved fine morphological details such as winged seeds inside a fruit
(Collinson 1983) and cellular structure such as the cast and moulds of starch grains within
mangrove hypocotyls (Wilkinson 1983). Robust structures such as twigs and seeds have not been
crushed by sediment overburden but some of the fruits have been compacted. Thus the seed case
of the large palm fruit Nipa burtini is compacted and slightly flattened (PI. 94, figs. 1, 3, 5). Pyrite
is most common as an infilling of cellular cavities but also occurs as a replacement of original
plant material.

The development of pyrite is controlled by the anatomy of the original organic material.
Mineralization preferentially selects the spring or early wood leaving the late or summer wood as
a coalified layer (text-fig. 6a) with infillings of pyrite. Pyrite forms as a response to the activity
of sulphate-reducing bacteria (see section on pyrite paragenesis). It is possible that the large, thin-
walled cells of the spring wood were more susceptible to decay (by sulphate-reducing bacteria)
than the small, thick-walled lignified cells of the late wood. Thus in the case of the wood figured
in text-fig. 6a, the action of microbial sulphate reducers led to the decay of spring wood prior to
the precipitation of pyrite. However, the late wood being more decay resistant, was preserved as
a coalified residue which includes cellular detail. Individual cells in this instance are commonly
infilled with pyrite. Similarly the tough outer cuticle of fruits such as Nipa is carbonaceous whilst
the internal structures are pyritized. It is likely that this too is due to increased decay resistance
within the cuticle.

Kenrick and Edwards (1988) have described a similar distribution of pyritized and coalified
portions of plant anatomy from the Lower Devonian of Wales.

Calcification. Calcification of plant material occurs as a permineralization (as opposed to a
replacement) and is restricted to larger woody fragments which have in some cases formed the
nuclei of large carbonate concretions. Calcification led to the preservation of fluid-bearing vessels
and individual cells (text-fig. 6b, c).

Phosphatization. Phosphatized plant material is rare and restricted to a few isolated specimens of
N. burtini. The phosphate has impregnated the carbonaceous outer cuticle of the fruit and also
occurs as a mineralized sediment infill within the fruit cavity (PI. 94, figs. 2, 4, 6). Nipa from

ALLISON: TAPHONOMY OF AN EOCENE BIOTA

1085

text-fig. 6. a, pyritized wood cut perpendicular to grain and viewed in reflected light, note preservation of
late wood (a) as coalified residue and early wood (b) as pyrite, x 5. b, SEM of calcified wood, x 20. c, thin-
section of calcified wood showing cellular structure and fluid-bearing vessels (v), note geopetal infill of latter
with framboidal pyrite, x 30. d, pyritized gastropod shell (s) coated in overpyrite (o) with internal cavity
filled with pyritized sediment (i), note vertical rods (r) of pyrite thought to represent original shell structure,
viewed in reflected light, x 20. e, SEM of calcareous spire of otherwise pyritized gastropod, x 1200. f, close-
up of e showing fine detail of lamellae structure of shell, x 2000. G, SEM of pyritized gastropod shell (s)
showing shell laminae (I), and pyritized sediment infill (i), x 30. h, thin-section of phosphatized bone (b) with
pyrite infill (p) of vesicles, x 40. i, thin-section of phosphatized crab cuticle including preservation of exo-
cuticle (x) and endo-cuticle (n) x 35.

1086

PALAEONTOLOGY, VOLUME 31

text-fig. 7. Diagram of polished sections of Nipa burtini. a, pyritized crushed specimen (PI. 94, fig. 5). b,

phosphatized/pyritized specimen (PI. 94, fig. 6).

Sheppey displays considerable biological variation. Some forms are sterile and without fruit, others
have been fertilized and have aborted and still others are fertile fruit-bearing forms (M. E.
Collinson, pers. comm.). It is therefore important to identify the nature of each specimen and
compare like with like before interpreting the effects and timing of compaction. The Nipa depicted
in Plate 94, figs. 1, 3, and 5 is pyritized and is slightly flattened with compaction cracks (text-fig.
8) whereas the specimen depicted in Plate 94, figs. 2, 4, and 6 is part phosphatized/part pyritized
and is three-dimensional (text-fig. 7). Phosphatization prevented flattening of the fruits and is
therefore pre-compactional whereas pyritization occurred later after sediment overburden had
crushed the fruits. It is possible that the pyrite occurring with phosphate in the uncompacted fruits
is a replacement of original phosphate.

PRESERVATIONAL MINERALOGY

Most of the early diagenetic authigenic minerals are precipitated as a result of the biodegradation
of organic matter. This is due to changes in Eh, pH, and in the concentration of various anionic
and cationic species within pore waters brought about by bacterial respiration. Bacterial degradation
proceeds through a number of chemical steps of which the best known is aerobic decay. However,
following the depletion of oxygen, bacteria utilize a number of alternative oxidizing agents to
continue active biodegradation and respiration. These reactions are stratified within sediment (text-
fig. 8). Species liberating the greatest free energy yield occur highest in the sequence and only when
these have been exhausted do less energetic reactions (lower in the column) occur (Redfield 1958).

Not all oxidants are present within any given sediment. Sulphate reduction and methanogenesis
dominate in marine environments whilst methanogenesis alone dominates in a freshwater system.

EXPLANATION OF PLATE 94

Figs. I 6. Nipa burtini. I, 3 and 5, apex of crushed pyritized specimen. 1, outline view. 3, lateral view. 5,
polished section. 2, 4 and 6, apex of uncrushed phosphatized/pyritized specimen. 2, outline view. 4, lateral
view. 6, polished section, all x 2-6.

PLATE 94

ALLISON, Nipa burtini

1088

PALAEONTOLOGY, VOLUME 31

Aerobic Zone

Manganese reduction

<D

c

o

N

o

15

o

i—

CD

CD

C

<

Nitrate reduction

Iron reduction

Sulphate reduction

Carbonate reduction
(Methane generation)

text-fig. 8. Occurrence of bacterial reduction zones in sediment (after Redfield 1958). Note sulphate
reduction dominates in a marine system and methanogenesis and nitrate reduction dominates in a

freshwater system (from Allison 1988a).

The sequence of decomposition events will have a radical effect upon pore-water chemistry and
can therefore be used as a model for the paragenesis of several authigenic sedimentary minerals
(text-fig. 8; see Berner 1981 for a review).

Rationale. Concretionary calcite and phosphate precipitate within the pore spaces of the unconsoli-
dated clay. Crystallization of these minerals did not force the detrital minerals apart since delicate
structures such as burrows and faecal pellets are still intact and undeformed. Thus the volume of
nodule-forming mineral is approximately equal to original porosity of the sediment at the time of
mineral precipitation (Raiswell 1971). Sediment compaction in mud rocks is proportional to depth
of burial (Greensmith 1978), hence an assessment of original porosity within early cemented mud
rocks is also an indication of diagenetic timing. Therefore original porosity of the sediment is
approximately equivalent to the acid soluble fraction of the concretions (Raiswell 1971). This
approach is invalid for pyrite and phosphate concretions since the acids capable of digesting these
minerals will also dissolve detrital minerals such as some clays. For this reason calcium and
phosphate levels in the apatite concretions was determined analytically and volume of nodule-
forming mineral estimated. Even this method would be invalid for pyrite concretions since pyrite
may replace detrital grains.

Minor and trace element chemistry of diagenetic minerals is strongly dependent upon pore-
water chemistry at the time of crystal growth. Pore-water chemistry is in turn controlled by
bacterial degradation of organic carbon and burial. Therefore the relative abundances of certain
elements (particularly manganese) within the concretion can be analysed to provide an indication
of diagenetic trends.

ALLISON: TAPHONOMY OF AN EOCENE BIOTA

1089

Sampling. Sampling for original porosity determination and chemical analysis was achieved by
taking a thin slice of rock from the centre of a nodule and cutting it to produce a square sectioned
core. This was then cut into cubes to allow porosity determination and chemical analysis of
different parts of the concretion. Analysis was undertaken by atomic absorption spectroscopy
using a Pye UNICAM PU 9000 instrument with aqueous solutions prepared from the calcareous
and phosphatic samples. Phosphate analysis was achieved by colour density spectrophotometry
using a Phillips SP 38 Spectrophotometer.

Pyrite

Morphology. Pyrite exhibits a variety of morphological types (Hudson 1982). Studies of the
morphologies and occurrences of pyrite in the London Clay allow the different phases of its

text-fig. 9. a, unusual framboid of octahedral pyrite upon surface of calcified/pyritized wood, x 600. b,
pyrite stalactites in gastropod cavity, x 50. c, coarse cubic overpyrite (c) with slender crystals of gypsum (g)
as alteration product, x 80. d, calcified and pyritized wood (w) with prominent gypsum crystals (g), x 20. e,
pyritized wood (w) with Teredina-bonng viewed in reflected light, note infill of boring with original Teredina
calcite (t) and four coloured bands of authigenic calcite (C1-C4); pyritization (p) of authigenic calcite is
limited to the outer (Ci) zone, all four zones of calcite are cut by pyrite-filled fracture (f), x 20. F, pyritized
wood (w) with Teredina boring lined with Teredina calcite (t) and infilled with geopetal pyritized sediment
(s) and cavity-lining pyrite (1), x 20. G, gastropod shell (g, outer margin = o and inner margin = i), note that
pyrite-infilled compaction fracture (f) is contiguous with overpyrite (op) and cavity-lining (p), x 30.

1090 PALAEONTOLOGY, VOLUME 31

formation to be dated in relation to the formation of early diagenetic calcium carbonate and
calcium phosphate.

1. Isolated and conjoined framboids are randomly dispersed in concretions and the host
sediment. Individual framboids range between 10-100 ^ni in diameter (text-fig. 9a). Conjoined
framboids occur as either linear or clustered aggregates: in the former as few as four framboidal
crystallites are joined in a single plane, whereas the latter consists of a large number of framboids
which have coalesced in a sub-spherical ‘clot’. The linear forms are always parallel or sub-parallel
and appear to delineate original sediment lamination. In some cases the ‘clots’ form thin coatings
upon fossils. The framboidal-pyrite replacement of wood shown in text-fig. 4 for example, occurs
as a 1 mm surface layer. This pyrite layer is limited to what would have been a hollow on the
surface of the wood below a shell lag horizon. Such a ‘meniscus’ development of pyrite may be
due to the development of a localized anoxic micro-environment within a topographic hollow.

2. Pyritized fossils are concentrated by wave action as lag deposits on the modern beach. Pyrite
in plant material occurs as octahedral crystals approximately 10-20 ^.m in size (text-fig. 6a). Where
original shell material of molluscs and brachiopods has been pyritized the individual crystallites
adopt a bipyramidal form and vary in size from 4 at the inner margin of the shell to less than
1 /x m at the outer. In some instances where pyrite has pseudomorphed the prismatic layer of the
original shell, no pyrite crystal form is discernible.

3. Pyritized internal moulds occur within most organic cavities:

a. Within calcified wood pyrite, framboids infill the large fluid bearing vessels (text-fig. 6c).

b. Octahedral pyrite with a pore-lining habit occurs within the vesicles of spongy bone in
vertebrae (text-fig. 6h).

c. The internal cavities of gastropods contain geopetal pyritized sediment, and pyrite linings. In
some cases the latter extend downwards from the upper surface of the cavity to form what
Hudson (1982) referred to as pyrite stalactites (text-fig. 9b). The crystallites here adopt the
octahedral habit and are between 40 and 100 ^m across.

d. Within Teredina cavities, pyrite occurs as permineralized sediment and as a bore lining, the
combination of both habits in the same cavity producing a geopetal structure (text-fig. 9f).
In some cases the pyrite has obviously pseudomorphed calcareous linings within the Teredina
borings (text-figs. 9e and 10d), although it also occurs as a discrete coat of crystallites
dusting the surface of some calcified Teredina infills within pyritized wood. The crystals in
this case adopt a octahedral habit and are generally less than 10 ^m in size.

4. Pyrite overgrowths, termed overpyrite by Hudson (1982), occur on the surface of most fossils.
It adopts the following morphologies:

a. Mineralized burrow traces occur on the surface of all fossil types. Octahedral crystallites are
most common in the 40-100 ^.m size range. In most cases the pyrite extends beyond the zone
of burrowing onto the surface of the fossil.

b. A light dusting of bipyramidal euhedral 40 ,um crystallites occurs on the surface of some
eumalacostracan carapaces within phosphatic concretions (text-fig. 12e). The pyrite is limited
to the surface of the fossil and is patchy in extent.

c. Cauliform pyrite growths (Ramdohr 1980) upon the surface of fossils are frequently associated
with pyritized trace fossils and show two forms of crystal habit.

i. Octahedral in the 40-100 size range, identical to the ‘light dusting’ described above
although better developed and forming an uneven cauliform coat.

ii. Radiating crystals producing a very smooth cauliform growth 1-2 cm across covered with
euhedral cubic crystals 1-2 mm across (text-fig. 9c), although a vertical polished section
through the growth reveals a radiating cone in cone crystal structure.

5. Pyrite concretions occur as flattened to sub-spherical concretions 1-10 cm across which in
places coalesce to form an aggregated mass. In some specimens relict sediment is clearly visible

ALLISON: TAPHONOMY OF AN EOCENE BIOTA

1091

text-fig. 10. a, thin-section of syntaxial calcite overgrowths on Teredina fragment, x 24. b, CL of a showing
numbered luminescent calcite zones on shell fragment (c), x 24. c, CL fractured Teredina calcite (t) that has
been infilled with authigenic cavity-lining calcite (c), x 36. d, thin section of bore-lining Teredina calcite (t)
in wood (w); pore-lining calcite (c) syntaxial with Teredina calcite has been pseudomorphed by pyrite (p),

x 40.

and step-like striations along the horizontal axis of the nodule delineate original bedding. In hand
specimens of this type the pyrite has a granular appearance. Alternatively, the pyrite adopts a
cauliform growth habit. Horizontal sectioning of the concretions reveal a septarian fracture system
of radial and concentric cracks infilled with euhedral pyrite. Septarian cracks formed as tensile
fractures during burial and compaction of the host shale (Astin 1987). Pyrite septarian infills are
limited to concretions of pyrite (described above) and calcium phosphate, and have not been
recorded in carbonate concretions. Within the calcium phosphate concretions pyritization usually
extends beyond the septarian fracture to replace phosphatized sediment.

Paragenesis. Of all the early diagenetic sedimentary minerals, pyrite has attracted the greatest
interest through either geochemical studies of modern anaerobic systems or mineralogical studies
of the geological occurrence (Berner 1970, 1971, 1984; Curtis 1980). Under anaerobic conditions,
iron hydroxide within sediment pore waters is reduced during bacterial respiration to produce iron
ions. Following the depletion of iron hydroxide, bacterial respiration utilizes the sulphate ion
present in sea water forming hydrogen sulphide as a by-product. The subsequent reaction of these
two reactive species leads to the formation of an initial iron monosulphide which upon further
reaction with elemental sulphur forms pyrite.

The overall pyrite content of the London Clay is low. Burrowing organisms were probably able

1092

PALAEONTOLOGY, VOLUME 31

to supply sufficient oxygen to the decomposition system to facilitate prolonged aerobic decay. Thus
low concentrations of iron and sulphate ions were produced by anaerobic microbial reduction
and pyrite formation was hindered. Pyrite was precipitated at several stages in the diagenetic
history of the sequence (Table 1 and text-fig. 14). This sequence resembles that in pyrite-bearing
shales from the Jurassic of England and Germany (Hudson 1982).

Pyrite framboids within concretions are morphologically identical to those dispersed in the
sediment and were therefore the first pyrite phase to form. Framboid formation was concentrated
at sites of active anaerobic decay such as decomposing plants and animals in response to bacterial
decomposition and in some cases formed pyrite concretions and pyritized sediment infills of organic
cavities.

Some gastropods have been crushed by sediment overburden prior to the formation of cavity
lining pyrite and pyritized sediment infills. Others show no compaction cracks. This may be due
to structural variation in the shells and/or a timing variation in pyrite formation. Overpyrite also
formed at this time since it is contiguous with compaction fractures and internal pore-lining pyrite.
Pyritization of plant remains is also likely to have occurred at this time since some of the pyritized
plant remains have been partially flattened.

It is not possible to date the formation of some of the pyrite morphotypes, such as pyrite
concretions relative to other mineral phases, because they do not occur in conjunction with other
minerals or varieties of pyrite. However, concretion shape is strongly controlled by sediment
porosity. If permeability is equal in both the horizontal and vertical planes then the concretions
will tend to be spherical. After a degree of compaction sediment permeability is greatest in the
horizontal plane and the concretions that form tend to be discoidal. Thus it is likely that the
flattened pyrite concretions formed later than the more spherical forms.

Alteration of pyrite by calcium rich pore waters during weathering led to the formation of
gypsum crystals (text-fig. 9c, d).

Cal cite

Morphology. Calcium carbonate within the London Clay occurs as original skeletal fragments and
as mud-rock concretions. The latter form prominent bands of spherical to flattened ovate
concretions along the foreshore and cliff exposures at Sheppey and are commonly in the 20-80 cm
size range although some reach over 1 m in diameter. Fossils within the concretions include wood
and occasional shelly debris and appear to be most common within one band along the foreshore.
Elsewhere the majority of the concretions are unfossiliferous, although frequently intensively
bioturbated, with burrows being most prominent on the outside of the concretion. Septaria are
most common within non-fossiliferous concretions and are frequently filled with a green-yellow
calcite with occasional selenite. The concretions are also associated with depositional events, e.g.
the lowest nodule bank in the cliffs of Warden Point occurs in a dark blue-black band of clay
which flames into the overlying bed.

Much of the wood within concretions has been intensively bored by Teredina and the borings
have been infilled with geopetal sediment and calcite (text-fig. 5b, c). At least four colour zones
are visible in hand specimen in the calcite lining the borings (text-fig. 5c). Differences in chemical
composition of the calcite, particularly in manganese levels, render it amenable to cathodo-
luminescence microscopy (CL). CL clearly shows six dullish red zones (text-fig. 10) coating shell
fragments and Teredina- calcite lining the borings. The bore-lining calcite is brighter, indicating a
higher manganese/iron ratio, whereas the dull red pore-lining calcite is indicative of a lower
manganese/iron ratio. The initial phase of pore-lining calcite is frequently syntaxial with skeletal
fragments and adopts a radial habit, whereas later zones are typically sparry.

Chemistry. Both chemical analysis and original porosity determinations (Table 2) show a series of
clear trends between the rim and core of the concretion.

The decrease in porosity away from the centre of the concretion reflects the timing of mineral
formation. At the centre of the nodule carbonates were precipitated prior to appreciable sediment

ALLISON: TAPHONOMY OF AN EOCENE BIOTA

1093

table 2. Chemical analysis of pyrite, phosphate, and calcite concretions.

Phosphate concretion Calcareous concretion

Pi

Core

P2

P3

P+

• Rim

Ci

c2

c3

C4

- Rim

ooie

Ca (%)

40-8

34-2

39-7

36-7

35-7

33-3

30-2

31-4

Fe (%)

3-72

519

4-68

4-53

3-74

3-93

3-65

4.49

Mn (ppm)

888

734

656

540

7007

6880

5720

4063

Mg (%)

0-50

0-48

0-49

0-49

1 40

1-40

1-42

1 54

A1 (%)

1 -66

1-64

1 59

1-58

2-58

2-79

2-91

4-36

P(%)

35-8

28-2

34-3

25-8

2-34

2-54

1 97

2-21

Si (%)

9-8

113

12 4

23-8

14-5

1 5-7

14 9

27-6

Mn/Ca

0 0021

00021

00016

0-014

0 19

0-20

0 18

0 12

Internal porosity
(%)

82

78

76

54

72-1

73-7

68-8

54-7

(estimated from Si and A1 content) (acid soluble fraction)

compaction. Later phases of mineral growth occurred after the onset of compaction and consequent
reduction of sediment porosity. The decreasing calcium and increasing aluminium (from detrital
clay minerals) levels towards the rim of the nodule are a function of this control on growth. The
depletion of manganese away from the centre of the nodule is also partly a function of
original porosity, since the manganese occurs as a carbonate with the calcite. However, the
manganese/calcium ratio at the rim of the concretion is 35% lower than at the core.

Par agenesis. The high internal porosities (between 80-90%) recorded within calcite concretions
indicate an early diagenetic origin. Porosities of this value within mud rocks only occur within the
top 5 m of sediment (Greensmith 1978). Such an early diagenetic origin has been attributed by
Berner (1968) and Raiswell (1971, 1976) to the bacterial degradation of organic matter.

Berner (1968) showed experimentally that calcium may be concentrated as calcium stearate (a
soap) within the alkaline decay aureole of proteinaceous animals. Stearate is the stable calcium
salt in this case and Berner suggested that calcite may form following its depletion. The formation
of calcite would be promoted by a high pH microenvironment generated by the production of
ammonia from putrefying proteins. In the London Clay soft parts were scavenged and decomposed
prior to mineralization. An alternative process must therefore have been responsible for the
formation of the calcareous concretions.

In a marine enironment, sediment pore waters achieve a high level of calcium saturation. The
precipitation of a calcareous mineral phase is therefore controlled by anionic concentration
gradients. Bicarbonate ions produced during anaerobic decay may react with the calcium ion to
form calcite (Raiswell 1971, 1976). Bicarbonate ion concentration may be centred around discrete
pockets of decomposing organic matter (e.g. megafossils) or within whole beds of organic rich
sediment.

The occurrence of discrete layers of concretions and the dearth of contained fossils suggests that
the bicarbonate ions originated from the decomposition of disseminated organic carbon in the
London Clay. Further, the occurrence of concretion bearing horizons within flamed beds suggests
a depositional control upon nodule formation. Rapid burial of an organic-rich layer of sediment
would lead to anoxia and the production of large amounts of the bicarbonate ion, thereby
promoting carbonate formation. The irregular development of discontinuous bands of concretions
in the London Clay may therefore represent episodic burial events.

The introduction of manganese into the calcite lattice is due to microbial respiration within
sediment. Manganese reduction is the highest reduction zone in the sediment pile and occurs in

1094

PALAEONTOLOGY, VOLUME 31

the poorly oxygenated sediments below the sediment-water interface. Manganese ions liberated
by microbial respiration become concentrated within sediment pore waters and may, upon reaction
with bacterially produced anionic species, e.g. the bicarbonate ion, precipitate as a mineral phase.
In the London Clay manganous mineral species were incorporated into other mineral precipitates
such as calcite and phosphate. About 075% of manganese occurs within the carbonate fraction
of the concretions. This is because initial stages of concretion growth occurred at or near the zone
of manganese reduction which provided an abundant source of manganese ions for inclusion into
the calcite lattice. The reduced levels of manganese towards the rim of the concretion show that
this part of the nodule formed deeper in the sediment pile when ionic manganese within sediment
pore waters was depleted.

Thin-section CL of wood-bearing concretions shows that the wood luminesces at the same
intensity as the surrounding sediment. Since CL is sensitive to changes in manganese and iron
chemistry this shows that levels of these elements within the wood and surrounding concretion are
the same. Calcification of wood within the London Clay is therefore an early event and synchronous
with the early stages of concretion growth.

The development of pore-lining calcite in some Teredina borings occurred after sediment infill
and either during or after the development of calcareous concretions. Septarian cracks formed
during burial at a depth of up to 50 m (Astin 1987).

Calcium phosphate

Morphology . King (1981) documented bands of phosphatic concretions at Sheppey although they
are more commonly randomly dispersed. XRD shows the phosphate to be francolite, a calcium
fluor-apatite, which occurs in the form of 1-2 p m sized crystal aggregates. The concretions are
ovate and up to 60 cm in size. They are a light buff colour in contrast to the dark brown-black
fossil material which they occasionally contain. Organic remains in the concretions includes faecal
pellets, burrows and fragments of arthropods and vertebrates (text-fig. 11a, c). The fossils are
composed of francolite virtually identical to that forming the surrounding concretion, apart from
a slight deviation in minor element chemistry. The colour difference between fossils and the
enclosing concretion is probably a function of crystal size.

Most of the fossils encountered in phosphatic concretions had an original phosphate component.

text-fig. 1 1. Thin section of phosphate concretions, a, burrows with sprcite structure, x 18. b, faecal pellets
surrounded by pyrite, x 25. c, faecal pellets with pyrite veining, x 20.

ALLISON: TAPHONOMY OF AN EOCENE BIOTA

1095

However, in the case of the arthropod cuticle this was probably quite low since phosphate content
of living arthropod cuticle is in the region of 1-5% dry weight (Allison 1988u). Non-concretionary
phosphate is limited to isolated fish teeth which are most commonly encountered in the coarse
sands near the base of the cliff and may have originated from eroded concretions.

Chemistry. Original porosities were estimated by chemical analysis. Manganese levels and the
calcium/manganese ratio decrease towards the rim of the concretion (Table 2). Internal porosities
are slightly higher than those of carbonate concretions although the difference may be within
experimental error.

Paragenesis. Phosphate mineralization occurred after the formation of framboidal pyrite but before
pyritization of Nipa and some shelly fossils, and before the formation of flattened concretions.
This is evidenced by the presence of high internal porosities, by pyrite framboids in the concretions,
and by three-dimensional preservation of fruits and gastropods in phosphate concretions but not
in pyrite concretions.

In normal marine pore waters, concentrations of the bicarbonate ion exceed those of phosphate
(Gulbrandsen 1969) and the most stable calcium mineral phase is normally calcite. Pore-water
phosphate levels must therefore be considerably enriched to allow the precipitation of phosphate
minerals. Gulbrandsen (1969) suggested that such enrichment may result from the bacterial

text-fig. 12. SEMs of crab-bearing nodule. A, fine mass of phosphate crystallites with organic pockets, x 80.
b, smooth organic pocket with small phosphate crystallites, x 8000. c, ‘bacterial’? microsphere, x 18 000. d,
aggregated microspheres, x 10 000. e, octahedral pyrite upon surface of crab, x 7000. F, phosphate crystallites,
x 3800. G, ‘bacterial’? microsphere with possible cell wall, x 38 000.

1096

PALAEONTOLOGY, VOLUME 31

degradation of organic matter such as marine plankton. Further, Lucas and Prevot (1984) have
shown experimentally that phosphate can be liberated by the microbial breakdown of organic
compounds such as adenosine-tri-phosphate (ATP), used by most organisms for energy transfer.
Benmore et al. (1983) believe that phosphate liberated in this way would become adsorbed to ferric
hydroxides in aerobic sediment. Under anaerobic conditions these iron compounds would be
reduced and phosphate would be liberated into solution. Thus phosphate concentrations at the
anoxic-oxic boundary may be sufficiently increased to allow localized phosphate precipitation.

Phosphate precipitation may also be induced by bacteria, e.g. plaque bacteria precipitate apatite
crystals inside the cell when phosphate concentrations are high. Ennever et al. (1981 ) showed that
several strains of bacteria can phosphatize in this way when cultured in the correct medium.
Further, phosphate crystalites from London Clay concretions are generally rounded bacteria-sized
crystalline aggregates about 1-2 ^m in diameter (text-fig. 12a, d, g). In some cases the crystallites
are enclosed (text-fig. 12g) by what appears to be the cell wall of the bacterium. It is therefore a
possibility that these microspherical crystallites may be a bacterial precipitate formed at the anoxic-
oxic interface where phosphate concentrations are high. However, such a microspherical form
could also be a product of physico-chemical conditions, e.g. rapid precipitation around a multitude
of nuclei from a saturated solution could produce a similar crystal form.

Manganese concentrations at the outer edge of the concretions are much lower than at the centre
of the nodule (Table 2). This is because pore-water manganese levels were lower during this late
stage of growth. This depletion is considered to be due to bacterial manganese reduction, as
documented above (p. 1094), but it is not meaningful to compare manganese levels in phosphate
and calcite concretions, since each mineral has a different affinity for the element.

Many fossils in the phosphate concretions have been replaced by secondary francolite. It is not
possible to date this mineralization relative to the formation of the concretions. Lucas and Prevot
(1984) showed experimentally that phosphatization of calcareous skeletal material may occur
rapidly if pore-water phosphate levels are high enough.

In the London Clay, phosphatization of fossils appears to be restricted to organic remains with
an original phosphate content (Balson 1980) e.g. arthropods, vertebrates, and faecal pellets.

SUMMARY

The Eocene London Clay cropping out on the Isle of Sheppey contains a suite of early diagenetic
calcareous, phosphatic, and pyrite concretionary growths which are often fossiliferous (see text-
figs. 13 and 14 for diagenetic and taphonomic summary). Framboidal pyrite formed first in small
anaerobic pockets in an otherwise aerobic and bioturbated sediment. The bacterial decomposition
of organic matter in this aerobic zone liberated phosphate into pore-water solutions which became
adsorbed to ferric hydroxides in the sediment (Benmore et al. 1983). The reduction of iron at the
anoxic-oxic interface liberated phosphate into solution which was precipitated either abiotically or
by sediment bacteria to form small concretionary bodies around faecal pellets, burrows, arthropods,
and vertebrates. Phosphatization was specific to organisms with an original phosphatic content.

Calcareous concretions formed as a result of rapid burial of organic rich layers. Subsequent
anaerobic decay of the organic matter in these layers promoted an increase in pore-water
concentrations of the bicarbonate ion which led to the precipitation of calcium carbonate.
Precipitation occurred around sediment inhomogeneities and megafossils which may have func-
tioned as ‘seed crystals’.

Framboidal pyrite formed internal moulds in some organic cavities and laminated cauliform
concretions. Pyritic replacement of organic remains, such as shelly fossils and plant material,
occurred later in the diagenetic sequence after sediment compaction.

Concretionary phosphate was clearly the earliest preservational mineral to form in the London
Clay. Mineralization is the principal means of halting the information loss that occurs during
decay (Allison 1988a, b). Therefore organisms preserved during the first stages of the diagenetic

ALLISON: TAPHONOMY OF AN EOCENE BIOTA

1097

VERTEBRATES ARTHROPODS PLANTS SHELLY ANIMALS

text-fig. 13. Summary of biostratinomic sequence.

sequence exhibit a higher level of preservation than those mineralized later. Since phosphatization
can be restricted to taxonomic groups with an original phosphate content a diagenetic taphonomic
bias is seen to function.

From the recognition of this preservational bias and a knowledge of the geochemical switches
governing mineral precipitation it is possible to characterize the conditions necessary for exceptional
preservation and apply them to the fossil record. With further work it will also be possible to
predict the level of preservation which can be expected in a given sedimentary regime.

Since early diagenetic minerals are strongly indicative of depositional environment they can be
considered as mineralogical taphofacies (Brett and Baird 1986). Thus the phosphatic taphofacies
is commonly associated with the highest levels of fossil preservation (Allison 19886)- For example,
the highest level of fossil preservation is that of three-dimensional muscle-fibres. Such fossils are
more commonly encountered preserved in phosphate than any other mineral form (e.g. Cretaceous
fish from Brazil, Martill 1988; ichthyosaurs from the Jurassic of the English Midlands, Martill
1987; and the mantle of squid from the Jurassic of south-west England, Donovan 1983; Allison
1988c).

Acknowledgements. Dr D. E. G. Briggs provided guidance for this work and together with Dr J. D. Hudson
critically reviewed an earlier draft of the manuscript. Drs V. P. Wright, E. N. K. Clarkson, and D. M. Martill
also commented on various aspects of this work. I am especially appreciative of the guidance provided by
A. J. Kemp of Bristol on geochemical analysis. Joyce Smith at Friday Harbor kindly entered the text into a
word processor and Dr M. E. Collinson discussed aspects of the biology of Nipa hurt ini and kindly donated

1098

PALAEONTOLOGY, VOLUME 31

text-fig. 14. Summary of diagenetic sequence.

ALLISON: TAPHONOMY OF AN EOCENE BIOTA

1099

several specimens. I am also grateful to R. Gutterie of Wells Kauft Kramer for taking the X-ray radio-
graphs of the phosphate concretions. This work was carried out during the tenure of a NERC award at the
University of Bristol and completed during the tenure of a Friday Harbor Post-Doctoral Fellowship at the
University of Washington, USA.

REFERENCES

allison, p. a. 1988a. The role of anoxia in the decay and mineralization of proteinaceous macrofossils.
Paleobiology , 14, 139 154.

— 1 988Z?. Konservat-Lagerstatten: cause and classification. Ibid. In press.

— 1988c. Phosphatized soft-bodied squid from the Jurassic Oxford Clay. Lethaia , 21, 293-301.

astin, t. r. 1987. Septarian crack formation in carbonate concretions from shales and mudstones. Clay
Miner. Bull. 21, 617 631.

balson, p. s. 1980. The origin and evolution of Tertiary phosphorites from eastern England. J. geol. Soc.
Loud. 137, 723 729.

benmore, r. a., coleman, m. l. and mcarthur, j. m. 1983. Origin of sedimentary francolite from its sulphur
and carbon isotope composition. Nature , Loncl. 302, 516-518.
berner, r. a. 1968. Calcium carbonate concretions formed by the decomposition of organic matter. Science ,
NY, 159, 195-197.

1970. Sedimentary pyrite formation. Am. J. Sci. 268, 1 23.

1971. Principles of chemical sediment ology, 240 pp. McGraw Hill, New York.

— 1981. Authigenic mineral formation resulting from organic matter decomposition in modern sediments.
Fortschr. Miner. 59, 117 135.

— 1984. Sedimentary pyrite, an update. Geochim. Cosmochim. Acta , 48, 605-615.

brett, c. e. and baird, G. c. 1986. Comparative taphonomy: a guide to paleoenvironmental interpretation
based on fossil preservation. Palaios. 1, 207-227.

briggs, d. e. g. and williams, s. h. 1981. The restoration of flattened fossils. Lethaia , 14, 157-164.
casier, e. 1966. Fame Icthyologique du London Clay , 496 pp. British Museum (Natural History), London.
chandler, m. e. j. 1961. The Lower Tertiary Floras of Southern England. I Palaeocene Floras. London Clay
Flora ( supplement ). Text and Atlas. 354 pp. British Museum (Natural History), London.

— 1964. The Lower Tertiary Floras of Southern England IV. A survey of findings in the light of recent
botanical observations. 151 pp. British Museum (Natural History), London.

COLLINSON, M. E. 1983. Fossil plants of the London Clay: Palaeontological Association field guide to fossils 7,
121 pp.. Palaeontological Association, London.

conway morris, s. 1979. Middle Cambian polycheates from the Burgess Shale of British Columbia. Phil.
Trans. R. Soc. B285, 227-274.

curry, d. 1965. The English Palaeogene pteropods. Proc. malac. Soc. Loud. 36, 357-371.
curtis, c. d. 1980. Diagenetic alteration in black shales. J. geol. Soc. Land. 137, 189-194.
davis, A. G. 1936. The London Clay of Sheppey and the location of its fossils. Proc. geol. Ass. 47, 328 345.

— and elliot, G. f. 1957. The palaeogeography of the London Clay sea. Ibid. 68, 255-277.
donovan, D. T. 1983. Mastigophora Owen 1856: a little-known genus of Jurassic coleoids. Neues Jb. Geol.

Paldont. Abh. 165, 484 495.

ennever, j., streckfuss, j. l. and Goldschmidt, m. c. 1981. Calcifiability comparison among selected micro-
organisms. J. Dent. Res. 60, 1793-1796.

gilkes, r. j. 1967. Clay Mineral provinces in the Tertiary sediments of the Hampshire Basin. Clay Minerals ,
7, 351-361.

greensmith, j. t. 1978. Petrology of the sedimentary rocks. 241 pp. Allen & Unwin, London.
gulbrandsen, r. a. 1969. Physical and chemical factors in the formation of marine apatite. Econ. Geol. 64,
365-382.

harrison, c. J. o. and walker, c. a. 1977. Birds of the British Lower Eocene. Tert. Res. Spec. Pap. 3, 1-96.
hooker, j. j., insole, a. N., moody, r. j. j., walkden, c. A. and ward, d. a. 1980. The distribution of
cartilagenous fish, turtles, birds and mammals in the British Palaeogene. Tert. Res. 3, 1 45.

Hudson, j. d. 1982. Pyrite in the ammonite bearing clays from the Jurrasic of England and Germany.
Se diment ology , 29, 639 667.

keen, m. c. 1978. The Tertiary Palaeogene. In bate, r. and robinson, e. (eds. ). A Stratigraphical Atlas of the
British Ostracoda , 385-450. Seel House Press, London.

1100

PALAEONTOLOGY, VOLUME 31

kenrick, p. and Edwards, d. 1988. The anatomy of Lower Devonian Gosslingia breconensis Heard based
upon pyritized axes, with some comments on the permineralization process. Bot. J. Linn. Soc. 97,
95-123.

king, c. 1981. The stratigraphy of the London Clay and associated deposits. Tert. Res. Spec. Pap. 6, 11 58.
and king, a. d. 1976. A London Clay section at Waterworks Corner, Woodford, Essex. Tert. Res. 1,
21-24.

knox, r. w. o’b. and harland, r. 1979. Stratigraphical relationships of the Early Palaeogene Ash Series of
N.W. Europe. J. geol. Soc. Lond. 136, 463-470.

lucas, j. and prevot, l. 1984. Apatite synthesis by bacterial activity from phosphatic organic matter and
several calcium carbonates in natural fresh water and seawater. Chem. Geol. 42, 101-118. [In French.]
martill, d. m. 1987. A taphonomic and diagenetic case study of a partially articulated ichthyosaur.
Palaeontology , 30, 543-555.

— 1988. The preservation of fishes in concretions from the Santana Formation (Cretaceous) of Brazil.
Ibid. 31, 1-18.

Murray, j. w. and wright, c. a. 1974. Palaeogene foraminiferida and palaeoecology, Hampshire and Paris
Basins and the English Channel. Spec. Pap. Palaeont. 14, 171 pp.
quayle, w. j. and collins, j. s. 1981. New Eocene crabs from the Hampshire Basin. Palaeontology , 24, 733-
758.

ramdohr, p. 1980. The ore minerals and their intergrowths (2nd edn.), 1202 pp. Pergamon, Oxford.
raiswell, r. 1971. The growth of Cambrian and Liassic concretions. Sedimentology, 17, 147-171.

— 1976. The microbiological formation of carbonate concretions in the Upper Lias of N.E. England.
Chem. Geol. 18, 227-244.

redfield, a. c. 1958. The biological control of chemical factors in the environment. Am. Scient. 46, 206-226.
rowell, a. j. and rundle, a. j. 1967. Lophophore of the Eocene Brachiopod Terebratulina wardenensis,
Elliot. Paleont. Contr. Univ. Kans. Paper 15, 18 pp.

rundle, a. j. and cooper, j. 1970. Occurrence of a fossil insect larva from the London Clay of Herne Bay,
Kent. Proc. geol. zlss. 82, 239-296.

seilacher, a. 1970. Begriff and Bedeutung der Fossil-Lagerstatten. Neues Jb. Geol. Palaont. Abh. 1970,
34-39.

reif, w. E. and westphal, f. 1985. Sedimentological, ecological and temporal patterns of fossil
Lagerstatten. Phil. Trans. R. Soc. Lond. B311, 5-23.

venables, e. m. and taylor, h. e. 1963. An insect fauna of the London Clay. Proc. geol. Ass. 73, 273-279.
ward, d. j. 1979. Additions to the fish fauna of the English Palaeogene. I. Two specimens of Aliopas (thresher
shark) from the English Eocene. Tert. Res. 2, 23-28.

wills, L. j. 1951. A palaeogeographical atlas of the British Isles and adjacent parts of Europe , 1 14 pp. Blackie,
London.

wilkinson, h. p. 1983. Starch grain casts and moulds in Eocene (Tertiary) fossil mangrove hypocotyls. Ann.
Bot. 51, 39-45.

zangerl, r. 1971. On the geologic significance of perfectly preserved fossils. Proc. First North Am. Paleont.
Conv. 1, 1207-1222.

and richardson, E. s. 1963. The paleoecological history of two Pennsylvanian black shales. Fieldiana ,
Geol. Mem. 4, 1-352.

PETER A. ALLISON
Friday Harbor Laboratories
University of Washington
620 University Road

Typescript received 24 April 1987 Friday Harbor

Revised typescript received 3 May 1988 Washington 98250, USA

THE LLANDOVERY ENTELETACEAN
BRACHIOPODS OF THE CENTRAL OSLO
REGION, NORWAY

by B. GUDVEIG BAARLI

Abstract. Details of the precise change in brachiopod faunas from the Ordovician to the Silurian systems
are elusive. A rich shelly fauna spans the Ordovician/Silurian boundary in the Oslo Region. The brachiopods
are the most prominent group represented in the lower Silurian Solvik Formation and among these the
enteletaceans are most abundant. Several of these earliest occurring enteletaceans show a Bohemian affinity.
Many of the genera described mark their first or last world-wide occurrence in this region. One genus and
nine new species are described out of a total of twenty enteletacean species; the remainder are reviewed and
redescribed as necessary. The new genus is Kampella and the new species are: Resserella matutina , Mendacella
bleikeriensis , Marklandella markesi , Kampella gut tula, ‘l Paw or this inopinatus , IDiorthelasma semotum, Salopina
pumila, Chrustenopora askeriensis , and Jezercia rongi. One family, the Chrustenoporidae, is elevated from
subfamily rank and transferred from the Orthacea to the Enteletacea.

Taxonomic treatment of Llandovery (Lower Silurian) faunas in the Oslo Region was first
undertaken by the author as a basis for local palaeoecological studies. Subsequently, it became
clear that these faunas shed light on the timing and character of Late Ordovician extinctions and
the ensuing response by the surviving Silurian fauna. The work also gave more precise ranges for
many genera. The largest group, the Enteletacea, with twenty out of about 1 10 brachiopod species,
is treated in this paper. A detailed description is given of nine new species. The remainder have
been reviewed and new information added where available. The material described is housed in
the type and auxiliary collections of the Paleontologisk Museum, University of Oslo.

Previous taxonomic work in the Llandovery of the Oslo Region is sparse. Thomsen and Baarli
(1982) and Cocks and Baarli (1982) gave an overview of earlier work and a preliminary list of
species found in the Llandovery of the central Oslo Region. The only enteletacean described from
local sequences before 1982 is Dicoelosia osloensis Wright, 1968a.

More general work on the Llandovery faunas of the Oslo and Asker districts includes Baarli
(1987), which provides a palaeoecological treatment of associations from the systemic boundary
up to mid-Aeronian strata. Baarli and Harper (1986) reviewed the Rhuddanian fauna in
consideration of the extinction event near the systemic boundary. Also of interest are studies of
the uppermost Ordovician fauna of the Oslo-Asker districts, which were treated both ecologically
(Brenchley and Cocks 1982) and taxonomically (Cocks 1982). There are, however, very few taxa
in common between the Upper Ordovician and Lower Silurian.

STRATIGRAPHY AND SAMPLING

The lithostratigraphy, biostratigraphy, and sedimentology of the Llandovery Series in the Oslo Region has
been described by Worsley et al. (1983), Baarli (1985), and Baarli and Johnson (in press). The sediments
were mixed siliciclastic and carbonates in the lowermost Solvik Formation, carbonates in the overlying
Rytteraker Formation, and shales with minor carbonate nodules in the uppermost Vik Formation (text-
fig. 1). The age of the sequence is also shown in text-fig. I. The Myren and Spirodden members of the Solvik
Formation in the Asker District and most of the Myren Member of the Solvik Formation in the Oslo District
are Rhuddanian of age. The rest of the Solvik Formation and parts of the Rytteraker Formation in both

(Palaeontology, Vol. 31, Part 4, 1988, pp. 1101-1129, pis. 95-99.|

© The Palaeontological Association

1102

PALAEONTOLOGY, VOLUME 31

text-fig. 1. Stratigraphic sections from the Asker and
Oslo districts showing variation in silt/sandstone, shale,
and limestone contents per metre section, lithostrati-
graphic and chronostratigraphic correlation.

the Asker and the Oslo districts are Aeronian in age (text-fig. 1). Upper parts of the Rytteraker Formation
and the Vik Formation are Telychian in age.

Most of the fossil material was retrieved from bulk samples collected at 5 nr intervals through the lower
Solvik Formation in the central Oslo Region. This material was treated with hydrochloric acid and yielded
good moulds. The other formations were spot-sampled. The Rytteraker Formation is much poorer in
enteletaceans. Those present are under-represented due to the difficulty of retrieval from very hard,
metamorphosed limestone. Although collecting is easier in the Vik Formation, they are not common in that
unit either. The very base of the Vik Formation is an exception, where several thin shale horizons contain
extremely rich faunas from mixed communities. All sampled localities mentioned in the text are indicated on
text-fig. 2.

PALAEOECOLOGY OF THE ENTELETACEANS

An analysis of seven faunal associations from the Ordovician/Silurian boundary layers to mid-
Aeronian strata (Baarli 1987) showed that enteletacean brachiopods were not an ecologically
homogenous group. One of the most common species, Isorthis prima, occurs in all associations
present in the Oslo-Asker districts throughout its long range, from layers at the boundary to a
mid-Aeronian position. It was the dominant taxon in two associations, and common in all the
others except one. In terms of the classical Silurian communities described by Ziegler et al. (1968),
this enteletacean was found in the Clorinda through Stricklandia communities and also the
Cryptothyrella community. Thus it must have been very eurytopic. Another common element
occurring in several associations was Dicoelosia osloensis. This species, however, showed very clear
preferences for the Clorinda related associations with rare membership in the Stricklandia related
associations. Mendacella bleikeriensis is common in the Cryptothyrella community, but sometimes

BAARLI: LLANDOVERY ENTELETACEAN BRACHIOPODS

1103

text-fig. 2. Map of the central Oslo Region
with sampled sections marked. 1, Avlos; 2,
Bleikerveien; 3, Chr. Skredsvik vei; 4, Gjettum;
5, Jongsasveien; 6, Kampebraten; 7, Kon-
glungo; 8, Leangbukta; 9, Malnroykalven; 10,
Malmoya; 11, Ostoya; 12, Skytterveien; 13,
Solhaugveien; 14, Spirodden; and 15, Vettre
Brygge.

occurs in most of the other associations. IDiorthelasma semotum appears rarely in three different
associations, from a Dicoelosia through a Stricklandia to a Cryptothyrella related association.
Other locally common enteletaceans (e.g. Dalmanella cf. pectinoides, Marklandella markesi sp.
nov., and Resserella matutina sp. nov.) seem to be restricted separately to different communities,
and hence are considered stenotopic.

The enteletaceans found near the systemic boundary will be treated in detail below. They are
minor components of their associations, however, and most disappear gradually through the first
40 m of the section, without a change in the dominant associated taxa or accompanying lithological
changes.

THE FAUNA AT THE ORDOVICIAN-SILURIAN BOUNDARY

The fauna spanning the boundary beds between the Ordovician and Silurian systems is not as well
known as that in the immediately underlying or overlying beds. This is due mostly to the world-
wide eustatic fall in sea-level at the end of the Ordovician and the following rapid rise in sea-level
which continued into the earliest Silurian. In many places, this sea-level change caused an
unconformity above shallow epicontinental strata marking the systemic boundary. The problem
is compounded by the widespread occurrence of early Silurian graptolitic shales devoid of shelly
fossils.

The sediments of the Asker District in Norway were deposited in a depression on an epicontinental
shelf where tectonic uplift or tilting partly counteracted the rapid rise in sea-level during early
Silurian time (Baarli, in press). Thus, there is here a nearly continuous section of shelf sediments
from the Upper Ordovician to the Lower Silurian, with a rich, shelly fauna present on both sides
of the systemic boundary. A break is present near the top of the Ordovician, but it seems to be
of limited duration in the Asker District. The base of the Silurian, defined at the base of the
Akidograptus acuminatus Zone, probably occurs in the first 10 m above this diastem in the Asker
District. There is therefore a continuous record across the boundary.

Levenea sp., IPaurorthis inopinatus sp. nov., Epitomyonia sp., Drabovia sp., Chrustenopora
askeriensis sp. nov., Jezercia rongi sp. nov., and Ravozetina cf. honorata are part of a fauna only
found near the systemic boundary and studied by Baarli and Harper (1986). These elements are
interpreted as Ordovician relicts belatedly suffering extinction (Baarli and Harper 1986). Baarli
and Harper (1986) made only a preliminary taxonomic treatment of the fauna. Enteletaceans and
orthaceans make up the main part of the relict taxa. Redesignations of the enteletacean taxa herein
revealed that some of them (e.g. the genera Drabovia , Jezercia , and Chrustenopora , and the species
R. honorata) are taxa known mainly or only from Bohemia. The three last, together with Epitomyonia ,
are found in the uppermost parts of the deep-water Kraluv Dvur Formation of uppermost

1104

PALAEONTOLOGY, VOLUME 31

Rawtheyan age in Bohemia (Havlicek 1982; Havlicek and Mergl 1982). This supports the
assumption of Baarli and Harper (1986) that these were elements of an immigrating deep-water
fauna. It also indicates that the origin of at least many of them was Bohemian. A northward
migration, because of cooling during glaciation (Spjeldnaes 1961), is possible. Deep-water faunas,
however, are often relatively unaffected by climate and temperature (Cocks and Rong 1988) and
cooling is probably therefore not the main reason for their migration. Such migration was
stimulated instead by connections via a nearby sea-way and the development of deep-water
conditions in the Oslo Region. Final extinction is interpreted as a product of competition with
newly developed and more eurytopic Silurian species, like I. prima (Baarli and Harper 1986).

SYSTEMATIC PALAEONTOLOGY

Repository abbreviations. GSC, Geological Survey of Canada; GSM, British Geological Survey; PMO,
Paleontologisk Museum i Oslo; UM, Museum of Paleontological Institute, University of Uppsala; USNM,
United States National Museum.

Order orthida Schuchert and Cooper, 1932
Superfamily enteletacea Waagen, 1884
Family dalmanellidae Schuchert, 1913
Subfamily dalmanellinae Schuchert, 1913
Genus dalmanella Hall and Clarke, 1892

Type species. By original designation, Orthis testudinaria Dalman, 1828, p. 115, pi. 2, fig. 4.

Dalmanella cf. pectinoides Bergstrom, 1968

Plate 95, figs. 1-5, 10, 11, 15, 22-24

1968 Dalmanella pectinoides Bergstrom, pp. 8-9, pi. 2, figs. 6-9.

1982 Dalmanella sp.; Thomsen and Baarli, pi. 1, fig. 10.

1977 Dalmanella pectinoides Bergstrom; Havlicek, p. 138, pi. 32, figs. 21 and 22.

Holotype. UM Vg873: from Dalmanitina Beds (upper Ashgill), Bestorp, Vastergotland, Sweden.

explanation of plate 95

Figs. 1-5, 10, 11, 15, 22-24. Dalmanella cf. pectinoides Bergstrom, 1968. Top of Myren Member, 17 m below
base of Padda Member, Solvik Formation (basal Aeronian) at Malmoykalven. 1 and 22, PMO 20969,
dorsal and frontal view of whole shell. 2, PMO 109747, internal mould of pedicle valve. 3, PMO 109753,
internal moulds of two brachial valves. 4, PMO 108284, internal mould of brachial valve, figured as
Dalmanella sp. by Thomsen and Baarli (1982, pi. 1, fig. 10). 5, PMO 109742, internal mould of brachial
valve. 10, 11, 15, PMO 109754, latex cast of brachial valve. 23 and 24, PMO 109743, latex and internal
mould of pedicle valve. All x 2.

Figs. 6-9, 12, 13, 16, 17, 20, 21, 25. Isorthis (Protocortezorthis) prima Walmsley and Boucot, 1975. 6-8,
PMO 1 1 1678, Spirodden Member (top Rhuddanian), Ringeriksveien, Sandvika; ventral, dorsal, and side
view of whole shell, x 3-5. 9, 12, 13, 17, from base of Leangen Member, Solvik Formation (basal

Aeronian), Skytterveien, Asker; 9, PMO 109730, internal mould of brachial valve; 12, PMO 103487,
internal mould of brachial valve, refigured from Thomsen and Baarli (1982, pi. 1, fig. 20); 13 and 17, PMO
111673, ventral and side view of internal mould of pedicle valve; all x3. 16 and 20, PMO 105215, 10m

above base of Myren Member, Solvik Formation (basal Rhuddanian), Spirodden, Asker; latex cast of
mould of brachial valve, x3. 21 and 25, PMO 111725, 9 m above base of Myren Member, Solvik
Formation (basal Rhuddanian), Konglungo, Asker; latex cast of internal mould of brachial valve, x3.

Figs. 14, 18, 19. Isorthis ( Ovalella ) mackenziei Walmsley in Boucot et al., 1966. PMO 108269, from base of
Vik Formation (Telychian) at Kampebraten, Sandvika; latex casts (14, 19) and mould (18) of brachial
valve, refigured from Cocks and Baarli (1982, pi. 3, fig. 8), x 2.

PLATE 95

BAARLI, Dalmanella , Isorthis

1106

PALAEONTOLOGY, VOLUME 31

Material. PMO 108284, 109742 109747, 109749, 109753, 109754, 111681 111685, 116840-116844, 116860-
116870, 20968-20976, 20978, 20980-20982: internal moulds of twenty brachial, eighteen pedicle valves, and
thirteen whole shells from uppermost 50 m of Solvik Formation (early Aeronian) on Malmoya (Grid Ref.
NM 981376) and Malmoykalven (NM 976377) of Oslo District.

Discussion. The species is close to Dalmanella pectinoides Bergstrom, 1968, but has a lower convexity
in the pedicle valve. It rarely shows the sulcus in the brachial valve and has small socket pads
rather than fulcral plates, as in D. pectinoides. In other characteristics it agrees well, including the
peculiar, almost parvicostellate ribbing. The Swedish material of D. pectinoides is not well preserved
and shows no details of the dental system. The present material shows small teeth without
cruralfossetts. Basally the teeth are supported by strong, short, and erect dental plates.

Genus ravozetina Havlicek, 1974

Type species. Orthis honorata Barrande, 1879, p. 53, from Upper Ordovician (Kraluv Dvur Formation) of
Bohemia.

Ravozetina cf. honorata (Barrande, 1879)

Plate 96, figs. 3-6, 9, 10, 13, 14

1879 Orthis honorata Barrande, p. 53, pi. 68, case III, figs. 1 and 2.

1950 Parmorthis (Dedzetinal) honorata (Barrande, 1879); Havlicek, pp. 34 35 and 105, pi. 1 1, fig. 9.

1977 Ravozetina honorata (Barrande, 1879); Havlicek, pp. 145-147, pi. 29, figs. 7 14.

Material. PMO 111710, 11 1735, 1 16769, 1 16775: three brachial and one pedicle valve from Myren Member
and basal 110 m of Solvik Formation of Konglungo (NM 849347), Spirodden (NM 841338), and Vakas
(NM 828357).

Discussion. The species is rare and the material fragmentary. It seems to fall within the description
of the type material. However, the one well-preserved brachial valve (PMO 1 1 1735) has a higher
convexity than the type material.

EXPLANATION OF PLATE 96

Fig. 1. Isorthis (Ovaleila) mackenziei Walmsley in Boucot et al., 1966. PMO 108270, from base of Vik
Formation (Telychian) at Kampebraten, Sandvika; mould of of pedicle valve, refigured from Cocks and
Baarli (1982, pi. 3, fig. 7), x 2.

Figs. 2, 7, 8, 11, 12, 16-18. Levenea sp. nov. 2, 7, 8, II, 12, PMO 109733, 10 m above base of Myren Member
(Hirnantian/Rhuddanian), Solvik Formation, Vakas, Asker; internal mould and latex casts of brachial
valve. 16-18, from 15 m above base of Myren Member, Solvik Formation, Spirodden, Asker; 16 and 18,
PMO 109761, internal mould of pedicle valve; PMO 116846, internal mould of pedicle valve. All x 3.

Figs. 3-6, 9. 10, 13, 14. Ravozetina cf. honorata (Barrande, 1879). 3, PMO 111710, 15 m above base of
Myren Member (earliest Rhuddanian), Spirodden, Asker; internal mould of pedicle valve. 4-6, 9, 10, 13,
14, from 9 m above base of Myren Member (Hirnantian/Rhuddanian), Solvik Formation, Konglungo; 4
and 6, PMO 111735, internal mould and latex cast of brachial valve; 5, 9, 10, 13, 14, PMO I 16769. latex
casts and internal mould (10) of brachial valve. All x 3.

Figs. 15, 19-26. Kampella guttula gen. et sp. nov. From base of Vik Formation (Telychian), Kampebraten,
Sandvika. 15, PMO 1 11701, internal mould of brachial valve, x 3. 19, 21, 23-25, PMO 109758, holotype,
mould (21), and latex cast of brachial valve, x 3. 20, PMO 111731, mould of exterior of brachial valve,
x 3. 22, PMO 1 16734, internal mould of pedicle valve, x 2-5. 26, PMO 1 1 1702, internal mould of pedicle
valve, x2-5.

Figs. 27-30. Mendacella sp. From base of Vik Formation (Telychian), Kampebraten, Sandvika. 27, 29, 30,
PMO 109757, latex cast and internal mould (29) of brachial valve, x 2-5. 28, PMO 109756, internal mould
of pedicle valve, x 2-5.

PLATE 96

BAARLI, Isorthis, Levenea, Ravozetina , Kampella , Mendacella

1108

PALAEONTOLOGY, VOLUME 31

Subfamily isorthinae Schuchert and Cooper, 1931
Genus isorthis Kozlowski, 1929

Type species. By original designation, Dalmanella ( Isorthis ) szajnochai Kozlowski, 1929, p. 75, pi. 2, figs. 24-
41, from Borszczow Stage (early Gedinnian) of Podolia.

Isorthis ( Ovalella ) mackenziei Walmsley in Boucot et al., 1966
Plate 94, figs. 14, 18, 19; Plate 95, fig. 1

1966 Isorthis mackenziei Walmsley in Boucot et al., p. 17, pi. 4, figs. 17-20.

1975 Isorthis ( Ovalella ) mackenziei Walmsley; Walmsley and Boucot, p. 77, pi. 7, figs. 4 12.

1976 Isorthis ( Ovalella ) mackenziei Walmsley; Walmsley and Basset, p. 203, pi. 1, figs. 9-12.

1982 Isorthis ( Ovalella ) mackenziei Boucot, Johnson, Harper and Walmsley; Cocks and Baarli,
pi. 3, figs. 7 and 8.

Holotype. GSC 189589 (Boucot et al. 1966. pi. 4, fig. 18): internal mould of brachial valve from Long Reach
Formation (Late Llandovery C6 to Early Wenlock); GSC locality 55061, southern New Brunswick, Canada.

Material. PMO 108269, 109750, 109752, I 11653 111656: internal moulds of five pedicle and four brachial
valves from base of Vik Formation (Telyclnan) at Kampebraten (NM 846402) and Chr. Skredsviks vei (NM
863428), in Sandvika.

Discussion. The material seems to fall well within the description by Walmsley and Boucot (1975)
and nothing can be added from the relatively sparse and fragmentary material found here.

Isorthis (Protocortezorthis) prima Walmsley and Boucot, 1975

Plate 94, figs. 6-9, 12, 13, 16, 17, 20, 21, 25; text-fig. 3; table 1

71917 Orthis ( Dalmanella ) crassa Lindstrom; Reed, 1917, pp. 849-850, pi. 9, figs. 8-10 (non Lindstrom).
1964 Dalmanella aff. testudinaria (Dalman); Boucot and Johnson, p. 3, pi. 1, figs. I 12; pi. 2,
figs. 1-7.

1975 Isorthis ( Protocortezorthis ) prima Walmsley and Boucot, p. 63, pi. 3, figs. 1 8.

1982 Isorthis prima Walmsley and Boucot; Thomsen and Baarli, pi. 1, fig. 20.

1986 Isorthis prima Walmsley and Boucot; Baarli and Harper, pi. 2, figs. d,f.

1987 Isorthis prima Walmsley and Boucot; Baarli, fig. 5 d.

Holotype. USNM 204883: pedicle valve from Mulloch Hill Formation (Rhuddanian), Rough Neuk Quarry,
Craighead Inlier, Girvan, Strathclyde, Scotland (Grid Ref. NS 270040).

Material. PMO 103487, 103510, 103550, 103551, 105209, 105213-105215, 105878, 109726, 109727, 109729,

109730, 111672, 111673, 111675, 111676, 111678, 111687, 111703-111706, 111721, 111723-111726, 111728,

111729, 111738-111745, 111751, 116723-116733, 116736-116745, 116766 116768, 116845, 116856 116859,

116865, 117035, 117037, 117039, 117040, 117368-117375, 1 17419: internal moulds of sixty-three brachial,

thirty-three pedicle, and nine whole valves found throughout Solvik Formation (Rhuddanian to early
Aeronian) of Asker, Sandvika, and Malmoya areas.

Discussion. Isorthis prima is a very variable species, both from the type locality and in the present
study area. The Norwegian material differs in that the brachial muscle scars tend to be bigger than
in the type material and the brachiophores less divergent and strong. The overall variation,
however, is so great that there is full overlap from region to region. Baarli (1987) listed two species
from the Solvik Formation: a small one, I. prima , in the two lower members; and a bigger one,
Isorthis sp., in the upper member. However, closer examination showed that all characters
overlapped and I now regard both as belonging to I. prima. The following changes from small to
larger specimens were found: 1, the hinge line changes from three-quarters to one-half of the
maximum width (text-fig. 3); 2, the shape tends to change from clearly biconvex to more
planoconvex; 3, the sulcus in the brachial valve is better developed in small specimens; 4, the
brachiophores change from thin, slender, and sharply triangular, to ponderous and bluntly

BAARLI: LLANDOVERY ENTELETACEAN BRACHIOPODS

1109

maximum width

text-fig. 3. Measurements of Isorthis (Protocortezorlliis) prima Walmsley
and Boucot, 1975, showing how the proportion of the lunge line to maximum
width varies relative to size of maximum width (see Table 1).

table I. Measurements of Isorthis prima used in text-fig. 3 in millimetres.

PMO no.

Maximum

width

Hinge

line

PMO no.

Maximum

width

Hinge

line

103510

111

60

111737

10 5

6-8

105209

4-7

3-8

111738

6-7

4-0

105213

5-2

31

111744

8-2

4-5

105215

10-9

4-5

111751

7-0

5-7

105878

12-0

6-5

116723

60

5-6

109726

10-4

5-8

116729

9-2

6-0

109728

7-7

50

116731

40

3 0

109730

81

5-4

116732

12-2

5-0

111672

12-5

60

116845

90

4-5

111672

110

6-8

116855

7-8

4-5

111673

13-5

7-5

117035

9-3

7-2

111675

10-5

5-2

117037

10-8

6-0

111678

91

5-3

117039

7-8

5-8

111687

11-8

7-2

1 1 7040

6-5

6-0

111703

8-8

5-8

117368

9-5

6-5

111704

81

51

117370

9-0

5-5

111705

91

5-5

1 17371

10-0

5-4

111706

9-8

6-5

117372

7-0

6-0

111723

80

5-2

1 17373

7-8

6-3

111724

6-6

4-2

117374

7-0

5-0

111725

8-5

6-8

117375

7-0

4-5

111726

12-5

7-2

117377

9-5

7-2

111728

6-8

4-8

117378

7-0

5-0

111729

11-2

5-6

117419

13-5

6-8

111736

8-7

5-5

1110

PALAEONTOLOGY, VOLUME 31

triangular; 5, the fulcral plates are more pronounced in the larger specimens; and 6, the muscle
fields are better impressed with muscle bounding ridges tending to curve on to the median ridge.

Genus levenea Schuchert and Cooper, 1931
Type species. By original designation, Orthis subcarinata Hall, 1857, p. 43, figs. 1 and 2.

Levenea sp. nov.

Plate 96, figs. 2, 7, 8, 1 1, 12, 16-18

Material. PMO 109733, 109760, 109761, 1 16846, 1 17400: internal moulds from three pedicle valves and two
brachial valves from basal Solvik Formation at Spirodden (NM 841338), Konglungo (NM 849347), and
Vakas (NM 828357) in Asker area.

Description. Exterior: subcircular biconvex to ventribiconvex and four-fifths as long as wide. Hinge line
straight and three-quarters as long as maximum width, which occurs at mid-length. Cardinal angle obtuse,
commissure gently and evenly curved. Beak straight and delthyrium and notothyrium open. Finely costellate
with four or five rounded ribs per mm. Weak, broad sulcus in brachial valve.

Interior of pedicle valve: delthyrial chamber narrow and high with relatively flat floor. Simple, triangular,
and very strong teeth. Very short, curved dental plates continue into well-developed muscle-bounding ridges,
which delimit narrowly bilobate and strongly impressed muscle scars. Muscle scars pentagonal, occupy two-
fifths of total length, and divided anteriorly by narrow and diverging median ridge originating at half length
of muscle scars. Vascula media initially nearly parallel, then diverges and strongly branches.

Interior of brachial valve: brachiophores ponderous and swollen medially with supporting plates more
divergent at base than at top. Socket pads present. Cardinal process bulbous and bilobed with long broad
shaft. Muscle field well impressed with strong muscle-bounding ridges. Broad parallel-sided and shallow
myophragm ends at anterior end of muscle scars. Muscle scars occupy two-fifths of total length, divided by
straight transverse ridges in small subquadrate posterior scars and larger suboval anterior scars.

Discussion. This material is close to Levenea media (Shaler, 1865), as redescribed by Walntsley and
Boucot (1975). It is distinguished by less well-developed socket pads, a smaller cardinal process.

EXPLANATION OF PLATE 97

Figs. 1-5, 10, 11, 14, 15, 18, 19. Resserella matutina sp. nov. I, 2, 5, PMO 52924, Leangen Member, Solvik
Formation (early Aeronian), Vettre Brygge, Asker; dorsal, ventral, and side views of whole shell, x2-5.
3, 4, 10, 11, 14, 15, 18, 19, from top of Feangen Member, Solvik Formation (middle Aeronian). 3, 14, 15,
18, PMO 111716, Avloes, Baerum; internal mould of brachial and pedicle valve (3), and latex cast of the
brachial valve, x 2. 4, PMO 111719, Gjettum, Baerum; internal mould of pedicle valve, x2. 10, PMO
108288, Gjettum, Baerum; holotype, internal mould of brachial valve, refigured from Thomsen and Baarli
(1982, pi. 1, fig. 22). 11 and 19, PMO 111717, Avloes, Baerum; ventral and posterior view of internal

mould of pedicle valve, x 2.

Figs. 6-9, 13, 21-26. Dicoelosia osloensis Wright, 1968c/. 6, PMO 103510, 42 m above base of Feangen
Member, Solvik Formation (middle Aeronian), Skytterveien, Asker; internal mould of pedicle valve. 7 and
8, from basal 10 m of Leangen Member, Solvik Formation (early Aeronian), Skytterveien, Asker; 7. PMO
105890, internal mould of brachial valve, refigured from Baarli (1987, fig. 5c); 8, PMO 105222, internal
mould of pedicle valve, refigured from Thomsen and Baarli (1982, pi. 1, fig. 15). 9, 13, 21, 23-26, basal
10 m of Leangen Member, Solvik Formation (early Aeronian), Leangbukta, Asker; 9, PMO 1 16763, ventral
view of external mould; 13, PMO 1 1 1732, internal mould of brachial valve; 21, 23-26, PMO 116756, latex
and internal mould (24) of brachial valve. All x 4.

Figs. 12, 16, 17. Dicoelosia a/ticavata (Whittard and Barker, 1950). From base of Vik Formation (Telychian),
Kampebraten, Sandvika. 12, PMO 111671, internal mould of pedicle valve. 16, PMO 1 11698, internal
mould of brachial valve. 17, PMO 108261, internal mould of pedicle valve, refigured from Cocks and
Baarli (1982, pi. 3, fig. 6). All x4.

Fig. 20. Epitomyonia sp. PMO 108287, base of Myren Member, Vakas, Asker; internal mould of pedicle
valve, refigured from Thomsen and Baarli (1982, pi. 1, fig. 16) and Baarli and Harper (1986, pi. 2k), x3.

PLATE 97

BAARLI, Resserella , Dicoelosia , Epitomyonia

1112

PALAEONTOLOGY, VOLUME 31

and larger muscle scars in the brachial valve. The pedicle valve shows a rounder and narrower
median ridge between the diductor scars. L. media was the oldest known species of Levenea (Idwian
time). This material belongs to a new species and also predates it. Additional material is needed,
however, to formally erect a new species.

Subfamily resserellinae Lazarev, 1970
Genus resserella Bancroft, 1928

Type species. By original designation, Orthis canalis J. de C. Sowerby in Murchison, 1839, p. 630 pars, pi.
13, fig. 12 a, non pi. 20, fig. 8.

Resserella matutina sp. nov.

Plate 97, figs. 1-5, 10, 11, 14, 15, 18, 19

1982 Resserella sp. Thomsen and Baarli, pi. 1, fig. 22.

Holotype. PMO 108288 (PI. 97, fig. 10): brachial valve from upper part of Leangen Member, Solvik Formation
(middle Aeronian) at Gjettum in Baerum (NM 856421).

Material. 52924 (twelve specimens), 53203, 53204, 53206, 108288, 109755, 111652, 111680, 111716-111719,
1 1 3669- 1 1 3676, 1 1 6848 1 1 6854: seventeen whole shells, eighteen pedicle, and six brachial valves from top of
Solvik Formation (early to middle Aeronian) at Gjettum (NM 856421), Jongsaskollen (NM 846404), Avlos
(NM 868431) in Sandvika and Vettre Brygge in Asker.

Diagnosis. Transversely elliptical, planoconvex to concavoconvex with fine ribbing of elegantu-
loid type. Ventral beak overhanging straight hinge line. Ventral muscle field well impressed and
pentagonal with vascula media divergent. Teeth strong with deep cruralfossetts. Anacline interarea
and shallow, broad sulcus in brachial valve. Brachiophore processes strong, tusk-like, and with
relatively small sockets for the genus. Muscle scars well impressed with posterior adductor scars
larger than anterior.

Description. Exterior: planoconvex to concavoconvex with very deep pedicle valve relative to brachial valve.
Outline transversely shield-shaped, with length three-quarters of width. Maximum width varies, 15-22 mm.
Hinge line long and straight, equal in length to three-quarters to four-fifths of maximum width. Cardinal
angle rounded, lateral and anterior margins evenly rounded. Commissure evenly rounded and anterior
commissure gently sulcate. Brachial valve bears broad, shallow sulcus originating near hinge line and widening
anteriorly. Ventral beak incurved, most often overhangs hinge line, and projects posteriorly about one-fifth
of total length of valve. Ventral interarea apsacline and gently concave. Delthyrium open with delthyrial
angle c. 60°. Dorsal interarea plane, anacline. Notothyrium open and occupied by myophore of cardinal
process. Ornament multicostellate, fine with four or five ribs per mm at 5 mm growth stage of brachial valve.
Costellae low, rounded, and evenly spaced. Ribbing style of elegantuloid type.

Interior of pedicle valve: delthyrial chamber deep with wide flat floor anteriorly. Very strong teeth with
cruralfossetts and triangular dorsal faces supported basally by short dental plates which diverge at c. 60°
from each other anterolaterally. Dental plates continue anteriorly into low ridges which define muscle field
laterally and converge slightly but do not define muscle field anteriorly. Muscle field variably impressed,
elongate pentagonal in outline, and occupies about one-third of total length of pedicle valve and one-quarter
of width. Adductor scars raised anteriorly on broad low ridge that descends at angle to floor of valve,
producing a distinct edge. Diductor scars narrow and elongate. Vascula media divide anteriorly.

Interior of brachial valve: cardinal process bulbous with short shaft. Brachiophores strong, tusk-like, and
erect. Tops of supporting plates diverge at 80° to one another and relative to bases. Posterolateral sides of
brachiophores with supporting plates triangular, anterior sides normal to floor of valve. Sockets supported
by small socket pads. Muscle field raised, quadripartite to subelliptical in outline; occupies about one-halt
of valve length and one-quarter of width; and bound laterally by low, curved ridges, which are continuations
of brachiophore bases and which converge anteriorly but do not meet myophore. Muscle field bisected by
rounded, broad myophore, which continues anteriorly beyond muscle field. Faint transverse ridges separate
triangular anterior muscle field from much smaller subquadrate posterior muscle field.

BAARLI: LLANDOVERY ENTELETACEAN BRACHIOPODS

1113

Discussion. This species is the oldest known Silurian Resserella and it occurs from the lowest
Aeronian. It seems to be a form transitional between Karlicium Havlicek, 1974 (an Ordovician
genus previously proposed as its predecessor by Havlicek 1977), and Resserella (thus supporting
Havlicek’s claim). The fine resserellid ribbing, the strongly impressed muscle-bounding ridges, and
the anacline interarea of the brachial valve connect it with Resserella. The transverse outline,
diverging vascula media, and the anterior pair of adductor scars which are larger than the posterior
pair, are features typical of Karlicium. Except for the diverging vascula media, all these features
are also found in some other species of Resserella , so the assignment to Resserella is reasonable.
R. matutina sp. nov. is distinguished from the slightly younger R. sefinensis Walmsley and Boucot,
1971, by finer costellae, less diverging brachiophores, and its resupinate form. R. concavoconvexa
(Twenhofel, 1928), which first occurs in a C3 to C4 position, has the same outline but differs in
its vascula media and hypercline interarea (as opposed to the anacline interarea of the brachial
valve of R. matutina sp. nov.).

Family rhipidomellidae Schuchert, 1913
Subfamily rhipidomellinae Schuchert, 1913
Genus mendacella Cooper, 1930

Type species. By original designation, Orthis uberis Billings, 1866, from Ellis Bay Formation (Ashgill),
Anticosti Island, Quebec, Canada.

Mendacella bleikeriensis sp. nov.

Plate 98, figs. 1-6, 9, 11, 13-15

1982 Mendacella cf. mullockiensis (Davidson); Thomsen and Baarli, pi. I, figs. 13 and 14.

1987 Mendacella sp.; Baarli, fig. 5c.

Holotype. PMO 105221 (PI. 98, figs. 1 -3, 9): internal mould of brachial valve from 60 70 nr above base of
Leangen Member (mid-Aeronian), Solvik Formation, Skytterveien (NM 820339), in Asker.

Material. PMO 103476, 105217 105219, 105221, 108286, 109740, 109741, 111688, 111689, 111730, 111746,
113715-1 13719, 11 6747: moulds of nine pedicle and eight brachial valves and four whole shells. Species found
rarely in Spirodden Member (Rhuddanian), but occurs frequently in upper 30 m of Leangen Member (early
Aeronian), Solvik Formation, both in Asker and Sandvika.

Diagnosis. Transverse subelliptical Mendacella with deep, nonsulcate brachial valve and low pedicle
valve with variably developed sulcus. Ribbing finely multicostellate. Triangular teeth with
cruralfossetts. Strongly impressed nonflabbelate to flabbelate muscle field tending to enclose
adductor muscle scars. Slender, widely divergent, and curved brachiophores with short supporting
bases. Variably developed crural pits. Weakly impressed dorsal muscle scars.

Description. Exterior: dorsibiconvex, transverse subelliptical, and about five-sixths as long as broad with
width 1 1-20 mm. Pedicle valve bears variably developed, often strong and broad sinus, originating at length
about one-third from posterior margin and occupying half of width. Hinge line straight, about half as long
as maximum width, which occurs at three-quarters length from posterior margin. Commissure crenulated.
Cardinal angle gently rounded, lateral margins rounded for posterior half of valve, then nearly parallel before
passing into anterior sulcus. Ventral beak short, low, and gently incurved. Delthyrium and notothyrium
open. Ventral interarea gently apsacline; dorsal interarea anacline and very low. Ornament finely multicostellate
with three or four rounded costellae per mm at 5 mm growth stage of brachial valve.

Interior of pedicle valve: delthyrial chamber low with flat floor anteriorly. Pedicle callist present. Well-
developed triangular teeth with small cruralfossetts. Low shallow ridges extend anteriorly from short dental
plates and bound muscle fields laterally. Muscle fields strongly impressed, occupying half of total length.
Clearly impressed small, elliptical, elongate adductor scars on rounded median ridge bounded and enclosed
by broad cordate diductor scars which may be faintly flabbelate.

Interior of brachial valve: brachiophores with supporting plates widely divergent at 90 110°, bluntly

1114

PALAEONTOLOGY, VOLUME 31

triangular, and inclined anterolaterally. May have receding socket pads. Cardinal process consists of long,
slender shaft with bilobed, sometimes expanded myophore. Very weakly impressed muscle-bounding ridges.
Dorsal muscle fields subcircular, about equally quadripartite, and occupy two-fifths of total length and width.
Low, broadly rounded median ridge tapers and extends anteriorly to anterior of muscle scars. Periphery of
dorsal valve bears low, rectangular crenulations complementary to those of pedicle valve.

Discussion. Boucot ei al. (1965) defined the differences between Mendacella and Dalejina as: non-
enclosed adductor scars and non-flabbelate diductor scars in the pedicle valve of the former; and
enclosed adductors with flabbelate diductors in the latter. Some of the material discussed above
should thus fall within Dalejina. However, some of the material figured as Mendacella by Boucot
et al. (1965) have the same enclosed adductor scars and faintly flabbelate diductor scars as the
Norwegian specimens.

The new species is close to M. mullockiensis (Davidson, 1869). This species was recently
reinvestigated by Temple (1987). He assigned all the abundant early Llandovery enteletaceans from
Wales to M. mullockiensis , including: M. mullockiensis (Davidson, 1869), M. crassiformis Bancroft,
1949, R. llandoveriana Williams, 1951, and /. prima Walmsley and Boucot, 1975. The new species
lies close to his M. mullockiensis morph, mullockiensis which I regard as a separate species.
However, M. bleikeriensis differs from M. mullockiensis in having more slender and curved
brachiophores, shorter brachiophore bases, lack of sulcus in dorsal valve (which most specimens
of M. mullockiensis have), generally smaller ventral muscle scars, occurrence of crural pits, and
slightly coarser ribbing.

Mendacella sp.

Plate 96, figs. 27-30

1982 Mendacella sp.; Cocks and Baarli, pi. 2, figs. 7 and 8.

Material. PMO 109756, 109757, 111659, 111663: internal moulds of three brachial valves and one pedicle
valve, from base of Vik Formation (Telychian) at Kampebraten (NM 848404) in Sandvika.

Discussion. The material is so scarce and badly preserved that the species is best left undetermined.

EXPLANATION OF PLATE 98

Figs. 1-6, 9, 11, 13-15. Mendacella bleikeriensis sp. nov. 1-6, 9, 15, from 60-70 m above base of Leangen
Member, Solvik Formation, Skytterveien, Asker; 11, 13, 14, middle of Leangen Member, Sol vik Formation
(lower parts of Aeronian), E6 near Vakas, Asker. 1-3, 9, PMO 105221, holotype, internal mould and latex
cast of brachial valve; 4, PMO 103476, internal mould of brachial valve, refigured from Baarli (1987, fig.
6c); 5, PMO 108286a, internal moulds of pedicle valves (deposited with 1082866); 6, PMO 1 1 1688, exterior
of pedicle valve; 15, PMO 1082866, side and posterior view of internal mould of pedicle valve, refigured
from Thomsen and Baarli (1982, pi. 1, fig. 14). II, 13, 14, PMO 1 1 1730, side, dorsal, and ventral views of
valve. All x 2.

Figs. 7, 8, 12, 16, 20, 24. Drabovia sp. 7, 8, 12, 16, 20, PMO 111708, 17 nr above base of Myren Member,
Solvik Formation (earliest Rhuddanian), Vakas, Asker; internal mould (7) and latex of brachial valve, x4.

24, PMO 109709, 9 m above base of Myren Member, Konglungo, Asker; internal mould of brachial valve,
refigured from Baarli and Harper (1986, fig. 2c), x4.

Figs. 10, 17 19, 21-23, 25-28. Marklandella markesi sp. nov. From middle (30 40 m above base) of Leangen
Member, Solvik Formation (early to mid-Aeronian), Leangbukta, Asker. 10, 17, PMO 105206, holotype,
lateral and posterior view of internal mould of pedicle valve, refigured from Baarli (1987, fig. 66). 18, 23,

25, 27, PMO 105208, internal mould and latex cast of brachial valve. 19, PMO 105205, internal mould of
brachial valve, refigured from Baarli (1987, fig. 6f). 21, PMO 103545, internal mould of pedicle valve.
22, PMO 109751, latex cast of brachial valve. 26, PMO 103533, internal mould of brachial valve. 28,
PMO 103504, internal mould of juvenile brachial valve. All x 3.

PLATE 98

BAA RLE Mendacella, Drabovia , Marklandella

1116

PALAEONTOLOGY, VOLUME 31

Family heterorthidae Schuchert and Cooper, 1931
Genus marklandella Harper, Boucot and Walmsley, 1969

Type species. By original designation, Marklandella giraldi Harper, Boucot and Walmsley, 1969, p. 83,
pi. 17, figs. 1-10, from Silurian (Wenlock or Ludlow) of Freshwater East Bay, Dyfed, Wales.

Marklandella markesi sp. nov.

Plate 98, figs. 10, 17-19, 21-23, 25-28
1987 Marklandella sp.; Baarli, fig. 6 f h.

Holotype. PMO 105206 (PI. 98, figs. 10 and 17): internal mould of pedicle valve from bulk sample taken
40 m above base of Leangen Member, Solvik Formation (mid-Aeronian), at innermost part of Leangbukta
in Asker.

Material. PMO 103504, 103533, 103545, 105206 105208, 109751, 113720 113723, 116748-116752: moulds of
sixteen brachial and nine pedicle valves from Leangen Member (mid-Aeronian) of Solvik Formation at
Leangbukta (NM 825341) in Asker.

Diagnosis. Planoconvex and subcircular with fine multicostellae. Adductor scars of pedicle valve
visible and not enclosed by the diductor scars. Brachiophores small, slender, and very divergent.
No socket pads. Dorsal muscle scars very faintly or not impressed, with posterior muscle scars
smaller than anterior. Costellae impressed over the entire interior of the valve.

Description. Exterior: planoconvex with circular to transversely elliptical outline, 6-15 mm wide. Hinge line
straight, three-fifths maximum width (which occurs posterior to midlength). Brachial valve may bear shallow
sulcus anteriorly. Cardinal angles rounded. Lateral and anterior margins evenly rounded. Ventral beak low,
gently curved, and projects slightly beyond hinge line. Ventral interarea apsacline; dorsal interarea anacline,
very short. Delthyrium and notothyrium open. Ornament fine multicostellate with three or four costellae per
mm at 5 mm length growth stage of brachial valve. Costellae rounded, equally spaced, and radiate with some
curvature.

Interior of pedicle valve: delthyrial chamber shallow with concave floor. Pedicle callist present. Teeth small
and supported by slender, faintly inclined, and curved dental plates. Dental plates extend anteriorly to about
midlength of muscle field, where they join a shallow ridge which curves and defines diductor muscle field
anteriorly. Muscle field weakly impressed and cordate in outline, occupies nearly half total valve length, and
is about four-fifths as wide as long. Adductor scars low and elongate in outline and not bound anteriorly by
diductor scars. Diductor scars subtriangular and elongate. Internal surface of valve impressed by costellae.

Interior of brachial valve: brachiophores small and slender; tops of supporting plates diverge slightly more
posterolaterally than bases. Lateral view of brachiophores with supporting plates is bluntly triangular and
concave. Supporting plates diverge at c. 80 100 from one another at top. Cardinal process varies in outline
from very thin, simple ridge to expanded and possibly three- to four-lobed myophore. Sockets are level with
floor of valve. Muscle fields most often not impressed, but when visible they occupy nearly half of total valve
length; they are quadripartite with posterior muscle scars bigger than anterior. Muscle scars are divided by
low, broad median ridge which tapers anteriorly. Anterior crenulations rounded in cross-section; impression
of costellae extend posteriorly to muscle field.

Discussion. This species is close to the other two known species: M. giraldi Harper et a/., 1969 and
M. macadamia Harper et a /., 1969. The new species differs in having slightly coarser ribbing, very
slender and commonly more divergent brachiophores, and faintly impressed muscle scars which
exhibit anterior scars larger than the posterior.

Genus kampella gen. nov.

Type species. Kampella guttula sp. nov.

Diagnosis. Shell biconvex and subcircular with relatively short hinge line. Ribbing costellate and
costellae situated in posterior portion of shell recurving posteriorly to intersect posterior valve
margin. Small ventral umbo. Ventral muscle scars widely cordate with diductor lobes extending

BAARLI: LLANDOVERY ENTELETACEAN BRACHIOPODS

1117

beyond and not enclosing adductor scars. Vascula media moderately divergent. Brachiophores
with bases widely diverging relative to top and to each other. No fulcral plates. Cardinal process
expanded in semiovoid structure. Vascula terminalia occurring along hinge line are posteriorly
directed.

Discussion. This genus is close to both Onniella and Heterorthina. It is distinguished from Onniella
by the possession of features that characterize the Heterorthidae, as redefined by Havlicek (1970),
e.g. its posteriorly recurving costellae, the vascula terminalia directed posteriorly on the posterior
part of the brachial valve, and fairly large muscle scars. It is distinguished from Heterorthina
by the fact that the ridges limiting the ventral diductor scars do not unite anteromedialy to
enclose the adductor scars, and because the vascula media is only moderately divergent (like
Onniella). In addition, the brachiophore bases are so greatly divergent that their posterior
surfaces acted as walls to the sockets, as seen typically in Bancroftina and rarely in Onniella , but
not in Heterorthina.

Kampella guttula sp. nov.

Plate 96, figs. 15, 19-26

Holotype. PMO 109758: mould of brachial valve (PI. 96, figs. 19, 21, 23-25) from base of Vik Formation
(Telychian) of Kampebraten in Sandvika (NM 848404).

Material. PMO 109758, 1 1 1657, 1 1 1660, 111661, 111669, 111697, 1 1 1699, 1 1 1701, 1 11702: seven brachial and
five pedicle valve moulds from base of Vik Formation (Telychian) of Kampebraten (NM 848404) and Chr.
Skredsvik vei (NM 863428) in Sandvika.

Diagnosis. Biconvex, with strongest convexity posterior, gradually flattening out anteriorly.
Fascicostellate. Both muscle fields faintly impressed. Ventral muscle field broadly cordate, occupying
less than half of length of valve and displaying broad, slightly elevated median ridge. Hinge line
straight and relatively short. Strong triangular teeth with well-developed cruralfossets. Strong
brachiophores with bases greatly divergent relative to their tops and to each other. Well-developed
sockets. Cardinal process expanded in ovoid structure. Quadripartite muscle scars with small
posterior scars relative to anterior scars.

Description. Exterior: transversely elliptical, biconvex, three-fifths to five-sixths as long as wide, with maximum
width 7 15 mm. Hinge line straight, two-fifths to three-fifths of maximum width. Commissure crenulated.
Cardinal angles and lateral commissure rounded, anterior commissure gently rounded. May have shallow
sulcus. Ventral beak incurved and slightly overhanging hinge line. Delthyrium and notothyrium open.
Fascicostellate with two or three first order costellae per mm at 5 mm growth stage of brachial valve and
much finer second order costellae superimposed. Exopunctate. Costellae in posterior part of valve recurve
posteriorly to intersect posterior margin of shell. Both pedicle and brachial valve with strongest convexity
posterior, and clear change in convexity one-third in length from posterior end.

Interior of pedicle valve: delthyrial chamber strongly concave. Strong, triangular teeth with very well-
developed cruralfosetts. Short dental plates which continue into shallow ridges and define muscle field
laterally; they curve but do not close completely anteriorly. Variably impressed muscle field occupying less
than half of valve length. Narrow, triangularly elongate adductor scars on broad, slightly raised median
ridge, bounded laterally by broad, triangular diductor scars not enclosing adductor scars. Greatly divided
and moderately divergent vascula media. Quadripartite muscle scars with posterior scars smaller than anterior
scars.

Interior of brachial valve: brachiophores erect. Tops of supporting plates diverge at 30-45 relative to each
other. Bases much more diverging than tops. Sockets large and triangular with brachiophore bases acting as
anterior walls supported by secondary shell deposits. Muscle field faintly impressed, especially anteriorly; it
occupies less than half of valve length and one-third of maximum valve width. Posterior adductor fields
subquadrate and small, while anterior scars seem larger and weakly defined. The myophragm is nonexistent
to broad, shallow and tapering. Transverse muscle-bounding ridges straight and faintly impressed. Cardinal
process has short shaft, is erect, and widest at middle. Vascula terminalia occurring along hinge line are
posteriorly directed.

1118

PALAEONTOLOGY, VOLUME 31

text-fig. 4. a d, Paurorthis inopinatus sp. nov. Earliest Rhuddanian, Solvik Formation, Asker, a, c, PMO
109731, holotype, 10 m above base of Myren Member, Vakas, Asker; internal mould and latex of brachial
valve, x 3. b, d, PMO 1 1 1733, 9 m above base of Myren Member, Konglungo; latex and internal mould of
pedicle valve, x 3. e-h, Enteletacea indet. sp. A. Solvik Formation, Asker, e g, PMO 103498, middle of
Leangen Member (early Aeronian), Skytterveien; internal mould and latex cast of brachial valve, x 2. h,
PMO 105216, middle of Spirodden Member (late Rhuddanian), Spirodden; internal mould of pedicle valve,

x 2.

Family paurorthidae Opik, 1933
Genus paurorthis Schuchert and Cooper, 1931

Type species. By original designation, Orthambonites parva Pander, 1830, from Lower Ordovician of Estonia.

? Paurorthis inopinatus sp. nov.

Text-fig. 4a -d

Holotype. PMO 109731: mould of brachial valve (text-fig. 4a, c), from 10 m above base of Myren Member
(earliest Rhuddanian), Solvik Formation, Vakas (NM 828357) in Asker.

Material. PMO 109731, 111712, 1 11733, 1 16735: moulds of three brachial and one pedicle valves from
Myren Member (earliest Rhuddanian), Solvik Formation, Konglungo (NM 849347) and Vakas (NM
828357) in Asker.

Diagnosis. Paurorthis with deep notothyrial cavity, ridge-like cardinal process, moderately developed
median ridge, and long elongate posterior muscle scars lying laterally to anterior scars in brachial
valve. Pedicle valve with broad median ridge extending to anterior margin, and very strong teeth.

Description. Exterior: biconvex, subcircular equal to nine-tenths as long as wide. Hinge line straight, two-
thirds to three-quarters as long as maximum width. Maximum width 6-10 mm occurring posterior of
midlength. Cardinal angles obtuse, margins evenly rounded. No clear folds or sulcus. Beak small and straight.
Ventral interarea apsacline. Short anacline dorsal interarea. Delthyrium and notothyrium open. Fascicostellate,
four or five costellae per mm at anterior margin.

Interior of pedicle valve: no pedicle callist. Strong, ponderous, and bluntly triangular teeth with obsolete
dental plates. Faint muscle-bounding ridges define muscle scars laterally and verge on to median ridge.
Lanceolate adductor scars situated on broad median ridge. Ridge continues to anterior margin and is bounded
by narrow, triangular diductor scars, and more anteriorly by vascula media, which thus is parallel.

Interior of brachial valve: brachiophores simple and ridge-like. Supporting plates, with bases strongly
diverging relative to tops, act as anterior wall of sockets. Angle between tops of supporting plates 40°. Large
specimens have well-developed notothyrial platforms. Thin, ridge-like cardinal process. Moderately developed
median ridge extending to anterior end of muscle scars. Muscle scars large and faintly impressed with

BAARLI: LLANDOVERY ENTELETACEAN BRACHIOPODS

1119

posterior scars situated laterally to anterior scars. They are divided by strongly oblique ridges diverging
anterolaterally. Muscle scars occupy half of total valve length and width.

Discussion. P. inopinatus undoubtedly belongs to the Paurorthidae. It has all the features of
Paurorthis, including the simple ridge-like cardinal process, obsolete dental plates, and dorsal
muscle scars divided by oblique ridges. However, the median ridges both in the pedicle and in the
brachial valves are weakly developed; likewise, the notothyrial platform is not as high as in typical
Paurorthis. These differences might justify erecting a new genus, but the available material is not
adequate. Havlicek et al. (1986) found what they called ‘probably the last occurring Paurorthis ’ at
Fluminimaggiore, Sardinia, of uppermost Berounian (= lowermost Ashgill) age. However, the
new species described here occurs considerably later.

Family dicoelosiidae Cloud, 1948
Genus dicoelosia King, 1850

Type species. By original designation, Anomia biloba Linnaeus, 1758, p. 703.

Dicoelosia osloensis Wright, 1968c/

Plate 97, figs. 6-9, 13, 21 26

1968 a Dicoelosia osloensis Wright, pp. 309-31 I, pi. 5, figs. 6-1 1; pi. 6, figs. 7-10.

1968 Dicoelosia verneuliana (Bucher); Anrsden, pi. 8, fig. 1.

1971 Dicoelosia aff. osloensis ; Rubel, p. 51, pi. 8, figs. 18-25.

1971 Dicoelosia osloensis ; Rubel, p. 53, pi. 9, figs. I 5, 14.

1982 Dicoelosia osloensis Wright; Thomsen and Baarli, pi. 1, fig. 15.

1986 Dicoelosia osloensis Wright; Baarli and Harper, pi. 2i.

1987 Dicoelosia osloensis Wright; Baarli, fig. 5c.

1987 Dicoelosia osloensis Wright; Temple, p. 49, pi. 3. figs. 10- 12.

Holotype. PMO 74564: conjoined valves, from Bilobites biloba shale, Myren Member, 20-50 m below base
of Padda Member (early Llandovery), Solvik Formation, Malmoya, Inner Oslo Fjord.

Material. PMO 87609, 103468, 103510, 105886, 105890, 1 1 1732, 113724, 113677 113714, 116756-116765,
117381 117386: twenty-nine brachial and forty-seven pedicle valves and four whole shells from Solvik
Formation, Asker and Malmoya.

Description. Exterior. Wright (1968c/) defined the species as typically broader than long. Maximum width,
however, seems to vary allometrically with length. Small specimens often broader than long (see PI. 97,
fig. 9). Relationship between maximum length and length to lobe is apparently constantly 4:3, as found at
Malmoya. Great variation in external shape of Dicoelosia osloensis is in accordance with what Rubel (1971)
and Musteikis and Puura (1983) found for Estonian and Lithuanian material.

Wright (1968c7) noted that the species probably did not have well-developed ribs on the sulcus. This present
material clearly shows ribs, although not as strongly developed as on top of lobes.

Interior of pedicle valve: form of delthyrial chamber seems to vary ontogenetically. Small forms tend to
have flatter and shallower floor in chamber than larger specimens, thus producing change in profile posteriorly
from gently curving in small forms to steeply descending in large forms. No pedicle callist. Teeth small with
cruralfossetts. Dental plates very short, continuing in low ridges that run parallel, delimiting muscle scars
laterally. Muscle field occupying one-quarter to one-fifth of total valve length and often situated on thickened
floor. Thickening sometimes bounded anteriorly by wedge giving bilobed impression. Median ridge present
and varies in width from width of one diductor scar to thin ridge. In former case, ridge longitudinally striated.

Interior of brachial valve: myophore short, bilobed, and often greatly expanded and crenulated; it fills
notothyrium. Myophore shaft continues in variably developed median ridge which may be high and sharp
and extend to anterior invagination, or may appear as broad elevated median area not clearly delimited.
Brachiophore bases continue from thickened notothyrial walls and are greatly divergent relative to tops of
supporting plates. Small sockets and small ‘swellings’ or hook-like brachiophores directed posterolaterally.
Muscle scars very well-developed and occupy most of space posterior to invagination. Both scars narrowly
elongate and divided by faint, very oblique ridge diverging towards lateral margins. Posterior scar may extend

1120

PALAEONTOLOGY, VOLUME 31

nearly equally far anteriorly as anterior scar, and slightly past invagination. Adductor muscle scars engulf
lanceolate area and start very far posteriorly where they are delimited by brachiophore bases. Follicular
eminences and embayments very well developed.

Discussion. The type species was described from the Solvik Formation at Malmoya in the Oslo
District. The material, however, was distorted and only exterior features were described. The Asker
District provides rich material of the species from the Solvik Formation, with well-preserved
internal features and structures. Additional information and a partial redescription of the species
therefore is presented above.

Dicoelosia alticavata (Whittard and Barker, 1950)

Plate 97, figs. 12, 16, 17

1950 Bilobites alticavatus Whittard and Barker, p. 577, pi. 8, figs. 16-18.

1968a Dicoelosia alticavata (Whittard and Barker); Wright, p. 31 1, pi. 2, figs. 11 15.

1974 Dicoelosia alticavata (Whittard and Barker); Bassett and Cocks, p. 11.

1982 Dicoelosia alticavata (Whittard and Barker); Cocks and Baarli, pi. 3, fig. 15.

Holotype. GSM 82551: conjoined valves from Purple Shales (Telychian), 200 m north-north-east of Hughley
Bridge, Shropshire (SO 562979).

Material. PMO 108261, 111667, 111671, 111691, 111693, 111695, 111696, 111698, 117410-117412: one
external mould and nine pedicle and two internal moulds of brachial valves from basal Vik Formation
(Telychian) at Kampebraten in Sandvika.

Discussion. The species is rather rare and not well preserved, so it provides little new information.
Wright (1968a), however, remarks that ribs are probably only weakly developed in the sulcus. This
material show the feature is well developed.

Genera epitomyonia Wright, 19686

Type species. By original designation, Epitomyonia glypha Wright, 19686, p. 128, pi. 1, figs. 1-5.

Epitomyonia sp.

Plate 97, fig. 20

1982 Dicoelosia cf. inghami Wright; Thomsen and Baarli, pi. 1, fig. 16.

1986 Epitomyonia sp.; Baarli and Harper, pi. 2, fig. k.

Material. PMO 108287, 1 17407 117409: five pedicle valves (two moulds and three exterior shells) from 0-3 m
above base of Myren Member (Hirnantian/Rhuddanian), Solvik Formation, Vakas in Asker (NM 828357).

Description. Exterior: only pedicle valves known. These are high, convex, and subquadrate. Width varies
from half to five sixths of total length. Hinge line long and straight, five-sixths of maximum length. There
might be a slight invagination anteriorly. Very shallow sulcus best seen in moulds. Ears very small. Ribbing
costellate, three ribs per mm, measured 5 mm in front of umbo.

Interior of pedicle valve: small, pentagonal, and faintly impressed muscle scars occupy one-fifth of total
length. Broad median ridge separates triangular diductor muscle scars. Small teeth and weakly developed
muscle-bounding ridges.

Family draboviinae Havlicek, 1950
Genus drabovia Havlicek, 1950

Type species. By original designation, Orthis redux Barrande, 1848, pi. 18, fig. la. from Drabov Quartzite
(Llandeilo), Czechoslovakia.

BAARLI: LLANDOVERY ENTELETACEAN BRACHIOPODS

1121

Drabovia sp.

Plate 98, figs. 7, 8, 12, 16, 20, 24
1986 Fascifera sp. Baarli and Harper, pi. 2, fig. e.

Material. PMO 109709, 1 1 1708: two moulds of brachial valves from lower 20 m of Myren Member, (earliest
Rhuddanian), Solvik Formation, at Konglungo (NM 849347) and Vakas (NM 828357) in Asker.

Description. Exterior: only brachial valves known. These are subcircular, nine-tenths as wide as long, convex,
and small (5 7 mm). Hinge line straight, nine-tenths of maximum width at mid-length. Evenly rounded
margins. Gentle sulcus in brachial valve. Multicostellate with three or four ribs per mm at anterior margin.

Brachial valve: fairly erect brachiophores with high supporting plates. Angle between tops of plates 40°.
Bases short and subparallel to convergent. Small fulcral plates present. Broad, very shallow median ridge
equal in width to one muscle scar. Myophore bilobed, crenulated, and small with short shaft. Small, well-
impressed, quadripartite muscle scars occupy one-third of maximum width and one-quarter to one-third of
total valve length. Muscle-bounding ridges, continuous with the brachiophore bases, delimit muscle scars
laterally and curve towards median ridge without meeting.

Discussion. The material is close to Drabovia westrogoetbica Bergstrom, 1968, and may belong to
that species.

Genus diorthelasma Cooper, 1956

Type species. By original designation, Diorthelasma parvum Cooper, 1956, p. 998, pi. 146, figs. 5-23, from
Pratt Ferry Formation (Llandeilo) of Alabama, USA. Gender neuter.

? Diorthelasma semotum sp. nov.

Plate 99, figs. 17, 22, 23, 27-29

Holotype. PMO 1 1 1670 (A) (PI. 99, fig. 17): mould of brachial valve from the upper part of Leangen Member
(early to middle Aeronian), Solvik Formation, Skytterveien (NM 820339) in Asker.

Material. PMO 103518, 105197-105199, 109759, 111670, 117417, 117418: internal moulds of eight brachial
and one pedicle valves, from upper parts of Spirodden Member and throughout Leangen Member (latest
Rhuddanian to middle Aeronian) of Asker District.

Diagnosis. Ventribiconvex to biconvex, minute Diorthelasma with relatively coarse ribbing and
long hinge lines. Broadly triangular pedicle muscle scars and strong, diverging dental plates with
minute teeth. Brachiophores slender and relatively long, situated on supporting plates running
parallel to each other. May show well-developed crural pits and has small fulcral plates.

Description. Exterior: ventribiconvex to biconvex, minute brachiopods, 3-5 mm wide. Outline subelliptical
to subquadrate, commonly half to two-thirds as long as wide. Hinge line wide, about five-sixths of maximum
width which occurs one-third of length from hinge line. Commissure crenulated with broad, shallow sulcus.
Cardinal angle obtuse, lateral margins evenly rounded, anterior margin gently rounded to straight. Ventral
beak short, projecting slightly beyond hinge line. Notothyrium open. Ornament costellate, four to six gently
rounded costellae per mm at anterior margin, separated by spaces.

Interior of pedicle valve: relatively shallow delthyrial chamber with flat floor anteriorly. Minute teeth
supported by short and divergent dental plates. Muscle field broadly triangular to cordate occupying one-
quarter of total valve length and situated on thickened floor. No median or muscle-bounding ridges.

Interior of brachial valve: brachiophores relatively long and slender. Together with supporting plates they
form prominent, erect, triangular plates with continuations of bases varying around parallel. Small fulcral
plates present. Crural pits may be well developed, especially in larger specimens. Cardinal process ridge-like
on short shaft. Muscle field small and very faintly impressed, often with ribbing superimposed. Muscle-
bounding ridges start at angle to brachiophore bases, and continue laterally parallel to muscle scars. No
clear myophragm, but it may have very broad, thickened, triangular area between brachiophore bases. Muscle
scars occupy one-third of total width.

1122

PALAEONTOLOGY, VOLUME 31

Discussion. The only earlier described species is D. parvum Cooper, 1956, from middle Ordovician
of Alabama, USA and Girvan, Scotland. The two species are therefore separated by a long time
span. Unfortunately, no Norwegian specimen shows well-impressed dorsal muscle scars that would
reveal how the anterior and posterior muscle scars are situated relative to each other. Other
features, however, seem to agree with Diorthelasma. Another possible genus is Saukrodictya, which
is common in the uppermost Ordovician to lowermost Silurian. However, there is no trace of its
peculiar honeycomb ribbing in the present material, in spite of a fairly well-preserved exterior.
Also, the brachiophores in ?£). semotum seem to be longer and more slender, and the hinge line
too long to be a Saukrodictya. A sure designation to genus, however, cannot be made before either
the configuration of the dorsal muscle scars are revealed or the honeycomb meshwork eventually
recognized.

Genus salopina Boucot, 1960

Type species. By original designation, Orthis lunata J. de C. Sowerby in Murchison, 1839, p. 611, pi. 3,
fig. 12c/; pi. 5, fig. 15.

EXPLANATION OF PLATE 99

Fig. 1. Resserella matutina sp. nov. PMO 111680, top of Leangen Member, Solvik Formation (middle of
Aeronian), Jongsaskollen, Sandvika; dorsal view of whole shell, x 2.

Figs. 2-8, 18, 19, 21. Jezercia rongi sp. nov. From Solvik Formation, Asker. 2 and 6, PMO 1 11727, 15 m
above base of Myren Member (earliest Rhuddanian), Spirodden; posterior and dorsal views of whole shell,
x 2. 3, 4, 7, 8, PMO 109732, 1 1 nr above base of Myren Member (earliest Rhuddanian), Vakas; holotype,
internal mould (3) and latex cast of brachial valve, refigured from Baarli and Harper (1986, pi. 1, fig. m),
x 2. 5, PMO 111713, 6 m above base of Myren Member (Hirnantian/Rhuddanian), Konglungo; latex of
internal mould of brachial valve, refigured from Baarli and Harper (1986, pi. 1, fig. o), x 3. 18, PMO

1 1 1707, 9 m above base of Myren Member (earliest Rhuddanian), Konglungo; internal mould of brachial
valve, x 2. 19, PMO 1 1 1748, 6 m above base of Myren Member, Vakas; internal mould of pedicle valve,
x 2. 21, PMO 1 1 1738, 9 nr above base of Myren Member (Hirnantian/Rhuddanian), Konglungo; internal
mould of pedicle valve, refigured from Baarli and Harper (1986, pi. 1, fig. n) and Baarli (1987, fig. 5/),
x 2-5.

Figs. 9-16, 20, 26. Chrustenopora askeriensis sp. nov. From Solvik Formation, Asker. 9, PMO 109737, 9 m
above base of Myren Member (earliest Rhuddanian), Konglungo; internal mould of pedicle valve, x 3.
10-12, 14, 15, from 17 m above base of Myren Member (earliest Rhuddanian), Vakas; 10, 12, 14, 15, PMO
109734, internal mould and latex cast of brachial valve, x 2; 11, PMO 109735, holotype, internal mould
of brachial valve, refigured from Baarli and Harper (1986, pi. 1, fig. /), x 2. 13, PMO 105891, basal metres
of Myren Member, Ostoya; internal mould of pedicle valve. 16, PMO 1 1 1734, 1 1 m above base of Myren
Member (earliest Rhuddanian), Konglungo; external cast of brachial valve, x2-5. 20, PMO 108281, 6 m
above base of Myren Member (Hirnantian/Rhuddanian), Vakas; holotype, internal mould of brachial
valve, refigured from Thomsen and Baarli (1982, pi. 1, fig. 2), x 3. 26, PMO 117379, 15 m above base of
Myren Member (earliest Rhuddanian), Spirodden; external mould of brachial valve showing punctae, x6.

Figs. 17, 22, 23, 27-29. ? Diorthelasma semotum sp. nov. Solvik Formation, Asker. 17, 22, 23, 27, 28, top of
Leangen Member (mid-Aeronian); 17, PMO 111670(A), Skytterveien, Asker, holotype, internal mould of
brachial valve; 22, 23, PMO 1 1 1670(B), latex cast of brachial valve; 27, PMO 109759 (counterpart of PMO
1 1 1670), Skytterveien, Asker, exterior of brachial valve; 28, PMO 105199, Bleikerveien, internal mould of
brachial valve. 29, PMO 105198, 1 m above base of Leangen Member (earliest Aeronian), Skytterveien;
internal mould of pedicle valve. All x 6.

Figs. 24, 25, 30 33. Salopina pumila sp. nov. From base of Vik Formation (Telychian), Kampebraten,
Sandvika. 24, PMO 109763, internal mould of pedicle valve, x 3. 25, 31, 32, PMO 109765, internal mould
and latex cast of brachial valve, x 3. 30, PMO 1 1 1691, holotype, internal mould of brachial valve, x 4.
33, PMO 1 1 1692, internal mould of pedicle valve, x 3.

PLATE 99

BAARLI, Resserella , Jezercia , Chrustenopora , IDiorthelasma, Salopina

1124 PALAEONTOLOGY, VOLUME 31

Salopina pumila sp. nov.

Plate 99, figs. 24, 25, 30 33

Holotype. PMO 111691 (PI. 99, fig. 30): mould of brachial valve from base of Vik Formation (Telychian)
at Kampebraten in Sandvika (NM 848404).

Material. PMO 109763 109765, 111664, 111665, 111668, 111669, 111691, 111692, 111694, 116773, 1 17412
1 17416: internal moulds of thirteen brachial and three pedicle valves from base of Vik Formation (Telychian)
in Sandvika area of Oslo Region.

Diagnosis. Planoconvex, nonsulcate, and subcircular small Salopina with very fine ribbing. Ventral
muscle field faintly impressed and divided by shallow, median elevated area. Thin, erect, triangular,
and slightly divergent brachiophores with small sockets floored by fulcral plates.

Description. Exterior: planoconvex to weakly ventribiconvex small Salopina. Outline subcircular. Hinge line
straight, two-thirds of maximum width, which occurs posterior to mid-length. Cardinal angle obtuse and
rounded, lateral and anterior margins evenly rounded. Ventral beak gently rounded, may be nearly straight,
and projects slightly past hinge line. Ventral interarea apsacline, dorsal interarea anacline. No sulcus.
Ornament very fine with seven costellae per mm at 5 mm growth stage. Costellae low, rounded, and equally
spaced.

Interior of pedicle valve: delthyrial chamber fairly low and flat. No pedicle callist present. Teeth strong,
triangular without cruralfossetts and with deep and short lateral cavities. Short, concave dental plates, which
may continue in very weak, short muscle-bounding ridges. Muscle field weakly impressed, most often divided
by low, broad median ridge anteriorly. The cordate muscle field occupies one-third to half of total length
and two-fifths of width.

Interior of brachial valve: brachiophores thin and erect. Bases of supporting plates slightly divergent
relative to tops. Brachiophores with support are thin, delicate, and sharply triangular in lateral view; they
descend normal to floor anteriorly and diverge at c. 50° from one another. Cardinal process has very short,
thin shaft and small, slightly bulbous myophore. Small sockets floored by fulcral plates. Muscle field weakly
impressed, bounded laterally by elevated ridges that continue from brachiophore bases without sharp flexure.
Muscle field quadripartite, with small posterior muscle fields and bigger anterior scars; may be longitudinally
bisected by broad, low median ridge occupying one-third of width of muscle scars. Muscle scars half as long
and wide as total length and width of valve.

Discussion. S. pumila sp. nov. is close to S. conservatrix (McLearn, 1924) and S. shelvensis
Walmsley, Boucot and Harper, 1969, of equivalent age. It is distinguished from the former by the
lack of a dorsal sulcus, and diverging rather than subparallel brachiophore bases; from the latter
by an impressed median ridge in the pedicle valve; and from both by finer ribbing.

Family chrustenoporidae fam. nov.

Diagnosis. Biconvex to unequally biconvex, transverse elliptical Enteletacea. Low, apsacline ventral
and anacline dorsal interarea. Fascicostellate to multicostellate. Pedicle muscle field oval to
pentagonal with elongate, simple diductor scars bordering, but not enclosing or going beyond
broad adductor tract anteriorly. Very short cardinalia consist of simple ridge-like to lobed cardinal
process, strongly divergent brachiophores. Very short brachiophore bases extend anteriorly parallel
to subparallel with continuation and do not contain adductor fields anteriorly. Fulcral plates often
present, concave and flooring small dental sockets. Variably developed dorsal muscle field with
oblique transverse ridges. Differs from both Dalmanellidae and Draboviidae in its very short
cardinalia. The latter possesses long brachiophores, directed anteroventrally with high perpendicular
supporting plates, as opposed to short brachiophores, very divergent anteroposteriorly with
extremely short supporting plates. Dalmanellidae differs in having less convex dorsal valve and
cordate ventral muscle field, with diductor scars usually extending beyond adductor scars.
Continuation of brachiophore bases is variably disposed.

BAARLI: LLANDOVERY ENTELETACEAN BRACHIOPODS

1 125

Genera assigned to Chrustenoporidae. Dysprosorthis Rong, 1984, late Ashgill, China, England, and ?Anti-
Atlas, Morocco; Chrustenopora Havlicek, 1968, Upper Ordovician to early Llandovery, Bohemia, Norway;
Jezercia Havlicek and Mergl, 1982, Ashgill to Llandovery, Bohemia, Norway, and England.

Discussion. Chrustenopora and the closely related Jezercia were placed in the subfamily Chrustenopo-
renae under the family Plectorthidae by Havlicek and Mergl (1982). Since the Norwegian material
is endopunctate and Havlicek described C. imhricata as possibly punctate, I choose to move them
to Enteletacea, together with the closely related Dysprosorthis , and to elevate the subfamily to
family rank. Havlicek and Mergl (1982) used the presence of concentric perforations to erect the
subfamily Chrustenoporenae. Such perforations are not found in the Norwegian material, which
might be due to poor preservation — but this seems unlikely since punctae are observed.

Genus chrustenopora Havlicek, 1968

Type species. By original designation, Chrustenopora imhricata Havlicek, 1968, p. 123, pi. 1, figs. 5-8, 10.

Chrustenopora askeriensis sp. nov.

Plate 99, figs. 9-16, 20, 26

1982 Ptychopleurella sp.; Thomsen and Baarli, pi. 1, figs. 2 and 3.

1986 Kinnella sp.; Baarli and Harper, pi. 1, figs. i,j, p, q.

Holotype. PMO 108281 (PI. 99, fig. 20): brachial valve from 6 m above base of Myren Member
(Hirnantian/Rhuddanian), Solvik Lormation at Vakas in Asker (NM 828357).

Material. PMO 103494, 105891, 108281, 109734, 109735, 109738, 111714, 111734, 111747, 111749, 111750,
117379, 117380, 117388-1 17399: moulds of fourteen pedicle and ten brachial valves and two whole shells
from Myren Member (Hirnantian/Rhuddanian), Solvik Lormation, at Vakas (NM 828357), Spirodden (NM
841338), and Konglungo (NM 849347) in Asker.

Diagnosis. Ventribiconvex and transversely subcircular Chrustenopora with moderately fine,
imbricated costellae. No fold or sulcus. Pedicle valve with thin, long dental plates and small teeth.
Broad, triangular adductor area flanked by narrow, simple diductor scars. Dorsal valve has large
sockets and sharp triangular dorsal median ridge with variable position of brachiophore bases
about parallel. Muscle-bounding ridges in brachial valve strongly impressed, curved, and start near
hinge line.

Description. Exterior: ventribiconvex and transverse, subcircular to subquadrate, from two-thirds as long as
wide to nearly equal. Hinge line straight and nine-tenths of to equal maximum width, which varies between
10 and 14 mm. Cardinal angle normal to obtuse. Lateral margins straight to slightly curved for half of length,
then curve evenly with anterior margin. May be faintly sulcate. Beak straight, overhanging lunge line. Ventral
interarea high, dorsal interarea relatively low. Delthyrium open. Costellate to fascicostellate imbricated
ribbing with three to five ribs per mm at 5 mm growth stage. Endopunctate shell.

Interior of pedicle valve: thin, convex, and long dental plates continue in strong ridges. These meet anterior
margin of thickened floor of delthyrial chamber, and together delimit the semicircular to oval muscle scars.
Anterior thickening sometimes crenulated. Teeth small. No pedicle callist. Muscle scars occupy one-third of
length of valve. Diductor scars long and narrow. Adductor area long, triangular, and somewhat wider than
one diductor scar. Adductor area may be coarsely striated. Vascula media widely divergent.

Interior of brachial valve: brachiophores with supporting plates are thick, erect, and very divergent. Sockets
well developed and floored by strong fulcral plates that converge on to brachiophores. Brachiophore bases
very short, and vary from nearly parallel to convergent towards median septum. Crural pits often prominent.
Cardinal process thin and bilobed with shaft continuing into sharp, triangular median septum. Septum
continues to anterior end of muscle scars in most specimens. Muscle-bounding ridges broad and strongly
impressed, and start outside brachiophore bases near hinge line; they follow outline of long, subtriangular
posterior muscle scars and diverge anterolaterally in long curved extensions. Larger anterior scars more
faintly impressed, long and flabellate, and lie well inside above mentioned extensions and posterior scars.
Muscle scars occupy one-third to half of total valve length. Laint transverse ridges slope obliquely
anterolaterally.

1126

PALAEONTOLOGY, VOLUME 31

Discussion. Except for the concentric perforations, all characteristics of the genus are present. The
differences which justify erection of a new species include: 1, larger sockets; 2, deeper median ridge
in the brachial valve; and 3, more triangular adductor area in the pedicle valve compared with C.
imbricate i Havlicek, 1968. Brachiophore bases vary from parallel to convergent, and are not
divergent as in C. imbricata. The lateral muscle-bounding ridges in the brachial valve are curved
in the new species, not straight.

Genus jezercia Havlicek and Mergl, 1982

Type species. By original designation, Jezercia ostiaria Havlicek and Mergl, 1982, Kraluv Dvur Formation,
Bohemia.

Jezercia rongi sp. nov.

Plate 99, figs. 2-8, 18, 19, 21

1986 Reuchella sp.; Baarli and Harper, pi. I, fig. 1 1 m, n.

1987 Reuchella sp.; Baarli, fig. 5/.

1987 Ravozetina rava silvicola (Temple, 1970); Temple, pi. 3, fig. 13.

Holotype. PMO 109732 (PI. 99, figs. 3, 4, 7, 8): internal mould of brachial valve from 1 1 m above base of
Myren Member (earliest Rhuddanian), Solvik Formation at Vakas in Asker (NM 828357).

Material. PMO 103495, 109732, 111707, 111713, 111715, 111727, 111738, 111748, 117401 117406: moulds
of seven brachial, four pedicle, and one of exterior valves, and one whole specimen, from the lowest 20 m
of the Solvik Formation at Vakas (NM 828357), Konglungo (NM 849347), and Spirodden (NM 841338).

Diagnosis. Biconvex and transverse elliptical outline with angular fold in pedicle and sulcus in
dorsal valve. Coarsely subangular ribs in fascicostellate pattern. Thin, short dental plates and
faintly impressed pentagonal muscle scars. Divergent vascula media. Brachiophores very short and
supported by even shorter, variably directed supporting plates. Cardinal process with shaft that
continues into myophore and further into median sulcus. Subcircular dorsal muscle field divided
by oblique ridges.

Description. Exterior: subequally biconvex to ventribiconvex. Transverse elliptical to subquadrate outline.
Maximum width 6 14 mm measured slightly in front of hinge line. Cardinal angle normal to obtuse, lateral
margins straight to slightly curved in posterior half of valve, then evenly curved like anterior margin. Valve
half as thick and eight-tenths as wide as long. Hinge line straight and eight-tenths of maximum width. Small
angular sulcus in brachial valve varies in depth and width. Ventral interarea apsacline with low anacline
interarea dorsally. Delthyrium open and widely triangular. Notothyrium open. Ribbing coarsely fascicostellate
with one or two subangular ribs per mm at 5 mm growth stage and separated by deep grooves. Punctate.

Interior of pedicle valve: dental plates short, thin, and divergent at c. 60°; they continue into muscle-
bounding ridges that delimit subcircular to subpentagonal muscle fields. No pedicle callist. Muscle scars very
faintly impressed and occupying one-third of total length. Narrow diductor scars separated by very broad
adductor area that continues slightly anterior to diductor scars.

Interior of brachial valve: brachiophores short, erect, and diverge at 90-100°. Very short brachiophore
bases vary about parallel from clearly divergent to strongly convergent. Small sockets floored by strong,
concave fulcral plates attached to hinge line. Cardinal process seems to vary from simple, thin ridge to
possibly trilobed form, with shaft continuing in strong myophragm extending half valve length or continuing
in sharp fold out to edge of valve. Muscle field quadripartite with elongate quadrate posterior scars slightly
smaller than subtriangular anterior adductor scars. Muscle scars occupy one-third of maximum width and
less than half of maximum length. Short transverse ridges run obliquely anterolaterally to each side. Muscle-
bounding ridges variably impressed; they curve and bound muscle scars anteriorly.

Discussion. The new species differs from J. ostiaria Havlicek and Mergl, 1982, in its lack of coarse
concentric perforation, its less transverse shape, and its coarser, more angular ribbing and fold.
The adductor field in the pedicle valve is broader relative to the diductor scars, and the dorsal
muscle scars are less elongate.

BAARLI: LLANDOVERY ENTELETACEAN BRACHIOPODS

1127

Entelatacea indet. sp. A

Text-fig. 4e-h

Material. PMO 103498, 105216: one pedicle and one brachial valve from basal part of Spirodden Member
(late Rhuddanian) at Spirodden (NM 841338) and middle part of Leangen Member (early Aeronian) at
Skytterveien (NM 820339) in Asker.

Discussion. These specimens may belong to different species, but they agree in size, shape, and
ribbing.

Acknowledgements. I thank the staff and students of the Palaeontological Museum (University of Oslo),
especially David Worsley, for their help in the field and laboratory. I am very grateful to Rong Jia-Yu
(Nanjing Institute of Geology and Palaeontology, Academica Sinica) who offered constructive criticism and
useful suggestions regarding the taxonomy. Financial support from the Norwegian Research Council for
Science and Humanities (NAVF) is gratefully acknowledged. I thank Markes E. Johnson (Williams College)
for help with the English and general encouragement. This paper is Paleontological Contributions from the
University of Oslo, no. 343.

REFERENCES

amsden, t. w. 1968. Articulate brachiopods of the St Clair limestone (Silurian), Arkansas, and Clarita
Formation (Silurian), Oklahoma. J. Paleont. 42 (pt. 2 of 2, suppl. to No. 3), 1 117, pis. 1 20.

baarli, b. G. 1985. The stratigraphy and sedimentology of the early Llandovery Solvik Formation, central
Oslo Region, Norway. Norsk geol. Tidsskr. 65, 229-249.

1987. Benthic faunal associations in the Lower Silurian Solvik Formation of the Oslo Asker Districts,
Norway. Lethaia, 20, 75-90.

In press. Bathymetric co-ordination of proximality trends and level-bottom communities: a case study

from the Lower Silurian of Norway. Palaios.

and harper, d. a. t. 1986. Relict Ordovician brachiopod faunas in the Lower Silurian of Asker, Oslo

Region, Norway. Norsk geol. Tidsskr. 66, 87-98, pis. 1 and 2.

— and Johnson, m. e. In press. Biostratigraphy of selected brachiopods from the Llandovery Series (Lower
Silurian) of the Oslo Region. Ibid.

BANCROFT, B. B. 1928. On the notational representation of the rib-system in Orthacea. Mem. Proc. Manchr
lit. phil. Soc. 72, 53-90, pis. 1-3.

1949. In lamont, a. (ed.). Welsh Valentian brachiopods and the Strophomena antiquata group of fossil

brachiopods, 16 pp., 3 pis. Privately printed, Mexborough.

barrande, j. 1848. Uber die Brachiopoden der silurischen Schichten von Bohmen. Naturwiss. Abh. Berl. 2,
155-256.

1879. Systeme silurien du Centre de la Boheme, V, 226 pp., 153 pis. Praha and Paris.

bassett, m G. and cocks, l. r. m. 1974. A review of Silurian brachiopods from Gotland. Fossils Strata , 3,
1-56, pis. 1-11.

bergstrom, J. 1968. Upper Ordovician brachiopods from Vastergotland, Sweden. Geologica Palaeont. 2,
1-21, pis. 1-7.

billings, e. 1866. Catalogues of the Silurian fossils of the Island of Anticosti, with descriptions of some new
genera. Geol. Surv. Can. 99 pp.

boucot, a. j. 1960. Lower Gedinnian brachiopods of Belgium. Mem. Inst. geol. Univ. Louvain, 21, 279-344,
pis. 9-18.

and Johnson, j. g. 1964. Brachiopods of the Ede Quartzite (Lower Llandovery) of Norderon, Jamtland.

Bull. geol. Instn Univ. Upsala, 42, 79-101, pis. 1-4.

harper, c. w. and walmsley, v. g. 1966. Silurian brachiopods and gastropods of southern New

Brunswick. Bull. geol. Surv. Can. 140, 1-45, pis. 1-18.

and walmsley, v. g. 1965. Revision of the Rhipidomellidae (Brachiopoda) and the affinities of
Mendacella and Dalejina. J. Paleont. 39, 331-340, pis. 45 and 46.

brenchley, p. j. and cocks, l. r. m. 1982. Ecological associations in a regressive sequence: the latest
Ordovician of the Oslo Asker district, Norway. Palaeontology , 25, 783-815.

cloud, p. e. 1948. Dicaelosia versus Bilobites. J. Paleont. 22, 373-374.

1128

PALAEONTOLOGY, VOLUME 31

cocks, L. R. m. 1982. The commoner brachiopods of the latest Ordovician of the Oslo- Asker District,
Norway. Palaeontology , 25, 755-781, pis. 78-84.

and baarli, b. G. 1982. Late Llandovery brachiopods from the Oslo Region. In worsley, d. (ed.). Field
meeting, Oslo Region 1982. Paleont. Contr. Univ. Oslo , 278, 79-89, pis. 1-3.

— and rong jia-yu 1988. A review of the late Ordovician Foliomena brachiopod fauna with new data
from China, Wales, and Poland. Palaeontology , 31, 53-67, pis. 8 and 9.

cooper, G. a. 1930. New species from the Upper Ordovician of Perce. Am. J. Sci. 20, 265-288, pis. 1-3.

1956. Chazyan and related brachiopods. Smithson, misc. Colins , 127, 1-1245, pis. 1-269.
dalman, j. w. 1858. Upstalning och Beskriftning af de i Sverige funne Terebratuliter. K. svenska VetenskAkad.
Hand!, (for 1857), 85 155, pis. 1-6.

davidson, t. 1869. A Monograph of the British Fossil Brachiopoda. Part 7. No. 3. The Silurian Brachiopoda.

Palaeontogr. Soc. [ Monogr .], 169-248, pis. 23-37.
hall, j. 1857. Description of Paleozoic fossils. Ann. Rep. NY St. Cabinet nat. Hist. 10, 411 86.

— and clarke, j. m. 1892. An introduction to the study of the genera of Paleozoic Brachiopoda.
Paleontology of New York , 8(1), 1 16, I 367, 41 pis.

harper, c. w., boucot, a. j. and walmsley, v. G. 1969. The rhipidomellid brachiopod subfamilies
Heterorthinae and Platyorthinae (new). J. Paleont. 43, 74 92, pis. 15-17.
havlicek, v. 1950. The Ordovician Brachiopoda from Bohemia. Rozpr. ustred. Ust. geol. 13, 1 135, pis.
113.

— 1968. New brachiopods in the Lower Caradoc of Bohemia. Vest, ustred Ust. geol. 43, 123-125.

— 1970. Heterorthidae (Brachiopoda) in the Mediterranean Province. Sb. geol. Ved ., Rada G, 12, 7-39,
pis. 111.

— 1974. New genera of Orthidina (Brachiopoda) in the Lower Paleozoic of Bohemia. Vest, ustred. (Jst.
geol. 49, 167-170.

— 1977. Brachiopods of the order Orthida in Czechoslovakia. Rozpr. ustred. Ust. geol. 49, 1-327, pis.
1-56.

— 1982. Ordovician in Bohemia: development of the Prague Basin and its benthic communities. Sb. geol.
Ved. Geol. 37, 103-136, pis. 1-8.

— kriz, j. and serpagli, e. 1986. Upper Ordovician brachiopod assemblages of the Carnic Alps, Middle
Carinthia and Sardinia. Boll. Soc. paleont. ital. 25, 277-31 1.

— and mergl, m. 1982. Deep water shelly fauna in the latest Kralodvorian (upper Ordovician, Bohemia).
Vest, ustred. Ust. geol. 47, 37-46.

king, w. 1850. A Monograph of the Permian Fossils of England. Palaeontogr. Soc. [Monogr.]. xxxvii + 258
pp., 29 pis.

kozlowski, r. 1929. Les brachiopodes gothlandiens de la Podolie polonaise. Palaeont. pol. 1, 1-254, pis. I 12.
lazarev, s. s. 1970. Morfologija i sistema brachiopod nadsemejstva Enteletacea. Paleontologiya Stratigr. 128,
31 pp.

LINNAEUS, c. 1758. Systema naturae , sive Regna tria Naturae systematice proposita per Classes , Ordines , Genera
et Species. 10th edn. Tom. I: Regnum animale, 824 pp. Stockholm.
mclearn, f. h. 1924. Palaeontology of the Silurian rocks of Arigsaig, Nova Scotia. Mem. geol. Surv. Can.
137, 1 180, pis. 1-30.

Murchison, R. I. 1839. The Silurian System, founded on geological researches in the counties of Salop , Hereford ,
Radnor , Montgomery, Caermarthen, Brecon , Pembroke, Monmouth, Gloucester, Worcester, and Stafford;
with descriptions of the coalfields and overlying formations, xxxii + 768 pp., 37 pis. London.
musteikis, p. and puura, i. 1983. The brachiopod genus Dicoelosia from the Baltic Silurian. Eesti NSV Tead.
Akad. Toim. (Geol.), 32, 138-146.

pander, c. h. 1830. Beitrdge zur Geognosie des Russichen Reiches, 165 pp., 31 pis. St Petersburg.

OPIK, a. A. 1933. Uber Plectanrboniten. Acta Comment. Univ. tartu. (A) 24, 1 79, pis I 12.
reed, f. r. c. 1917. The Ordovician and Silurian Brachiopoda of the Girvan District. Trans. R. Soc. Edinb.
51, 795 998, pis. 1-24.

rong jia-yu. 1984. Brachiopods of latest Ordovician in the Yichang district, Western Hubei, Central
China. In Strat. and Palaeont. of Systemic Boundaries in China. Ordovician- Silurian Boundary , 1, 111 176,
pis. 1-14.

rubel, m. 1971. Taxonomy of dicoelosiid brachiopods from the Ordovician and Silurian of east Baltic.
Palaeontology , 14, 34-60, pis. 8 10.

salter, j. w. 1851. List of some of the Silurian fossils of Ayrshire. Q. Jl geol. Soc. Lond. 7, 170 178, pis.
8-10.

BAARLI: LLANDOVERY ENTELETACEAN BRAC'HIOPODS

1 129

schuchert, c. 1913. Systematic paleontology of the Lower Devonian deposits of Maryland. Brachiopoda.
Md geol. Surv. stratigr. Mem. 290 449, pis. 53-74.

and cooper, g. a. 1931. Synopsis of the brachiopod genera of the suborder Orthoidea and Pentameroidea,
with notes on the Telotrenrata. Am. J. Sci. 22, 241 251.

1932. Brachiopod genera of the suborder Orthoidea and Pentameroidea. Mem. Peabody Mas. Yale ,
4, I -270, pis. A, 1-29.

spjeldnaes, N. 1961. Ordovician climatic zones. Norsk geol. Tidsskr 41, 45 77.

temple, J. T. 1987. Early Llandovery brachiopods of Wales. Palaeontogr. Soc. [Monogr.], 137 pp., 15 pis.
thomsen, e. and baarli, b. G. 1982. Brachiopods of the Lower Llandovery Sselabonn and Solvik formations
of the Ringerike, Asker and Oslo districts. In worsley, d. (ed . ). Lield meeting, Oslo Region 1982. Paleont.
Contr. Univ. Oslo , 278, 63-78.

twenhofel, w. H. 1928. Geology of Anticosti Islands. Mem. geol. Surv. Can. 154, 1 481, pis. 1 60.
waagen, w. H. 1884. Salt Range fossils. Part 4. Brachiopoda. Mem. geol. Surv. India Palaeont. indica , 1,
547 610, pis. 50-57.

walmsley, v. G. and bassett, M. G. 1976. Biostratigraphy and correlation of the Coralliferous Group and
Gray Sandstone Group (Silurian) of Pembrokeshire, Wales. Proc. Geol. Ass. 87, 191 220, pis. I 3.

— and boucot, a. 1971. The Resserellinae— a new subfamily of Late Ordovician to Early Devonian
dalmancllid brachiopods. Palaeontology , 14, 487 531, pis. 91 102.

1975. The phylogeny, taxonomy and biogeography of Silurian and early to mid Devonian Isorthinae
(Brachiopoda). Palaeontographica, (A), 148, 34 108, pis. 1 10.

and harper, c. w. 1969. Silurian and Lower Devonian salopinid brachiopods. J. Paleont. 43, 492
516, pis. 71-80.

whittard, w. F. and barker, g. h. 1950. The Upper Valentian brachiopod fauna of Shropshire, Part I.

Inarticulata: Articulata, Protremata, Orthoidea. Ann. Mag. nat. Hist. 3, 553-590, pis. 5-8.
williams, a. 1951 Llandovery brachiopods from Wales with special reference to the Llandovery district. Q.
Jl geol. Soc. Lond. 107, 85 136, pis. 3 7.

worsley, d., aarhus, n., bassett, m. g., howe, m. p. a., mork, a. and olaussen, s. 1983. The Silurian
succession of the Oslo Region. Norg. geol. unders. 384, 1 -57.
wright, a. d. 1968m The brachiopod Dicoelosia biloba (Linnaeus) and related species. Ark. Zool. 20, 261
319, pis. I 7.

— 1968/). A new genus of dicoelosiid brachiopod from Dalarne. Ibid. 22, 127 138. pi. 1.

ziegler, a. m., cocks, L. r. m. and mckerrow, w. s. 1968. The Llandovery transgression of the Welsh
Borderland. Palaeontology , II, 736 782.

B. GUDVEIG BAARLI

Department of Geology
Williams College
Williamstown
Massachusetts 01267
USA

Typescript received 1 1 September 1987
Revised typescript received 18 May 1988

THE PALAEONTOLOGICAL ASSOCIATION

ANNUAL REPORT OF COUNCIL FOR 1987

Membership and Subscriptions. Membership totalled 1,313 on 31 December 1987, a decrease of 55 over the
previous year. There were 830 Ordinary Members, a decrease of 39; 71 Retired Members, an increase of 1;

1 18 Student Members, a decrease of 5; and 294 Institutional Members, a decrease of 12. Total individual and
institutional subscriptions to Palaeontology through Basil Blackwell’s agency numbered 395, a decrease of 47.

Subscriptions to Special Papers in Palaeontology numbered 118 individuals, an increase of 5; and 114
Institutions, a decrease of 8. Standing orders for Special Papers through Blackwell’s agency numbered 51
volumes (for each of Special Papers 37 and 38).

Sales of backparts of Palaeontology via the Membership Treasurer realized £594. Sales of backparts of
Palaeontology to Institutional Members yielded £93 50. Sales of back numbers of Special Papers in Palaeon-
tology to individuals yielded £1,168-50, and to Institutions, £158 plus US $228. Twenty copies of various
Special Papers were sold by the Marketing Manager yielding £237-65 plus US $14. Sales of Fossil Plants of
the London Clay through him numbered 40 copies, yielding £228-95 and 28 copies of the Fossils of the Chalk
were sold, yielding £278-80. He also sold 2 copies of the PADS Guide, yielding £4 and 4 copies of the Malta
Field Guide, yielding £8-80. Sales of the Atlas of Invertebrate Macrofossils yielded £1 ,610 nett. Sales of Palstat
yielded £118 in royalties. Many of the above sales by the Marketing Manager were made at the Annual
Meeting at Bristol, where his stall yielded a healthy total of £500-70.

Finance. During 1987 the Association published Volume 30 of Palaeontology at an estimated cost of £62,590
(including postage and distribution). Special Papers 37 and 38 were published at a cost of £4,570 and £6,020
respectively. The Association is grateful to all those who made donations to offset the cost of publishing
Palaeontology.

Publications. Volume 30 of Palaeontology , published in four parts during 1987, contained 868 pages and 91
plates. Special Papers 37 (Biology and revised systematics of some late Mesozoic stromatoporoids , by Rachel
Wood), and 38 ( Taxonomy , evolution , and biostratigraphy of late Triassic-early Jurassic calcareous nannofossils,
by Paul R. Bown) were published in November. In addition to the journals and Circular, publications have
included the latest addition to the Field Guides to Fossils, on Fossils of the Chalk , published in December;
and, earlier in the year, the Palstat software package for statistical handling of palaeontological data was
produced by Lochee Publications on behalf of the Association.

Meetings. Twelve meetings were held in 1987. The Association is indebted to the organisers, hosts and field
leaders of these.

a. Joint meeting with the Society for Experimental Biology on ‘Biomechanics in Evolution’, held on 6 7
January, at the University of Manchester, and convened by Dr J. M. V. Rayner. This was one of several
thematic symposia at the SEB general meeting.

b. Review Seminar on ‘Major events in the evolution of land vegetation’, held on 1 1 February at King’s
College, London (KQC), and convened by Dr M. E. Collinson. About 80 people attended.

c. Lyell Meeting , held jointly with the Geological Society, on ‘Catastrophes and the history of life’, at
Burlington House, Piccadilly, London, on 25 February, convened by Dr L. B. Halstead. Over 250 people
attended the meeting.

d. Thirtieth Annual General Meeting , held in the Lecture Theatre of the Geological Society of London, on 9
March. Dr Alan J. Charig delivered the Annual Address, on ‘Ornithischian dinosaurs evaluate cladistic
method— the proof of the pudding’ The Sylvester-Bradley Award was made to Dr S. K. Donovan.
Following the meeting, a 30th anniversary dinner was held at the Savile Club, London, W. 1 .

e. Joint meeting with the Systematics Association and the Linnean Society , on ‘The phylogeny and classifica-
tion of the tetrapods’, held on 26-27 March, at the Linnean Society, Burlington House, Piccadilly,
London, and convened by Dr M. J. Benton. Over 100 people attended.

1 132

THE PALAEONTOLOGICAL ASSOCIATION

f. Field Meeting, organized by the Carboniferous Group to Cork and Waterford counties, Ireland, led by
Professor G. D. Sevastopulo and Dr A. D. Sleeman, during 3-6 April. A party of 36 people participated.

g. ‘ Progressive Palaeontology meeting, held at University College London, on 27 May, was convened by
Gloria Lee. Some 40 people were present for this open meeting for presentations by research students.

h. Field Meeting to localities and museums exhibiting extraordinary fossil biotas, and the Nordlinger Ries
impact crater, in West Germany, led by Dr M. E. Collinson and Dipl.-Geol. Kurt Goth, during 25-29
September. There were 31 participants.

i. Joint meeting with the Geological Society and the Geological Curators' Group, on ‘The use and conservation
of palaeontological sites’, held on 1-2 October at Burlington House, convened by Dr P. R. Crowther,
Dr W. A. Wimbledon, and M. F. Stanley. The meeting was attended by about 80 people. The proceedings
are to be published in 1988 as a Special Paper in Palaeontology.

j. Review Seminar on ‘Triassic vertebrates and faunal replacement’, held on 1 1 November at the University
of Cambridge, and convened by Dr M. J. Benton and Dr N. C. Fraser. The meeting was attended by
some 50 people.

k. Joint meeting with the Linnean Society and the Geological Society, on ‘Biota and palaeo-atmospheres’,
held on 17 December at Burlington House, Piccadilly, London, and convened by Professor W. G.
Chaloner, and Dr L. R. M. Cocks. Over 70 people attended the meeting.

/. The Annual Conference, held at the University of Bristol, on 17-20 December, was an open meeting,
attended by over 160 people. The Local Secretary was Dr D. E. G. Briggs, and the field trips were led by
Professor R. J. G. Savage, Professor D. T. Donovan, and Dr V. P. Wright. The President’s Award was
shared by A. King and A. Williams.

Council. The following members served on Council following the Annual General Meeting on 9 March 1987:
President , Dr L. R. M. Cocks; Vice-Presidents , Dr D. E. G. Briggs, Dr L. B. Halstead; Treasurer, Dr M.
Romano; Secretary, Dr P. W. Skelton; Membership Treasurer , Dr A. T. Thomas; Institutional Membership
Treasurer, Dr A. W. Owen, Editors, Dr D. Edwards, Dr P. A. Selden, Dr P. R. Crowther, Dr T. J. Palmer,
Dr C. R. C. Paul, and Dr M. J. Benton (also Public Relations Officer); Marketing Manager, Dr V. P. Wright;
Circular Reporter, Dr Derek J. Siveter; Other Members, Dr M. E. Collinson, Dr H. A. Armstrong, Dr P. D.
Taylor, Dr J. A. Crame, Dr G. B. Curry, and Professor B. M. Funnell (co-opted).

Circulars. Four Circulars, numbers 127-130, were distributed to Ordinary, Student, and Retired Members,
and, on request, to over 90 Institutional Members.

Council Activities. The Association’s 30th Anniversary has been marked by a busy and eventful year.
Council has been actively promoting the longer term interests of palaeontology in the UK in submissions to
the continuing University Grants Committee review on university provision for the Earth Sciences, and the
Royal Society/Natural Environment Research Council survey on Earth Science research priorities to the year
2000. Complementary submissions have also been made by the Joint Committee for Palaeontology. Council
members have also been active in corresponding with suitable bodies over several matters of palaeontological
interest, including the fate of the Messel pit in West Germany, with its exceptionally well-preserved biota, and
the genuineness of the Archaeopteryx fossils.

A large number of meetings has been arranged, many of them in conjunction with other bodies, such as the
Geological Society, Linnean Society, Geological Curators’ Group, Society for Experimental Biology, and
Systematics Association, testifying to the healthy level of liaison between palaeontologists and other specialists.
The longer than usual autumn field trip, to West Germany, was a particular highlight of the anniversary year,
providing the appreciative participants with an opportunity to visit the Messel pit mentioned above, as well
as many other spectacular sites.

Financial assistance was given to the Fourth International Fossil Algae Symposium held in Cardiff, in July,
in the form of a loan for the preparation of a field guidebook, and a further contribution from the Conservation
Fund was awarded to the West London Wildlife Group for their work at the Lower Writhlington tip site in
Avon.

A full programme of future meetings is planned, including the Lyell Meetings on ‘Palaeocomputing:
Keyboard to the past’ in 1988, and on palaeoclimatology, in 1989, as well as other joint meetings with the
Geological Society, on ‘Origins and evolution of the Antarctic biota’ and ‘Palaeozoic biogeography and
palaeogeography’, both in 1988, and a variety of Review Seminars and Field Meetings. Current publishing
projects include an ‘Encyclopaedia of Palaeobiology', to be published in conjunction with Blackwell Scientific,
and further volumes in the Field Guides to Fossils series, on Permian Zechstein Reef fossils and Coal Measures
plants.

BALANCE SHEET AND ACCOUNTS FOR THE
YEAR ENDING DECEMBER 1987

Balance Sheet as at 31 December 1987

1986

£

Investments at Cost (see schedule)
Current Assets

£

2,829

Sundry debtors ......

2,087

76,330

Cash at bank .......

83,369

1,974

Sylvester-Bradley Fund ....

1,885

450

Loans ........

450

81,583

87,791

Current Liabilities

2,854

Subscriptions received in advance
Provision for cost of publication of:

3,450

16,000

Palaeontology ......

17,324

16,291

Special Papers ......

11,465

2,973

Sundry creditors ......

3,669

38,118

35,908

43,465

£

52,359

51,883

£95,382

£104,242

69,917

Represented by:

Publications Reserve Account

Balance brought forward ......

90,777

20,860

Excess of income over expenditure for the year transferred
from Income and Expenditure Account

8,949

Sylvester-Bradley Fund

2,029

Balance brought forward .......

1,974

145

Interest ..........

1 1 1

(200)

Grant awarded ........

(200)

Meeting Reserve ........

99,726

1,885

2,631

£95,382

£104,242

Income and Expenditure Account

INCOME

1986

£

Subscriptions

1987

1986

41,840

Palaeontology
Sales .....
Donations ....

35.180

Special Papers

Sales .....
Donations ....

8,901

Burgess Shale Portfolio
Sales .....

169

Fossil Plants of the London Cla y
Sales .....

306

Atlas of Invertebrate Macrofossils
Sales .....

1,954

Palstat

Income .....
Expenses .....

303 Offprints .....

4,209 Profit on Sales of Investments

1 2. 1 80 Investment Income (see schedule) .

30 Sundry Income ....

£ £

41,996

673

42,669

33,799

831

34,630

7,048

7,048

88

88

340

340

1,560

1,560

118

(710)

(592)

(73)

480

12,617

50

£98,817

£105,072

for the Year Ended 31 December 1987

EXPENDITURE

1986

£ £ £

Cost of Publication of Palaeontology

Volume 30— Part 1 1 5,023

Part 2 16,286

Part 3 14,781

Part 4 1 7,324

Under provision for Volume 29, Part 4 . 3,267

57,744 67,681

Cost of Publication of Special Papers

No. 37 4,776

No. 38 6,689

Over provision for No. 36 ....... (44)

16,291 11,421

2,280 Warehousing of Publications ....... 2,666

200 Grants ........... 100

Cost of Circulars

Preparation .......... 2,298

Postage ........... 1,686

Credit (582)

3,481 3,402

Administrative Costs

Postage and stationery ......... 868

Editorial expenses ......... 424

Meeting expenses ......... 2,996

Audit fee .......... 250

Membership of Societies ........ 60

4,423 4,598

£84,212

£89,868

£20,860

Excess of Income over Expenditure for
to Publications Reserve Account

the Year

Transferred

£8,949

Income and Expenditure Account

INCOME

1986

£

Subscriptions

1987 ... -

1986 ....

41,840

Palaeontology

Sales

Donations ....

35.180

Special Papers
Sales

Donations

8,901

Burgess Shale Portfolio
Sales .....

169

Fossil Plants of the London Cla y
Sales ....

306

A tlas of Invertebra te Macrofossils
Sales

1,954

Palstat

Income .....
Expenses ....

303 Offprints

4,209 Profit on Sales of Investments

1 2. 1 80 Investment Income (see schedule)

30 Sundry Income ....

£105,072

£ £

41,996

673

42,669

33,799

831

34,630

7,048

7,048

88

88

340

340

1,560

1.560

118

(710)

(592)

(73)

480

12,617

50

£98,817

for the Year Ended 31 December 1987

EXPENDITURE

1986

£ £ £
Cost of Publication of Palaeontology
Volume 30— Part I ......... 15,023

Part 2 16,286

Part 3 14,781

Part 4 1 7,324

Under provision for Volume 29, Part 4 3,267

57,744 67.681

Cost of Publication of Special Papers

No. 37 4,776

No. 38 6,689

Over provision for No. 36 ....... . (44)

16,291 11.421

2,280 Warehousing of Publications ....... 2,666

200 Grants 100

Cost of Circulars

Preparation .......... 2,298

Postage ........... 1,686

Credit (582)

3,481 3,402

Administrative Costs

Postage and stationery ......... 868

Editorial expenses ......... 424

Meeting expenses ......... 2,996

Audit fee 250

Membership of Societies ..... 60

4,423 4,598

£84,212 £89,868

Excess of Income over Expenditure for the Year Transferred
£20,860 to Publications Reserve Account £8,949

Schedule of Investments and Investment Income as at 31 December 1987

Gross Income

Cost

£

1,991

for Year

£2,000

1 1% Exchequer Stock 1991 .........

t

220

£12,000

13^% Exchequer Stock 1987— Redeemed February 1987

-

795

£1,000

9% Treasury Stock 1992/1996 ........

992

90

£1,000

9% Treasury Stock 1994

955

90

£9,000

9% Treasury Stock 1994— Purchased February 1987 ....

8,988

405

£4,000

8% Treasury Stock 2002/2006 ........

2,192

320

£5,357

13£% Treasury Stock 1997 .........

5,000

710

£3,280

13^% Exchequer Stock 1996 .........

3,000

435

5,270

M. & G. Charifund Units

4,073

1,422

10,000

New Throgmorton Trust (1983) p.l.c. 25p Income Shares

1 ,706

474

1,400

Clarke, Nicholls & Coombs p.l.c. 25p Shares

668

61

6,180

M.E.P.C. p.l.c. 6j% Convertible Unsecured Loan Stock 1995/2000

4,943

402

374

M.E.P.C. p.l.c. 25p Shares

703

61

1,140

National Westminster Bank p.l.c. £1 Ordinary Shares ....

3,929

342

10,150

Agricultural Mortgage Corporation Ltd. 7|% Debenture Stock 1991/1993

8,251

787

1,460

Saatchi & Saatchi 6-3% Convertible Cumulative Redeemable Preference

£1 Shares ............

1,994

126

2,100

Hanson Trust p.l.c. 10% Convertible Unsecured Loan Stock 2007/2012

2,974

210

7,1 1 1

Bank Interest ...........

5,506

52,359

12,617

Market Value at 31 December 1987 (1986— £83,255) . . . .

£91,931

Report of the Auditor to the Members of
The Palaeontological Association

In my opinion, the Accounts as set out on pages 1 133-1 136 give a true and fair view of the state of the affairs
of the Association at 31 December 1987 and of its income and expenditure for the year ended on that date.

March 1988

Market Harborough, Leicestershire

G. R. Powell
Chartered Accountant

INDEX

Pages 1-235 are contained in Part 1; pages 237-566 in Part 2; pages 567-903 in Part 3; pages 905 1 141 in Part 4. Figures

in Bold Type indicate plate numbers.

A

Acanthostega , 699
Achomosphaera cmdalousiensis , 81
Acritarchs: Ordovician, China, 109
Adelosaunts, 958
Aegironetes sp., 8
Agraulos citicephalus , 594, 55
Ainoceras kamuy , 7
Alga: Carboniferous, England, 741

Allison, P. R. Taphonomy of the Eocene London Clay
biota, 1079

Amar aleator sp. nov., 139
Amhitryorf! sp., 642, 61
Amiculosphaera umbracula , 81

Ammonoids, analysis of heteromorphs by differential
geometry, 35; life orientation and ontogeny of hetero-
morphs, 281; Jurassic, Tibet, 295; heterochronic evolu-
tionary trends in the Namurian, 1033
Amphibian: Triassic, Australia, 857
Ampyx sp., 625, 58
Ampyxoides inermis , 19
Angiosperm: Palaeocene, Scotland, 503
Antarctica: Cretaceous bivalves, 341
Anthracosaur, 85
Aphelaspis sp. indet., 596
Archeria crassidisca , 85

Arthropod: Ordovician, Czechoslovakia, 611; Cambrian
chelicerate, Canada, 779
Aryballomorpha grootaertii, 113, 13, 14
Ash, S. R. See Miller, G. L. and Ash, S. R.

Astralochoma helenae sp. nov., 560, 50, 51
Athabascaella penika sp. nov., 114, 14; playfordii , 116, 15;
rossii , 118, 15

Australia: Triassic capitosaurid amphibian, 857; Cambrian
and Ordovician trilobites, 905

B

Baarli, B. G. The Llandovery enteletacean brachiopods of
the central Oslo Region, Norway, 1101
Bancroft, A. J. Palaeocorynid-type appendages in Upper
Palaeozoic fenestellid Bryozoa, 665
Bathynotus holopygus , 581, 52
Bathyuriscusl sp., 17
Bengtson, P. Open nomenclature, 223
Bird: Pleistocene 'swan-goose', Malta, 725
Birmanites lotus , 638, 61, 62

Bishop, J. D. D. Disarticulated bivalve shells as substrates
for encrustation by the bryozoan Cribrilina puncturata
in the Plio-Pleistocene Red Crag of eastern England,
237

Bivalve: Cretaceous, Antarctica, with a review of Mesozoic
Pholadidae, 341; Pliocene, Japan, 419; Isognomonidae,

Inocerarmdae, and Retroceramidae, 965; Crassostrea ,
Plio-Pleistocene and Recent, Jamaica, 1013
Bohemilla ( BohemiUa ) sp., 690, 68
Bolaspidella sp. indet., 594

Brachiopods: Late Ordovician, China, Wales and Poland,
53; Silurian, Norway, 1101
Brachyphyllum punctatum, 92
Brannerion vestitum , 2

Brazil: Cretaceous fish, 1; Silurian or Devonian fish, 771
Briggs, D. E. G. and Collins, D. A Middle Cambrian
chelicerate from Mount Stephen, British Columbia,
779

Brinkman, D. A weigeltisaurid reptile from the Lower
Triassic of British Columbia, 951
Bryozoa: encrusting bivalve shells, Plio-Pleistocene, Eng-
land, 237; Cretaceous, growth pattern and astogenetic
gradients, 519; Carboniferous, USA, 551; Palaeocorynid-
type appendages in fenestellids, 665; parasitism of Ordo-
vician trepostomes, 939

C

Calymenesun longinasuta sp. nov., 633, 60
Cambrian: trilobite hypostomes and ventral cephalic sutures,
577; chelicerate, Canada, 779; trilobites, New South
Wales, 905

Canada: Cambrian chelicerate, 779; Triassic reptile, 951
Carboniferous: lycophyte, England, 69; fish trails, England,
255; bryozoans, USA, 551; alga, England, 741; ammonoid
evolution, 1033

Catellocaula vallata ichnogen. et ichnosp. nov., 947, 87
Cephalopod: new Silurian genus with reinforced frilled shell,
651

Chatterton, B. D. E. See Fortey, R. A. and Chatterton,
B. D. E.

Cherns, L. Faunal and facies dynamics in the Upper Silurian
of the Anglo-Welsh Basin, 451
China: Ordovician Foliomena brachiopod fauna, 53; Ordo-
vician acritarchs, 109; Palaeocene and Eocene mammals,
129

Chlupac, I. The enigmatic arthropod Duslia from the Ordo-
vician of Czechoslovakia, 61 1
Choffatia cf. balinensis , 334, 25; cf. funata , 334, 25; cf.
madani , 333, 25

Choffatia (Grossouvria) propinqua , 336, 25
Choffatia (Indosphinctes) aff. urbana , 336; sp., 336
Chondrites , 41

Chordate: stem-group chordate, Ordovician, Northern Ire-
land, 1053

Christiania nilssoni , 8, 9 «

Chrustenopora askeriensis sp. nov., 1125, 99
Clack, J. A. New material of the early tetrapod Acanthostega
from the Upper Devonian of East Greenland, 699

1138

INDEX

Clack, J. A. and Holmes, R. The braincase of the anthraco-
saur Archeria crassidisca with comments on the inter-
relationships of primitive tetrapods, 85
Cocks, L. R. M. and Rong Jia-Yu. A review of the late
Ordovician Foliomena brachiopod fauna with new data
from China, Wales, and Poland, 53
Collins, D. See Briggs, D. E. G. and Collins, D,

Conifer: Cretaceous, Wealden, 1029
Conocoryphe sulzeri , 591, 55

Cope, J C. W. A reinterpretation of the Arenig crinoid
Ramseyocrinus , 229

Crampton, J. S. Comparative taxonomy of the bivalve
families Isognomonidae, Inoceramidae, and Retroceram-
idae, 965

Crane, P. R., Manchester, S. R and Dilcher, D. L.
Morphology and phylogenetic significance of the angio-
sperm Platanites hebridicus from the Palaeocene of Scot-
land, 503

Crassostrea virginica , 91

Cretaceous: fish, Brazil, 1; wood, Alaska, 19; bivalves,
Antarctica, 341; pollen, Egypt, 373; bryozoan, 519;
mosasaur, Nigeria, 747; dragonfly, England, 763; bi-
valve families Isognomonidae, Inoceramidae, and
Retroceramidae, 965; conifer Brachyphyllum punctatum ,
1029

Crinoid: Ordovician, Wales, 229

Cripps, A. P. A new species of stem-group chordate from
the Upper Ordovician of Northern Ireland, 1053
Crustacea: oldest freshwater decapod, Triassic, Arizona, 273
Cryptolithus tesselatus , 19
Cybeloides sp., 19
Cygnus equitum , 725, 69, 70
Czechoslovakia: Ordovician arthropod, 611

D

Dalmanella cf. pectinoides , 1 104, 95

Dashzeveg, D. and Russell, D. E Palaeocene and Eocene
Mixodontia (Mammalia, Glires) of Mongolia and China,
129

Dean, W. T. and Zhou Zhiyi. Upper Ordovician trilobites
from the Zap Valley, south-east Turkey, 621
Dedzetina sp., 8

Devonian: tetrapod Acanthostega , Greenland, 699; acantho-
dian fish, Brazil, 771
Diacanthaspis sp., 644, 61, 62
Dicoelosia alticavata , 1120, 97; osloensis , 1119, 97
Dicranopeltis sp., 644, 62

Dilcher, D. L. See Crane, P. R., Manchester, S. R. and
Dilcher, D. L.

Dindymenel sp., 631, 58
Dinoflagellates: Quaternary, North Sea, 877
IDiorthelasma semotum sp. nov., 1121, 99
Dolerolenus sp. indet., 585

Donovan, S. K. See Littlewood, D. T. J. and Donovan,
S. K.

Dorsetensia cf. romani , 314, 20

Drabovia sp., 1121, 98

Dragonfly: Cretaceous, England, 763

Duftonia sp., 632, 59

Durranella ? sp., 8

Duslia insignis , 612, 56, 57

E

Echinoderms: patterns of diversification and extinction in
the early Palaeozoic, 799
Egypt: Cretaceous pollen, 373

Elliott, G. F. A new alga from the Carboniferous Frosterley
Marble of northern England, 741
England: Plio-Pleistocene Bryozoa, 237; Carboniferous fish
trails, 255; Triassic tetrapod, 567; Tremadoc trilobites,
677; Carboniferous alga, 741; Cretaceous dragonfly, 763;
Caradoc trilobites, 829; Permian reptile, 957; Eocene
taphonomy, 1079
Enneles sp., 4

Enoploclytia porteri sp. nov., 275
Enteletacea indet. sp. A, 1127
Eobostrychoceras japonicum, 7; muramotoi , 7
Eocene: Mammals, China and Mongolia, 129; taphonomy
of London Clay biota, 1079
Eohoploceras cf. subdecoratum , 312
Eomylus borealis, 138; zhigdenensis sp. nov., 133
Eopleclodonta ( Kozlowskites ) nuntius, 9
Epitomyonia sp., 1120, 97
Erymnoceras sp. nov. aff. coronatum, 332
Eurekia ulrichi, 597

Evans, S. E. The Upper Permian reptile Adelosaurus from
Durham, 957

F

Fieldaspis celer, 588, 54

Fish: Cretaceous, Brazil, 1; trails, Carboniferous, England,
255; Silurian or Devonian, Brazil, 771
Fisher, H. L. See Watson, J., Fisher, H. L. and Hall, N. A.
Foliomena folium , 8, 9
Fontannesia hay deni, 316; kiliani, 316, 20
Fontannesial n. sp. aff. arabica, 317
Fontannesial (n. subgen.?) cf. arabica, 317, 20
Fortey, R A. and Chatterton, B. D. E. Classification of the
trilobite suborder Asaphina, 165
Fortipecten takahashii, 419, 39, 40

Fraser, N. C. Rare tetrapod remains from the late Triassic
fissure mfillings of Cromhall Quarry, Avon, 567
Frosterleyella diaspora sp. nov., 743

G

Geragnostus callavei, 680, 66; sp., 66
Globampyx trinucleoides , 19
Glyptorthis sp., 8
Goose, 69, 70

Goronyosaurus nigeriensis, 748

Grayicerasl gucuoi sp. nov., 330, 24; Waageni, 324, 23, 24;

aff. waageni, 328, 23
(l)Grayiceras nepaulense, 323, 22, 23
Greenland: Devonian tetrapod Acanthostega, 699

H

Hall, N. A. See Watson, J., Fisher, H. A. and Hall, N. A.
Harland, R. Quaternary dinoflagellate cyst biostratigraphy
of the North Sea, 877
Harpidella sp., 640, 60, 62

Hayami, I. and Hosoda, I. Fortipecten takahashii, a reclining
pectinid from the Pliocene of north Japan, 419

INDEX

1139

Hedinaspis sp., 929

Helenopora duncanae sp. nov., 556, 46, 47, 48, 49
Herpetopora anglica , 44, 45; laxata , 42, 43, 44, 45
Hibbertia sp., 628, 58, 59

Higgs, R. Fish trails in the Upper Carboniferous of south-
west England, 255

Holmes, R. See Clack, J. A. and Holmes, R.

Holmia kjerulfi , 579

Hosoda, I. See Hayami, I. and Hosoda, I.

Hutchinson, M. N. See Warren, A. A. and Hutchinson,
M. N.

Hypantoceras orienlale , 7
Hysterolenus furcatus sp. nov., 932, 86

I

Ingham, J. K. See Owen, A. W. and Ingham, J. K.
Inoceramus australis , 989, 90; bicorrugatus , 989, 90; concen-
tricus , 990, 90; fyfei, 990, 90; matotorus , 990; ope tins, 990,
90; pacificus , 991; rangatira , 991, 90; tawhanus , 991; sp
A, 991, 90; ? sp. B, 991, 90

Isognomon (Isognomon) rekohuensis sp. nov., 978, 89;

wellmani sp. nov., 974, 88, 90; sp., 986, 89
Isognomon (Mytiloperna) sp. A, 987, 89; (A/.?) sp. B, 987,

89

Isorthis ( Ovalella ) mackenziei , 1 108, 94, 95
Isorthis (Protocortezorthis) prima , 1 108, 94

J

Jamaica: Plio-Pleistocene and Recent bivalves, 1013
Janvier, P. and Melo, J. H. G. Acanthodian fish remains
from the Upper Silurian or Lower Devonian of the
Amazon Basin, Brazil, 771
Japan: Pliocene bivalve, 419

Jarzembowski, E. A. A new aeshnid dragonfly from the
Lower Cretaceous of south-east England, 763
Jeanneticeras cf. anomalum , 318
Jerercia rongi sp. nov., 1 126, 99
Jurassic: ammonites, Tibet, 295

K

Kampella guttula gen. et sp. nov., 1 1 17, 96
Kassinella anisa , 9; incerta, 9

Kelly, S. R. A. Cretaceous wood-boring bivalves from
western Antarctica with a review of the Mesozoic Pholadi-
dae, 341

Khaychina elongata sp. nov., 141

L

Leptestiina prantli , 8
Levenea sp. nov., 1110, 96
Lichas aff. laciniatus , 643, 62
Lingula lata , 41

Littlewood, D. T. J. and Donovan, S. K. Variation of
Recent and fossil Crassostrea in Jamaica, 1013
Lonchodomas , 626, 58
Lua erdaopuziana sp. nov., 120, 16

Lucas, S. G. and Oakes, W. A late Triassic cynodont from
the American south-west, 445
Lycophyte: Carboniferous, England, 69

M

Macrocephalites cf. macrocephalus , 319, 21
Macrocephalitesl cf. etheridgei , 319, 21
Malta: Pleistocene bird, 725

Mammal: Palaeocene and Eocene, Mongolia and China,
129

Manchester, S. R. See Crane, P. R., Manchester, S. R. and
Dilcher, D. L.

Marklandella markesi sp. nov., 1 1 16, 98
Martill, D. M. Preservation of fish in the Cretaceous Santana
Formation of Brazil, 1

Martin, F. and Yin Leiming. Early Ordovician acritarchs
from southern Jilin Province, north-east China, 109
Melo, J. H. G. See Janvier, P. and Melo, J. H. G.
Mendacella bleikeriensis sp. nov., 1 1 13, 98; sp., 1 1 14, 96
Micragnostus sp., 911

Miller, G. L. and Ash, S. R. The oldest freshwater decapod
crustacean, from the Triassic of Arizona, 273
Mills, K. J. See Webby, B. D., Wang Qizheng and Mills,
K. J.

Mimotona Hi sp. nov., 145

Miraspinidae genus and species undetermined, 645, 62
Miraspis sp., 645, 62

Mongolia: Paleocene and Eocene mammals, 129
Mosasaur: Cretaceous, Nigeria, 747
IMultispinula minuta , 82
Muramotoceras yezoense , 7

N

Nematosphaeropsis labyrinthea , 78, 82

Nigeria: Cretaceous mosasaur, 747

Nileid sp. I, 693, 68; sp. 2, 693, 68

Nileus orbiculatoides svalbardensis, 18, 19; porosus , 18, 19

Niobina davidis , 692, 68

Nipa burtini , 94

Nipponites mirabilis , 7

Nomenclature: use of open taxonomic nomenclature, 223
North Sea: Quaternary dinoflagellates, 877
Northcote, E. M An extinct 'swan-goose' from the Pleisto-
cene of Malta, 725

Northern Ireland: Ordovician stem-group chordate, 1053
Norway: Silurian brachiopods, 1101
Notelops brama , 1 , 4

O

Oakes, W. See Lucas, S. G. and Oakes, W.

Okamoto, T. Analysis of heteromorph ammonoids by differ-
ential geometry, 35; changes in life orientation during the
ontogeny of some heteromorph ammonoids, 281
Olenellus gilberti , 58 1

Onnia gracilis , 850, 77; superba superba , 844, 74; superba
cobboldi , 844, 75; superba creta subsp. nov., 845, 76
Operculodinium centrocarpum , 78; israelianum , 82
Opertochasma psyche sp. nov., 349, 26, 27
Ordovician: Foliomena brachiopod fauna, China, Wales,
and Poland, 53; acritarchs, China, 109; crinoid, Wales,
229; arthropod, Czechoslovakia, 611; trilobites, Turkey,
621; trilobites, English Lake District, 677; trilobites,
Shropshire, 829; trilobites. New South Wales, 905; bryo-
zoans, parasitism of, 939; stem-group chordate. Northern
Ireland, 1053

1140

INDEX

Ovalocephalus tetrasulcatus , 632, 59

Owen, A. W. and Ingham, J. K. The stratigraphical distri-
bution and taxonomy of the trilobite Onnia in the type
Onnian Stage of the uppermost Caradoc, 829
Oxycerites n. sp. A, 317, 20

P

Pagetia ocellata , 579

Palaeocene: mammals, China and Mongolia, 129; angio-
sperm, Scotland, 503
Palaeocorynid appendages, 65

Palmer, T. J. and Wilson, M. A. Parasitism of Ordovician
bryozoans and the origin of pseudoborings, 939
Parabolinella triarthroides , 686; sp. indet., 596
Paradoxides davidis , 586, 53
Paraphillipsinella pilula sp. nov., 637, 60
Pareuloma aculeatum sp. nov., 915, 83; expansion sp. nov.,
684, 66

Parotosuchus aliciae sp. nov., 860

Parrish., J. T. and Spicer, R. A. Middle Cretaceous wood
from the Nanushuk Group, central North Slope, Alaska,
19

IPaurorthis inopinatus sp. nov., 1118
Peltocare olenoides , 688, 67

Penny, J. H. P. Early Cretaceous acolumellate semitectate
pollen from Egypt, 373
Peraspis erugata, 18, 19
Permian: reptile Adelosaurus , Durham, 957
Phorocephala sp., 642, 60
Platanites hebridicus , 505
Pleistocene: ‘swan-goose’, Malta, 725
Pliocene: bivalve, Japan, 419

Plio-Pleistocene: bryozoans encrusting bivalve shells. Red
Crag, England, 237; bivalve Crassostrea , Jamaica, 1013
Poland: Ordovician Foliomena brachiopod fauna, 53
Pollen: Cretaceous, Egypt, 373
Polyptychoceras sp., 7

Poronileus fistulosus, 18, 19; isoteloides, 18, 19
Prionocheilus cf. obtusus , 636, 60
Proboscisambon quaesitus, 8

Proceratopvge ocella sp. nov., 926, 85; sp., 85; sp. indet.,
595

Prosaukia? sp., 918
Prospectatrix brevior sp. nov., 694, 68
Protoperidinium conicoides , 82; conicum , 78, 82; leonis , 78,
82; pentagonum, 78, 82

Pseudoborings: Catellocaula vallata ichnogen. et ichnosp.

nov. parasitic on Ordovician bryozoans, 939
Pseudocybele nasuta , 19
Pseudotoitesl cf. sphaeroceroides , 319
Pseudotriconodon chatterjeei sp. nov., 445
Pseudoyuepingia lata sp. nov., 923; whitei sp. nov., 922, 84
Ptychoparia striata , 590, 54

Q

Quaternary: dinoflagellates, North Sea, 877
R

Ramseyocrinus cambriensis , 234
Raphiophorus ? sp., 626, 58

Ravozetina cf. honorata , 1 106, 96
Redlichia sp. indet., 584
Remopleurides aff. R. eximius , 17

Reptile: Triassic, USA, 445; Triassic, British Columbia, 951;

Permian, Durham, 957; Triassic prolaceritiform, 997
Resserella matutina sp. nov., 1112, 97
RETIMONO-BASKET, 388, 31; -BIGHOLE, 406, 38; -HAIRY, 406,
38; -hedgehog, 398, 34; -knobble, 390, 32; -necklace,
382, 28; -pimple, 412, 38; -ridged, 402, 36; -smallhole,
402, 35; -spinerow, 382, 29; -spotspines, 390, 33; -typesix,
384, 30; -walnut, 406, 37
Retroceranms (Fractoceramus) inconditus , 989
Retroceramus ( Retroceramus ) aff. everesti , 988; galoi, 988,
90; haasti , 988, 90; marwieki , 988; cf. subhaasti , 989, 90
Rhacolepis sp., 1, 2, 3, 4
Rliaptagnostus leitchi sp. nov., 913
Rhopaliophora cf. R. palmata, 122, 16: pilata , 122, 16
Rong Jia-Yu. See Cocks, L. R. M. and Rong Jia-Yu.

Ross, J. R. P. New chaetetiform trepostome Bryozoa from
the Upper Mississippian of the western United States, 551
Rowe, N. P. A herbaceous lycophyte from the Lower
Carboniferous Drybrook Sandstone of the Forest of
Dean, Gloucestershire, 69

Rushton, A. W. A. Tremadoc trilobites from the Skiddaw
Group in the English Lake District, 677
Russell, D. E. See Dashzeveg, D. and Russell, D. E.

S

Salopina pumila sp. nov., 1124, 99
Sanctacaris uncata sp. nov., 781, 71, 72, 73
Sao hirsuta, 591, 53
Saukiid? gen. et sp. indet., 919
Scalarites scalaris, 7
Scotiaecystis collapsa , 1055, 93
Scotland: Palaeocene angiosperm, 503
Selaginellites resimus sp. nov., 80, 10, 11, 12
Shumardia (Conophrys) sp., 682, 66

Silurian: faunal and facies dynamics, Anglo-Welsh Basin,
45 1 ; cephalopod with reinforced frilled shell, 65 1 : acantho-
dian fish, Brazil, 771; enteletacean brachiopods, Norway,
1101

Sinocybylel fluminis sp. nov., 628, 59
Smith, A. B. Patterns of diversification and extinction in
early Palaeozoic echinoderms, 799
Soliar, T. The mosasaur Goronyosaurus from the Upper
Cretaceous of Sokoto State, Nigeria, 747
Sonninia s. 1. sp., 312
Spencellal sp., 17

Spicer, R. A. See Parrish, J. T. and Spicer, R. A.
Spiniferites elongatus , 80; lazus , 79; membranaceus , 82;

mirabilis, 80; ramosus, 79
Stenopareia sp., 643, 62

Stridsberg, S. A Silurian cephalopod genus with a reinforced
frilled shell, 651
Subkossmatia cf. opis, 332, 23

Swan, A. R. H. Heterochronic trends in Namurian am-
monoid evolution, 1033

T

Taxon A wood, 6
Taxon B wood, 6

INDEX

1141

Taylor, P. D. Colony growth pattern and astogenetic gradi-
ents in the Cretaceous cheilostome bryozoan Herpetopora ,
519

Tectatodinium pellitum , 82

Teredina jeffersoni sp. nov., 355

Tetrapod: Triassic, England, 567; Devonian, Greenland, 699

Tibet: Jurassic ammonites, 295

Torquatoceras auritum sp, nov., 660, 64; undulatum sp. nov.,
658, 63

Triassic: decapod crustacean, Arizona, 273; cynodont reptile,
USA, 445; tetrapod, Avon, 567; capitosaurid amphibian,
Australia, 857; weigeltisaurid reptile, British Columbia,
951; prolacertiform reptile, 997

Trilobites: classification of Suborder Asaphma, 165; Cam-
brian, hypostomes and ventral cephalic sutures, 577;
Ordovician, Turkey, 621; Ordovician, English Lake Dis-
trict, 677; Ordovician, Shropshire, 829; Cambrian and
Ordovician, New South Wales, 905

Tschanz, K. Allometry and heterochrony in the growth of
the neck of Triassic prolacertiform reptiles, 997

Tscherskidium , 9

Turkey: Ordovician trilobites, 621

Turnus kotickensis sp. nov., 359; sp., 361

U

USA: Cretaceous wood, 19; Triassic decapod crustacean,
273; Triassic cynodont reptile, 445; Carboniferous trepo-
stome bryozoans, 551; new pseudoboring parasitic on
Ordovician bryozoans, 939

Undichna bina , 257; britannica sp. nov., 257; consulca sp.
nov., 262; simplicatus , 267

V

Valdaeshna surreyensis sp. nov., 765

W

Wales: Ordovician Foliomena brachiopod fauna, 53

Wang Qizheng. See Webby, B. D., Wang Qizheng and
Mills, K. J.

Wang Yi-Gang. See Westermann, G. E. G. and Wang Yi-
Gang.

Wapitisaurus problematic us gen. et sp. nov., 952

Warren, A. A. and Hutchinson, M. N. A new capitosaurid
amphibian from the early Triassic of Queensland, and the
ontogeny of the capitosaurid skull, 857

Watson, J., Fisher, H. L. and Hall, N. A. The holotype of
the Wealden conifer Brachyphyllum punctatum Michael,
1029

Webby, B D., Wang Qizheng and Mills, K. J. Upper
Cambrian and basal Ordovician trilobites from western
New South Wales, 905

Welleraspis swartzi , 596

Westermann, G. E. G. and Wang Yi-Gang. Middle Jurassic
ammonites of Tibet and the age of the Lower Spiti Shales,
295

Whittington, H. B. Hypostomes and ventral cephalic sutures
in Cambrian trilobites, 577

Whooper Swan, 69, 70

Wilson, M. A. See Palmer, T. J. and Wilson, M. A.

Witchellia cf. australica , 313; cf. sutneri , 314, 20; tibetica ,
313

Wood: Cretaceous, Alaska, 19

X

Xenoxylon latiporosum , 5
Xylophagella truncata , 362
Xystridura sp. indet., 588

Y

Yin Leiming. See Martin, F. and Yin Leiming.

Z

Zhou Zhiyi. See Dean, W. T. and Zhou Zhiyi.

VOLUME 31

Palaeontology

1988

PUBLISHED BY THE
PALAEONTOLOGICAL ASSOCIATION
LONDON

Dates of Publication of Parts of Volume 31

Part l,pp. 1-236, pis. 119
Part 2, pp. 237-566, pis. 20-51
Part 3, pp. 567-904, pis. 52-82
Part 4, pp. 905 1 141, pis. 83-94

28 January 1988
20 May 1988
9 September 1988
16 December 1988

THIS VOLUME EDITED BY M. J. BENTON, P. R. CROWTHER, J. E. DALINGWATER,
D. EDWARDS, L. B. HALSTEAD, T. J. PALMER, C. R. C. PAUL, AND
P. A. SELDEN

Dates of Publication of Special Papers in Palaeontology

Special Paper No. 39, I September 1988
Special Paper No. 40, 2 December 1988

© The Palaeontological Association, 1988

Printed in Great Britain
at the University Printing House, Oxford
by David Stanford
Printer to the University

CONTENTS

Part

Allison, P. A. Taphonomy of the Eocene London Clay biota 4

Ash, S. R. See Miller, G. L. and Ash, S. R.

Baarli, B. G. The Llandovery enteletacean brachiopods of the central Oslo region, Norway 4

Bancroft, A. J. Palaeocorynid-type appendages in upper Palaeozoic fenestellid Bryozoa 3

Bengtson, P. Open nomenclature 1

Bishop, J. D. D. Disarticulated bivalve shells as substrates for encrustation by the bryozoan
Cribrilina puncturata in the Plio-Pleistocene Red Crag of eastern England 2

Brinkman, D. A weigeltisaurid reptile from the Lower Triassic of British Columbia 4

Briggs, D. E. G. and Collins, D. E. A Middle Cambrian chelicerate from Mount Stephen,
British Columbia 3

Chatterton, B. D. E. See Fortey, R. A. and Chatterton, B. D. E.

Cherns, L. Faunal and facies dynamics in the Upper Silurian of the Anglo-Welsh Basin 2

ChlupaC, E The enigmatic arthropod Duslia from the Ordovician of Czechoslovakia 3

Clack, J. A. New material of the early tetrapod Acanthostega from the Upper Devonian of
East Greenland 3

Clack, J. A. and Holmes, R. The braincase of the anthracosaur Archeria crassidisca with
comments on the interrelationships of primitive tetrapods I

Cocks, L. R. M. and Rong Jia-Yu. A review of the late Ordovician Foliomena brachiopod
fauna with new data from China, Wales, and Poland 1

Collins, D. E. See Briggs, D. E. G. and Collins, D. E.

Cope, J, C. W. A reinterpretation of the Arenig crinoid Ramseyocrinus 1

Crampton, J. S. Comparative taxonomy of the bivalve families Isognomonidae, Inoceramidae,
and Retroceramidae 4

Crane, P, R., Manchester, S. R. and Dilcher, D. L. Morphology and phylogenetic significance
of the angiosperm P/atanites hebridicus form the Palaeocene of Scotland 2

Cripps, A. P. A new species of stem-group chordate from the Upper Ordovician of Northern
Ireland 4

Dashzeveg, D. and Russell, D. E. Palaeocene and Eocene Mixodontia (Mammalia, Glires) of
Mongolia and China I

Dean, W. T. and Zhou Zhiyi. Upper Ordovician trilobites from the Zap Valley, south-east
Turkey 3

Dilcher, D. L. See Crane, P. R., Manchester, S. R. and Dilcher, D. L.

Donovan, S. K. See Littlewood, D. T. J. and Donovan, S, K.

Elliott, G. A new alga from the Carboniferous Frosterley Marble of northern England 3

Evans, S. E. The Upper Permian reptile Adelosaurus from Durham 4

Fisher, H. L. See Watson, J., Fisher, H. L., and Hall, N. A.

Fortey, R. A. and Chatterton, B. D. E. Classification of the trilobite suborder Asaphina 1

Fraser, N. C. Rare tetrapod remains from the late Triassic fissure infillings of Cromhall Quarry,
Goucestershire 3

Hall, N. A. See Watson, J., Fisher, H. L. and Hall, N. A.

Harland, R. Quaternary dinofiagellate cyst biostratigraphy of the North Sea 3

Hayami, H. and Hosoda, E Fortipecten takahashii, a reclining pectinid from the Pliocene of
north Japan 2

Higgs, R. Fish trails in the Upper Carboniferous of south-west England 2

Holmes, R. See Clack, J. A. and Holmes, R.

Hosoda, I. See Hayami, H. and Hosoda, I,

Hutchinson, M. N. See Warren, A. and Hutchinson, M, N.

Ingham, J. K. See Owen, A. W. and Ingham, J. K.

Janvier, P. and Melo, J. H. G. Acanthodian fish remains from the Upper Silurian or Lower
Devonian of the Amazon Basin, Brazil 3

Page

1079

1101

665

223

237

951

779

451

61 1

699

85

53

229

965

503

1053

129

621

741

957

165

567

877

419

255

771

IV

CONTENTS

Jarzembowski, E. A. A new aeshnid dragonfly from the Lower Cretaceous of south-east England 3

Kelly, S. R. A. Cretaceous wood-boring bivalves from Western Antarctica with a review of the
Mesozoic Pholadidae 2

Littlewood, D. T. J. and Donovan, S. K. Variation of Recent and fossil Crassostrea in Jamaica 4
Lucas, S. G. and Oakes, W. A Late Triassic cynodont from the American south-west 2

Manchester, S. R. See Crane, P. R., Manchester, S. R. and Dilcher, D. L.

Martill, D. M. Preservation of fish in the Cretaceous Santana Lormation of Brazil 1

Martin, F. and Yin Leiming. Early Ordovician acritarchs from southern Jilin Province, north-
east China 1

Melo, J. H. G. See Janvier, P. and Melo, J. H. G.

Miller, G. L. and Ash, S. R. The oldest freshwater crustacean, from the Triassic of Arizona 2
Mills, K. J. See Webby, B. D., Wang Qizheng and Mills, K. J.

Northcote, E. M. An extinct ‘swan-goose’ from the Pleistocene of Malta 3

Oakes, W. See Lucas, S. G. and Oakes, W.

Okamoto, T. Analysis of heteromorph ammonoids by differential geometry 1

Okamoto, T. Changes in life orientation during the ontogeny of some heteromorph ammonoids 2
Owen, A. W. and Ingham, J. K. The stratigraphic distribution and taxonomy of the trilobite
Onnia in the type Onnian Stage of the uppermost Caradoc 3

Palmer, T. J. and Wilson, M. A. Parasitism of Ordovician bryozoans and the origin of
pseudoborings 4

Parrish, J. T. and Spicer, R. A. Middle Cretaceous wood from the Nanushuk Group, central
North Slope, Alaska 1

Penny, J. H. J. Early Cretaceous acolumellate semitectate pollen from Egypt 2

Rong Jia-Yu. See Cocks, L. R. M. and Rong Jia-Yu.

Ross, J. R. P. New chaetetiform trepostome Bryozoa from the Upper Mississippian of the
western United States 2

Rowe, N. P. A herbaceous lycophyte from the Lower Carboniferous Drybrook Sandstone of
the Forest of Dean, Gloucestershire 1

Rushton, A, W. A. Tremadoc trilobites from the Skiddaw Group in the English Lake District 3
Russell, D. E. See Dashzeveg, D. and Russell, D. E.

Smith, A. B. Patterns of diversification and extinction in early Palaeozoic echinoderms 3

Soliar, T. The mosasaur Goronyosaurus from the Upper Cretaceous of Sokoto State, Nigeria 3
Spicer, R. A. See Parrish, J. T. and Spicer, R. A.

Stridsberg, S. A Silurian cephalopod genus with a reinforced frilled shell 3

Swan, A. R. H. Heterochronic trends in Namurian ammonoid evolution 4

Taylor, P. D. Colony growth pattern and astogenetic gradients in the Cretaceous cheilostome
bryozoan Herpetopora 2

Tschanz, K. Allometry and heterochrony in the growth of the neck of Triassic prolacertiform
reptiles 4

Wang Qizheng. See Webby, B. D., Wang Qizheng and Mills, K. J.

Wang Yi-Gang. See Westermann, G. E. G. and Wang Yi-Gang.

Warren, A. and Hutchinson, M. N. A new capitosaur amphibian from the early Triassic of
Queensland and the ontogeny of the capitosaur skull 3

Watson, J., Fisher, H. L. and Hall, N. A. The holotype of the Wealden conifer Brachy-
phyllum punctatum Michael 4

Webby, B. D., Wang Qizheng and Mills, K. J. Upper Cambrian and basal Ordovician
trilobites from western New South Wales 4

Westermann, G. E. G. and Wang Yi-Gang. Middle Jurassic ammonites of Tibet and the age
of the Lower Spiti Shales 2

Whittington, H. B. Hypostomes and ventral cephalic sutures in Cambrian trilobites 3

Wilson, M. A. See Palmer, T. J. and Wilson, M. A.

Yin Leiming. See Martin, F. and Yin Leiming.

Zhou Zhiyi. See Dean, W. T. and Zhou Zhiyi.

763

341

1013

445

1

109

273

725

35

281

829

939

19

373

551

69

677

799

747

651

1033

519

997

857

1029

905

295

577

NOTES FOR AUTHORS

The journal Palaeontology is devoted to the publication of papers on all aspects of palaeontology. Review articles are
particularly welcome, and short papers can often be published rapidly. A high standard of illustration is a feature of the
journal. Four parts are published each year and are sent free to all members of the Association. Typescripts should conform in
style to those already published in this journal, and should be sent to Dr. Dianne Edwards, Department of Plant Science,
University College, P.O. Box 78, Cardiff CF1 1XL, who will supply detailed instructions for authors on request (these are
published in Palaeontology 1985, 28, pp. 793-800).

Special Papers in Palaeontology is a series of substantial separate works conforming to the style of Palaeontology.

SPECIAL PAPERS IN PALAEONTOLOGY

In addition to publishing Palaeontology the Association also publishes Special Papers in Palaeontology. Members may
subscribe to this by writing to the Membership Treasurer: the subscription rate for 1988 is £37-50 (U.S. $66) for Institutional
Members, and £18-50 (U.S. $32) for Ordinary and Student Members. A single copy of each Special Paper is available to
Ordinary and Student Members only, for their personal use, at a discount of 25% below the listed prices. Non-members
may obtain copies, but at the listed prices, from Marston Book Services, P.O. Box 87, Oxford 0X4 1LB, England.

RECENT PALAEONTOLOGICAL ASSOCIATION

PUBLICATIONS

Special Papers in Palaeontology

Numbers 1-23 are still in print and are available (post free ) together with those listed below:

24. (for 1980): Dinoflagellate Cysts and Acritarchs from the Eocene of Southern England, by j. p. bujak, c. downte, g. l.
eaton and G. l. williams. 100 pp., 24 text-figs., 22 plates. Price £15 (U.S. $23).

25. (for 1980): Stereom Microstructure of the Echinoid Test, by a. b. smith. 81 pp., 20 text-figs., 23 plates. Price £15
(U.S. $23).

26. (for 1981): The Fine Structure of Graptolite Periderm, by p. r. crowther. 119 pp., 37 text-figs., 20 plates. Price £25
(U.S. $38).

27. (for 1981): Late Devonian Acritarchs from the Carnarvon Basin, Western Australia, by G. playford and r. s. dring.
78 pp., 10 text-figs., 19 plates. Price £15 (U.S. $23).

28. (for 1982): The Mammal Fauna of the Early Middle Pleistocene cavern infill site of Westbury-sub-Mendip, Somerset, by
m. j. bishop. 108 pp., 47 text-figs., 6 plates. Price £25 (U.S. $38).

29. (for 1982): Fossil Cichlid Fish of Africa, by j. a. h. van couvering. 103 pp., 35 text-figs., 10 plates. Price £30 (U.S. $45).

30. (for 1983): Trilobites and other early Arthropods. Edited by d. e. g. briggs and p. D. lane. 276 pp., 64 text-figs., 38 plates.
Price £40 (U.S. $60).

31. (for 1984): Systematic palaeontology and stratigraphic distribution of ammonite faunas of the French Coniacian, by w. j.
Kennedy. 160 pp., 42 text-figs., 33 plates. Price £25 (U.S. $38).

32. (for 1984): Autecology of Silurian organisms. Edited by m. g. bassett and j. d. lawson. 295 pp., 75 text-figs.. 13 plates.
Price £40 (U.S. $60).

33. (for 1985): Evolutionary Case Histories from the Fossil Record. Edited by i. c. w. cope and p. w. skelton. 202 pp., 80 text-
figs., 4 plates. Price £30 (U.S. $45).

34. (for 1985): Review of the upper Silurian and lower Devonian articulate brachiopods of Podolia, by o. i. Nikiforova,
t. l. modzalevskaya and m. g. bassett. 66 pp., 6 text-figs., 16 plates. Price £10 (U.S. $15).

35. (for 1986): Studies in palaeobotany and palynology in honour of N. F. Hughes. Edited by d. j. batten and
d. e. g. BRIGGS. 178 pp., 29 plates. Price £30 (U.S. $50).

36. (for 1986): Campanian and Maastrichtian ammonites from northern Aquitaine, France, by w. j. Kennedy. 145 pp., 43
text-figs., 23 plates. Price £20 (U.S. $35).

37. (for 1987): Biology and revised systematics of some late Mesozoic stromatoporoids, by rachel wood. 89 pp., 31 text-
figs., 7 plates. Price £20 (U.S. $35).

38. (for 1987): Taxonomy, evolution, and biostratigraphy of late Triassic-early Jurassic calcareous nannofossils, by
p. r. bown. 1 18 pp., 19 text-figs., 15 plates. Price £30 (U.S. $50).

39. (for 1988): Late Cenomanian and Turonian ammonite faunas from north-east and central Texas, by w. J. Kennedy.
131 pp., 39 text-figs., 24 plates. Price £30 (U.S. $50).

Field Guides to Fossils

1 . (1983): Fossil Plants of the London Clay, by m. e. collinson. 121 pp., 242 text-figs. Price £7-95 (U.S. $12).

2. (1987): Fossils of the Chalk, compiled by e. owen; edited by a. b. smith. 306 pp., 59 plates. Price £11-50 (U.S. $18).

3. (1988): Zechstein Reef fossils and their palaeoecology, by n. hollingworth and t. Pettigrew, iv + 75 pp. Price £4-95
(U.S. $8).

Other Publications

1982. Atlas of the Burgess Shale. Edited by s. conway morris. 31 pp., 24 plates. Price £20 (U.S. $30).

1985. Atlas of Invertebrate Macrofossils. Edited by j. w. Murray. Published by Longman in collaboration with the
Palaeontological Association, xiii + 241 pp. Price £13-95. Available in the USA from Halsted Press at U.S. $24.95.

© The Palaeontological Association, 1988

Palaeontology

VOLUME 31 ■ PART 4

CONTENTS

Upper Cambrian and basal Ordovician trilobites from
western New South Wales

B. D. WEBBY, WANG QIZHENG and K. J. MILLS 905

Parasitism of Ordovician bryozoans and the origin of pseudo-
borings

T. J, palmer and M. A. WILSON 939

A weigeltisaurid reptile from the Lower Triassic of British
Columbia

D. BRINKMAN 951

The Upper Permian reptile Adelosaurus from Durham

S. E. EVANS 957

Comparative taxonomy of the bivalve families Isognomon-
idae, Inoceramidae, and Retroceramidae

J. S. CRAMPTON 965

Allometry and heterochrony in the growth of the neck of
Triassic prolacertiform reptiles

K. TSCHANZ 997

Variation of Recent and fossil Crassostrea in Jamaica

D. T. J. LITTLEWOOD and S. K. DONOVAN 1013

The holotype of the Wealden conifer Brachyphyllum puncta-
tum Michael

J. WATSON, H. L. FISHER and N. A. HALL 1029

Heterochronic trends in Namurian ammonoid evolution

A. R. H. SWAN 1033

A new species of stem-group chordate from the Upper Ordo-
vician of Northern Ireland

a. p. cripps 1053

Taphonomy of the Eocene London Clay biota

P. A. ALLISON 1079

The Llandovery enteletacean brachiopods of the central Oslo
region, Norway

B. G. BAARLI 1101

Printed in Great Britain at the University Printing House , Oxford
by David Stanford , Printer to the Univ

ISSN 0031-0239

*' § 3

x o

> 'W*^ S > "W S ^ > Xi£Uu5^/ “> ’■V >

2 CO 2 CO ‘2 lo fc- Z

S3iavaan libraries Smithsonian institution NouniiisNi nvinoshiiiais saiavaan lie

z CO ^ CO ™ . „ . CO “p

. , / , - \ 1) , B ✓VUSAjbX 9 S 9 S-- . .. ddl XlX Ft . 11 “■

o x^y pc^x — o x2LEi>^ — x^nns^X o

INSTITUTION2 NOlinillSNI^NViNOSHlIlAIS SBIdVaan^LIBRARI ES2 SMITHSONIAN^ INSTITUTION 2 NOI

2: □ > 2 r- z z

° m ^ 2 m x<fl2£qX 2 A 03 ^T5gVoX o

#^> ll

#/• f va,

m '' ^ xjvasi^x^ rn Xfoos vQjZ ^ m x^vosiOx'

co * — co £ co S:

S3IHVM9I1„LIBRARIES SMITHSONIAN INSTITUTION NOIiniliSNI NVINOSHIIIAIS S3 I H VB 0 II LIE

^ i ^ I /*^\ i V. I jfe, I I 4/toe* I

x

co AW&1' UJ [ps;; rS8<| co .

VJS*^.' r-> AK/.f- -AW. t l&E, -311 O #/ aK*/; X

, v V Z

5 '" 2 '^W >"’ 2 x^vosvjix >’ ^

to N Z CO •-*■*' z CO Z co

NSTITUTION NOIiniliSNI NVIN0SH1IWS S3IMVH8n LIBRARIES SMITHSONIAN INSTITUTION NOI

co z \ co — co 5 co

'^TiO>v rn V CU X^VT^N. £ UJ ^ x-fooT^x UJ

S3 1 H VH 8 IT LIBRARIES SMITHSONIAN INSTITUTION NOIiniliSNI NVINOSHIIIAIS $3IIJViJan U

03,

53

** Ivrfo* Sflo > X , loj£. -SIo) "ZL ...

— \cx4ikxiv t“ K^^hWJ/ _ ^ H Vd%tJJ^y — v

pi xii^oc^ ~ w/ m xJJLSS-^ co x^®^/ m ^ x^om^jx m

co £ co ± co \ £ co

INSTITUTION ^ NOIiniliSNI NVINOSHlUAIS^SB I MVHB 11 LIB RAR I ES ^SMITHSONIAN INSTITUTION, n NO

Z < x<<v5sov7x S , <

> «w* 2 > *«r 2 ^ > X^QiisiTX 2 ^. >

*, z CO 2 CO * z CO -Z

S3iavaan libraries Smithsonian institution NouniiisNi nvinoshiiiais S3iavaan_Li

X CO X __ CO - CO

*<1- i . o x<T<<;nT7X i iS -a— iXX , atJ

CO

z

CO

c

$C4l*S>’

o x^/ydc^x 2 o py' _ '<an^y o " — o

’institution ■“ NOliniliSNI^NVINOSHimS^Sl lavaail^LIBRARIES SMITHSONIAN INSTITUTION NO

2 r* > Z r- Z ^ ._ 2

/\ ° ro X 2 5 2 ^ ra g >

5A ►- /MWXA “ ^ »- A«A ” t; Wf/Jb,™ fmWh^ H *’

^ r W

rn x w - m ^ ' m z

CO ■ ' ' — CO

C/l

sjiavaan libraries Smithsonian institution noiiiuiisni NviN0SHUWs,_S3iavaan_i

A 1 X . 1 I V . 1 A. I /S3S£x 1 1

IMOTITIITIOM klnimillOMI MUIMTVCUIIUIC C5 I UWVIQ II I I O D A D 1 C C ITUCOW I A M INSTITUTION

_ _ | %A

S ^ > XgUa’i^X § >

ES^SMITHSONIAN^ INSTITUTION NOiiniliSNI NVINOSHIIWS^SB I U VH a 11 LI B R AR I ES^SMITHSONIAN

co 5 co 5 . co z ,n

w co “ /sSjSgv^ W W XSjggk

O 'N^YDV^ _ X^atllS^X O

NI^NVINOSHilWS^SS I avy a ll^LI BRAR I E$2 SMITHSONIAN”* IN ST I TUT SON NOlinillSNI^NVINOSHillAls'
t“ v Z r- z £ Z ^ > <

co ^ 2 x^£x ro x^^ox ° ™ xrx£3x p ^

PO

>

73

rn <* vw zi x^vasv^' cn Xi^osv'^X' 'ts. m X'yosv^x ^ xjvas»2>^ m

co £ in £ co £ co ■

ES SMITHSONIAN INSTITUTION NQIlfUlISN! NVINOSHilWS SSIdVHaH LIBRARIES SMITHSONIAN

co z > co z co z > ^

< 2 < V i xs^ox

^ z

X^jv ASV^>

:NI NVINOSHillAIS^Saiavaail^LIBRARIES^SMITHSONIAN INSTITUTION NQlinillSNI_NV!NQSHllWS
<0 — to ” co ^ \ <

w w X®X ^ ^ ^ "

z

co

O "'W “ Xir^SX o

z «s z

ES SMITHSONIAN INSTITUTION NOiiniliSNI NWIN05HISSAJS S31HVH011 LIBRARIES SMITHSONIAN
z r~ Z r- z 9“ z f

I

m ^ rn 2 m 'XS'Tdc^

cn £ — to \ £ to

5NI NVIN0SH1IINS S31HVaai1 LIBRARIES SMITHSONIAN INSTITUTION NOiiniliSNI NVIN0SH1UAIS
w z .,. co 2 s>- co z

,< i , . i < .J&s. s

_ > ___ _

ES^SMITHSONIAN INSTITUTION NOiiniliSNI NVINOSHIIWS^SB I d VH a n\l B RAR I ES SMITHSONIAN

CO 2 <0 = ... CO 5 co

lu £ /S^IX iJ-* x^cv^X £ ,^5\K o uj xCo^vvJJx £ Xvtw»7X lu

Airsr®KV>

\ _ /.viwi irAM =j » \x vex

cc
<

,'*°r 5 N«3KjS/ 5 5 V “ 5

SNI^NVINOSHlIWS S3 ! HVd 0 11 LI B RAR I ES^SMITHSONIAN^INSTiTUTION^NOlinillSNI^NVINOSHilWS
?“ > z r- z r~ 2 _ «“ >

m c'\-w xw^x m X^vo^x ^ m u> m

to £ to £ co ± co

ES SMITHSONIAN INSTITUTION NOiiniliSNI NVIN0SH11WS S3iavaail LIBRARIES SMITHSONIAN
z \ co z co z \ ^

2 ^ILr<oX < V S 2 ^b. < x 2

m tsrr" I