O O

TUTION NOIinillSNl“’NVINOSHiWS S3lbVHan LI B RAR I ES^SMITHSONIAN~INSTiTUTION'^NOIini!l

r- 2: r- 2:

DO

^ /i'i

> Ft
XI \^\

rn ^ rn

CO _ CO _

vaan libraries Smithsonian institution NoiiruiiSNi nvinoshiiims saiavaan librari

~ CO Z CO z ,

rUTION NOIiniliSNI_NVINOSHilWS SHIHVaail LIBRARIES SMITHSONIAN INSTITUTION NOlinili

v-^

O ~ O

Z _t Z -I

vaan libraries smithsonian institution NoiiniiiSNi nvinoshiiins saiavaan librar

" /SiT^ § ,/ x-,*i I " .vCs I

CO

TUTION NOliniliSNI NVINOSHillNS SHiavaaH LIBRARIES SMITHSONIAN INSTITUTION NOlinili

CO

A

-c
I
CO

o
z

> W" 5 '* >

* Z CO z CO *Z cO

Vaail LIBRARIES SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHIIWS S3iavaa

^ CO CO — CO

CO

TUTION NOliniliSNI NVINOSHIIINS S3iavaail LIBRARIES SMITHSONIAN INSTITUTION NOlinili

Z r- z r- Z r- Z

rn '' rn ^ m

CO ' ± CO £ CO ^

vaan libraries smithsonian institution NoiiniiiSNi NvmosHims saiavaan librari

rUTlON NOliniliSNI NVINOSHIIWS S3iavaai3 libraries smithsonian institution NOlinili

U>

0/

~:S-y — ■'

^VASVA^/ O ■■ Q

^ ‘ -I :z _

vaan libraries smithsonian institution NoiiniiiSNi nvinoshiiiais sBiavaan librar

[: ^ ^ H ^ ^ z r- ^ z z

i?V4rX m ^ • m O v<CIsvXj\ .. O m

IN0SHimS^S3 IdVdan^LIBRAR ! ES^ SMITHSONIAN^ INSTITUTION ^NOlinillSNI^NVINOSHimS^S 3 I d V

CD

;o
>

iV ^ ' t— ^ • ' ' ■ ^ f— ' r '

C/) _ c/5 £ c/5 ' _

ITHSONIAN INSTITUTION NOliniliSNI NVINOSHIIWS SSiaVHan LIBRARIES SMITHSONIAN INSTITl

c/5 z CO z: , CO 2:

— v;Ns»»^>r‘>

O Xi^ost5jiZ
z

ITHSONIAN INSTITUTION NOliniliSNI NVINOSHlIkMS S3iaVaail LIBRARIES SMITHSONIAN INSTITl

— — ^ — — —

o

m ^ m pcjz ^ m X^i;^

CO ± \ Z ^ (/,

iNOSHims saiavyan libraries Smithsonian institution NoiiniiiSNi nvinoshiiins saiav

^ 5

ITHSONIAN INSTITUTION NOlinillSNl'^NVINOSHlIWs'^Sa I B Va a

CC5 = CO

>

I^LIBRARIES SMITHSONIAN INSTITl
r: CO ::

«

O 15^^ _ x^ociiS^ Q '''' Z X^OiiiS^ Q XO;v_dC>^ “ O

2 -J Z* _J 2 _| 2

N0SHimS_S3 iavaan_LIBRARI ES_SMITHS0N1AN_ institution NOIinillSNI_NVINOSHlllMS S3 I av

o

z

o

— i/y — (/) ± CO

ITHSONIAN INSTITUTION NOliniliSNI NVINOSHIIIMS S3iavaan LIBRARIES SMITHSONIAN INSTITL

^ Z » CO z CO

> 2 X^os>5^ > 2 ' V^5v'- > \^^A^ 2

CO •'* Z CO Z CO Z CO

NosHiiws S3 lava a 11 libraries smithsonian institution NoiiniiiSNi nvinoshiiws S3iav

co_= CO — ,n - ■, (n

- x,h:a^ 5 ■• 5 !f o '

z

ITHSONIAN INSTITUTION NOliniliSNI NVINOSHII/JS S3iavaai1 LIBRARIES SMITHSONIAN INSTITl

z
o

z

o

CD

Z

o

/fc"SO;

THE PALAEONTOLOGICAL ASSOCIATION

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

COUNCIL 1987-1988

President: Dr. L. R. M. Cocks, Department of Palaeontology, British Museum (Natural History), Cromwell Road,

London SW7 5BD

Vice-Presidents: Dr. D. E. G. Briggs, Department of Geology, University of Bristol, Bristol BS8 IRJ
Dr. L. B. Halstead, Department of Geology, University of Reading, Reading RG6 2AB
Treasurer: Dr. M. Romano, Department of Geology, University of Sheffield, Sheffield SI 3JD
Membership Treasurer: Dr. A. T. Thomas, Department of Geological Sciences, University of Aston, Birmingham B4 7ET
Institutional Membership Treasurer: Dr. A. W. Owen, Department of Geology, The University, Dundee DDl 4HN
Secretary: Dr. P. W. Skelton, Department of Earth Sciences, The Open University, Milton Keynes MK7 6AA
Circular Reporter: Dr. D. J. Siveter, Department of Geology, University of Hull, Hull HU6 7RX
Marketing Manager: Dr. V. P. Wright, Department of Geology, University of Bristol, Bristol BS8 IRJ
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. P. R. Crowther, City of Bristol Museum and Art Gallery, Bristol BS8 IRL
Dr. D. Edwards, Department of Plant Science, University College, Cardiff CFl IXL
Dr. T. J. Palmer, Department of Geology, University College of Wales, Aberystwyth SY23 2AX
Dr. C. R. C. Paul, Department of Geology, University of Liverpool, Liverpool L69 3BX
Dr. P. A. Selden, Department of Extra-Mural Studies, University of Manchester, Manchester M13 9PL

Other Members

Dr. H. A. Armstrong, Newcastle upon Tyne Dr. G. B. Curry, Glasgow

Dr. M. E. Collinson, London Professor B. M. Funnell, Norwich

Dr. J. A. Crame, Cambridge Dr. P. D. Taylor, 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 1987 are:

Institutional membership £50 00 (U.S. $79)

Ordinary membership £21 00 (U.S. $38)

Student membership . £1 1 50 (U.S. $20)

Retired membership £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 L>niversily, Dundee DDl 4H1N. 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. A. T. Thomas, Department of Geological Sciences, University of Aston, Aston Triangle,
Birmingham B4 7ET. Subscriptions cover one calendar year and are due each January; they should be sent to the Membership
Treasurer, All members who join for 1987 will receive Palaeontology. Volume 30, Parts 1-4. Back numbers still in print
may be ordered from Marston Book Services, P.O. Box 87, Oxford 0X4 ILB, England.

Cover: Pedicle valve of the brachiopod Strophonella euglypha (Dalman, 1828) from the Wenlock Limestone of Dudley,
West Midlands; x 2. Photography by Harry Taylor of the British Museum (Natural History) Photographic Studio. One of
the specimens illustrated in the Atlas of Invertebrate Macrofossils published by the Association.

PROTOCYSTITES MENEVENSIS —

A STEM-GROUP CHORDATE (CORNUTA) FROM
THE MIDDLE CAMBRIAN OF SOUTH WALES

by R. P. S. JEFFERIES, M. LEWIS (md S. K. DONOVAN

Abstract. Protocvstiles menevensis Hicks, 1872, from the Hypagnostus pcirvifrons Zone of the Middle Cam-
brian, near St David’s, Dyfed, South Wales, is reconstructed and redescribed. It proves to be a cornute, and
therefore a stem-group chordate, representing a plesion intermediate between that of Ceratocystis peineri (the
most primitive known chordate) and that of Nevadaecystis aniericana. For purposes of reconstruction, tectonic
distortion of the fossils was corrected by means of a computer program. The positions of oesophagus, stomach,
and intestine are suggested in P. menevensis on the basis of skeletal evidence. The locomotory cycle of the
animal, which probably crept rearwards over the sea floor pulled by its tail, is deduced.

It is argued that, on a practical definition, every plesion is fundamentally paraphyletic. The term ‘more
crownward’ is used to signify that a plesion is more closely related to the relevant crown group than is some
other plesion. The term ‘nodal group’ is proposed for all those members of a group which possessed all the
autapomorphies of the crown group but none of the autapomorphies of any of the subgroups of the crown
group.

A comparison between stem chordates and the echinoderms shows that echinoderm ‘dorsal’ is homologous
with chordate ventral and vice versa, so in echinoderms the use of the terms ‘dorsal’ and ‘ventral’ should be
abandoned.

The aims of this paper are to redescribe Protocystites menevensis Hicks, 1872, from the Middle
Cambrian near St David’s, Dyfed, Wales; to locate it stratigraphically; to reconstruct its skeletal
anatomy, soft parts, and functional morphology; and to determine its systematic position. The
species proves to be a stem-group chordate of the group Cornuta. It is the second most primitive
cornute known and, at present, the oldest chordate known from Britain. Other interpretations
regard P. menevensis, and all other cornutes, as echinoderms, e.g. Ubaghs (1967, 1981) and Philip
(1979). One of us has argued elsewhere why these various views are mistaken and the arguments
will not be repeated here (see Jefferies 198 k/, h, 1986).

The present study of P. menevensis began in 1981 when the two junior authors independently
examined the lectotype A. 1021 in the Sedgwick Museum, Cambridge, and E432 in the British
Museum (Natural History). They immediately recognized them as belonging to a species of cornute
and several small expeditions to Porth-y-rhaw in 1982 to 1984 yielded much more material. The
senior author devoted most of 1984 to reconstructing the animal in detail. The stratigraphical part
of this paper results from the work of M. Lewis.

PHYLOGENETIC METHODOLOGY

The terms ‘stem group’ and ‘crown group’ as applied in this paper still require explanation (although
their use seems to be spreading: Jefferies 1979; Patterson 1981; Smith 1984a; Paul and Smith 1984;
Thulborn 1984). Given two sister groups (1 and 2) with still extant members (text-fig. 1), there are
two obvious ways of delimiting both groups when extinct forms are taken into account. The
narrower delimitation of group 2, for example, would include the latest common ancestor of all
the living members, plus all its descendants, whether living or dead. This delimitation can be called
the crown group as proposed in Jefferies (1979)— a term that corresponds to the *group of Hennig

jPalaeontology, Vol. 30, Part 3, 1987, pp. 429-484, pis. 54-60.| © The Palaeontological Association

430

PALAEONTOLOGY, VOLUME 30

(1969, 1981). The wider delimitation of group 2 would include all descendants of the latest common
ancestor of groups 1 and 2 except members of group 1. This delimitation can be called the total
group (‘Gesamtgruppe’ of Hennig). Now, if the crown group of 2 is subtracted from the total group
of 2, a paraphyletic ancestral grouping remains which can be called the stem group of 2.

Within this stem group, it is theoretically important to distinguish the stem lineage of 2 (Ax, in
press; ‘Stammlinie’ of Ax 1984)— this, for us, is the lineal sequence of ancestors and descendants
which led from the latest common ancestor of [1 +2] up to, but not including, the first member of
crown group 2, which latter was the latest common ancestor of all living members of 2. (Our usage
differs slightly from that of Ax, who includes the latest common ancestor of the living forms in the
stem lineage as he defines it.) However, the stem group does not include the stem lineage alone: it
also contains the ‘side branches’— all descendants of the stem lineage which do not belong to the

TEXT-FIG. 1. Stem group, crown group, and stem lineage. Crown groups 1 and 2 are sister groups with

some members still living.

Present day

I I n’rn 1 1 I 1

Crown group

Crown group 2

Stem group 2

Stem lineage 2

crown group. Within the stem group, it is possible to conceive of different degrees of relationship
to the crown group, and Patterson and Rosen (1977, p. 165) proposed the term ‘plesion’ for ‘fossil
groups of species sequenced in a classification according to the convention that each such group is
the plesiomorph sister-group of all those, living and fossil, that succeed it . . .’. By thus arranging
plesions within the stem group in order of increasing relationship with the crown group, it is possible
to reconstruct the sequence of origin of the autapomorphies of the crown group as these evolved
within the stem lineage.

Patterson and Rosen implied, by using the term ‘sister-group’ in the definition quoted, that
plesions are monophyletic. We should like to redefine the word ‘plesion’, however, to accord more
with practical realities, as follows: a plesion includes all, and only, those members of a stem group
which, so far as can be discerned, are equally closely related to the crown group. As discussed
below, a plesion so defined, if fully known, is necessarily paraphyletic.

JEFFERIES, LEWIS AND DONOVAN: CAMBRIAN CORNUTE

431

TEXT-FIG. 2. The paraphyletic nature of plesions and the position of the nodal group, a, the phytogeny of
stem group 2 as it actually happened, h, cladograms of plesions 1, 2, and 3 so far as reconstructable.

The term ‘crownward’ has been proposed by Jeflferies (1986, p. 13) to mean ‘more closely related
to the crown group’. Thus, within the stem group of group 2, plesion 3 (of text-fig. 2) is crownward
of plesion 2. It is less ambiguous than ‘higher’, which may indicate stratigraphy, increase in
complexity in any direction, or even moral approval; it is also clearer than ‘more advanced’ since it
implies advance towards the crown group along the stem lineage only; and it is better than ‘later’
which ought to refer only to time. As opposite to crownward, we use ‘less crownward’ to signify
position, or ‘anti-crownward’ to indicate direction.

To split a stem group into plesions we usually search for a feature shared with the crown group
by some members of the stem group, but not by others. However, the feature sought need not be
shared with all members of the crown group, since some may secondarily have lost it. And, indeed,
a feature shared with more crownward plesions may be used, even when it was later lost within
the stem lineage and is therefore primitively absent in the crown group itself. Thus the cornute
Nevadaecystis americana (Ubaghs, 1963) possesses a strut, in common with more crownward
plesions of the chordate stem group such as those of Cothwnocystis etizae Bather, 1913 and
Galliaecystis lignieresi Ubaghs, 1969 (text-fig. 26), but unlike less crownward plesions such as those
of P. menevensis and Ceratocystis penieri Jaeckel, 1900. The strut is a legitimate reason for putting
the plesion of N. americana more crownward than that of P. menevensis. The strut is absent,
however, from all mitrates, which are primitive members of the chordate crown group, and is

432

PALAEONTOLOGY, VOLUME 30

secondarily incomplete in the crownmost members of the chordate stem group, such as the cornute
Reticulocarpos hamisi Jefferies and Prokop, 1972.

The plesion concept has some difficulties. When a palaeontologist begins to divide a stem group
into plesions, each plesion will probably be monospecific, and therefore monophyletic as regards
its only known constituent, and this monophyly accords with Patterson’s and Rosen’s concept of
the term plesion. As study proceeds, however, this early false clarity will be lost, because more
than one species will come to be assigned to each plesion, and often these species will not share
synapomorphies with each other such as would show them to form, on their own, a monophyletic
group. Indeed, the stem group can be divided into plesions only to the extent which changes in the
stem lineage will allow. The smallest theoretically recognizable segment of a stem lineage will be
the sequence of ancestors and descendants lying between one evolutionary novelty and the next
more crownward one evolved in the stem lineage, e.g. the segment of the stem lineage within plesion
1, between novelties 1 and 2 in text-fig. 2. All side branches from this segment, together with the
segment itself, will necessarily be members of the same plesion, in so far as this term is usable in
practice. If, therefore, all members of such a plesion (all its constituent individuals) had come to be
known, the plesion would include part of the stem lineage as well as any side branches. And this
segment of the stem lineage would be ancestral to forms which did not belong to the plesion, i.e. to
more crownward plesions and to the crown group itself. But such a fully known plesion would be
paraphyletic by Hennig’s definition of paraphyly since, being in part ancestral to non-members, it
would have no ancestor common to it alone (Hennig 1966, p. 146). This would remain true, even if
the relevant part of the stem lineage was only a single generation. As knowledge advances, therefore,
a plesion will change from being monophyletic as regards its single known member species, to being
possibly paraphyletic as regards all its known members. On the other hand, as soon as a formerly
recognized plesion can be shown to be paraphyletic by demonstrating that an evolutionary novelty
evolved within it in the stem lineage, then it will be split into two plesions, one of which will be
more crownward than the other. The paradox is therefore reached that a plesion is by its nature
paraphyletic, but as soon as it can be shown to be so in the particular case, it splits. Also it is
possible to show that a particular fossil is not a member of the stem lineage if it possesses features
which never existed in that lineage. But it is never, or almost never, possible to show that a fossil
a member of the stem lineage.

We use the term ‘intermediate category’ (Hennig’s ‘Zwischenkategorie’) for an overtly para-
phyletic grouping of two or more adjacent plesions within a stem group. Such provable paraphyletic
groupings may sometimes be convenient to recognize. An example is the group Cornuta for the
crownward part of the chordate stem group. (It is still unknown what fossils should be placed less
crownward than the cornutes in the chordate stem group.)

We propose the term ‘nodal group’ for all those members of a monophyletic group which possess
all the autapomorphies of the crown group, but are primarily lacking any of the autapomorphies
of any of the subgroups of the crown group. Thus, with reference to group 2 in text-fig. 2, the nodal
group will show novelty 4 (the last one evolved in the stem lineage of group 2) but will lack novelties
5 and 6 (the first ones evolved in the respective stem lineages of the major subgroups of 2, i.e. 2a
and 2b). Thus the nodal group will include the latest common ancestor of the extant members of
group 2, and this gives it particular importance, but it will also contain the most crownward parts
of the stem group of 2 and the least crownward parts of the stem groups of subgroups 2a and 2b.

The word ‘calcichordate’ was proposed by one of us (Jefferies 1967) for any chordate with a
calcite skeleton of echinoderm type, and in particular for the cornutes and the mitrates. On these
definitions, P. menevensis is a calcichordate. However, in the light of Hennig’s work (1969, 1981)
the ‘Calcichordata’ form an ‘invalid stem group’ since the cornutes are stem-group chordates while
the mitrates are primitive crown-group chordates (Jefferies 1 979, 1 986). Consequently, the word ‘calci-
chordate’ is best abandoned or, at most, used informally. The word ‘Stylophora’ (Gill and Caster
1960) is coextensive in meaning with Calcichordata. It should be abandoned for the same reasons,
and also because the workers who use it mistakenly regard the cornutes and mitrates as echinoderms
and it wrongly implies that the cornute stylocone is homologous with the mitrate styloid.

JEFFERIES, LEWIS AND DONOVAN: CAMBRIAN CORNUTE

433

METHODS OF STUDY

P. menevensis was reconstructed by one of us (R.P.S.J.) on a drawing board, several projections being plotted
simultaneously as with previous such studies (e.g. Jefferies 1968, 1969). The specimens were examined by
means of latex casts to reconstruct the skeleton, and by direct observation of internal moulds to reconstruct
the soft parts. Sometimes pyrite and limonite were removed from the fossils by soaking them overnight in
10% thioglycollic acid; this cleaning allowed much better latex casts to be made.

Correction of distortion

Tectonic distortion made great difficulties. These were partly overcome by means of a computer-graphical
method based on suggestions by Appleby and Jones (1976) and Ramsay (1967). The bedding planes of the
shale in which the specimens occur are crossed by stretching lineations which give the rock a slight graininess
like that of wood. All these lineations run parallel to tight parallel folds in the thinner shelled trilobites and
represent the long axis (.v-axis) of the strain ellipse for the bedding plane; the direction at right angles to them
(j^-axis) is the direction of greatest compression in the bedding plane. The original outline of a plate would
therefore correspond to the observed outline expanded by some definite factor along the v’-axis, perpendicular
to the lineations.

L/yr

1.0-

0.9-
0.8-
0.7-
0.6- -
0.5-
0.4-

I 1 1 1 1 1 1 1 1 1

1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9

TEXT-FIG. 3. Correction of distortion on the basis of two specimens of P.
menevensis Hicks plate g (see text-fig. 10) compressed approximately perpen-
dicular to each other. The .v-axis is the presumed major axis of the strain ellipse.
Specimen a, BM(NH) E62952; specimen h, BM(NH) E62930; proportionate
increase of unit length along the v-axis, relative to the .v-axis (further expla-
nation in text).

434

PALAEONTOLOGY, VOLUME 30

1.60 1.90

TEXT-FIG. 4. Correction of distortion of a hind-tail ossicle of P. menevensis Hicks,
BM(NH) E432. The numbers show (see text).

A computer program was devised by Mr A. J. Paterson of the Biometrics Section, British Museum (Natural
History), to modify the observed outlines in the manner required. To use this program, camera-lucida drawings
of the specimen were placed on a digitizer with the .v-axis (assumed to be parallel to the stretching lineations)
arranged parallel to the .\-axis of the digitizer. The shape in the drawing was transformed into x-y coordinates
by tracing the outline with the cross-wires on the ‘puck’ (follower) of the digitizer. The r coordinates were
then multiplied by a factor n, while the corresponding .v coordinates were multiplied by \jn. The results of
these multiplications were displayed on a visual display unit and simultaneously drawn, as needed, on a
plotter. It was also possible to multiply both sets of coordinates by a uniform factor M so that the visual
output was magnified to a convenient size. The proportionate increase of the v coordinate relative to the .v
coordinate was (since n^( I//?) = iP).

To decide the appropriate value of and therefore «, was not easy. With initially symmetrical structures,
such as the obliquely distorted heads or tails of trilobites or the tail ossicles of Protocystites, the presumed’
correct value restored the initial symmetry. With asymmetrical structures, such as the head plates of
Protocystites, it was necessary to find two specimens of the same plate compressed in different directions,
preferably at right angles. The appropriate value of would then be the one that gave the same shape
to the two specimens. In fact, this ideal agreement was never achieved, and it was therefore necessary to use
some index of shape, such as the ratio of the length of two chosen lines on the plates or the angle of
some prominent corner. The appropriate value of tP was then the one that gave identical values for the
index.

When comparing two specimens of corresponding plates compressed in different directions, a series of
computer plots was made for both plates with tP increasing at increments of 0 05 from 105 to 1-70. Graphs
were then drawn of iP against the measured value for the chosen index for both plates. The appropriate value
of )P was that at which the indexes of both plates were equal, i.e. where the lines for the two plates crossed
each other on the graph. At this value, the computer plot was presumed to show the original shape of the
plate.

The deduced values for iP are not uniform. Three head plates (e, g, and k on text-fig. 10) gave values of
close to 1-20 (text-fig. 3). The same value was found to hold for an isolated hind-tail ossicle (text-fig. 4) and
for the hind-tail ossicles of the lectotype (designated below). On the other hand, two specimens of the left

1.00 1.45 1-60 1.90

b

TEXT-FIG. 5. Correction of distortion for two trilobite
cephalons. a, Ptychagnostus. b, Eodiscus. The num-
bers show (see text).

X axis

JEFFERIES, LEWIS AND DONOVAN: CAMBRIAN CORNUTE

435

process (the process on plate 1), which was thin and is sometimes visibly crumpled, required a higher value of
at about 1-60. The same value was required to restore symmetry to the trilobite Eoclisciis (text-hg. 5h)
whereas a specimen of the cephalon of Ptycliagnostus (text-hg. 5a) became symmetrical at about iP = 115.
The highest values of iP, as shown by the trilobite Eodiscus, probably approach the distortion of the
matrix itself, whereas the lower values usually shown by the skeleton of Protocystites meneveusis suggest
that stereom calcite or its pyritic replacement was, because of its strength, less distorted than the surround-
ing rock. Indeed many specimens of P. menevensis seem to have responded to tectonic pressure by the
plates sliding over or against each other, as well as by changing their outline. Sometimes adjacent tail
ossicles have been pushed against each other, causing very high local pressures and non-homogeneous
distortion, as testified by highly asymmetrical outlines. It is impossible to make proper allowance for these
variations.

The reconstructions are therefore based on computer plots with iP= 1-20 (n = I 095) which seems to be
correct for the tougher parts of the head and tail skeleton of P. menevensis (cf. text-fig. 6). The results are
probably better than those that would be obtained using uncorrected drawings of the specimens. Nevertheless,
there is uncertainty about the shapes of the plates and the relative sizes of different parts of the animal, and
this must be remembered in considering the reconstructions.

n z 1.00

n :1.20

TEXT-FIG. 6. Protocystites menevensis Hicks; uncorrectcd and corrected drawings of
BM(NH) E62963 (ventral aspect). The numbers show iP (see text).

Distortion perpendicular to the bedding planes is an even bigger problem than distortion in the plane of
bedding. Thin parts of the skeleton, which originally stood almost vertical, and thus perpendicular to the
bedding plane, have sometimes been squashed flat on to the bedding plane. This is particularly true of the
posterior wall of the head in plates f, g, j, and k (see text-fig. lO). To obtain some idea of the original shape of
these plates, replicas based on computer plots with iP = I -20 were cut in aluminium sheet and bent to the
likely original shape in three dimensions. Once again, the results are uncertain, so that the vertical dimension
of the reconstructions is not reliable.

Tectonic distortion can thus be partly, but not totally, corrected; better reconstructions will require undis-
torted material (which may never be found). It is remarkable that tectonic distortion has not destroyed the
histology of the plates, for the superficial features of different types of stereom can readily be recognized. (The
three-dimensional structure of the stereom is not usually deducible.)

436

PALAEONTOLOGY, VOLUME 30

SYSTEMATIC PALAEONTOLOGY

Superphylum deuterostomia Grobben, 1908
Subsuperphylum dexiothetica Jeflferies, 1979
Phylum CHORDATA Bateson, 1886
[Stem group of the Chordata]

Intermediate category cornuta Jaekel, 1901
Plesion of Protocystites menevensis herein
Genus protocystites Hicks, 1872

Type species. P. menevensis Hicks, 1872, by monotypy.

Systematic position. The above statement of systematic position is unorthodox. We deliberately have not
placed P. menevensis in a family because it is at present alone in its plesion. Any family which included it,
therefore, would either be: 1, coextensive with the species P. menevensis and therefore redundant; or 2, it
would include one or more of the adjacent plesions of the chordate stem group— it would thus be overtly
paraphyletic and (unlike the overtly paraphyletic grouping Cornuta, for example) would, in our opinion,
never be a useful grouping in practice. Another unorthodoxy is that the intermediate category Cornuta is here
given no conventional Linnaean rank. This omission is likewise deliberate and is based on the fact that nobody
has yet explained how Linnaean rank can objectively be assigned, particularly to paraphyletic groupings of
fossils (Ax, 1984, Ch. K; Ax, in press). Those ranks which are assigned above are either hallowed, though not
validated, by long usage (superphylum, phylum) or else are obtained by interpolation (subsuperphylum).

Protocystites menevensis Hicks, 1872

Plates 54-60; text-figs. 6, 10, 15-19, 23-25

1866 Protocystites Salter in Hicks and Salter, p. 285 {nom. mid.).

1871 Protocystites menevensis Hicks in Harkness and Hicks, p. 396 {nom. mid.).

1872 Protocystites menevensis Hicks, pi. 5, fig. 19; p. 180 (lower illustration only).

1873 Protocystites menevensis Hicks; Salter, p. 3.

1887 Protocystites meneviensis Hicks; Barrande, p. 10.

1900 Protocystis Hicks; Bather, p. 48.

1943 Protocystites meneviensis Hicks; Bassler and Moody, p. 1 84.

1967 Protocystites Hicks; Ubaghs, p. S493.

1967 Protocystites meneviensis Hicks; Paul in Jefferies et at., p. 567.

1979 Protocystites meneviensis Hicks; Paul, p. 453.

Comments on synonymy. The only previous figure and description of Protocystites menevensis was that of
Hicks (1872). His illustration shows two specimens which can be identified with specimens A. 1021 and A. 1022
now held in the Sedgwick Museum, Cambridge. Of these, A. 1022 is the operculum of a hyolithid, whereas
A. 1021 is a fairly good specimen of the species described in this paper and is here designated as lectotype
(text-figs. 17 and 18; PI. 56, figs. 1-3; PI. 57, figs. 1 and 2). Hicks’s figure is very poor and the species is not
recognizable from it. Subsequent references to the species or genus, therefore, add nothing to knowledge and
are of purely bibliographic interest.

The name Protocystites must not be confused with Proteocystites Barrande, 1887 which is a diplorite cystoid
(see e.g. Kesling 1967, p. S248) and an echinoderm.

Material, horizon, and locality. The material examined is as follows;

(a) Sedgwick Museum, Cambridge: A. 1021 (here chosen as lectotype) ex Hicks Collection, locality Porth-
y-rhaw near St David’s, Dyfed, horizon Menevian. Figured by Hicks (1872, pi. 5, fig. 19, lower figured
specimen only) (text-figs. 17 and 18; PI. 56, figs. 1-3; PI. 57, figs. 1 and 2).

(h) British Museum (Natural History), London; i, old material, E432 ex Hicks Collection, locality St
David’s, horizon Menevian (text-fig. 19; PI. 58, fig. 1); ii, new material, E62912-E62921, E62923-E62925, all
from loc. 1, Porth-y-rhaw (text-figs. 7 and 8), horizon middle part of Hypagnostus parvifrons Zone, Middle
Menevian; iii, also new material, E62926-E62934, E62937-E62939, E62942-E62945, E62950, E62952, E62955-
E62966, E62968-E62981, E63006-E63056, all from loc. 2, Porth-y-rhaw (text-fig. 7), horizon nearly or exactly
the same as for loc. 1 .

JEFFERIES. LEWIS AND DONOVAN: CAMBRIAN CORNUTE

437

TEXT-FIG. 7. Geology and topography of Porth-y-rhaw, Dyfed, Wales. I and 2 arc the
localities that produced the new material of Protocystites menevensis Hicks. X and Y are
respectively the lower and upper ends of the composite profile shown in text-fig. 8.

438

PALAEONTOLOGY, VOLUME 30

(c) National Museum of Wales, Cardiff: NMW. 80. 34G. 948-958, all from loc. 1, Porth-y-rhaw, same
horizon as BM(NH) material from that locality.

Most of the material consists of dissociated plates. All of it is tectonically distorted. Articulated specimens
include SM A. 1021 (lectotype) and BM(NH) E432, E62950, E62952, E62963 (the most instructive specimen;
text-figs. 6, 15, f 6; PI. 54, figs, f -3; PI. 55, figs. 1 and 2), E62977, E62979.

Porth-y-rhaw, from which all the recently found material came (our Iocs. 1 and 2; text-figs. 7 and 8) (as also
did one, or perhaps both, of the two nineteenth-century specimens), is a small inlet situated on the coast of
Dyfed, Wales, about 3-6 km east-south-east of the cathedral of St David’s and about 1-5 km west of Solva
Harbour (text-fig. 7). The east side of Porth-y-rhaw (including our loc. 1 ) is the type section (text-fig. 8) of the
Menevian Group of Hicks and Salter (1866). It was on this eastern side that Salter discovered Paracloxides
davidis Salter, 1863 and its associated fauna in 1862, by chance, as a result of misnavigation. Text-hg. 9 shows
the now accepted stratigraphical divisions for the Middle Cambrian near St David’s.

Loc. 1 is on the eastern cliff section, Porth-y-rhaw (NGR SM 78596 24235), middle part of H. parvifrons
Zone, at shore level, stratigraphically c. 22-24 m below the basal contact of two, almost vertical, 4 m thick
sills and approximately 10 12 m stratigraphically below the local base of the Ptychagnostus pimctiiosus Zone
(text-figs. 7 and 8). The nature and distribution of the fauna is shown in text-fig. 8.

Loc. 2 (text-fig. 7) is on the western cliff section, also in the H. parvifrons Zone at or near the same horizon
as loc. 1, just above the cliff top near the southern end of Porth-y-rhaw (NGR SM 78492 24252). Here the
beds are less cleaved and more suitably weathered for yielding fossils than at loc. 1 . In addition to Protocyslites
menevensis, the fauna here comprises the trilobites Ptychagnostus ciceroides (Matthew, 1896), P. davidis (Hicks,
1872), P. punctiiosus affinis (Brogger, 1878), P. pimctiiosus s.l. (Angelin, 1851), Cotalagnostus lens (Grdnwall,
1902), Phalacroma hibullatum (Barrande, 1846), Peronopsis scutalis scutalis (Hicks, 1872), P. falla.x depressa
Westergard, 1946, Phalagnostus cC. nudus {Bey rich, 1845), Eodiscus punctatus punctatus (Saher, 1864), Agraulos
longicephalus (Hicks, 1872), Jincella applanata (Hicks, 1872), Hartshillina spinata (Illing, 1916), Clarella salteri
(Hicks in Salter, 1865), Acontheiis n. sp. aff. A. inarmatus Hutchinson, 1962, and a pagetiid (new genus); and
the non-trilobite fossils Linnarssonia sagittalis (Davidson, 1871), Lingidella sp., Hyolithes corrugatus (Salter,
1864), Stenotheca cornucopia Hicks, 1872, Protospongia fenestrata Salter, 1864, and Ctenocystis sp. The record
of Ctenocystis sp. is of interest as being the only known occurrence of the genus in Britain.

In the western cliff section, where loc. 2 is situated, the stratigraphically overlying punctuosus Zone seems to
be unrepresented and the base of the parvifrons Zone could not be located exactly. However, evidence of the
underlying Tomagnostus fissus- Ptychagnostus ataviis Zone was found 40 m stratigraphically below loc. 2. As
already stated, some 12 m of the parvifrons Zone exists in the eastern cliff section above the horizon of loc. 1
(text-fig. 8). If the horizons of Iocs. 1 and 2 are identical, therefore, the greatest possible thickness of the
parvifrons Zone at Porth-y-rhaw is 12-1-40 = c. 52 m.

In biostratigraphic terms, ‘Menevian’ (text-fig. 9) conventionally refers to the traditional zones of Paradox-
ides hicksii and P. davidis and possibly other zones. References covering the most important faunas in the
Menevian Group include Salter (1863, 1864, 1865), Salter and Hicks (1867, 1869), Hicks in Harkness and
Hicks (1871), and Hicks and Jones (1872). Most of the trilobites were redescribed by Lake (1906-1946) and
they indicate the presence of ihc fissus, parvifrons, and punctuosus zones of Scandinavian terminology.

The term ‘Lower Menevian’ is equivalent to the hicksii Zone of authors, which can be equated approximately
with the fissus-atavus Zone of Sweden. P. hicksii is here considered a probable senior subjective synonym of

TEXT-FIG. 8. Stratigraphic log of the beds exposed on the foreshore on the eastern side of Porth-y-rhaw,
between points X and Y of text-fig. 7, and a complete list of the fossils found in that section. The profile is a
composite, built up from three separate sections as shown in text-fig. 7. Locality I is at 22-24 m below the
lower dolerite sill and is just above mean high-water mark. Abbreviations: M, mud; S, silt; LS, fine sand; MS,
medium sand; CS, coarse sand; MHW, mean high water; MLW, mean low water; 1, Hyolithes sp.; 2,
Ptychagnostus punctuosus affinis', 3, Protocystites menevensis', 4, pagetiid gen. et sp. nov.; 5, Linnarssonia
sagittalis', 6, centropleurine fragments; 7, Meneviella sp. indet.; 8, Ptychagnostus punctuosus s.l.; 9, Acontheus
sp. nov.; 10, P. davidis', 11,//. corrugatus', 12, Jincella applanata', 13, Peronopsis fallax depressa', 14, Protospongia
sp.; 15, conocoryphid gen. et sp. indet.; 16, Peronopsis sp.; 17, M. cf. vemdosci', 18, Heperditkf hicksi', 19, P.
scutalis scutalis", 20, Eodiscus punctatus punctatus', 21, Cotalagnostus lens (s.l.); 22, Ptychagnostus ciceroides',
23, M. venuloscr, 24, Clarella sp.; 25, P. punctuosus punctuosus', 26, Anopolenus henrici', 27, Peronopsis ex gr.
fallax', 28, Pleuroctenium cf. hifurcatum', 29, Holocephalina cf. primordialis', 30, Phalagnostus cf. nudus', 31,
Solenopleurina variolaris', 32, Pseudoperonopsis sp.; 33, Paracloxides davidis.

' 0

1

2-

3-

4-

V)

(b 5'

|e-

7-

8-

9-

1 0-

1 1-

1 2-

13-

14-

1 5-

1 6-

1 7-

18-

19-

20-

2 1-

22-

2 3-

2 4-

2 5-

26-

2 7-

28-

2 9-

JEFFERIES, LEWIS AND DONOVAN: CAMBRIAN CORNUTE

439

I I I I

1 \

\

\ ^

/ -
s

/ /

/

\ /

/

V

-s

^ /

33

20“=.' 25 32

2223 • 262728293031 '

Tentative base of P. punctuosus Zone

1516 17’.®, g

10 1.1

■MHW

I -i-r»-|i^-|-i:i MLW

M S MS
FS CS

440

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 9. Stratigraphical subdivisions of
the Middle Cambrian of St David’s (mo-
dified after Cowie et al. 1972). The zones of
Paradoxides hicksii and P. davidis are tra-
ditional for the South Welsh area; those
shown to their right are the equivalent zones
of the Scandinavian succession. The
traditional zones are shown here with their
traditional extent, although P. davidis at
Porth-y-rhaw is in fact unknown outside the
Ptychagnostus punctiiosus Zone.

P. aurora Salter in Salter and Hicks (1869), and its range is extended downwards to the base of the Upper
Solva Beds. The Lower Menevian Beds, estimated by Hicks (1881, p. 299) to be 300 ft (91 m) thick, consist of
light-grey and dark-grey laminated mudstones which suffered soft sediment deformation in the lower part.
Towards the top of the division are some greenish mudstone units, lithologically similar to the Upper Caered
Mudstones (Nicholas 1916) of St Tudwal’s Peninsula, North Wales.

The Middle Menevian Beds were estimated by Hicks (1881, p. 299) and by Stead and Williams (1971,
p. 181) to be 350 ft (107 m) thick, although a precise boundary between these and the Lower Menevian Beds
was not defined by these authors. The Middle Menevian Beds are darker and more uniform in colour than
the Lower Menevian Beds and consist of cleaved pyritic mudstones with occasional thin, sometimes lenticular,
sandy horizons and several thin ( < 10 cm) pale beds which, according to Nicholas (1916, p. 99) are composed
of ashy material. Certain beds contain numerous small flattened phosphatic nodules, and these and the
Lower Menevian strata appear to have been deposited in a euxinic environment (Rushton 1974, p. 90). In
biostratigraphic terms ‘Middle Menevian’ is equivalent to the davidis Zone of certain authors, which corre-
sponds approximately to the parvifrons and punctuosus zones of Scandinavian nomenclature (text-fig. 9). The
species P. davidis, however, seems at Porth-y-rhaw to be confined to the punctuosus Zone.

The Upper Menevian Beds at Porth-y-rhaw abruptly succeed the Middle Menevian beds and comprise
coarse, dark-grey sandstones with shaly interbeds. These sandstones, exposed at the tip of the eastern headland
of Porth-y-rhaw (text-fig. 7), are massive. They form beds up to 1 m thick at the base of the unit and seem to
mark the initiation of deposition from current-agitated water or from turbidity currents. This style of depo-
sition continued into "Lingula Flags’ times (Rushton 1974, p. 90).

The Upper Menevian Beds of Porth-y-rhaw were said to be 100 ft (30 m) thick by Hicks (1881, p. 299) and
Stead and Williams (1971, p. 81) but these authors did not define an upper limit. The sandstones contain
" BiUingsella' liicksi (Davidson) and other brachiopods and are commonly referred to as 'Orthis' hicksi Beds.

JEFFERIES, LEWIS AND DONOVAN: CAMBRIAN CORNUTE

441

Hicks (1892, p. 22) collected Paradoxides and a new species oi "Conocoryphe from them. These identifications
suggest that the Upper Menevian of this locality may partly correspond to the P . forchhammeri ‘Stage’ of the
Scandinavian sequence.

Stead and Williams (1971, p. 188) believed that the junction between Middle and Upper Menevian beds
was conformable. However, Taylor and Rushton (1972, p. 9) suggested a widespread non-sequence at this
level in England, Wales, and south-eastern Newfoundland caused by a regression during the time of the
Solenopleitra brachymetopa Zone.

As to correlation, the beds of the parvifrons Zone are correlated with: the upper part of the Nant-pig
Mudstones of St Tudwal’s Peninsula, North Wales (Rushton 1974, p. 72; after Nicholas 1916); Illing’s (1916)
horizons F1-F3, representing the parvifrons Zone in the Abbey Shales of Warwickshire (Rushton 1979,
p. 43); beds in the lower part of the Clogau Formation in the Harlech Dome, North Wales (Allen et al. 1981,
p. 303); the lower part of the davidis Zone in the Manuels River Formation, south-eastern Newfoundland
(Hutchinson 1962); and the parvifrons Zone of Scandinavia (summarized by Martinsson 1974). Thomas et al.
(1984, p. 888) discussed the distribution of the various trilobite species in these rocks.

The lithostratigraphical terminology used for the rocks of Porth-y-rhaw in this paper is in need of revision,
but it is not appropriate to make this revision here.

Thus the total stratigraphical range of Protocystites menevensis is not known, although all of the abundant
new material comes from the middle part of the parvifrons Zone of the Middle Menevian. The horizon of the
lectotype cannot be ascertained exactly, beyond the fact that it came from the Menevian of Porth-y-rhaw.
Harkness and Hicks (1871, p. 396) recorded the species from both Lower and Middle Menevian Beds but this
does not tally with our experience.

Preservation and conditions of deposition. The plates of P. menevensis and Ctenocystis have been completely
replaced by pyrite (often converted to limonite), or are represented by air-filled holes. No trace of calcite
remains. The superficial histological detail is well preserved so that it is possible to recognize the surface
features of different types of stereom. On the other hand, all the plates have been distorted tectonically and
are often squashed on to the bedding planes.

On the sea-floor, the black shale matrix was probably a stinking black mud. However, there is abundant
benthos which indicates that the bottom water was usually oxygenated. Conditions may have resembled those
of the German Lower Devonian Hunsriickschiefer, which are similarly black with abundant pyritized benthic
fossils. Seilacher and Hemleben (1966) have argued, on the basis mainly of trace fossils, that the Hunsriick-
schiefer were normally laid down in oxygenated bottom water, but that sometimes the bottom water lost its
oxygen, asphyxiating the benthos. The same may have been true of the sea in which P. menevensis lived. As
we show later, the species is adapted to staying up on extremely soft mud.

Anatomical description

Introduction. In describing the anatomy of P. menevensis we frequently refer, in passing, to related species,
since without comparison morphological features have little significance. The particular species compared are:
Ceratocystis perneri Saekd from the Middle Cambrian of Bohemia, Czechoslovakia (text-fig. 1 1 ); Nevcidaecystis
americana (Ubaghs) from the Upper Cambrian of Nevada (text-fig. 12); Cothurnocysiis elizae Bather from the
Upper Ordovician of Scotland (text-fig. 14); and 'C.' fellinensis Ubaghs, 1969 from the Lower Ordovician of
the South of France (text-fig. 13).

To anticipate the arguments given below under ‘Systematic Position’, Ceratocystis perneri is less crownward
(less closely related to the chordate crown group) than P. menevensis, and in many respects shows the
most primitive condition among known cornutes; N. americana is more crownward than P. menevensis',
'Cothurnocystis' fellinensis is more crownward than N. americana', and C. elizae is more crownward than ‘C.’
fellinensis (text-fig. 26). These conclusions are mentioned early so that the anatomical description, which is
comparative, will be easier to understand. It is unfortunate that N. americana is known only from one
specimen and that the floor of the head is not visible from beneath, though partly visible from above.

As regards plate notation, one of us used to employ an objective system (Jefferies 1968) in which marginal
plates were numbered from the anterior end of the tail and were given suffixes for left and right, dorsal and
ventral (e.g. Mild was the first left dorsal marginal plate). This system was explicitly intended not to imply
homology, so that marginal M3L of C. elizae, for example, was not homologous with M3L of N. americana—
they correspond, respectively, to plates t and I of the comparative notation. In earlier studies of cornutes and
mitrates, such an objective system was necessary. However, comparative anatomy depends on recognizing
and naming homologies, accepting the risk of thereby making mistakes. Accordingly, a comparative, non-
objective plate notation was proposed by Jefferies and Prokop (1972) when describing the crownward cornutc

442

PALAEONTOLOGY, VOLUME 30

k-spike

narrow groove (eer)

ventral branchial integument

l-appendage

right integument

right posterior
Integument

branchial slit
posterior U“plate

dorsal branchial
Integument

gonopore

anus

dorsal groove
(median eye)

TEXT-FIG. 10. Protocystites menevensis Hicks; reconstructed external morphology, a, dorsal; b, ventral; c,
posterior; and d, right lateral aspects. Letters a w, e, ii, and numbers 1, 2, 3, 5, 6 indicate plate homologies as
explained in text. The exact number of branchial slits is not known.

JEFFERIES, LEWIS AND DONOVAN: CAMBRIAN CORNUTE

443

> ventral prominence
keel

incipient hinge-line

TEXT-FIG. 1 1 . Ceratocystis perneri Jaekel; reconstructed external morphology, a, dorsal; h, ventral; c, posterior;

and t/, right lateral aspects.

Reliculocarpos hcmiisi. In this notation, marginal plates are given lower-case letters of the roman alphabet.
The series starts at plate a, which occupies a position near, but not at, the anterior left part of the head, and is
followed by plate b just left of the mouth, and so forth in clockwise order in dorsal aspect around the head to
finish at plate 1 in R. hanusi. In applying this scheme to species other than R. hantisi, additional letters are
interpolated as necessary and, since the number of homologous series of marginal plates in cornutes and

444

PALAEONTOLOGY, VOLUME 30

J fractures
matrix

internal surface of

ventral plates

keel

l-appendage

anterior U
flap

posterior U

integument
of buc. cavity

e-spike

integument
of pharynx

f-spike

?gonopore-anus

gonorectal groove

TEXT-FIG. 12. Nevadaecyslis americana (Ubaghs); Upper Cambrian, Nevada, United States
National Museum 143237; diagrammatic drawing of the dorsal aspect of the only known
specimen and holotype (redrawn after Jefferies 1969, text-hg. 4).

mitrates now exceeds twenty-six, the English roman alphabet is supplemented by letters from German and
French (ii, e, etc). (We do not imply that plate ii has any special relation to plate u, nor plate e to plate e; nor
do the U-plates of the branchial slits form part of this notation, for with them the U is upper case and
describes the shape of the plates.) For centrodorsal plates, in which homologies can be recognized between
Ceratocystis periieri, P. menevensis, and N. americana, the numbers 1 to 6 are employed, starting from the
situation in C. perneri where the sequence is once again clockwise in dorsal aspect. Suffixes are used when
several plates correspond to a single dorsal plate in C. perneri. This comparative notation is difficult to
memorize, but becomes clear in the text-figures.

Various protuberances were borne on the heads of cornutes and, like the plates, these can now be homolo-
gized from species to species. Such processes were of two main types— appendages and ventral spikes— though
the distinction between the two is not precise. Appendages made up parts of the anterior margin of the head
and projected horizontally, or horizontally and downwards; spikes, by contrast, projected downwards from
the ventral surface. Jefferies (1968) used an objective notation for the spikes (Sir, S2R, etc.) much like the
notation used for the plates, whereas the appendages were referred to as ‘left’, ‘left oral’, and ‘right oral’. In
this paper, we replace these terms by a new notation which refers to the plate on which the process was
carried, e.g. e-spike, l-appendage. As with the plates, the new notation is not objective but implies homology.
It is convenient, because a plate in cornutes seldom bears more than one individualized process (though that
process is often complicated in shape) and such a process is never constituted from more than one plate.

The nomenclature of different types of stereom follows Smith (19846, fig. 3.2).

General external features and the plates of the head. P. menevensis (text-fig. 10) consisted of a head and a tail,
like every other cornute and mitrate. The outline of the head was boot-shaped, as was common among
primitive cornutes and it particularly resembled that of C. perneri in that plate k extended further leftward
than plate 1. The right margin of the head, as reconstructed, was almost straight and at right angles to the
posterior margin. This could be a mistake in the reconstruction caused by tectonic distortion, but if there
is no mistake the right margin of the head was different in shape to that of C. perneri in which it ran

JEFFERIES, LEWIS AND DONOVAN; CAMBRIAN CORNUTE

445

leftwards and forwards. In P. menevensis it was in this respect probably more like N. ainericana, "Cothunw-
cystis' fellinensis, and C. elizae.

The ventral surface of the head of P. menevensis was made up, except for a small oral integument, of large
plates and would have been rigid (PI. 55, fig. 1). The dorsal surface also contained some large plates, but was
mainly covered with plated integument (PI. 54, fig. 1). This combination of rigid floor and flexible roof is
otherwise known in cornutes only in N. aniericana. It was morphologically intermediate between Cercitocystis
perneri, where the floor and roof were both rigid, and Cothwnocystis elizae, for example, where the floor
and roof were both flexible though the floor was crossed by a strut. The large plates forming the floor of
P. menevensis can be called a ventral shield. We shall describe this shield first, before discussing the skeleton
attached to it dorsally.

TEXT-FIG. 13. 'Cotinirnocysris' fellinensis Ub-
aghs; Lower Ordovician (probably Lower Ar-
enig), Montagne Noire, France; dorsal aspect
of marginal skeleton of head (redrawn after
Ubaghs 1969, fig. 19).

The right edge of the ventral shield, and of the head, of P. menevensis was made up of plates g, f, e, and c.
Plate g (PI. 55, figs. I and 3; PI. 58, fig. 1; PI. 60, figs. 3, 4, 9, 10) was the first right ventral marginal (M|rv iri
the old notation) and included the right half of the tail attachment. Much of its posterior part was almost
vertical, forming part of the posterior wall of the head. Its anterior part was horizontal, however, forming
part of the rigid floor. Where ventral and horizontal portions joined, there was a distinct fold (oesophageal
fold), concave dorsally and convex ventrally, which ran rightwards into plate f. Another fold, similarly convex
downwards but wider and resembling half of a crescent in plan view, was situated just anterior to the tail
attachment and underlay the right half of the posterior coelom (see below, under The chambers and soft
anatomy of the head’).

Plate f (PI. 54, figs. 1 and 3; PI. 55, figs. 1, 4, 5; PI. 56, fig. 1; PI. 57, figs. 1 and 4; PI. 58, figs. 1; PI. 60, figs. 5,
6, 7) formed the posterior right corner of of the ventral shield, i.e. the ‘heel’ part of the ‘boot’. Like plate g, it
was divided into a horizontal portion, which was part of the floor, and a vertical portion, which formed part
of the posterior and right lateral walls of the head. Plate f was drawn out horizontally into a sharp-edged
peripheral flange and the posterior end of this flange was turned downwards to form the f-spike. Plate f carried
the right end of the oesophageal fold, the rest of which was on plate g.

Plate e (pi. 54, figs. 1-3; PI. 55, fig. 1; PI. 56, fig. 1; PI. 57, fig. 3; PI. 58, fig. 4) formed the middle part of the
right edge of the ventral shield. It had a sharp-edged peripheral flange which was continuous with that of
plate f and which ran forward into that of plate c. In dorsal aspect it showed, on the right, a distinct marginal
frame, the middle part of which articulated with dorsocentral plate 5. On the ventral surface of plate e there
was an e-spike, homologous with that of other cornutes. We use the term ‘spike’ for consistency with other
cornutes, though the word is not totally appropriate in P. menevensis since the process was rounded in shape
and, in particular, had the same slope in all directions, whereas most spikes in cornutes have the anterior
slope steeper and sharper than the posterior one. The e-spike of P. menevensis was hollow, corresponding to a
circular concavity in the internal dorsal face of plate e.

446

PALAEONTOLOGY, VOLUME 30

l-appendage

b-appendage b

mouth c-appendage

muscle
articulation (d

branchial slits

gonopore-anus

•• • • •. «Vf

fore tail
mid tail
hind tail

t-spike

k-spike

stylocone

TEXT-FIG. 14. Cotlnmiocysfis elizae Bather; Upper Ordovician (Ashgill), Girvan, Scotland (redrawn after
Jefferies 1968, text-fig. 1), with most of the hind tail omitted, a, dorsal; b, ventral; c, posterior; and </, right

lateral aspect of right side of frame only.

Plate c (PI. 54, figs. 1 and 3; PI. 55, fig. 1; PI. 58, fig. 1), which included the c-appendage (right oral
appendage), was the most anterior plate on the right side of the ventral shield. It had a sharp-edged keel
which was a forward continuation of the peripheral flange of plate e. It was ventrally convex and dorsally
plane. Dorsally it articulated with, or was lightly sutured to, dorsocentral plate 3.

The left edge of the ventral shield was formed of plates j, k, 1, x, and b. Plate], which is not well shown in
any specimen (but see PI. 54, fig. 4; PI. 55, fig. 1; PI. 56, fig. 3; PI. 57, fig. 1 ) was the first left marginal (Milv of
the old notation). It had a vertical portion, forming part of the posterior wall of the head, and a horizontal
portion included in the floor. Plate] carried the left half of the tail insertion. A ventrally convex portion of it,
resembling half a crescent in plan view, lay ]ust anterior to the tail and was the floor of the left half of the
posterior coelom. In the leftward portion of the plate the ]unction between the horizontal and the vertical
parts was abrupt and presented an almost rectangular keel on the external surface. Just dorsal to this keel was
an opening (PI. 55, figs. 1 and 2) which corresponded in general position to the narrow groove of Ceralocystis
perneri, though it penetrates plate] and not plate i as in C. perneri. The opening probably represented an ear,
functioning as lateral line, and was thus a primitive representation of the acustico-lateralis system (Jefferies
1969, p. 522; 1986, Ch. 7). Dorsally and towards the median line, plate] of P. menevensis would have articulated
with plate i and was notched by the ventral margin of the gonopore-anus.

Plate k (PI. 54, figs. 1, 2, 4; PI. 55, fig. 1; PI. 56, fig. 1; PI. 57, fig. 1; PI. 58, fig. 1; PI. 60, fig. 1) of the ventral
shield of P. menevensis was triangular in plan view and was the leftmost plate in the head. Its horizontal
portion made part of the floor and was sharply distinct from the vertical portion which formed part of the
posterior and left lateral walls of the head. There was a ventral k-spike (PI. 60, fig. 1) which was V-shaped,

JEFFERIES, LEWIS AND DONOVAN: CAMBRIAN CORNUTE

447

with the limbs of the V running respectively along the left boundary and posterior boundary of plate k. The
smooth, outward-facing surfaces of the spike were continuous, respectively, with the posterior and left lateral
walls of the plate and were almost vertical. The inward-facing surfaces of the spike were rough and made an
angle of c. 45° to the horizontal. (The corresponding k-spike of C. perneri (text-Iig. 1 1 ) was similarly V-shaped,
though much less well marked than that of P. menevensis.) In dorsal aspect, plate k of P. menevensis showed a
keel which ran forward and rightward from the left corner of the head up towards the dorsal surface. The
dorsal edge of plate k (PI. 60, fig. 2), unlike that of C. perneri, was entire and not notched for branchial slits.

Plate I (PI. 55, fig. I; PI. 57, fig. I; PI. 58, figs. 1 and 3) lay anterior to plate k and consisted of two portions
abruptly separated from each other— the marginal frame and the horizontal lamina. The frame made part of
the left wall of the head while the lamina formed part of the floor. The frame extended into a sort of wing—
the l-appendage (left appendage)— which was approximately horizontal in disposition, or possibly sloped
downwards and forwards to judge by the situation in C. perneri. The edge of the l-appendage was continuous
posteriorly with the keel that separated the horizontal portion from the left vertical portion of plate k.
Anteriorly the edge of the l-appendage passed into a keel which became less and less distinct, to disappear
approximately at the anterior border of the plate.

Plate X (PI. 55, fig. 1; PI. 58, fig. I ) was elongate and only its anterior third contributed to the left margin of
the head and ventral shield and to the left lateral wall. Posteriorly it was horizontal and ran rearward to abut
against plates g and j.

Plate b (PI. 54, fig. I; PI. 55, fig. 1; PI. 58, fig. I; PI 59, figs. 2-4) included the b-appendage (left oral
appendage) and was the most anterior plate of the left side. Like plate c on the other side of the mouth, it was
convex ventrally and flat dorsally. Unlike plate c, however, there was a thin laminar portion of plate b which
extended rightwards to form part of the floor.

Other plates made part of the ventral shield but did not contribute to its right and left edges in ventral
aspect. These were plates d, w, and v (PI. 55, fig. 1; PI. 58, fig. I ) behind the oral integument, and plates a and
e whose positions were more central. Plates d, w, and v made a post-oral series and their anterior outlines
were concave where they abutted against the oral integument. The position of plate d thus resembled that of
the same plate in C. perneri (text-fig. 116), while in Nevadaecystis and all cornutes more crownward than it,
plate d had become a marginal plate inserted between plates e and c (text-figs. 12 14). Plate a was a large
polygonal plate of central position in P. menevensis, whereas plate e was a smaller plate situated between plates
a and f. (C. perneri likewise had a plate e, but also a single large plate corresponding to plates w-l-a-l-x of
P. menevensis.) In 'Cotinirnocystis' fellinensis (text-fig. 13) and many more crownward cornutes, on the other
hand, plate a was a marginal plate forming the anterior part of the strut and the adjacent parts of the frame.
In many crownward cornutes, including C. elizae (text-fig. 14), plate x did not exist, whereas in C. fellinensis it
was a marginal plate overlying parts of plate a.

Concerning the strut, which crosses the head from front to rear in most cornutes, the situation in P.
menevensis is complicated. The strut of C. elizae (text-fig. 14), for instance, was formed of parts of plates g
and a and it divided the ventral integument into two parts. Such was the usual situation in cornutes, as seen
also in "C.' fellinensis (text-fig. 13), for example. In Nevadaecystis, which so far as can be seen had a rigid
floor, the strut was represented by a thickening of plate g which ran forward and presumably continued on to
plate a (though that plate is not visible on the only known specimen). There was no trace of such a single
thickening on plates g or a of P. menevensis. Ribs existed and ran more or less radially out from the growth
centres on the thinner parts of plates g, a, j, k, 1, x, and b (the laminar part of this plate). However, the ribs
were irregular in position and numerous, and none of them can certainly be identified with portions of the
strut of other cornutes. Perhaps they represented a primitive, unfixated condition, two apposed ribs of which
alone survived to give the strut of more crownward cornutes.

An alternative possible evolutionary beginning for the strut is represented in P. menevensis by an internal
process on plate g (cleft on the internal mould, text-fig. 23a; PI. 55, fig. 3; PI. 60, figs. 3, 9, 10) situated between
the posterior coelom and the right anterior coelom (see below, under ‘The chambers and soft anatomy of the
head’). If it grew forward during evolution, this process would produce a thickening in the same position as
the strut of N. americana. In broad terms, however, it is true to say that P. menevensis had no strut, or at
least, no strut that can certainly be homologized with the strut of cornutes more crownward than P. menevensis.
The absence of such a strut is probably primitive and likely to be homologous with its absence in Ceratocystis
perneri. Mechanically speaking, the strut of cornutes would have served mainly to prevent the head being
crushed under anteroposterior compression. This function would have been carried out in P. menevensis by
the triple arch of the dorsal surface (see below), in collaboration with the rigid floor of the head. This also is
probably primitive, since the head of P. menevensis to this extent retained the box-like construction of the
head of C. perneri.

448

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 15. Half-scale, outline drawing of Plate 54, fig. 1.
For single letters or numbers, see plate notation shown in
text-fig. 10. Other labels: 1/2, surface on plate 2 representing
articulating junction with plate series 1; cb, cerebral basin;
dbi, dorsal branchial integument; ga, gonopore-anus; li, left
anterior integument; pu, posterior U-plate; oc, oral cone;
stc, stylocone; ri, right integument; vo, ventral ossicle of
hind tail.

As to the stereom of the ventral shield, the central part, which was stiffened by ribs, consisted of extremely
thin, retiform stereom, except for the ribs themselves which were of labyrinthic stereom and were more convex
externally than internally. This type of structure, consisting of a web with struts, would be described by
architects as a space frame. It would have combined lightness with stiffness and suggests that P. menevensis
was adapted for low weight. More peripherally, the ventral shield tended to be thicker. Sometimes it was of
perforate stereom externally, but smooth and almost imperforate internally, as in the posterior right-hand
part of the floor (plate e and the adjacent parts of plates e, a, and g) where it underlay the cavity of the right
anterior coelom (see below, under ‘The chambers and soft anatomy of the head’), and also in plates g and j
beneath the posterior coelom. The peripheral parts of the ventral shield tended to be formed of labyrinthic
stereom, though the sharp edges were sometimes of fibrillar stereom. Thus the oral appendages of plates b
and c were covered on their convex ventral surfaces with a sheet of fibrillar stereom (PI. 55, fig. 1; PI. 59, figs.
3 and 4), the fibrils being more or less radial in arrangement. But the flat dorsal surfaces of these plates
(PI. 54, fig. 1; PI. 59, fig. 2) were formed of labyrinthic stereom, except peripherally; this labyrinthic stereom filled,
so to speak, the dorsal concavity of the ventral fibrillar layer. Again, in plate 1, the posterior parts of the
plate, forming part of the floor, were of space-frame type (PI. 55, fig. 1; PI. 57, fig. 1; PI. 58, fig. 3; PI. 60,

EXPLANATION OF PLATE 54

Figs. 1-4. Protocystites menevensis Hicks. 1-3, BM(NH) E62963 from loc. 2, Porth-y-rhaw (see text-fig. 7);
Hypagnostus parvifrons Zone, Middle Cambrian. This is the best specimen known. Plate e has been dislocated
and rotated approximately 90° from its original position. 1, general dorsal aspect of latex cast (see text-fig.
15 on facing page), x 6-9. 2, latex of posterior left part of head at an earlier stage of dissection than in fig.
1; note the series of plates 1, x 1 1-9. 3, plate 3 at a higher magnification than in fig. 1 and oriented as in the
reconstruction in text-fig. 10, x 12-8; is, imperforate stereom of plate e; rs, retiform stereom of plate e; pbbc,
callus marking posterior boundary of buccal cavity. 4, BM(NH) E62979, also from loc. 2, Porth-y-rhaw.
Latex in dorsal aspect; note posterior coelom (pc) clearly defined by an arcuate ridge on the dorsal surface
of plates g and j; note also the vertical wall (vw) which represented the posterior limit of the posterior
coelom and which formed the front wall of the cerebral basin, x 4-2.

When not otherwise explained, isolated letters and numbers are the notations of plates (see text-fig. 10); all
figs, whitened with ammonium chloride sublimate.

PLATE 54

JEFFERIES, LEWIS and DONOVAN, Pwtocystites

450

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 16. Half-scale, outline drawing of Plate 55, fig. 1. For
single letters or numbers, see plate notation shown in text-fig.
10. Other labels: esp, e-spike; fsp, f-spike; ft, fore tail; kj, keel
on plate j; If, fibrillar part of 1; Ir, retiform part of 1; ng, narrow
groove (ear) in plate j; oc, oral cone; oi, oral integument; stc,
stylocone; vo, ventral ossicle of hind tail.

fig. 8), but the vertical wall was of labyrinthic stereom and the l-appendage was fibrillar stereom, with the fibrillae
running parallel to the anterior edge of the appendage. The stereom of plate f was particularly complicated.
The spike and peripheral flange seem to have been built of labyrinthic stereom which was particularly dense
at the sharp edges. Most of the lateral wall of plate f, however, was of very thin retiform stereom crossed by
horizontal folds (pw in PI. 57, fig. 4; PI. 60, figs. 5-7). This thin part of the wall corresponds to the presence of
a pharynx inside the head, as argued below.

The dorsal surface of the head, as already mentioned, contained some regular, individualized plates separ-
ated by plated integuments. The first group of individualized plates to be described is the dorsal arch of the
tail insertion, consisting of plates h, y, and i (M|rd, Mp^, and M]ld of the old notation). Of these, plate y
(PI. 58, fig. 2) formed the keystone of the arch, being median and dorsal in position; its posterior, external
surface was notched from below by a groove (the dorsal groove) which in life probably carried a median eye
that arose from the dorsal surface of the brain. Plate i (PI. 54, fig. 1), forming the left part of the arch, ran
between plates y and j; at its ventral end it had a large, almost circular embayment which in life carried the
gonopore-anus. Plate h (PI. 54, fig. 4) ran between plates g and y and formed the right part of the dorsal arch.

A comparison of the plates and openings near the tail attachment with those of other cornutes is of interest.
Compared with C. perneri (text-fig. 1 1), in P. menevensis: 1, the gonopore-anus was left of the tail, not right
of it; 2, there was no plate o (Mpy of the old notation); 3, plate g formed part of the tail attachment, had a
much larger ventral posterior portion, and did not abut against the gonopore-anus; 4, plate j formed a larger
part of the tail attachment, had a larger vertical wall, included the narrow groove (ear) (which in C. perneri was
in i), and formed the lower boundary of the gonopore-anus; 5, plate i was much smaller, contained the

EXPLANATION OF PLATE 55

Figs. 1-5. Protocystites menevensis Hicks. All specimens from loc. 2, Porth-y-rhaw. 1, BM(NH) E62963, latex
of ventral aspect of best specimen (see text-fig. 16 on facing page), x 6-9. 2, detail of same to show plate j
with ear (ng, narrow groove), x 8-8. 3, BM(NH) E62958, natural moulds of internal surfaces of two
specimens of plate g, in ventral aspect, suggesting the soft parts (cf text-fig. 22a), x 8-4; br, brain; cl, cleft
between posterior coelom and right anterior coelom; gr, gonoduct-rectum; oes, oesophagus; pc, posterior
coelom; st, change in nature of stereom from almost imperforate at right posterior to thin and retiform at
left anterior (retiform stereom has a granular appearance). 4, 5, BM(NH) E63034, latexes of plate f in
ventral and dorsal aspect; oesf, oesophageal fold; fsp, f-spike; 4, x 10-3; 5, x 1 1-8.

When not otherwise explained, isolated letters and numbers are the notations of plates (see text-fig. 10); all
figs, whitened with ammonium chloride sublimate.

PLATE 55

JEFFERIES, LEWIS and DONOVAN, Protocystites

452

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 17. Half-scale, outline drawing of Plate 56, fig.
1 . For single letters or numbers, see plate notation shown
in text-fig. 10. Other labels; ai, almost imperforate ster-
eom in plate a; D, dorsal plate of fore tail; ar, reticulate
stereom in plate a; dp, dorsal plate of hind tail; ef, frame
part of plate e; ei, almost imperforate stereom in plate e;
fsp, f-spike; h esp, hollow in interior of plate e, cor-
responding to e-spike; li, left anterior integument; pbbc,
callus marking posterior boundary of buccal cavity; ri,
right integument; stc, stylocone; vo, ventral ossicle of
hind tail.

dorsal boundary of the gonopore-anus, and was dorsolateral in position with respect to the tail attachment,
instead of lateral— indeed, much of plate i as seen in C. penieri seems to be replaced by integument in P.
menevensis; 6, plate y was much smaller and bounded anteriorly by integument; 7, plate h was likewise much
smaller and bounded anteriorly by integument, was perhaps partly replaced by integument, and did not abut
against the gonopore-anus nor the hydropore (which in P. menevensis was absent, see below). These differences
are in some cases interrelated: thus, the fact that plate g in P. menevensis made no contact with the gonopore-
anus is due to the position of the latter left of the tail. With one exception, however, all these differences from
C. perneri are resemblances to "Cothwnocystis' fellinensis (text-fig. 13) and probably to N. americana also
(text-fig. 1 2), whose morphology is less well known. The exception relates to the ear, whose position penetrating
plate j (text-fig. 1 16, c; PI. 55, fig. 2) could be a specialization of P. menevensis, though the fact that it was not
conflated with the gonopore-anus was a resemblance to Ceratocystis perneri. These differences from C. perneri
are also resemblances to Cothiirnocystis elizae (text-fig. 14)— the best studied Coiliurnocystissxcept that
C. elizae lacked plate y, the median eye, and plate x.

A three-legged arch occupied the central part of the dorsal surface of the head. The largest plate in this
arch was dorsocentral 2 (PI. 54, figs. 1 and 4) which lay to the right of the mid-line of the tail and probably

EXPLANATION OE PLATE 56

Figs. 15. Protocystites menevensis Hicks. 1-3, SM A. 1021, lectotype (see also PI. 57, figs. 1 and 2), e.x Hicks’s
Collection; Menevian, Porth-y-rhaw, Dyfed. [The specimen was cleaned in thioglycollic acid which had the
unintended effect of removing all pyrite. The gonorectal groove, which was filled with pyrite and showed
clearly when the specimen was first seen, is no longer preserved in the cleaned specimen.] 1, latex of dorsal
surface of cleaned specimen (general view), showing that it is dislocated; in particular, plate e is pushed
against plate f, and plate f has been inverted so that it here shows the ventral aspect with its f-spike (see
text-fig. 17 on facing page), x 5-8. 2, portion of hind tail, at higher magnification, to show features of the
dorsal plates; the left dorsal plates have rotated outwards and show their internal surfaces, whereas the
right plates show their external surfaces (cf. text-fig. 25), x 18-9; ap, articular process; erf, crescentic facet;
fp, fibrillar part of dorsal plate; if, inter-plate facet; ms, median suture (i.e. edge of the dorsal plate where it
met the median suture); pol, posterior lobe of dorsal plate; tg, transverse groove. 3, plaster cast of a latex
made before the specimen was cleaned with thioglycollic acid and which represents the original shape of
the natural mould; ventral aspect of region of the posterior coelom (pc) and of the infilling of the gonorectal
groove (gr), i.e. mould of the dorsal surfaces of plates g and j; note that the infilling of the gonorectal groove
crosses the posterior coelom and ends at the gonopore-anus (ga) to the left of the tail, towards the bottom
right of the figure, x 7-5; g/j, suture between plates g and j. 4 and 5, BM(NH) E62920, latexes of left and
right aspects of a specimen consisting of ossicles of the mid and hind tail, still articulated with each other
but separated from all other plates, both x 7-0.

When not otherwise explained, isolated letters and numbers are the notations of plates (see text-fig. 10); all
figs, whitened with ammonium chloride sublimate.

PLATE 56

JEFFERIES, LEWIS and DONOVAN, Prolocystites

454

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 18. Half-scale, outline drawing of Plate 57, fig. 1.
For single letters or numbers, see plate notation shown in
text-fig. 10. Other labels: dp, dorsal plate of hind tail; If,
fibrillar part of 1; li, left anterior integument (ventral surface
exposed); Ir, retiform part of 1; oesf, oesophageal fold; stc,
stylocone; vo, ventral ossicle of hind tail; V, ventral plate of
fore tail.

contained the summit of the triple arch. A blunt curved ridge on plate 2 followed its left posterior edge. The
posterior articulation of plate 2 is uncertain; it may have made direct contact with the marginal frame near
the suture of plates f and g; on the other hand, it may have linked with dorsocentral plate 6, which itself made
contact with the frame at plate f. The uncertainty is caused by the fact that the plate identified as dorsocentral
6 may in fact have been only an integument plate thickened along its posterior margin (PI. 54, fig. 1). Plate 2
made contact on its left with a series of plates ( 1 1, I2, 1 3, etc.; PI. 54, figs. 1,2,4; PI. 56, fig. 1 ) perhaps variable
in number, which passed leftwards to meet the vertical keel on the dorsal surface of plate k. The plates of this
series were convex upwards and formed the crest of a gentle ridge which stretched between plates 2 and k.
The third leg of the triple arch was plate ii (PI. 54, fig. 1): this had a distinct keel, being inverted L-shaped in
section. It contacted plate b anteriorly and plate x ventrally, and it probably articulated directly with plate 2.
On the other hand, an additional plate (U2; ?ii in text-fig. 15) may have been interpolated between plates ii
and 2 in life (although, if so, it has been considerably displaced after death in the one specimen where it was
seen). The same triple arch can be recognized in Ceratocystis perneri, where it was made up of plates 1, 2, ii,
and 6, and formed most of the rather solid roof of the head. It also existed in N. americana, though plate ii
was divided there into at least three plates and the equivalents of plate 1, which may have existed in life, are
not preserved in the only known specimen.

Three other individualized plates existed on the dorsal surface; plates u, 3, and 5. Plate u (PI. 54, fig. 1) lay
anterior to plates 1 1, I2, and I3 and has only been seen on one specimen. It carried an elongated ridge, as also
did plate u of C. perneri, but the orientation and outline of the plate are uncertain. Plates 3 and 5, likewise
only known in the specimen shown in PI. 54, fig. 1 , have already been mentioned as making contact respectively
with plates c and e of the ventral shield.

Plated integument formed the rest of the dorsal surface. The largest such stretch of integument lay to the
right of the triple arch, ran forward to the mouth and over the front margin of the head into the oral
integument of the ventral surface, and was limited on the right and posteriorly by the plates of the ventral

EXPLANATION OF PLATE 57

Figs 14. Protocystites menevensis Hicks. 1 and 2, SM A. 1021, latex of the lectotype, ex Hicks’s Collection;
Menevian, Porth-y-rhaw, Dyfed (see also PI. 56, figs. 1-3). 1, general view of ventral aspect (see text-fig. 18
on facing page), x 5-8. 2, portion of hind tail, x 18-9; fp, fibrillar portion of dorsal plate; vp, thick ventral
portion of dorsal plate; if, inter-plate facet. 3, BM(NH) E62958, from loc. 2, Porth-y-rhaw (see text-fig. 7);
latex of plate e in ventral aspect with boss-like e-spike (same plate as in PI. 58, fig. 4), x 5 0. 4, BM(NH)
E62965, also from loc. 2, Porth-y-rhaw; natural mould of inside of plate f in right posterior aspect (cf. PI.
60, fig. 8), X 6-3; fsp, f-spike; int, intestine; pw, pharyngeal wall; pvl, pharyngo-visceral line; sto, stomach.
When not otherwise explained, isolated letters and numbers are the notations of plates (see text-fig. 10); all
figs, whitened with ammonium chloride sublimate.

PLATE 57

JEFFERIES, LEWIS and DONOVAN, Protocystites

456

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 19. Half-scale, outline drawing of Plate 58, fig. 1 . For
single letters, see plate notation in text-fig. 10. Other labels:
dp, dorsal plate of hind tail; fsp, f-spike; ft, fore tail; oi, oral
integument; stc, stylocone; vo, ventral ossicle of hind tail.

b

shield (PI. 54, fig. 1); we refer to it, including the oral integument behind the mouth on the ventral surface, as
the right integument. The second largest patch of integument lay to the left of, and anterior to, the triple arch
and surrounded, or almost surrounded, plate u (PI. 54, figs. 1 and 2; PI. 56, fig. 1); we call it the left anterior
integument. The third largest patch of integument lay behind the triple arch, between it and the posterior wall
of the ventral shield and dorsal arch of the tail attachment; we call it the posterior integument. To the left of
the tail, the posterior integument was presumably penetrated by branchial slits and consisted of rather thick
plates, some of which were half-moon shaped and probably homologous with posterior U-plates (pu in text-
fig. 15). Well differentiated anterior and posterior U-plates surrounding cfeae-type slits (‘cothurnopores’) did
not exist in P. menevensis, unlike N. americana.

As to their stereom, the plates of the right integument were very thin near the posterior part of the head
and made up of a single layer of retiform stereom (PI. 54, fig. 1). Other integument plates were thicker and
made up of at least two layers of retiform stereom. The stereom of the dorsocentral plates, and also plates u
and u, was thicker again and tended to be labyrinthic.

EXPLANATION OF PLATE 58

Figs. 1-4. Protocystites menevensis Hicks. 1, BM(NH) E432, ex Hicks’s Collection; Menevian, from near St.
David’s; latex of ventral surface (see text-fig. 19 on facing page), x 12-8. 2, BM(NH) E6301 1, from loc. 2,
Porth-y-rhaw (see text-fig. 7); latex of plate y in external aspect to show dorsal groove (dg) for median eye,
X 19-2. 3, BM(NH) E63008, from loc. 2, Porth-y-rhaw; latex of plate 1 in dorsal aspect (same plate as in PI.
60, fig. 7), X 4-6. 4, BM(NH) E62958, from loc. 2, Porth-y-rhaw; latex of plate e, dorsal surface (same plate
as in PI. 57, fig. 3), x 8-3; is, imperforate stereom; pbbc, posterior boundary of buccal cavity; rs, retiform
stereom.

When not otherwise explained, isolated letters and numbers are the notations of plates (see text-fig. 10); all
figs, whitened with ammonium chloride sublimate.

PLATE 58

JEFFERIES, LEWIS and DONOVAN, Pwtocystites

458

PALAEONTOLOGY, VOLUME 30

The openings of the head. All of these have already been mentioned in describing the external features of the
head skeleton. They are crucial in understanding the anatomy, but we shall identify them without giving
reasons— the arguments for these identifications are to be found in Jefferies (1986, Chs. 7 and 8).

The mouth was near the anterior end of the head, between the oral-appendage plates (b and c) and anterior
to the approximately semicircular oral integument of the ventral surface. It seems to have been guarded by an
oral cone (oc of text-fig. 15, cf PI. 54, fig. 1; text-fig. 16, cf PI. 55, fig. 1) of radially disposed elongated plates.
This mouth would therefore be like that of Cothurnocystis elizae in position and in having an oral cone. It
differed from the mouth of Ceratocystis perneri in having an oral cone and in having flexible integument
posterodorsal to it, rather than a rigid upper lip. The mouth of N. americana is unfortunately not known.

The gonopore-anus was to the left of the tail in P. menevensis, between plates i and j. This was very similar
to the situation in Cothurnocystis elizae and "C.' feUinensis and differed strikingly from that in Ceratocystis
perneri where the gonopore and anus were right of the tail and sometimes, though not always, slightly
separated from each other. As discussed below, C. perneri probably represented the primitive condition in
these respects. When the gonopore-anus was to the left of the tail it lay in the outwash from the branchial
slits, so that faeces and gametes would have been flushed away (Jefferies 1986, Ch. 7). This was presumably
advantageous. As in most cornutes, a gonorectal groove in P. menevensis ran inside the head (PI. 54, fig. 4;
PI. 55, fig. 3; PI. 60, figs. 3, 9, 10) across the floor of the posterior coelom from the gonopore-anus to right of the
tail, i.e. to the place where these openings were situated in C. perneri. In N. americana the gonopore-anus was
probably to the left of the tail, for the right part of a gonorectal groove can be seen in the only known
specimen and is disposed as in Cothurnocystis elizae and most other cornutes. However, the specimen is too
damaged to prove the existence of a gonopore-anus between plates i and j.

There was almost certainly no hydropore in P. menevensis. This opening existed in Ceratocystis perneri
where it was a slit crossing the suture of plates f and h. In P. menevensis there was no such suture, for plate f
did not contact plate h, but the suture of plates f and g had a corresponding position in the head and had no
such notch (PI. 57, fig. 4). It is theoretically possible that a hydropore existed and emerged in the dorsal
integument, but this is unlikely since some sort of special organization for it would, in that case, be expected.
C. perneri is the only cornute known to have had a hydropore, and by outgroup comparison with echinoderms,
this is a primitive condition. The absence of a hydropore is an advanced condition and a synapomorphy of P.
menevensis with all cornutes other than C. perneri, with mitrates, and with all other crown-group chordates
(including those still extant).

The branchial slits of P. menevensis were almost certainly situated just to the left of the tail, in the posterior
patch of plated integument, but were not well-defined skeletally. Indeed, the most obvious signs of them are
half-moon shaped plates, at least three in number, each of which probably represents the posterior U-plate of
a slit (pu in text-fig. 15, cf PI. 54, fig. 1). There is space, however, for several other such plates in the specimens.
Unlike N. americana, there is no definite sign in P. menevensis of U-plates framing the slits anteriorly and it
may be, as shown in the reconstruction (text-fig. 10), that each slit merely slightly notched the integument
anterodorsal to it (this integument is shown in PI. 54, fig. 1). Comparison with C. perneri suggests that the
integument anterodorsal to the slits was evolved by the breakup of plate 1, which formed the dorsal margin
of all seven slits in C. perneri. It is no surprise that the posterior U-plates of P. menevensis were differentiated
while anterior U-plates were not (or not clearly), since three of the seven gill slits of C. perneri had posterior
U-plates while none had anterior U-plates. The presumed position of the gill slits of P. menevensis is of
interest — they would all be posterior to the gentle ridge which ran from plate 2, along plates U, U, and U to
the keel on plate k. This was different from C. perneri where two of the slits were anterior to the ridge and
only five lay behind it.

EXPLANATION OF PLATE 59

Figs. U4. Protocystites menevensis Hicks. Latexes, all from loc. 2, Porth-y-rhaw (see text-fig. 7). 1, BM(NH)
E63022, stylocone and hind-tail ossicles in dorsal aspect; the ossicle marked ! is seen in anterior or posterior
aspect and indicates the original shape in transverse section, x 10-8; erf, crescentic facet; mg, median groove;
stc, stylocone; tg, transverse groove. 2 and 3, BM(NH) E62958, a single plate b in dorsal and ventral aspects,
X 5-2 and x 8-2 respectively. 4, BM(NH) E62958, a different plate b in ventral aspect, x 4-3.

When not otherwise explained, isolated letters and numbers are the notations of plates (see text-fig. 10); all
figs, whitened with ammonium chloride sublimate.

PLATE 59

JEFFERIES. LEWIS and DONOVAN, Protocvstites

460

PALAEONTOLOGY, VOLUME 30

The dorsal groove, which in life probably contained a median eye, was situated in plate y (dg in PI. 58,
fig. 2), at the summit of the dorsal arch of the tail insertion. It is identified as having contained a median eye by
comparison with C. perneri where it arose in the mid-line from the dorsal face of the brain (i.e. from a region
identified as optic by comparison with mitrates and through them with fish; see Jefferies 1986, Chs. 7 and 8)
and went up to the dorsal surface of the head where it could readily receive light (Jefferies 1969, p. 521). This
eye cannot be homologous with either of the median eyes of modern vertebrates (epiphysis, paraphysis). For
these occur only within the monophyletic group of the myopterygians (lampreys Tgnathostomes) whereas no
median eye exists in; 1, crownward cornutes, e.g. Corthurnocystis elizae, Scotiaecystis curvata Bather, 1913,
and the series of plesions leading crownward of these forms to the mitrates, i.e. Galliaecystis, Amygdalotlieca,
Reticulocarpos hanusi, and R. pissotensis Chauvel and Nion, 1977; 2, the mitrates which were primitive crown
chordates and therefore included the latest common ancestor of living chordates; and 3, the myxinoids, which
are the living sister group of the myopterygians. The situation of the dorsal groove of P. menevensis, notching
the ventral edge of plate y, is fundamentally like that of Ceratocystis perneri and even more similar, in the
small size of plate y, to ' Cothurnocystis' fellinensis and some other cornutes such as Phyllocystis blayaci Thoral,
1935, P. crassimarginata (Thoral, 1935), and Cbaiivelicystis iihaghsi (Chauvel, 1966) (see Ubaghs 1969). Among
known cornutes the presence of plate y with a dorsal groove is certainly a primitive feature.

The narrow groove (ear) of Protocystites menevensis is anomalous in some ways. In the one specimen where it
has been seen (text-fig. 16; PI. 55, figs. 1 and 2), it is grossly in the same position as that of Ceratocystis perneri,
penetrating the posterior wall of the head to the left of the tail. These two cornutes are the only species known to
possess such an opening. It seems highly probable that the openings are homologous in both species and, for
reasons explained by Jefferies ( 1 986, Ch. 7) with reference to C. perneri, that they represent the primitive begin-
nings of the acustico-lateralis system— or, more precisely, they contained an ear functioning as lateral line. The
exact position of the ear of P. menevensis is remarkable, however, for it penetrates plate j dorsal to the keel of
this plate (PI. 55, figs. 1 and 2), whereas in C. perneri it penetrates plate i ventral to the equivalent keel (which
is on plate i) and just dorsal to plate j (text-fig. 1 Ic; Jefferies 1969, pi. 95, fig. 1; pi. 96, fig. 10). Although C. perneri
and P. menevensis were the only known cornutes with a narrow groove, it is likely that all cornutes had an ear,
for there is evidence of a homologous ear in mitrates (Jefferies 1969, p. 521; 1986, Ch. 8). Probably the ear in most
cornutes had become undetectable in the fossils by being conflated with the gonopore-anus. If the gonopore-anus
in C. perneri is imagined to move from left of the tail to right of the tail, to take up a position between plates i and
j as in other cornutes, then the narrow groove would come to be situated in the ventral part of the gonopore-
anus. But the narrow groove in P. menevensis was separated from the gonopore-anus and is therefore different,
not only from that of C. perneri, but probably also from that of the ear in all other known cornutes. The fact that
the ear is not conflated with the gonopore-anus is therefore a primitive resemblance to C. perneri, but its exact
position is likely to be a specialization of P. menevensis alone.

Flexibility in the heads of Protocystites menevensis, Ceratocystis perneri, and Nevadaecystis americana. C.
perneri (text-fig. 1 1) shows some signs of flexibility between some of the plates of the head. These signs are of
two types: accessory gaps and transversely rounded plate contacts. The accessory gaps are small irregular

EXPLANATION OF PLATE 60

Figs. 1-10. Protocystites menevensis Hicks. Isolated plates, all from loc. 2, Porth-y-rhaw (see text-fig. 7). 1 and
2, BM(NH) E63013, latexes of plate k in posteroventral and posterodorsal aspect respectively, both
X 41; ksp, k-spike. 3 and 4, BM(NH) E62952, latexes of a plate g in dorsal and ventral aspects to show the
irregularly disposed ribs in the anterior part of the plate, x 3-9; cb, cerebral basin; oesf, oesophageal fold;
pc, posterior coelom; pr, process separating right anterior coelom from posterior coelom (equivalent to
cleft in fig. 9, etc.); pvl, pharyngo-visceral line. 5-7, BM(NH) E62958, all showing the same plate; 5 and 6,
latexes in ventral and dorsal aspect; 7, natural mould in ventral aspect (cf. PI. 57, fig. 4), all x 8-7; int,
intestine; sto, stomach; pw, pharyngeal wall; pvl, pharyngo-visceral line. 8, BM(NH) E62958, latex of plate
1 in ventral aspect (same plate as PI. 58, fig. 3), x4-6; Isp, 1-spike; rs, retiform stereom. 9 and 10, BM(NH)
E63034, a plate g as natural mould in ventral aspect and as corresponding latex in dorsal aspect, both
X 5-6; br, brain; cl, cleft between posterior coelom and right anterior coelom; gr, gonoduct-rectum; oes,
oesophagus; pc, posterior coelom; st, change in stereom from almost imperforate at right posterior to thin
and retiform at left anterior.

When not otherwise explained, isolated letters and numbers are the notations of plates (see text-fig. 10); all
figs, whitened with ammonium chloride sublimate.

PLATE 60

JEFFERIES, LEWIS and DONOVAN, Pwtocystites

462

PALAEONTOLOGY, VOLUME 30

voids situated along the dorsal margins of plates k and I (at their junctions with plates u and 1) and along the
ventral (or posterior) margins of the three posterior branchial U-plates, where these joined plates k, i, and y.
The accessory gaps were therefore situated at the edges of areas which in P. menevensis were flexible, i.e. the
left anterior integument and the integument posteroventral to the gill slits. This in turn suggests that the
accessory gaps in C. perneri represented the beginnings of flexibility in the head— meaning, in this species, the
ability slightly to raise and lower the roof.

Transversely rounded plate contacts in C. perneri occurred near the right side of the head, along the contacts
of plates e and f, plates 5 and e, and plates c and d (text-fig. II/?). The most striking was that between plate 5,
whose ventral edge was convexly cylindrical, and plate e, whose dorsal edge was concave to receive the opposed
cylindrical edge of plate 5. These cylindrically rounded plate contacts were almost continuous with each other
in C. perneri and would have formed a sort of hinge line near the right side of the head— a line along which
minimal rotation would have allowed the roof to be raised or depressed slightly. In P. menevensis the dorsal
border of plate e made contact partly with integument and partly with centrodorsal plate 5 (text-fig. 15;
PI. 54, figs. 1 and 3). The right integument of P. menevensis, being flexible, could presumably inflate and deflate;
in this motion, plate 5 would rotate about its contact with plate e, exactly as deduced for C. perneri but
probably to a greater extent. This situation in P. menevensis therefore tends to confirm that the supposed
right hinge line of C. perneri indeed acted as such. The rest of the right hinge line of C. perneri, however, does
not seem to have been specially flexible in P. menevensis, so far as can be deduced from the ill-preserved
fossils. Rather the line of flexibility between plates 5 and e would have passed forward between plates 3 and c
and rearward between the dorsal integument and the dorsal edge of plate f. To sum up, signs of slight flexibility
between the head plates of C. perneri largely correspond to signs of greater flexibility in P. menevensis, but
there are exceptions where the correspondence does not hold.

The dorsal surface of N. americana (text-fig. 12), though not completely known, seems to have been more
flexible than that of P. menevensis. Thus neither plate 3 nor plate 5 is definitely recognizable, presumably
being represented by plates of the dorsal integument of the right side of the head. Plate ii is represented by
three adjacent plates, not by one, or perhaps two, as in P. menevensis. On the other hand, plate 6 of
N. americana is larger and more recognizable than that of P. menevensis. Indeed, it is possible that the plate
labelled 2 in P. menevensis is in fact equivalent to 2-1-6 of C. perneri and N. americana. Concerning the dorsal
integument in the left part of the head in N. americana, there is little direct evidence. However, there is a
three-rayed keel on plate 2 and the left posterior ray of this keel points towards the left extremity of the head.
Perhaps this ray, as in P. menevensis, was continued leftward by a line of keeled plates disposed anterior to
the gill slits (though no such plates are preserved in the only known specimen). Unlike P. menevensis, such a
crest cannot have led to plate k, which is entirely posterior to the gill slits in N. americana and carries no crest.
To sum up, N. americana has the same three dorsal integuments as in P. menevensis (1, right; 2, posterior; 3,
left anterior), and these three integuments are larger relative to the individualized dorsal plates. It is possible
that the left anterior integument of N. americana, unlike P. menevensis, was not entirely separate from the
posterior integument.

The chambers and soft anatomy of the head. Considering cornutes as a whole, there is direct evidence for four
chambers in the head, and presumptive comparative evidence for a fifth (Jefferies 1986, Ch. 7). The chambers
for which there is direct evidence were the buccal cavity, the right anterior coelom, the posterior coelom, and
the pharynx (left or primary pharynx) (text-figs. 20-22). The chamber for which the evidence is purely
comparative is the left anterior coelom. The evidence for these chambers is different in different cornutes
but they probably all existed in all of them. A modification of the cornute arrangement occurred in mitrates,
but differed in several ways, chiefly in the existence of a right pharynx, with right gill slits, as well as a left
pharynx.

The evidence for these head chambers in cornutes can be exemplified from Cothurnocystis elizae (text-fig.
20). (For a complete discussion, see Jefferies 1986, Ch. 7.) In this species the buccal cavity filled the ‘ankle’
part (buccal lobe) of the boot-shaped head, immediately behind the mouth. It is defined posteriorly on the
right by a vertical ridge on the inside face of plate e and posteriorly on the left by the right-angle bend in plate
a which delimits the buccal lobe or, in one specimen, by a vertical ridge on the inside face of plate a just
anterior to that bend. In middle-sized specimens of C. elizae there is also a difference in the plates of the
dorsal integument as between the buccal cavity and regions posterior to it: over the buccal cavity, the plates
are polygonal without gaps between them, whereas behind the buccal cavity, over the pharynx, the plates are
circular and separated by gaps— perhaps suggesting, if there were muscles in the gaps, that the dorsal wall of
the pharynx was more muscular than that of the buccal cavity. The posterior coelom in C. elizae was a small,
approximately hemispherical chamber just anterior to the tail and convex anteriorly. It was defined

JEFFERIES, LEWIS AND DONOVAN: CAMBRIAN CORNUTE

463

buccal cavity
right ant. coelom
gonoduct-rectum
post, coelom
pharynx

TEXT-FIG. 20. Cothurnocystis elizae Bather; the chambers of the head (redrawn
after Jefferies 1968, text-fig. 4). I -5 are the high points of the pharyngo-visceral
line, a, left lateral; h, dorsal; c, right; and d, posterior aspects.

anteroventrally by an anteriorly convex ridge on the upper, interior surface of plates g and j and would have
made contact anterodorsally with the concave posteroventral surfaces of plates h and i. A gonorectal groove,
indicating the position of gonoduct and rectum, traversed the floor of the posterior coelom from right to left
and climbed vertically upwards near its end to finish at the gonopore-anus to the left of the tail. The posterior
coelom was homologous with the left epicardium of tunicates (Jefferies, in press, Ch. 7 and 8).

In the ‘foot’ part of the boot in C. elizae, there is direct evidence for two further chambers, one overlying
the other. The upper chamber is characterized in the larger specimens by a special ornament on the internal
surfaces of the marginal plates; a decussate ornament of strong horizontal lines crossed by weak vertical ones.
The branchial slits opened through the roof of this chamber, so it can be identified as the pharynx. The lower
chamber was mainly situated in the posterior right corner of the head and is characterized by the smooth
surface of the stereom where the chamber made contact with the inside faces of the marginal plates. A sharp
but undulating line (the pharyngo-visceral line) separated the smooth areas from the overlying ones of
decussate pharyngeal sculpture. The undulations of the pharyngo-visceral line are the same in different
specimens, and there are five high points where it made contact, or almost so, with the upper edge of the

464

PALAEONTOLOGY, VOLUME 30

buccal cavity
right ant. coelom
post, coelom

I 1 pharynx

^^§tail foramen, gaps in skeleton, integument
■Hgill slits, hydropore, gonopore, anus
|l I ! 1 1 1 nerves and brain
muscle

TEXT-FIG. 21. Ceratocystis perneri Jaekel; the chambers and other soft parts of the head
(redrawn after Jefferies 1969, text-fig. 10). a, ventral; b, dorsal; and c, posterior aspects.

marginal plates. These points are numbered 1 to 5 from left to right. The gonorectal groove, which passes
under the posterior coelom, emerges on the right from the lower chamber and approaches the posterior
coelom from anterior right. If the groove indeed carried the gonoduct and rectum, then the non-pharyngeal
gut (except the rectum) and the gonad were situated in this lower chamber, which therefore functioned as
coelom; we call it the right anterior coelom.

The fifth chamber in C. elizae, for which there is only comparative evidence, can be called the left anterior
coelom. It probably overlay the pharynx and buccal cavity and would probably have been virtual (i.e. with
no open cavity). It corresponds to the left somatocoel of echinoderms (which faces upwards in crinoids), the
left mandibular somite of vertebrates and acraniates, and the left metacoel of hemichordates. Once again, the
arguments for its existence can be found in Jefferies (1986, Chs. 7 and 8).

The same chambers probably also existed in Ceratocystis perneri (text-fig. 21). Here again, there is direct
evidence for buccal cavity, right anterior coelom, posterior coelom, and pharynx; and the left anterior coelom
is presumed to have existed. The main differences in the head chambers between C. perneri and Cotimrnocystis
elizae were; 1, in Ceratocystis perneri the right anterior coelom was extensively in contact with the right side
of the head, whereas in Cotimrnocystis elizae, so far as is determinable, it made contact only with the right
and posterior faces of the marginals; 2, in Ceratocystis perneri the hydropore, gonopore, and anus opened
direct from the cavity of the right anterior coelom, passing through the posterior wall of the head to the right
of the tail, whereas in Cotimrnocystis elizae there was no hydropore and the gonopore-anus was to the left of
the tail, connected with the right anterior coelom by gonoduct and rectum which traversed the floor of the
posterior coelom (as indicated by the gonorectal groove).

In P. menevensis the same chambers probably existed as in other cornutes. The buccal cavity presumably
occupied the ‘ankle’ part of the boot-shaped head, i.e. the buccal lobe. Its probable right posterior margin
was defined by a triangular wedge of callus on the internal face of plate e (pbbc in text-fig. 1 5, cf. PI. 54, figs. 1
and 3; text-fig. 17 and PI. 54, fig. 1; PI. 56, fig. 1; PI. 58, fig. 4), this wedge being comparable in position to the
ridge on the internal face of plate e in C. elizae and many other cornutes. The posterior coelom was well
defined ventrally by a crescentic area on the floor of the head just to the left of the tail. This crescent was
concave upwards (pc in PI. 54, fig. 4; PI. 55, fig. 3; PI. 56, fig. 3; PI. 57, fig. 4; PI. 58, fig. 4) and floored by

JEFFERIES, LEWIS AND DONOVAN: CAMBRIAN CORNUTE

465

buccal cavity

TEXT-FIG. 22. The soft parts of the head in: n, b, a mitrate (Mitrocystites)', and c, d, a
cornute (Cothurnocystis) (redrawn after Jefferies 1981a, fig. 5). a, dorsal aspect; b, section
through b-b of a; c, section through c-c of d\ d, dorsal aspect. The mitrate condition
can be derived from the cornute condition mainly by origination of a right pharynx,
which would have appeared at the point x of c.

466

PALAEONTOLOGY, VOLUME 30

a

cleft between post, coelom
and right ant. coelom

b

gonoducl-reclum oesophagus

pharyngo-visceral line

stomach

TEXT-FIG. 23. Protocystites menevensis Hicks; reconstructed internal natural mould (i.e. soft parts), a, ventral;
b, posterior; and c, right lateral aspects. The reconstruction stops arbitrarily at the anterior margins of plates

a, e, and x.

almost imperforate stereom (text-fig. 23n; PI. 55, fig. 3), by contrast with the floor anterior to it, which is made
of thin, retiform stereom stiffened with ribs. A distinct gonorectal groove crosses the floor of the posterior
coelom transversely and climbs upwards to the left of the tail to end at the gonopore-anus (PI. 56, fig. 3).
Thus the gonoduct and rectum would have crossed the floor of the posterior coelom, as in C. elizae and all
known cornutes except Ceratocystis perneri where, as already stated, the gonopore-anus was to the right of
the tail. A high wall rising from the floor of the head in plates g and j formed the posterior limit of the
posterior coelom (vw in PI. 54, fig. 4) and separated that chamber from the deduced position of the brain. The
latter was situated in a distinct cerebral basin (PI. 54, figs. 1 and 4; PI. 60, fig. 3) excavated in plates g and j
and faced with almost imperforate stereom. In having such a basin, with a high wall in front of it, P. menevensis
once again differed from C. perneri and resembled more crownward cornutes such as Cothurnocystis elizae.
Anterodorsally, the limit of the posterior coelom presumably coincided with the posteroventral surfaces of
plates h, y, and i, but these surfaces have not been seen in the available specimens.

Regarding the chambers in the remainder of the head— behind the buccal cavity but in front of the posterior
coelom— P. menevensis, like other cornutes, probably had a pharynx overlying a right anterior coelom, with
the latter lying mainly in the posterior right portion of the head. As to the evidence, in the posterior wall and
right wall of the head, on the inside faces of plates g and f, a clear pharyngo-visceral line is visible (text-fig.
23c; pvl in PI. 57, fig. 4; PI. 60, figs. 3, 6, 7). This line separates a smooth, almost imperforate, surface below,
with a thick wall of stereom, from retiform stereom above, forming a very thin wall. Indeed, in the anterior
part of plate f the stereom above the line is not merely retiform but thrown into horizontal folds, like those in
the pharyngeal ornament of C. elizae. Moreover, the wall of plate f is so thin above the line in this region that
the folds are visible from outside, and thus the pharyngo-visceral line can be detected externally (text-fig. lOr/;
pw in PI. 60, fig. 5). Also, the undulations of the line in the right part of the head are the same as in C. elizae
with high points 3, 4, and 5 clearly recognizable. It is reasonable to suppose that, as with C. elizae, the
marginal plates were in contact with the pharynx above the pharyngo-visceral line (or rather with virtual left
anterior coelom embracing the pharynx) and with right anterior coelom beneath it. Moreover, the left
boundary of the right anterior coelom, or at least the cavity of that coelom, is indicated in P. menevensis by a
change in the stereom in the floor of the head. Thus, in the right posterior portion of the head, in plates f and
e (PI. 54, fig. 1), in the posterior right parts of plates g (PI. 55, fig. 3; PI. 60, fig. 9) and a (PI. 56, fig. 1, for the
internal surface of the plate; PI. 55, fig. 1 for the external surface), and in the right posterior part of plate e.

JEFFERIES, LEWIS AND DONOVAN: CAMBRIAN CORNUTE

467

t 1 buccal cavity

i\\\\\S right ant. coelom
post, coelom
I [pharynx

TEXT-FIG. 24. Protocystites menevensis Hicks; the chambers and other soft parts of the head in

dorsal aspect.

the upper surface of the floor of the head is constructed of almost imperforate stereom, continuous with that
below the pharyngo-visceral line in the right and right posterior wall of the head. The left anterior limit of
such stereom is a line running forwards and rightwards from near the right side of the tail insertion (text-fig.
23fl); and to the anterior left of this line, the floor stereom is of different nature being thin and retiform, except
where it is thickened by ribs (space-frame construction). The region of almost imperforate stereom, whether
in the floor or in the posterior faces of the posterior and right walls below the pharyngo-visceral line, was
probably in contact with the right anterior coelom and indicates the extent of that chamber. The rest of the
floor of the head, left of the area of almost imperforate stereom, was presumably overlain by pharynx, though
perhaps there was, beneath the pharynx but above the stereom, a purely virtual extension of right or left
anterior coeloms in this region. The pharynx would also have made contact with the posterior and right walls
of the head dorsal to the pharyngo-visceral line (though again it would probably have been clothed with
virtual left anterior coelom). As in other cornutes, the branchial slits, just to the left of the tail insertion,
would penetrate the roof of the pharynx.

The pharyngo-visceral line cannot be traced to the left of the tail insertion. This is because no known
specimens show the internal face of the left part of plate j.

A process on the internal surface of plate g, and a corresponding cleft in the natural mould (text-fig. 23a;
PI. 55, fig. 3; PI. 60, figs. 3, 9, 10), is situated between the deduced positions of the posterior coelom and the
right anterior coelom. It varies in shape, but tends to be anteroposteriorly elongate. Probably this process
was intercameral and represented the stereomic infilling of a gap between the limiting epithelia of the posterior
and right anterior coeloms. As already mentioned, this process is situated in a position corresponding to the
posterior end of the strut in cornutes more crownward than P. menevensis. It may represent the evolutionary
beginnings of the strut.

The left anterior coelom probably covered the whole of the dorsal surface of the pharynx in P. menevensis
but had no cavity, being solely virtual. As with other cornutes, and also with mitrates, the arguments for its
existence are purely comparative.

As to the contents of the right anterior coelom, the most direct evidence comes from Cemtocystis penieri,
where the hydropore, gonopore, and anus opened directly out of it. In that species, therefore, the right anterior
coelom is likely to have contained the non-pharyngeal gut, the gonad, and whatever organs were associated
with the hydropore. Among the latter, the heart and pericardium were probably included for the pulsatile
madreporic vesicle (pericardium) and the contained head process of the axial organ (heart) are situated near
the hydropore in living echinoderms. Important evidence as to the contents of the right anterior coelom comes

468

PALAEONTOLOGY, VOLUME 30

also from mitrates such as Placocystites forhesicmus (Jefferies and Lewis, 1978) (Jefferies 1986, Ch. 8). In
mitrates, however, the position of the right anterior coelom, hanging from the ceiling of the head near the
mid-line, is governed by the right pharynx which in cornutes did not exist. Clear signs indicate that the
oesophagus of mitrates opened from the right posterior coelom into the pharynx to the right of the mid-line,
near the posterior right corner of the cavity of the right anterior coelom (as shown in text-fig. 22a). From
here, the non-pharyngeal gut ran forward to an unknown extent, along the right margin of the cavity of the
right anterior coelom, turned through an angle, and ran rearwards along the left margin of this cavity to join
the rectum, which debouched into the left atrium. The gonad was probably situated in this loop of the gut,
partly because the gonoduct seems to arise from this position in the mitrate Mitrocystella, and partly because
the gonad is situated in the loop of the gut in enterogonous tunicates. The contents of the right anterior
coelom of mitrates is further discussed by Jefferies (1986, Ch. 8).

The cornute condition of the right anterior coelom can partly be deduced by mentally subtracting the right
pharynx from the mitrate condition, reversing what happened in evolutionary history. This would allow the
cavity and contents of the right anterior coelom to fall to the floor of the head and to occupy the posterior
right portion of the head, as seems to have been true in C. perneri, Cothunwcyslis elizae, Protocystites
menevensis, and other cornutes. In that case, the non-pharyngeal gut would follow the periphery of the cavity
of the right anterior coelom (as suggested in text-fig. 22d). Part of the intestine would run leftwards and
rearwards along the left margin of the cavity to join the rectum in the posterior coelom. (There is direct
evidence in C. elizae that the intestine approached the rectum by running leftwards and posteriorly, but in
P. menevensis the internal moulds indicate that it probably joined the rectum by running almost vertically
downwards.) And the oesophagus of cornutes would run rightwards, from a position to the right of the tail
insertion, just in contact with the right posterior wall of the head.

Some direct indication of the non-pharyngeal gut inside the right anterior coelom of P. menevensis is
probably given by the internal sculpture of plates g and f. As already mentioned, a distinct fold in plate g,
concave upwards and convex downwards (PI. 55, fig. 3; PI. 60, figs. 3, 9, 10) and corresponding to a horizontal
hemicylinder on the surface of the natural mould, runs rightwards into plate f (PI. 55, fig. 4) and coincides
with the position of the oesophagus as reconstructed above for cornutes in general; we assume that the fold
carried the oesophagus and therefore call it the oesophageal fold. This fold is ventral to high point 4 of the
pharyngo-visceral line but makes up only a small portion of the internal surface beneath that high point.
Some other organ, as discussed later, presumably filled the area between the oesophageal fold and the
pharyngo-visceral line. More anteriorly, in the right wall of plate f (PI. 57, fig. 4; PI. 60, figs. 6 and 7), the
pharyngo-visceral line ascends towards high point 5. On the internal natural moulds, however, the surface
beneath this line is complex in this region. Ventrally the natural mould formed a horizontal hemicylinder
whose upper edge was distinct from, and much more ventral than, the pharyngo-visceral line and whose
radius was about the same as that suggested for the oesophagus by the oesophageal fold. This hemicylinder
corresponds to a groove on the inner face of plate f and presumably it too carried a portion of the gut. More
posteriorly, just anterior to the right posterior angle of the head, the hemicylinder expands into something
more inflated. This inflation may represent the stomach (PI. 57, fig. 4; PI. 60, fig. 7), in which case the gut
anterior to it would be part of the intestine.

The organ beneath high point 4 but dorsal to the presumed oesophagus may have been the pericardium
and heart. This conclusion, which is highly tentative, is based on two opposite approaches. First, the position
roughly coincides with that of the hydropore in Ceratocystis pernerp and the heart and pericardium of
echinoderms (head process of axial gland and madreporic vesicle) are located just internal to the hydropore.
Secondly, the heart of tunicates is located on the right side of the gut, as may have been true of mitrates also.
If so, then mental subtraction of the right pharynx to yield the cornute condition would bring the heart against
the right posterior wall of the head, next to the oesophagus, to occupy the observed space above the oeso-
phageal fold but beneath high point 4. We believe that the grounds for placing the heart, gonad, and non-
pharyngeal gut somewhere inside the right anterior coelom are reasonably strong. The reasons for putting
them in particular grooves and depressions in the inner face of the skeleton are necessarily weaker.

Not much can be said about the brain and cranial nerves of P. menevensis. The brain, as already mentioned,
would have occupied a well-defined cerebral basin excavated in plates g and j (PI. 54, fig. 1; PI. 55, fig. 3;
PI. 60, figs. 9 and 10) and faced with almost imperforate stereom. This basin formed part of the posterior surface
of the wall that stood behind the posterior coelom. The brain would have extended upwards into the median
eye (occupying the dorsal groove) in the posterior surface of plate y (PI. 58, fig. 2). The left and right pyriform
bodies (trigeminal ganglia), which were present in all other cornutes just anterior to the brain, could not be
located in P. menevensis. If they existed, they either made no contact with the skeleton or else the specimens
are too crumpled to show them.

JEFFERIES, LEWIS AND DONOVAN: CAMBRIAN CORNUTE

469

The tail. The tail of P. menevensis, like that of all other cornutes, is divided into fore, mid, and hind regions.
In text-fig. 10 it is shown as small and slender relative to the head, when compared with Cothurnocystis elizae
or Ceratocystis perneri, for instance. This may be correct but, because of tectonic distortion, it is impossible
to be sure. The abrupt end of the tail is usual, or perhaps universal, in cornutes and mitrates, and in general
probably means that more distal segments were regularly lost by autotomy, perhaps several times in the life
of the animal. In P. menevensis, however, none of the observed tails is well enough preserved to show how it
terminated. The reconstruction of the fore tail was particularly difficult, for its dorsal surface is not shown
adequately by any specimen.

The skeleton of the hind tail consisted of a median series of ventral ossicles, approximately hemicylindrical
in shape, and dorsal plates arranged in right and left series (PI. 56, figs. 1, 2, 4, 5; PI. 57, figs. 1 and 2; PI. 59,
fig. 1). So far as can be judged, the plates were not opposite each other on the right and left. Neither were
they uniformly related to the ventral ossicles beneath them: the number of plates in the left and the right
series, as indicated in particular by the repetition of segmental structures on the dorsal surface of the ossicles,
is greater than the number of ossicles, i.e. there is somewhat more than one plate per ossicle on each side
proximally, and about two plates per ossicle on each side distally.

The ossicles in life were probably about as wide as deep, as suggested by a specimen whose longitudinal
axis lies perpendicular to the bedding plane (PI. 59, fig. 1). The dorsal, interior surface of the series of ossicles
(PI. 56, fig. 2; PI. 59, fig. 1) bore a deep median groove which was rounded V-shaped in transverse section.
The walls which formed the right and left sides of the groove were intermittently notched by semicircular
transverse grooves, but the notches, like the plates, were not opposite each other, nor uniformly related to the
ossicles that bore them. Lateral to each notch was a depression in the dorsal surface of each ossicle, each
depression being separated from those more proximal and more distal by a transverse ridge. (Each notch is
located near the anterior end of the corresponding depression.) The dorsal surface of the ossicles varied in
width, being narrowest at the ridges and widest at the depressions. At the lateral end of each ridge there was a
flat crescentic facet, concave outwards, which received the articular process of the corresponding dorsal plate.

The dorsal plates, as seen in the specimens, are usually spread out on the bedding plane on either side of
the ossicles, being slightly concave on the original internal surface and slightly convex on the original external
surface (PI. 56, fig. 2; PI. 57, fig. 2). In our view, the outward spread of the plates is preservational and caused
by the explosive release of the gases of putrefaction from the soft tissues of the hind tail after death; and the
slightness of curvature of the plates probably results from post-mortem compaction perpendicular to the
bedding plane. (Ubaghs, who believes that the plates of cornutes could open outwards in life (1981), would
probably dispute both these views.) The plates in text-fig. 25 are reconstructed with greater curvature than
observed, on the assumption that the median groove of the ossicles contained the notochord and this, being
anti-compressional, would lie in the mechanical neutral axis of the hind tail, and therefore near the centre of
the hind tail in transverse section. The deduced curvature of the plates is similar to that seen in C. perneri, N.
americana, or Cothurnocystis elizae, for example.

Each dorsal plate consisted of a thick-shelled ventral portion and a thin-shelled dorsal portion. The thin-
shelled portions presumably met those of the opposite side of the tail at a median dorsal suture, and the edge
of such a plate where it would have met the suture is visible in Plate 56, fig. 2 (ms). However, such a median
dorsal contact between the dorsal hind-tail plates of the left and right sides has never been seen in the fossils.
As already mentioned, it is unlikely that a plate of the left side was ever exactly opposite one of the right, and
vice versa. The thick-shelled part of the plate was formed of labyrinthic stereom externally, and of smooth,
almost imperforate, stereom internally. The thin-shelled part of the plate consisted of fibrillar stereom, with
the fibres directed approximately parallel to the posterior and anterior edges. The thick-shelled part had a
convex posterior outline which formed a ventral lobe, and this lobe overlapped the posterior part of the next
plate behind. The external surface of the anterior part of the thick-shelled portion of the plate bore a smooth
facet (inter-plate facet) which would have slid against the inside face of the next plate in front. The thin-
shelled part of each plate probably slightly overlapped the corresponding part of the next plate behind. The
anteroventral angle of each plate was produced medianward into an articular process which made contact in
life with the crescentic facet on the ventral ossicle.

Three segmentation series, therefore, seem to have existed in the hind tail: 1, that of the ossicles mid-
ventrally; 2, the left series of dorsal half-segments; and 3, the right series of dorsal half-segments. The skeleton
of each half-segment contained; 1, a plate with its thin- and thick-shelled portions, ventral lobe, articular
process, and inter-plate facet; 2, a depression on the dorsal surface of the ossicles, between two successive
transverse ridges; and 3, a transverse groove (notch) leading into that depression from the median groove.
Each half-segment would have had corresponding soft parts whose nature will be discussed later. Eeft and
right half-segments were about equal in number, but not opposite each other and had no one-to-one

470

PALAEONTOLOGY, VOLUME 30

with attached dorsal plates; h, posterodorsal aspect of several ossicles and plates. The dorsal nerve cord is
shown as overlying the notochord by analogy with the mitrates.

relationship with the ventral ossicles. One specimen suggests that, when the dorsal plates and corresponding
soft parts had been lost, the column of ventral ossicles would curve downwards under its own elasticity
so as to be concave ventrally (PI. 56, figs. 4 and 5). This specimen suggests that such would be the relaxed
condition of the hind tail during life.

The fore tail would have had a large lumen. The skeleton of the ventral surface of the fore tail can be
restored with some confidence (cf. text-fig. 18 and PI. 57, fig. 1). There were about three ventral plates on the
left and right, and these overlapped each other in the ventral mid-line and alternated. A similar alternation
occurred in Ceratocystis perneri (text-fig. 1 \b) where, however, there were many more half-segments on each
side (about seventeen instead of about three). The dorsal surface of the fore tail is much harder to reconstruct.
In text-fig. 10 the left and right dorsal series of plates, three plates in each series, are shown and the plates of
left and right sides are shown as alternating. All this may have been true, partly by analogy with C. perneri,
but the relevant specimens are smashed to pieces in this region. One specimen (the lectotype, text-fig. 17 and
PI. 56, fig. 1), however, shows a single reasonably complete dorsal plate, so the reconstruction is plausible. It
is clear that, as in other cornutes, the skeleton of the fore tail was looser than that of the mid and hind tails
and would have allowed much movement. The mid-ventral alternation of left and right plates suggests that
most of the movement was flexion from side to side.

The skeleton of the mid tail consisted of a massive stylocone ventrally (text-fig. 15 and PI. 56, fig. 1; text-fig.
17 and PI. 57, fig. 1; PI. 58, fig. 1; PI. 59, fig. 1) and presumably there were left and right series of plates
dorsally. The stylocone was an approximately conical ossicle with a deep anterior excavation. It was probably
serially homologous to several ventral hind-tail ossicles. The dorsal surface of the stylocone has a sculpture
much like that of the hind-tail ossicles and there are indications of about three half-segments on each side.

As to soft parts, the lumen of the fore tail was probably largely filled with muscle which, to judge by the

JEFFERIES, LEWIS AND DONOVAN: CAMBRIAN CORNUTE

471

mid-ventral alternation of left and right plates, probably served mainly to swing the tail from side to side.
Concentration of muscle in the proximal part of an appendage is efficient because, for a given angle of bending,
the muscles require a smaller percentage contraction than if attached more distally, and muscles work most
efficiently at small percentage contractions (Gray 1957). To prevent telescoping, some anti-compressional
structure in the soft parts of the fore tail would have been required, which suggests that a notochord was
present. This notochord would have continued rearwards from the fore tail into the mid and hind tail where
it would have occupied the median groove. In the hind tail, however, the notochord would have had no anti-
compressional function; rather it would have served, like the string in a bead necklace, to keep the ossicles in
alignment. By analogy with mitrates, where there is direct evidence in the fossils, and with all living chordates,
the notochord was presumably overlain by a dorsal nerve cord. The transverse grooves (notches) probably
represent places where nerves went off from the dorsal nerve cord, and perhaps blood vessels from a longitudi-
nal vessel in the notochord, to more lateral parts of the tail. The imbrication of successive dorsal plates in the
hind tail indicates that the tail could have curled upwards. Such curling implies the presence of muscles in the
lumen of the hind tail which, on contraction, would have shortened the dorsal side of the tail and so caused
flexion. These muscles would have required an antagonist, and this was probably supplied by the elasticity in
the ligaments of the column of ventral ossicles which, as already noted, seems to have flexed downwards when
relaxed. The presence of serial depressions in the dorsal surface of the ossicles, separated by transverse ridges
and supplied by whatever filled the transverse grooves, suggest that these muscles were divided into blocks
(somites), each muscle block occupying a depression.

SYSTEMATIC POSITION

Some methodological remarks

Text-fig. 26 is a cladogram of the dexiothetes which lays particular stress on the less crownward
cornutes near P. menevensis. We regard a cladogram as a phylogeny conventionalized by placing
all the known forms under consideration at the ends of branches. The terminal branches of a

•2

C

Q>

O

v>

TEXT-FIG. 26. A cladogram of the Dexiothetica to show the position of Protocystites menevensis Hicks within
the chordate stem group and the origin of some important evolutionary novelties. For fuller explanation see

text.

472

PALAEONTOLOGY, VOLUME 30

cladogram are therefore in the first place conventional. They can sometimes be shown to have
existed, when they did exist, by two types of argument— morphological and stratigraphical. The
morphological argument requires that some feature has arisen within the terminal branch, as an
autapomorphy of the single species or the species group at the end of the branch. The stratigraphical
argument requires, in the present case, that the animal or animals at the end of the branch be
contemporaneous with, or preferably later than, their more crownward neighbour. To show that a
terminal branch did not exist is difficult or impossible. In the present instance this means that, even
if the morphological and stratigraphical arguments for the existence of a terminal segment failed, it
would be impossible to prove that any known form lay exactly in the chordate stem lineage, in the
sequence of even-numbered segments 6-24. Text-fig. 26 is not comprehensive crownward of
N. americana (segments 12-24) nor anti-crownward of C. penieri (segment 6), in the sense that
known fossils other than those named belong, or can be suspected of belonging, in those regions.
Moreover, the parts of the diagram crownward of Cothurnocystis elizae are not discussed in this
paper, since they are treated by Jefferies (1986, Chs. 7 and 9).

Particular difficulties arise in establishing the primitiveness of some features of Ceratocystis
pemeri—lhe least crownward cornute known. When a feature is found in C. penieri and also in
some related forms, then there is no dilemma. For example, C. penieri had a hydropore, uniquely
among cornutes; but this feature also occurs in crown echinoderms and in hemichordates (left
mesocoel pore), which indicates that it existed in segments 1, 2, 4, 6, and 7 of text-fig. 26 and
disappeared within segment 8. Again, C. penieri had a rigid floor to the head, and this also existed
in P. menevensis and N. americana., this indicates that a rigid floor existed throughout segments
7-11, in the crownward part of 6, and in the anti-crownward part of 12, and that it disappeared in
segment 12. Features unique to C. penieri among known forms, however, present difficulties; for
prima facie, they could either have arisen in segment 7, as autapomorphies of C. penieri, or they
could be primitive features of cornutes present throughout segment 7, in the crownward part of
segment 6, and in the anti-crownward part of segment 8.

The stratigraphical criterion of primitiveness is of no help in resolving such dilemmas, for the
Cambrian record of cornutes it too incomplete, on several grounds. First, only four cornutes have
been described from the Cambrian, i.e. C. penieri, P. menevensis, and the ‘stylophoran’ of
Sprinkle ( 1 976, pi. 1 , fig. 1 ), all of which are approximately contemporaneous and Middle Cambrian
in age, and N. americana which is Upper Cambrian in age. Secondly, these four forms are consider-
ably different from each other, while N. americana from the Upper Cambrian is in most ways
intermediate between P. menevensis and Sprinkle’s ‘stylophoran’ (which broadly belongs to Cothur-
nocystis) from the Middle Cambrian. Thirdly, new forms are being discovered in the Cambrian
(four new species of Cambrian cornute came to our attention in the years 1981 to 1985, one of which
is P. menevensis). In deciding what features are primitive among Cambrian cornutes, therefore,
stratigraphy is useless. (We do not deny that stratigraphy can indicate primitiveness in other groups
of fossils (Fortey and Jefferies 1982).)

Functional analysis can sometimes suggest which of two alternative feature states is the more
primitive. Thus, in having its gonopore-anus to the right of the tail, Ceratocystis penieri was less
efficiently laid out than other cornutes, whose gonopore-anus was to the left of the tail and therefore
in the branchial outwash, so that faeces and gametes could be flushed away. This suggests: that
having the gonopore-anus to the right of the tail was more primitive than the alternative; that this
primitive state existed in the crownward part of segment 6 of text-fig. 26, the anti-crownward part
of segment 8, and throughout segment 7; and that the change to the location left of the tail happened
in segment 8. Such arguments are risky, in as much as functional interpretations are always
uncertain.

Anti-crownward extrapolation of an evolutionary trend can also be invoked to indicate that a
unique feature of C. penieri was primitive. The anti-crownward sequence Cothurnocystis elizae,
^C.' fellinensis, N. americana, P. menevensis, Ceratocystis penieri is one of decreasing relationship to
the chordate crown group on several grounds. It is also one of decreasing flexibility of the dorsal
surface. This suggests that the rigid, or almost rigid, roof to the head seen in C. penieri, and only in

JEFFERIES, LEWIS AND DONOVAN: CAMBRIAN CORNUTE

473

it among known cornutes, is the evolutionary starting point from which the flexible roof, seen in
increasing degree crownward among other cornutes, developed— essentially in segments 8, 10, 12,
and 14 of the chordate stem lineage. However, this argument lacks logical rigour, to the extent that
evolution may reverse in direction. It can partly, but not completely, be subsumed as a series of
more rigorous sub-arguments in which C. perneri is not unique among known forms. For example,
C. perneri and its crownward neighbour P. menevensis are the only known cornutes showing
individualized plates 3 and 5 in the roof of the head. If C. perneri, because of its hydropore, is the
least crownward cornute known, then P. menevensis is its crownward neighbour because it retains
plates 3 and 5 (among other arguments) and all other cornutes lack these plates as distinguishable
elements, presumably having lost them. But sub-arguments of this sort can show only that a
condition seen in P. menevensis is primitive with respect to all cornutes except C. perneri. They
cannot show that a condition known uniquely in C. perneri is primitive. Anti-crownward extrapola-
tion of evolutionary trends goes further than these sub-arguments, but is less logically rigorous. It
can be expressed, in this instance, as a working rule: when P. menevensis shows some feature state
intermediate between that of C. perneri on the one hand, and N. americana on the other, then the
feature state in C. perneri is likely to represent the primitive condition from which the others were
derived.

The primitiveness, or otherwise, of a feature known uniquely in C. perneri therefore remains
undecided if its primitiveness for cornutes cannot be established in one of the following ways: 1, by
outgroup comparison with echinoderms (such a comparison, if successful, would imply its existence
in the crownward part of segment 1, and at least throughout segments 2, 4, 6, and 7 of text-fig. 26);
2, by its presence, perhaps in less marked form, in the least crownward of other cornutes; 3, by
functional arguments. Such unique features of C. perneri could be primitive for cornutes (present
at the junction of segments 6, 7, and 8) or could be autapomorphies of C. perneri evolved within
segment 7. Only the recognition of stem chordates less crownward than C. perneri, or of nodal-
group dexiothetes, will favour one or other of these alternatives. To assume that all features of C.
perneri, even those known uniquely in it, were primitive for cornutes and existed in the chordate
stem lineage, is not legitimate and leads into an intellectual trap. For it is like assuming, in view of
Ornithorhynchus, that all other mammals evolved from a toothless ancestor, with a duck-like beak;
or, in view of modern amphioxus, that the latest common ancestor of vertebrates and acraniates
was brainless, which cannot be true (Jefferies 1973).

The above methodological discussion assumes the relationships which will be discussed in the
rest of this section. This is legitimate, since the assignment of different cornutes to their places in
the cornute cladogram does not start from a blank (see Jefferies 1979; in press, Ch. 9). Our task is
to fit P. menevensis into a phylogenetic framework which is already partly known.

The cladogram of the dexiothetes and the position of Protocystites menevensis within the chordate
stem group

The cladogram shown in text-fig. 26 is, in our view, the most parsimonious and probable arrange-
ment of the Dexiothetica, so far as the facts at present available indicate. The evolutionary novelties
assignable to the various segments are as follows.

Segment 1. In this segment, which was the dexiothete stem lineage, a form like the pterobranch
Cephalodiscus lay down on its originally right side and lost the openings and tentacles of the right
side and probably also the pterobranch stalk. This process of lying on the right side, with all its
consequences, can be called dexiothetism. Henceforth, in the dexiothete stem lineage, the primitive
and hemichordate right became ventral in chordate terms and hemichordate left became dorsal.
Also a calcite skeleton of stereom mesh was acquired (Jefferies 1986, Chs. 2 and 7). No known
fossil forms have yet been assigned to the dexiothete stem group through which segment 1 would
have passed.

Segment 2. This is the least crownward part of the echinoderm stem lineage. The evolutionary
novelties acquired in the echinoderm stem lineage as a whole (segments 2 and 4) can now be

474

PALAEONTOLOGY, VOLUME 30

discussed much more fully than before because of two stimulating and perspicacious recent papers
(Paul and Smith 1984; Smith 1984u). In segment 2 fixation occurred, probably by extending the
lower surface of the animal (corresponding to the hemichordate right side and the chordate ventral
surface) down into the sea-floor. As seen from above, the mouth remained peripheral in position as
in a cornute. The ambulacra leading into the mouth (presumably connected with the water vascular
system and thus with the left mesocoel = left hydrocoel) became triradiate. Any gill slits on the
upper surface were lost— those on the lower surface would already have disappeared as a result of
dexiothetism. Thus segment 2 gave rise to the helicoplacoids as a plesion.

Segment 3. This is purely conventional, and perhaps did not exist. Three genera of helicoplacoid
are known.

Segment 4. In this segment the mouth moved into the centre of the upper surface by the process
known in crinoid embryology as torsion. The triradiate ambulacra became pentaradiate but retained
a distinct 2-1- 1 -E 2 pattern reflecting the primitive triradiality. The upper surface became flat. This
produced the form Camptostroma which may actually lie on the echinoderm stem lineage or even
be the latest common ancestor of living echinoderms (the first crown echinoderm). It thus belongs,
in the present state of knowledge, to the nodal group of the echinoderms.

Segment 5 is conventional and perhaps did not exist.

All known echinoderms, apart from helicoplacoids and perhaps Camptostroma, are probably
crown-group echinoderms, being assignable to one or other of the two primary echinoderm sub-
groups (Pelmatozoa and Eleutherozoa).

The meanings of the words ‘dorsal’ and ‘ventral’ require discussion. The upper surface of pelmato-
zoans and stem-group echinoderms is homologous, if the above account is correct, with the left
side of hemichordates and with the upper (i.e. dorsal) surface of chordates. Unfortunately, however,
the use of the terms ‘dorsal’ and ‘ventral’ in echinoderm literature is based on eleutherozoans such
as starfishes and sea-urchins which have inverted in evolution so that the primitive upper surface
faces downwards. Hence ‘dorsal’ and ‘ventral’ in echinoderms mean the exact opposite to what
they do in chordates. The chordate usage clearly has priority (Latin dorsum = back; venter = belly)
and is habitual to far more people than the echinoderm usage. The best solution to this nomenclator-
ial difficulty would be to eliminate the words ‘dorsal’ and ‘ventral’, henceforth, from echinoderm
terminology. The words ‘aboral’ and ‘oral’ have respectively the same meaning as the conventional
‘dorsal’ and ‘ventral’ in all echinoderm groups except helicoplacoids, for which the words ‘upper’
and ‘lower’ can fittingly be used with their obvious meanings.

Segments 6 and 7. A large number of important changes occurred in segment 6: the locomotory tail
was acquired and reached roughly the condition seen in Ceratocystis perneri with fore, mid, and
hind portions, while the soft parts of the tail probably included muscle blocks, notochord, and
dorsal nerve cord; the brain was developed at the anterior end of the tail; the plates of the head
evolved, probably to an almost rigid condition as seen in C. perneri, the water vascular system was
lost, but the hydropore was retained as outlet for the axial sinus (which in the early embryology of
crinoids it still is); the gill slits increased to seven in number (assuming that the single gill slit on the
left side of Cephalodiscus represents the primitive complement in Dexiothetica), and probably an
endostylar mucous filter developed inside the enlarged pharynx; the ear, paired trigeminal ganglia,
and median eye developed; and the layout of the head chambers seen in Ceratocystis perneri evolved,
with the viscera concentrated in the right anterior coelom to the right of the tail, and with a large
pharynx, a large buccal cavity, and a posterior coelom. A virtual left anterior coelom lay dorsal to
the other chambers.

Features known uniquely in C. perneri create special methodological problems, as already argued.
Many of them were probably primitive for cornutes, i.e. were present in the crownward part of
segment 6. Such include: the almost rigid surface of the head (by anti-crownward extrapolation);
the presence of a single plate (w-l-a-l-x) (by anti-crownward extrapolation since plates w, a, and x
are less differentiated from each other in P. menevensis than in "Cothurnocystis' fellinensis for

JEFFERIES, LEWIS AND DONOVAN: CAMBRIAN CORNUTE

475

example); the large number ( > 2 on either side) of dorsal segments per ventral ossicle in the hind
tail (by anti-crownward extrapolation); the presence of a hydropore (by outgroup comparison with
echinoderms); the position of the gonopore-anus to the right of the tail (by functional argument);
the absence of a high wall in front of the brain (by functional argument since this wall is associated
with the presence of the rectum in the posterior coelom, and this in turn depends on the gonopore-
anus being left of the tail).

Many features known uniquely in Ceratocystis perneri, however, are of indeterminate status. They
may have been primitive for cornutes and present in the crownward part of segment 6, or they may be
autapomorphies of C. perneri among known forms, and in that case evolved in segment 7. Such in-
clude: the absence of an oral pyramid; the position of two of the gill slits (nos. 1 -I- 2) anterior to the left
posterior dorsal crest; the ear penetrating plate i (it penetrates plate j in P. menevensis); the presence of
plate o; the separation, in some specimens, of gonopore from anus; the large number of segments in
the fore tail (about seventeen rings of successive plates on each side). The allocation of these features
as evolutionary novelties to their correct segment must await the recognition of plesions less crown-
ward than C. perneri, or perhaps between C. perneri and P. menevensis.

Segment 8. Most of the differences between C. perneri and P. menevensis probably arose as evolution-
ary novelties in this segment. The exceptions are features that arose in segment 7, none of which is
certainly identifiable, and those that arose in segment 9 as autapomorphies of P. menevensis.
Evolutionary novelties that probably arose in segment 8 include: loss of hydropore; migration of
gonopore and anus to the left of the tail and perhaps their unification to a single opening (if this
latter had not already happened in segment 6); acquisition of a high wall behind the posterior
coelom, in plates g and j, forming the cerebral basin; increase in the angle between the posterior
and right margins of the head (if correctly reconstructed in P. menevensis)-, reduction in the number
of segments in dorsal parts of the hind tail to only slightly more than one per ossicle on either side
proximally; acquisition of an oral cone (if this did not already exist in segment 6); better development
of f-spike and k-spike; reduction in size of plate y.

Segment 9. Features found exclusively in P. menevensis, but not in C. perneri nor in N. americana,
and which do not represent morphological intermediates between these two species, were probably
autapomorphies of P. menevensis, i.e. were evolved in segment 9. Such features are minor but do
seem to exist. They include: the very light build of the skeleton with two-dimensional retiform
stereom in large parts of the dorsal integument and ventral floor, strengthened in the floor by
irregularly placed struts to give a ‘space-frame’; the roundness of the e-spike; the bluntness of the
ridge on plates 1 and 2; perhaps the small number of segments in the foretail (only three, as
compared with about seventeen in C. perneri and five in N. americana). The lightly built stereom
strengthened by ribs, and the rounded dorsal ridges would have reduced weight and can probably
be seen as adaptations to life on a very soft sea-bottom. The small number of segments in the fore
tail perhaps suggests that P. menevensis was less motile than C. perneri or N. americana. The
autapomorphies of P. menevensis show that segment 9 was not purely conventional but really
existed.

Segment 10. Most of the differences between N. americana and P. menevensis are shared by
N. americana with more crownward plesions and therefore arose as evolutionary novelties in segment
10. They include: the strut as a thickening of plate g in N. americana— as already mentioned, the
strut may have begun from the internal process of plate g as seen in P. menevensis or by the
stabilization of apposed ribs in plates g and a in the ‘space-frame’ structure as seen in P. menevensis
(but in that case the space-frame type of construction would not be an autapomorphy of
P. menevensis)-, the clear differentiation of anterior U-plates in the branchial slits; the breakup of plate
ii into three pieces (though perhaps it was already in two pieces in P. menevensis)-, the disappearance
of plates 3 and 5 as recognizable entities and the smaller size of plate 2; the fact that plate d is part
of the frame instead of merely forming part of the floor of the head behind the mouth; the fact that
plate 1 extends more leftwards than plate k; the accurately opposite position of left and right hind-
tail plates. As already said several times, the only available specimen of N. americana is poor.

476

PALAEONTOLOGY, VOLUME 30

Segment II. Two features of N. americana are unique to it and are thus likely to be autapomorphies
of the species and evolved in segment 1 1. These are the rightward spread of the e-spike and f-spike
so that both are clearly visible in dorsal aspect. These autapomorphies show that segment 1 1
actually existed. This is also evident on stratigraphical grounds, for N. americana is less crownward
than the Cothurnocystis-\\ke species of Sprinkle (1976), particularly in having a rigid floor to the
head, but is stratigraphically later (Upper Cambrian rather than Middle Cambrian).

Segment 12. For purposes of text-fig. 26 we choose "C.' fellinensis to represent its particular plesion,
mainly because Ubaghs (1969) has described this species with his usual thoroughness (text-fig. 13).
Many other species of cornute belong near this position but cannot yet be accurately placed—
among them the undescribed Middle Cambrian ‘stylophoran’ of Sprinkle (1976). Evolutionary
novelties which can be ascribed to segment 12 are as follows: flexibility of the floor of the head
(apart from the strut); plate t and the t-spike; and breakdown of the triple-arch of the dorsal surface
so that only a line of spines anterior to the gill slits remains as a vestige. In ^C.' fellinensis plate k
extends somewhat further leftwards than plate 1. This could be a primitive feature, in which case
N. americana acquired the opposite condition in segment 11, or could be a secondary reversion
simulating a primitive condition.

Segment 13. The only evidence for this segment is stratigraphical: ' C fellinensis is contemporaneous
with Galliaecystis lignieresi and Amygdalotheca griffei (both from the Lower Arenig of the Montagne
Noire) and these two species are more crownward.

Segment 14. The chief changes in this segment are: the loss of plate y and the median eye; the loss
of plate x; and the loss of the spines which in ^C.’ fellinensis form a curved row anterior to the gill
slits. So far as can at present be determined, the family Scotiaecystidae {Scotiaecystis, Thoralicystis,
and Bobemiaecystis) belongs to the plesion of C. elizae.

Segment 15. This must have existed for stratigraphical reasons. C. elizae, from the uppermost
Ordovician (Ashgill) of Scotland is younger than all the cornute plesions crownward of it, i.e.
G. lignieresi and A. griffei, both from the Lower Arenig, Reticulocarpos hanusi from the Llanvirn,
and R. pissotensis from the Llandeilo. It is also younger than the earliest known members of the
chordate crown group (the mitrates Peltocystis corniita and Chinianocarpos thorali from the Lower
Arenig of the Montagne Noire).

Segments 16 to 24. We shall not discuss these here since there is nothing to add to the account given
by Jefferies (1986).

Thus Protocystites menevensis fits into a plesion between those of Ceratocystis perneri and
N. americana. It increases our knowledge of the evolution of the chordate stem lineage within the
less crownward cornutes.

LOCOMOTION IN PROTOCYSTITES MENEVENSIS

The strange shape of cornutes such as P. menevensis demands a functional explanation. Details of
their morphology suggest that they could move rearwards, at least occasionally, pulled by the tail.
Thus in Cotlmrnocystis elizae or S. curvata (Jefferies, in press, Ch. 7) the ventral spikes of the
head have points or sharp edges anteriorly but blunt terminations posteriorly, while the anterior
appendages would have sloped forwards and downwards into the mud. Both of these types of
anteroposterior asymmetry would tend to prevent forward movement and to facilitate rearward
movement. Also there is evidence that the tail was highly flexible in all cornutes, as is appropriate
to a locomotory organ. Neither the anteroposterior asymmetry of the spikes and appendages, nor
the flexibility of the tail, can be explained if the animal always rested immobile on the sea-bottom.
The tail of C. elizae would mainly have moved from side to side, as indicated by the existence of
gaps between the major plates of the fore tail on right and left but not in the vental mid-line. Also
the end part of the hind tail of C. elizae seems to be specially adapted for bending downwards. The

JEFFERIES, LEWIS AND DONOVAN: CAMBRIAN CORNUTE

477

external morphology of this species therefore suggests that the animal pulled itself across the sea-
floor by side-to-side wagging of the tail, gripping the sea-floor intermittently during the locomotory
cycle by pushing the distal part of the hind tail downwards into the mud. The fore-tail plates of
Ceratocystis perneri and P. menevensis probably likewise flexed from side to side, in view of the
overlap of the fore-tail plates across the ventral mid-line. Similarly, there is evidence that the
mitrates moved rearwards pulled by the tail, as suggested in particular by the presence of cuesta-
shaped ribs with the steep slope of the cuestas always morphologically anterior (Jefferies 1984,
1986). An asymmetrical shape, such as the head of a boot-shaped cornute, is easier to pull across a
surface than to push, since it is directionally stable when pulled but directionally unstable when
pushed. This is probably the fundamental reason why cornutes and mitrates moved rearwards.
With these thoughts in mind, we now reconstruct the locomotory cycle of P. menevensis in detail.
In what follows, we use the word ‘yaw’ in the standard sense for rotation about a vertical axis and
‘roll’ for rotation about a horizontal, anteroposterior axis.

The two largest ventral spikes of P. menevensis are situated near the left and right ends of the
head on plates k and f respectively. This suggests that movement involved yaw, with left spike and
right spike acting alternately as pivots; and such a yawing motion agrees with the presumed side-
to-side flexion of the tail as suggested by the ventral overlap of the fore-tail plates. The motion
would have been somewhat like one way of moving a heavy cupboard across a floor, by pivoting it
alternately about its left and right leading corners. The muscles moving the head of P. menevensis
would have been mainly those filling the large lumen of the fore tail; these would have represented
the motor of the animal. An adequate reconstruction of the locomoty cycle must, therefore, explain
how pressure was placed alternately on left and right spikes during crawling (on the k-spike and
then the f-spike, and then the k-spike again).

Our reconstruction of the locomotory cycle of P. menevensis is given in text-fig. 27. Text-fig. 21a
shows successive stages of the cycle in dorsal aspect and in absolute space, whereas text-fig. 276
shows the head and respective positions of the tail, always in exact posterior aspect. The drawings
in text-fig. 27 are based on an adjustable model where the notochord in the fore tail is represented
by a flexible ruler which allows side-to-side flexion, while the head and mid and hind tail are
represented by their outlines drawn on stiff white card, as described by Jefferies (1984). The dorsal
projections of the mid and hind tail in text-fig. 21a have been modified according to the reconstructed
inclinations of these parts in text-fig. 276. The outline of the head is shown in dorsal aspect,
neglecting the effects of inclination during roll. Ventrally prominent structures (ventral spikes, etc.)
are indicated by concentrations of dots, whose density suggests the degree of ventral prominence.

Each arrow in text-fig. 27 connects a particular anatomical point in one stage with its new position
in the next stage. An arrow thus indicates approximately the direction of travel, at the stage shown,
of the point in question. The lengths of the arrows for the different points of a given stage also
suggest their velocities relative to each other. However, successive stages shown are not supposed
to be separated by equal intervals of time. The relative lengths of arrows for the same point at
different stages therefore have no meaning. We have assumed that mud had strength and could
resist motion, whereas water did not. The force produced by the mud on the moving tail would
have been equal and opposite to the force exerted by the tail on the mud. Thus the arrows of
movement are opposite in direction to the forces that the movement provoked.

At stage 1 of text-fig. 21a, 6 the hind tail was buried in the sea-floor and curved ventrally (by the
relaxation of its dorsal muscles). This position represented the close of the previous locomotory
cycle. By stage 2 the hind tail had been raised out of the mud and straightened. Between stages 1
and 2, therefore, the tip of the tail had moved upwards and a resultant downward force would have
acted on the hind tail, tending to rotate the head in roll and to drive the right spike (f) into the sea-
bed while lifting the left spike (k) out of the sea-bed. The actual axis of rotation of the head would
have been approximately anteroposterior and probably located near the mid-ventral line of the tail
attachment, which protrudes ventrally below the general ventral surface of the head. This ventral
protrusion would have allowed the head to rock, see-saw fashion, alternately to right and left,
resting on the protrusion and on the mid-ventral line of the fore tail.

478

PALAEONTOLOGY, VOLUME 30

JEFFERIES, LEWIS AND DONOVAN: CAMBRIAN CORNUTE

479

13

TEXT-FIG. 27. Protocystites menevensis Hicks; reconstructed locomotory cycle, a, dorsal aspect; b, posterior
aspect (note that, unlike a, the viewpoint of successive diagrams is fixed relative to the animal but not in
absolute space). Locomotion probably involved yaw, with the f-spike (on the right) and the k-spike (on the
left) being used alternately as pivots. The head was pulled rearward by the tail when this penetrated the sea-
bottom on the side farthest from the pivot. The movements of the tail in the vertical dimension, as seen in b,
would automatically throw the weight of the head on to the intended pivot and lift the side of the head that

was to be moved.

480

PALAEONTOLOGY, VOLUME 30

Through stages 2 to 5 the tail swung leftwards in the sea water to a position just left of the mid-
line of the tail attachment. This movement, being in water, would have met little resistance and had
no effect on the position of the head.

By stage 5 the hind tail had moved down into the sea-floor. This would have tended to rotate the
head in roll with the same sense as previously and thus to lift the left spike further and throw
the weight of the head more on to the right spike (f). At stage 6, further movement of the tail
leftwards towards the head had the effect of rotating the head in yaw about the right spike
as pivot. As a consequence the left side of the head moved rearwards in space. This yaw continued
until stage 7.

At stage 8 the hind tail had been lifted out of the sea-floor. This movement was opposed by a
force exerted by the mud downward on the hind tail which rotated the head in roll, driving the left
spike down into the sea-floor and lifting the right spike. This rotation in roll was therefore opposite
in sense to that between stages 1 and 2. Once again, the axis of rotation would probably have been
approximately in the mid-line of the tail insertion, rocking on the ventral protrusion of the tail
insertion and of the ventral mid-line of the fore tail.

Through stages 8, 9, and 10 the tail moved rightward, but in the water. The movement would
therefore have met with little resistance and would have had no effect on the position of the head.

At stage 1 1 the mid and hind parts of the tail moved down into the sea-floor. Since these parts
were right of the mid-line of the tail insertion and their movement was opposed by an upward force
in the mud, the head would have tended to rotate in roll in the same sense as between stages 7 and
8, i.e. the left spike would have been driven further into the mud and the right spike lifted higher.
The same rolling rotation would have tended to push the left oral appendage (b-appendage)
downwards into the mud.

Through stages 11 to 13 the tail flexed rightwards and rotated the head in yaw about the left (or
k-) spike as pivot. Thus the right side of the head was moved rearwards. In this yawing rotation the
parts of the head furthest from the pivot would have been supported on the b-appendage (left oral
appendage). This was curved approximately, though not accurately, concentric to the k-spike and
therefore would scarcely have resisted the yawing rotation about that spike. At stage 13 a position
was reached exactly like stage 1, except that the head had moved rearwards and rightwards. Thus
the locomotory cycle was complete.

Locomotion, therefore, would probably have involved yaw alternately about the left and right
spikes, combined with roll so that left and right spikes were alternately pushed into the mud and
lifted clear of it. These motions of the head would have resulted automatically from the movements
of the tail from side-to-side and up-and-down into the mud. The rolling movement would have
been facilitated by the ventral protrusion of the head near the tail insertion. All boot-shaped
cornutes, since they have spikes and appendages concentrated at left and right of the head and the
same ventral protrusion of the head near the tail, probably crept rearwards somewhat in this
manner.

CONCLUSIONS

P. menevensis Hicks, 1872, from the Middle Cambrian of South Wales, is a cornute and therefore a
stem-group chordate. In the present paper it is described in detail for the first time and reconstructed.
It shared with N. americana the remarkable condition that the roof of the head was flexible but the
floor rigid.

Within the chordate stem group, P. menevensis belongs to a plesion between that of C. perneri
and that of N. americana. It is crownward of C. perneri (i.e. more closely related to the chordate
crown group) especially in lacking a hydropore, in having the gonopore-anus to the left of the tail,
and in the flexible, or partly flexible, roof to the head. It is less crownward than N. americana
especially in lacking the strut, in retaining a greater number of individualized plates in the roof of
the head, and in the absence of specialized U-plates framing each gill slit anteriorly.

JEFFERIES, LEWIS AND DONOVAN: CAMBRIAN CORNUTE

481

Specialized features (autapomorphies) of P. nienevensis included the very lightly built stereom
(particularly of the dorsal integument, of the right posterior wall of the head, and of parts of
the floor of the head) and the presence of irregularly placed ribs in the lightly built parts of
the floor. These autapomorphies were probably weight-saving adaptations favouring a life on very
soft mud.

As to soft parts, details of the superficial internal anatomy of P. nienevensis suggest the positions
of the oesophagus, stomach, and intestine in the right anterior coelom. The left boundary of that
coelom, or at least of its patent cavity, is indicated by a change in the stereom structure from
retiform to almost imperforate along a line in the floor of the head.

The locomotory cycle of P. nienevensis is reconstructed above. The animal probably crept rear-
wards by pivoting alternately around spikes near the left and right posterior corners of the head.
This movement was produced by waving the tail alternately to the left and right while lowering it
into, or raising it out of, the mud in particular parts of the cycle. This same locomotory cycle was
probably usual in boot-shaped cornutes and was therefore probably primitive for stem-group
chordates, so far as these are at present known.

Concerning morphological terminology, we have abandoned the objective notation of plates
which was formerly used for cornutes and mitrates (e.g. Jefferies 1968). Instead we apply the
comparative terminology proposed by Jefferies and Prokop (1972); this uses the same lower-case
letter (a, b, e, li, etc.) for all plates believed to be homologous in cornutes and mitrates. The terms
dorsal and ventral, as conventionally used in echinoderms, should be abandoned, since in that
phylum they signify the exact opposite to what they mean in chordates.

As to phylogenetic terminology, we argue that the plesion, though a useful concept in subdividing
a stem group, is inherently paraphyletic when completely known, i.e. when all its constituent species
are known. We use the term ‘crownward’ to mean ‘more closely related to the crown group’, with
its opposites ‘less crownward’ to indicate phylogenetic position and ‘anti-crownward’ to indicate
direction away from the crown group along the stem lineage. ‘Crownward’ is more restricted in
meaning than ‘advanced’, for it means ‘advanced along the stem lineage only’. Hennig’s term
‘intermediate category’ is used for a paraphyletic group which provably comprises two or more
adjacent plesions within a stem group. We propose the term ‘nodal group’ to comprise those forms
which could, on the basis of synapomorphies, be crownward members of the stem group of some
extant group, or could be primitive members of the crown group.

Tectonic distortion of the material made P. nienevensis difficult to reconstruct. This distortion
was corrected, to some extent, with the help of a computer program and a plotter. This correction
was a necessary preliminary to the normal process of reconstructing the animal in three dimensions
on a drawing board.

Thus P. nienevensis, a strange boot-shaped animal, throws light on the evolution of our early
ancestors and marks the oldest occurrence known of the chordate phylum in Britain.

Acknowledgements. We should like to thank Drs Chris Paul (University of Liverpool) and Adrian Rushton
(British Geological Survey) for invaluable help and advice in the early stages of the project. We are also
grateful to Drs David Price (Sedgwick Museum, Cambridge) and Michael Bassett and Bob Owens (National
Museum of Wales, Cardiff) for loan of material. Dr Andrew Smith critically read the text and made many
helpful comments. Christopher Griffiths (University of Liverpool) gave much valuable help in the field.

David Lewis (British Museum (Natural History)) carefully curated the abundant but scrappy material of
P. nienevensis now preserved in that museum. Alan Paterson (Biometrics Section, BM(NH)) devised the
computer program for correction of tectonic distortion and was always ready with help and advice. Dr John
Cosgrove (Geology Department, Imperial College, London) gave useful advice on the problem of correcting
tectonic distortion. Dr David Hardwick (Civil Engineering Department, Imperial College, London) kindly
discussed locomotion of P. nienevensis from an engineer’s standpoint.

Finally, we acknowledge our debt to the great Welsh amateur palaeontologist and collector Henry Hicks,
who discovered P. nienevensis, recognized its distinctiveness, and put it on record; without his pioneering work
the species would still probably be unknown.

482

PALAEONTOLOGY, VOLUME 30

REFERENCES

ALLEN, p. M., JACKSON, A. A. and RUSHTON, A. w. A. 1981. The Stratigraphy of the Mawddach Group in the
Cambrian succession of North Wales. Proc. Yorks, geol. Soc. 43, 295-329, pis. 16 and 17.

ANGELIN, N. p. 1851. Pcilaeontologia suecica. Pars 1. Iconographia criistaceorwn formatiouis transitionis. Ease.
1, pp. 1-24, pis. 119.

APPLEBY, R. M. and JONES, G. L. 1976. The analogue video-reshaper — a new tool for palaeontologists. Palaeon-
tology, 19, 565-586.

AX, p. 1984. Das phylogenetische System, 349 pp. Fischer, Stuttgart.

In press. The phylogenetic system. Wiley, Chichester and New York.

BARRANDE, j. 1 864. Notice preliminaire sur le systDne silurien et les trilohites cie Boheme, vi -I- 97 pp. Leipzig.

1887. Systeme silurien dii centre de la Boheme. Vol. VII. Classe des echinodermes, ordre des Cystidees,

233 pp. Prague.

BASSLER, R. s. and MOODY, M. w. 1943. Bibliographic and faunal index of Paleozoic pelmatozoan echinoderms.
Spec. Pap. geol. Soc. Am. 45, i vi, 1 -733.

BATJJER, F. A. 1900. Chs. 8 12 (pp. 1-216) in lankester, e. r. A treatise on zoology. Part III. Echinoderma,
344 pp. London.

1913. Caradocian Cystidea from Girvan. Trans. R. soc. Edinb. 49, 359-521, pis. 1 6.

BEYRICH, E. 1845. Ueher einige bohmischer Trilobiten, 47 pp., 1 pi. Berlin.

BROGGER, w. c. 1878. Om paradoxidesskifrene ved Krekling. Nyt Mag. Natwvid. 24, 18-88, pis. 16.
cowiE, j. w., RUSHTON, A. w. A. and STUBBLEFIELD, c. J. 1972. A Correlation of Cambrian rocks in the British
Isles. Spec. Rep. geol. Soc. Land. 2, I -42.

DAVIDSON, T. 1871 . A monograph of the British fossil Brachiopoda. Vol. 3. Devonian and Silurian. Palaeontogr.
Soc. [Monogr.], pt. 7 (4), 249-397, pis. 38-50.

FORTEY, R. A. and JEFFERIES, R. p. s. 1982. Fossils and phylogeny— a compromise approach. In joysey, k. a.

and FRIDAY, A. E. (eds.). Problems of phylogenetic reconstruction. Spec. Vol. Syst. 21, 1-442.

GILL, E. D. and CASTER, K. E. I960. Carpoid echinoderms from the Silurian and Devonian of Australia. Bull.
Am. Paleont. 41 (185), 171.

GRAY, J. 1957. How animals move, 144 pp. Penguin, London.

GRONWALL, K. A. 1902. Bomholms Paradoxideslag og deres fauna. Damn. geol. Unders. (2) 13, i -xi, I 230,
pis. 1-50.

HARKNESS, R. and HICKS, H. 1871. On the ancient rocks of the St. David’s promontory. South Wales, and their
fossil contents. With descriptions of the new species. Q. Jl geol. Soc. Lond. 27, 384-404.

HENNiG, w. 1966. Phylogenetic systematics, 263 pp. University of Illinois, Urbana.

1969. Die Stammesgeschichte der Insekten, 436 pp. Kramer, Frankfurt am Main.

1981. (Translated pont, a.) Insect phylogeny, 514 pp. Wiley, Chichester.

HICKS, H. 1881. The classification of the Eozoic and Lower Palaeozoic rocks of the British Isles. Pop. Sci. Rev.
(NS) 5, 289-308.

1892. The fauna of the Olenellus Zone in Wales. Geol. Mag. (Dec. 3), 9, 21-24.

and JONES, T. R. 1872. On some undescribed fossils from the Menevian Group. By Henry Hicks. With a

note on the Entomostraca from the Cambrian rocks of St. David’s. By Prof. Rupert Jones. Q. Jl geol. Soc.
Lond. 28, 173-185, pis. 5-7.

and SALTER, J. w. 1866. Report on further researches in the Lingula Flags of South Wales, by Henry

Hicks. With some notes in the sections and fossils by J. W. Salter, F.G.S, A.L.S. Rep. Br. zl55. Advmt Sci.
1865, 281-285.

1867. Second report on the ‘Menevian Group’ and the other formations at St. David’s, Pembroke-
shire. Ibid. 1866, 182-186.

HUTCHINSON, R. D. 1962. Cambrian stratigraphy and trilobite faunas of southeastern Newfoundland. Bull,
geol. Surv. Can. 88, 1 156, pis. 1-25.

ILLING, V. c. 1916. The Paradoxidian fauna of a part of the Stockingford Shales. Q. Jl geol. Soc. Lond. 71,
396-450, pis. 28-38.

JAEKEL, o. 1900. liber Carpoideen, eine neue Klasse von Pelmatozoen. Z. dt. geol. Ges. 52, 661 677.

JEFFERIES, R. p. s. 1967. Some fossil chordates with echinoderm affinities. Svmp. zool. Soc. Lond. 20,
163-208.

1968. The subphylum Calcichordata (Jefferies, 1967)— primitive fossil chordates with echinoderm affini-
ties. Bull. Br. Mus. nat. Hist. (Geol.), 16, 243 339.

JEFFERIES, LEWIS AND DONOVAN: CAMBRIAN CORNUTE

483

1969. Ceratocystis perneri Jaekel — a Middle Cambrian chordate with echinoderm affinities. Palaeon-
tology, 12, 494-535.

1973. The Ordovician fossil Lagynocystis pyramidalis (Barrande) and the ancestry of amphioxus. Phil.

Trans. R. Soc., B265, 409 469.

1979. The origin of chordates— a methodological essay, pp. 443-477. In house, m. r. (ed.). The origin of

major invertebrate groups. Spec. Vol. Sysl. Ass. 12, 1-515.

1981a. Fossil evidence on the origin of the chordates and echinoderms, pp. 487 561. In ranzi, l. (ed.).

Origine dei grand! phyla dei Metazoa. A/ti Conv. Accacl. naz. Lined, 49, 1 565.

1981b. In defence of the calcichordates. J. Linn. Soc. (Zool.), 73, 351 396.

1984. Locomotion, shape, ornament and external ontogeny in some mitrate calcichordates. J. vert.

Paleont. 4, 292-319.

1986. The ancestry of the vertebrates, 316 pp. British Museum (Natural History), London.

JOYSEY, K. A., PAUL, c. R. c. and RAMSBOTTOM, w. H. c. 1967. Echinodermata: Pelmatozoa, Ch. 21. In

HARLAND, w. B. (ed.). The fossil record— a symposium with documentation. Spec. Pnhl. geol. Soc. Land. 2,
1 827.

and LEWIS, d. n. 1978. The English Silurian fossil Pkicocystites forhesianiis and the ancestry of the

vertebrates. Phil. Trans. R. Soc. B282, 205-323.

and PROKOP, R. J. 1972. A new calcichordate from the Ordovician of Bohemia and its anatomy, adap-
tations and relationships. J. Linn. Soc. (Biol.), 4, 69 1 15.

KESLING, R. V. 1967. Cystoids, pp. S85-S267. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part
S. Echinodermata 1(1), pp. S1-S296. Geological Society of America and University of Kansas Press, New
York and Lawrence, Kansas.

KOBAYASHi, T. 1935. The Cambro-Ordovician formations and faunas of South Chosen with special study on
the Cambrian trilobite genera and families. J. Fac. Sci. Tokyo Univ. (sect. 2), 4 (2), 49-344.

LAKE, p. 1906-1946. A monograph of the British Cambrian trilobites. Palaeontogr. Soc. [Monogr.], 350 pp.,
47 pis.

MARTiNSSON, A. 1974. The Cambrian of Norden, pp. 185-283. In Holland, c. h. (ed.). Lower Palaeozoic rocks
of the World. Vol. 2. Cambrian of the British Isles, Norden, and Spitzbergen. Wiley, London.

NICHOLAS, T. c. 1916. The geology of the St. Tudwal’s Peninsula (Caernarvonshire). Q. Jl geol. Soc. Lond. 71,
83-143.

PATTERSON, c. 1981. Signihcance of fossils in determining evolutionary relationships. A. Rev. Ecol. Svst. 12,
194223.

and ROSEN, d. e. 1977. Review of the ichthyodectiform and other Mesozoic teleost fishes and the theory

and practice of classifying fossils. Bull. Am. Mas. nat. Hist. 158, 84 172.

PAUL, c. R. c. 1979. Caledonian echinoderms of the British Isles, pp. 453 456. In Harris, a. l., Holland, c. h.
and LEAKE, B. E. (eds.). The Caledonides of the British Isles— reviewed. Spec. Publ. geol. Soc. Lond. 8,
1 768.

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.

RAMSAY, A. c. 1866. The geology of North Wales. Mem. Geol. Surv. U.K. 3, i viii, 1-381.

RAMSEY, J. G. 1967. Folding and fracturing of rocks, 568 pp. McGraw-Hill, New York.

RUSHTON, A. w. A. 1974. The Cambrian of Wales and England, pp. 43 121. In Holland, c. h. (ed.). Lower
Palaeozoic rocks of the World. Vol. 2. Cambrian of the British Isles, Norden and Spitzbergen. Wiley,
London.

1979. A review of the Middle Cambrian Agnostida from the Abbey Shales, England. Alcheringa, 3,

43-61.

SALTER, J. w. 1 863. On the discovery of Paradoxides in Britain. Q. Jl geol. Soc. Lond. 19, 274-277, pi. 1 3.

1864. On some new fossils from the Eingula-Flags of Wales. Ibid. 20, 233-241, pi. 13.

1865. On some additional fossils from the Lingula-Elags. By J. W. Salter, Esq., A.L.S., F.G.S. with a

note on the genus Anapolenus', By Henry Hicks Esq., M.R.C.S. Ibid. 21, 476 482.

1873. A catalogue of the collection of Cambrian and Silurian fossils contained in the Geological Museum of

the University of Cambridge, 204 pp. Cambridge University, Cambridge.

and HICKS, h. 1867. On a new Lingulella from the red Eower Cambrian rocks of St. David’s. Q. Jl geol.

Soc. Lond. 23,339-341.

1869. On some fossils from the ‘Menevian Group’. Ibid. 25, 4 47, pis. 2 and 3.

SEiLACHER, A. and HEMLEBEN, c. 1966. Spurenfauna und Bildungstiefe der Hunsruckschiefer. Notizbl. hess.
Landesamt. Bodenforsch. Wiesbaden, 94, 40 53.

484

PALAEONTOLOGY, VOLUME 30

SMITH, A. B. 1984a. Classification of the Echinodermata. Palaeontology, 27, 431-459.

1984fi. Echinoid palaeobiology, 190 pp. George Allen & Unwin, London.

SPRINKLE, j. 1976. Biostratigraphy and paleoecology of Cambrian echinoderms from the Rocky Mountains.
Geology Stud. Brigham Young Univ. 23, 61-73.

STEAD, K. T. G. and WILLIAMS, B. p. J. 1971. The Cambrian rocks of North Pembrokeshire. In bassett, d. and
BASSETT, M. G. (eds.). Geological excursions in South Wales and the Forest of Dean. National Museum of
Wales, Cardiff.

TAYLOR, K. and RUSHTON, A. w. A. 1972. The pre-Westphalian geology of the Warwickshire Coalfield, with an
account of three boreholes in the Merevale area. Bull. geol. Surv. Gt Br. 35, i-vii, 1 150, pis. 112.

THOMAS, A. T., OWENS, R. M. and RUSHTON, A. w. A. 1984. Trilobites in British stratigraphy. Spec. Rep. geol.
Soc. Fond. 16, 1 78.

THULBORN, R. A. 1984. The avian relationships of Archaeoptervx, and the origin of birds. J. Finn. Soc. (Zool.),
82,119 158.

UBAGHS, G. 1963. Cothurnocystis Bather, Phyllocystis Thoral and an undetermined member of the order Soluta
(Echinodermata, Carpoidea) in the uppermost Cambrian of Nevada. J. Paleont. 37, 1133 1142, pis. 151
and 152.

1967. Stylophora, pp. S495 S565. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part S.

Echinodermata 1(2), pp. S297-S650. Geological Society of America and University of Kansas Press, New
York and Lawrence, Kansas.

1969. Les echinodermes carpoides de I’ordovicien inferieur de la Montagne Noire, 1 12 pp. Cah. Paleont.,

C.N.R.S. Paris.

1981. Reflexions sur la nature et la fonction de I’appendice articule des carpoides Stylophora (Echino-
dermata). Annls Pcdeont. (Invertebres), 67, 33-48.

R. P. S. JEFFERIES
Department of Palaeontology
British Museum (Natural History)
Cromwell Road, London SW7 5BD

M. LEWIS

Department of Geology
University College

PO Box 78, Cardiff CEl IXL

and

S. K. DONOVAN

Typescript received 21 February 1986
Revised typescript received 28 July 1986

Department of Geology
University of the West Indies
Mona, Kingston 7, Jamaica

A REVIEW OE EAVOSITID AEEINITIES

by COLIN T. SCRUTTON

Abstract. Although the favositids have been traditionally interpreted as a group of Palaeozoic tabulate
corals, there has been persistent speculation, particularly over the last decade, that they could be the massive
basal skeletons of sponges and should be transferred to the Porifera. Two recent papers, claiming respectively
the preservation of spicules and the fossilization of soft polyps, strongly focus the dispute. Here, all the
evidence relating to the affinities of favositids, including these recent claims, is reviewed. It is concluded that
this evidence strongly favours retention of the favositids within the Tabulata and assignment of the Tabulata
to the Cnidaria Anthozoa.

The favositids are an important group of extinct organisms with a massive or branched calcareous
coralline skeleton and are conventionally assigned to the Palaeozoic subclass Tabulata. Almost all
specialists classify the Tabulata as corals in the anthozoan Cnidaria (Hill 1981). That the favositids
might be sponges was first seriously suggested when the sclerosponges were discovered early this
century (Kirkpatrick 1912), an observation largely overlooked and ignored at the time. During the
last twelve years, however, the rediscovery of the sclerosponges has promoted a long-running debate
concerning the affinities of favositids and even of the Tabulata as a whole (Hartman and Goreau
1975; Fliigel 1976; Stel 1978; Oliver 1979, 1986; Scrutton 1979; Oekentorp and Stel 1985). Two
recent papers appear to polarize the argument: the claimed discovery of spicules in the favositid
Thamnopora (Kazmierezak 1984), and the report of fossilized polyps in Favosites itself (Copper
1985). All specialists regard Favosites and Thamnopora as closely related so both presumably cannot
be right. This paper sets these recent conflicting claims in context by reviewing all the evidence
relating to the problem of affinity of these structurally simple fossils.

Some supposed tabulate corals variously homoeomorphic with favositids have been reclassified
in the light of sclerosponge work. The Chaetetida, in whole or in part, are now widely considered
to be sponges (Hartman and Goreau 1972; Fischer 1977; West and Clark 1984; Vacelet 1985; but
see also Hill 1981) and some have yielded unquestionable spicules (Died et al. 1977; Gray 1981).
However, Sokolov (1962) had already argued strongly against the inclusion of this group in the
Tabulata before the new sclerosponge discoveries. In addition, a reassessment of Nodulipora and
Desmidopora, formerly classified as Favositidae (Hill and Stumm 1956), has established a good case
for their transfer to the sclerosponges (Hartman and Goreau 1975; Stel and Oekentorp 1981). It is
possible that some further tabulate taxa may also eventually require reassignment, but generally it
is considered less likely that tabulates other than favositids could be sponges (Hartman and Goreau
1975; Scrutton 1979), although transfer of the whole group to the Porifera has been proposed (Stel
and de Coo 1977). The definition of the Tabulata taken here is that outlined by Scrutton (1984), who
argued that the subclass essentially constitutes a monophyletic grouping. In this paper, discussion is
limited to Favosites and its close relatives, collectively and informally termed favositids and equiva-
lent in general terms to the Favositina of Hill ( 1981 ).

The sclerosponges are now considered to be a polyphyletic collection of various demosponges
(Vacelet 1977, 1985). They are united only by the possession of a massive ‘coralline’ basal skeleton
which in itself seems to be of little phylogenetic significance. Indeed, the basal skeleton in different
sponges shows a wide range in macroscopic form, microstructure, and mineralogy. The term
sclerosponge is retained here informally, as the group collectively represents the most coral-like
representatives of the sponges. It includes the Tabulospongida of Hartman and Goreau (1975)

IPalaeontology, Vol. 30, Part 3, 1987, pp. 485-492.)

© The Palaeontological Association

486

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. I. Comparison of basic morphology and characteristic corallite/calicle size in a favositid and a
sclerosponge, in transverse and longitudinal sections, a, b, Favosites multipora Lonsdale, BM(NH) R51936;
Silurian, Wenlock Series, Much Wenlock Limestone; road cut on B4378, 2-5 km north-north-east of Much
Wenlock, Shropshire, c, d, Tabulospongia japonica Mori, BM(NH) 1986:7;7:1a; Recent; Ishigaki-shima,

Ryukyu Islands, Japan. All x 8.

as well as the morphologically very similar Chaetetida and these two orders together represent the
sclerosponges most closely homoeomorphic with favositids (text-fig. 1).

ANALYSIS OF FAVOSITID STRUCTURE

Favosites is a colonial organism consisting of closely appressed polygonal tubes interconnected by
mural pores (text-fig. 1a, b). Conventionally, the skeletal tubes in favositids are called corallites (as in
corals generally) whilst those in sclerosponges are calicles; it will be convenient to use these terms
here although 1 do so without prejudice. In favositids, corallite ‘diameters’ vary from c. 0-5-5-0 mm,
comparable to corallite size in many other tabulate corals and some scleractinian corals, although

SCRUTTON: FAVOSITID AFFINITIES

487

TEXT-FIG. 2. Three different methods of origin of new units within a colony, a, b, lateral increase in favositids;
the point of communication with the parent corallite is so small that the critical stage is seldom seen in random
sections; for this reason, increase in favositids was earlier thought to be intermural increase, c, o, intramural
increase in sclerosponges; the new calicle arises within the wall with no communication with surrounding
calicles. e, f, longitudinal hssion in sclerosponges; an existing calicle is subdivided subec]ually by the growth and
fusion of pseudosepta from opposite walls; occasionally, subdivision may be effected by a single pseudoseptum
growing from one wall across the calicle. a, c, e are cross-sections, with an arrow indicating the critical stage
of increase; b, d, f are ‘perfect’ longitudinal sections through the corresponding critical stages.

to few rugose corals (most of which have larger corallites). Of the living sclerosponges so far known,
none has calicle diameters in excess of 0-6 mm (text-fig. Ic, d), and those with the larger diameters
have a functional relationship between the calicles and ostia in which the tissue enclosed by each
calicle is the unit supported by a single ostium (Hartman and Goreau 1975). Meiiia, however, with
very small calicles (01 2-0- 15 mm), has no such relationship. The fossil tabulosponges and the
chaetetids have calicle diameters not exceeding c. 1-2 mm, with most in the range 01 5-0-50 mm,
significantly smaller than the corallites in the majority of favositids.

New corallites in Favosites are now known always to arise by lateral increase (Oliver 1968; Stel
1978; Scrutton 1979), equivalent to the peripheral intracalicular increase of Hill ( 1981 ) (text-fig. 2a,
B). The process is structurally comparable with lateral increase in other corals. In sclerosponges,
however, new calicles arise either by longitudinal fission or by apparently true intramural increase
(Hartman and Goreau 1972, 1975) (text-fig. 2c-f). There appears to be no overlap between the two
groups.

The determination of original composition and microstructure in fossil material is a more debat-
able area but the favositid skeleton was most probably originally calcitic (Richter 1972; Sandberg
1975). Corallite walls are considered to have had a fibronormal microstructure, although some
argue that lamellar microstructure was primary in certain genera (Lafuste 1962). More critically.

488

PALAEONTOLOGY, VOLUME 30

there is no doubt that this fabric is based on an epithecal surface bounding individual corallites
which is expressed as a median dense band when corallite walls are fused back to back (Oekentorp
and Sorauf 1970; Schouppe and Oekentorp 1974; Stel 1978; Hill 1981) (text-fig. 3a-c). Precisely
similar walls are known in cerioid rugose corals such as Actinocyathus, Hexagonaria, Lithostrotion,
and many others (Hill 1981), as well as in other tabulate corals (Flower 1961). This indication of
the individuality of the component corallites within the favositid colony is buttressed by two
additional features. First, by the occurrence of subcerioid growth, in which irregular intercorallite
cavities are formed within an otherwise cerioid morphology (Philip 1960). Secondly, by the manner
of formation of pseudoperculae, plates with concentric growth lines, often with an excentric origin,
that individually close off abandoned calices in some specimens (Dunbar 1927; Swann 1947). The
presence of intermural spaces (Swann 1947; Ross 1953) would represent further evidence, although
some structures so described are due to commensal organisms and others are at least enhanced
diagenetically if not wholly of diagenetic origin (Oekentorp 1969). Even so, the distribution of the
commensal structures themselves often follows a pattern related to the corallite walls which suggests
that the latter defined individual units of soft tissue.

Sclerosponges are mainly aragonitic with spherulitic or trabecular structure, although calcitic
lamellar skeletons are known in tabulosponges (Hartman and Goreau 1972, 1975; Vacelet 1985)
(text-fig. 3d, e). In neither, however, is there any indication of an axial zone in the wall representing
fused epitheca, or of any individuality of the component calicles (Hartman and Goreau 1975).
Intercalicular walls have a unitary microstructure. The ‘epitheca’ of sclerosponges (Hartman and
Goreau 1972, 1975) is equivalent to the holotheca of non-cerioid tabulate and massive rugose corals
(Hill 1981). In most, if not all, cerioid tabulates and rugosans the outer wall of the colony is the
sum of the free epithecal walls of adjacent, peripheral corallites.

The septal spines and tabulae of Favosites can be matched by both other corals and some
sclerosponges in gross morphology and possibly microstructure, although if favositid septal spines
are trabecular, as Hill (1981) speculated, then they are uniquely cnidarian (text-fig. 3c, e). Although
favositid spines may be arranged in regular vertical rows and there may be twelve such rows, other
configurations may occur and their distribution may also be irregular (Schouppe and Oekentorp
1974; Oekentorp 1976; Oekentorp and Stel 1985). A very similar range of variation appears to be
possible in some sclerosponges. However, some favositids, the Agetolitidae, have unusually well-
developed septa for which rugosan septal insertion has been claimed (Kim 1974), thus strongly
supporting anthozoan affinities.

The mural pores of favositids (text-figs. 2a, b, 3a, b) are structurally comparable with, and were
presumably identical in function to, the horizontal tubules of syringoporoids. There is no difference
between the appearance of syringoporoid intercommunication when the corallites become com-
pressed and contiguous and the mural pores of favositids (see, for example. Hill 1981). Some other
tabulate coral groups also possess mural pores of supposed similar function to those of favositids,
if of different structure in detail, but pores are not present in all tabulate corals. No similar structures
are known in living sclerosponges and the chaetetids, although they are present in the probable
fossil sclerosponges Nodulipora and Desmidopora. Stel and Oekentorp (1981) suggested a relation-
ship between the presence of pores and larger calicle size in sclerosponges, although they noted an
exception to this themselves. Fliigel’s (1976) suggested analogy between mural pores in favositids
and astrorhizal systems in sclerosponges is ingenious but unconvincing (Scrutton 1979; Stel and
Oekentorp 1981 ).

Mural pores are as equally unknown among rugose and scleractinian corals as among the bulk
of sclerosponges. However, in favositids they can be interpreted most convincingly as a device
allowing interpolypal communication and thus a higher level of integration of the colony than in
unmodified cerioid morphologies (Coates and Oliver 1973). In other corals this is achieved by the
wholesale loss of the epithecal barrier to integration, or by pervasively and finely perforate walls in
some Scleractinia. It is clear, however, that the presence or absence of mural pores cannot be taken
as a criterion of great significance in determining the affinity of the favositids.

Neither favositids nor any other tabulate corals show any sign of astrorhizal structures like those

SCRUTTON; FAVOSITID AFFINITIES

489

TEXT-FIG. 3. Microstructural characteristics of a favositid and a sclerosponge. a c, Favosites midtipora Lonsdale
(specimen details as for text-fig. 1a, b). The dark mid-line of the wall, representing fused epithccae of adjacent
corallites, and its fibro-normal coating is clear in a and c, and growth lines on the epithecal surface in section
can be distinguished in b; mural pores are present in a and b and septal spines are sectioned (particularly top
left) in c. D, E, Tahidospongia japonica Mori (specimen details as for text-fig. Ic, d). The skeleton is high-Mg
calcite with lamellar microstructure, clearly seen in e; the undulose surfaces of the lamellae are responsible for
the concentric patterns in the wall in cross-section d; calical spines arc well developed, formed of sharply
peaked extensions of lamellar tissue, but spicules are not incorporated into the calcitic skeleton. All x 50.

490

PALAEONTOLOGY, VOLUME 30

of sclerosponges and stromatoporoids. As many sponges do not reflect the system of exhalant
canals in their skeletons, this may not be particularly significant. However, the individuality of
favositid corallites strongly argues against the former presence of continuously integrated tissue
across the colony surface as in sponges. Under these circumstances, some skeletal reflection would
be expected of a sponge-scale exhalant current system— hence Fliigel’s interpretation of mural pores
(Fliigel 1976). Mural pore distribution, however, seems to have no pattern to it that would support
such an interpretation. Indeed, Nodulipora may possess both astrorhizae and mural pores with no
specific relationship between them (Stel and Oekentorp 1981 ).

Turning now to recent developments: first, Kazmierczak (1984) has claimed the preservation
of desma-like spicules in a Devonian Thanmopora. These are rare, approximately parallel-sided,
irregularly branched structures preserved in the peripheral part of the skeleton in microgranular
low-Mg calcite and lined with micrite. They do not have a convincing spicular morphology. Their
appearance, location, and mode of occurrence, however, strongly suggests that they are endolithic
borings (Oekentorp 1985; Finks 1986). No convincing records of spicules in favositids are known.
On the other hand, this is not a strong argument in itself against sponge affinites as several
sclerosponges, particularly tabulosponges (Hartman and Goreau 1975; Mori 1976, 1977), do not
incorporate spicules into their calcareous skeletons (text-fig. 3d, e).

The second recent development is the report by Copper (1985) of presumed polyps of cnidarian
character preserved in Silurian Favosites from Anticosti Island. Six well-preserved colonies show
small dome-like structures in the centres of calices, with axial pits surrounded by normally twelve
concentrically wrinkled radiating segments. In the specimen figured, their development is strikingly
wide and uniform. A possible diagenetic origin for the structures has been suggested by Oekentorp
and Stel (1985); their reference to a Protrochiscolithus figured by Flower (1961, pis. 14 and 16) is
misleading, however, as the polypoid appearance of the silicified surface in that case is a direct
reflection of the septal structure of the genus. No such interpretation seems possible with Copper’s
Favosites. Oliver (1986) tentatively suggested an origin related to pseudoperculae; this may be
important in understanding how calcification might have occurred, but the ‘polyps’ are most unlike
known pseudoperculae in detailed form and regularity. It is a remarkable case of preservation and
one not easy to explain, but it is difficult to avoid the conclusion that the structures seen are indeed
the remains of polyps. The twelve tentacles of Copper’s favositid polyps are intriguing in view
of the frequency with which twelvefold septal distributions occur in tabulates (taken here to include
the heliolitids: Hill 1981; Scrutton 1979, 1984). However, septal layouts and patterns of insertion in
tabulates require further study (Oliver 1986) and it is premature at the moment to set aside the
Tabulata as a group characterized by dodecal symmetry (Copper 1985; Oekentorp and Stel 1985).

CONCLUSIONS

There appears to be no single item of evidence in favour of the favositids being sponges, other than
a very gross morphological similarity with the tabulosponges. Mural pores and septal structures in
favositids both show features more strongly related to other tabulates and to the Rugosa respect-
ively, whilst neither in the broadest sense is exclusively cnidarian. On the other hand, corallite size,
and particularly mode of increase and evidence of corallite individuality are all positive cnidarian
features. The polyps described by Copper (1985) appear to represent one additional, if spectacular,
item on the cnidarian side. The weight of evidence is strongly in favour of favositids being cnidarians
and a subgroup of the monophyletic subclass Tabulata (as defined by Scrutton 1984). To maintain
any claim for sponge affinities for these extinct organisms, not only must Copper’s polyps be
explained away, but some new and convincing positive evidence must be forthcoming.

Acknowledgements. I am grateful to Kei Mori (Tohoku University, Japan) for supplying me with a specimen
of Tabulospongia japonica. I thank Bill Oliver (United States Geological Survey, Washington, DC) and Kevin
Goodger (University of Newcastle upon Tyne) for helpful comments on the manuscript.

SCRUTTON: FAVOSITID AFFINITIES

491

REFERENCES

COATES, A. G. and OLIVER, w. A., JR. 1973. Coloniality in zoantharian corals. In boardman, r. s., cheetham,
A. H. and OLIVER, w. A., JR. (eds.). Animal colonies, 3 27. Dowden, Hutchinson and Ross, Stroudsbcrg,
Pennsylvania.

COPi’ER, p. 1985. Fossilized polyps in 430-Myr-old Favosites corals. Nature, Land. 316, 142 144.

DiECi, G., RUSSO, A., RUSSO, E. and MARCHi, M. s. 1977. Occurrence of spicules in Triassic chaetetids and
ceratoporellids. Boll. Soc. paleont. ital. 16, 229 238.

DUNBAR, c. o. 1927. Pseudopcrcula in the tabulate coral Favosites. Am. J. Sci. 13, 101 114.

FINKS, R. M. 1986. ‘Spicules’ in Thamnopora. Fossil Cnidaria, 15 ( 1 .2), 22.

FISCHER, j.-c. 1977. Biogeographie des Chaetetida et des Tabulospongida post-paleozoi'ques. Minn. Bur. Reel),
geol. minier. 89, 530 534.

FLOWER, R. H. 1961. Montoya and related colonial corals. Mem. Inst. Min. Technol. New Mex. 7, I 97.

FLUGEL, H. w. 1976. Ein Spongienmodell fiir die Favositidae. Letliaia, 9, 405 419.

GRAY, D. I. 1981. Spicule pseudomorphs in a new Palaeozoic chaetetid, and its sclerosponge affinities. Palaeon-
tology, 23, 803-820.

HARTMAN, w. D. and GOREAU, T. F. 1972. Ceratoporellc) (Porifera; Sclerospongiae) and the chaetetid ‘corals’.
Trans. Conn. Acad. Arts Sci. 44, 133 148.

1975. A Pacific tabulate sponge, living representative of a new order of sclerosponges. Postilla, 167,

1 13.

HILL, D. 1981. Rugosa and Tabulata. In teichert, c. (ed.). Treatise on invertebrate paleontology. Part F (Suppl.

1) , xl + 762 pp. Geological Society of America and University of Kansas Press, Boulder, Colorado and
Lawrence, Kansas.

and STUMM, e. c. 1956. Tabulata. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part F, 444

477. Geological Society of America and University of Kansas Press, New York and Lawrence, Kansas.
KAZMiERCZAK, J. 1984. Favositid tabulates: evidence for poriferan affinity. Science, 225, 835-837.

KIM, A. I. 1974. O filogenii i polozhenii v sisteme nekotorykh tabulyatomorfnykh korallov. In sokolov, b. s.

(ed.). Drevnie Cnidaria, 1, 118 122. Nauka, Moscow. [In Russian.]

KIRKPATRICK, R. 1912. On the nature of stromatoporoids. Natio-e, Fond. 89, 607.

LAFUSTE, J. 1962. Note preliminaire sur la microstructure de la muraille chez Favosites Lamarck (Coelenterata,
Tabulata). C. somni. Seanc. Soc. geol. Fr. 1962, 105 106.

MORI, K. 1976. A new recent sclerosponge from Ngargol, Palau Islands and its fossil relatives. Sci. Rep. Tohoku
Univ., 2nd ser. (Geol.), 46, 1 9.

1977. A calcitic sclerosponge from the Ishigaki-shima coast, Ryukyu Islands, Japan. Ibid. 47, I 5.

oekentorp, k. 1969. Kommensalismus bei Favositiden. Miinst. Forseb. Geol. Paldont. 12, 165-216.

1976. Beschreibung und Systematik devonischer Favositidae asturiens und Betrachtungen zur Biogeogra-
phie nord-spanischer Korallenfaunen. Ibid. 37, 1 129.

1985. Spicules in favositid Tabulata— remarks to J. Kazmierezak’s interpretation. Fossil Cnidaria, 14 (1),

34-35.

and SORAUF, j. e. 1970. Uber Wandporen bei Favosites (Fav.) gothlandiciis Lamarck, 1816 (Coelenterata,

Tabulata). Neues Jb. Geol. Paldont. Abb. 134, 283 298.

and STEL, J. H. 1985. Favosites—a. true coral. Remarks to P. Copper’s discoveries of fossilized polyps.

Fossil Cnidaria, 14 (2), 28-29.

OLIVER, w. A., JR. 1968. Some aspects of colony development in corals. J. Paleont. 42 (5) (Paleont. Soc. Mem.

2) , 16-34.

1979. Review: Sponges they are not. Paleobiology, 5, 188 -190.

— - 1986. Favositids are corals— further remarks. Fossil Cnidaria, 15(1 .2), 19 21 .

PHILIP, G. M. 1960. The Middle Palaeozoic squamulate favositids of Victoria. Palaeontology, 3, 186 207.
RICHTER, D. K. 1972. Autliigenic quartz preserving skeletal material. Sedimentology, 19, 21 1 218.

ROSS, M. H. 1953. The Favositidae of the Hamilton Group (Middle Devonian of New York). Bidl. Buffalo Soc.
nat. Sci. 21, 37-89.

SANDBERG, p. A. 1975. Bryozoan diagenesis: bearing on the nature of the original skeleton of rugose corals. ./.
Paleont. 49, 587-606.

SCHOUPPE, A. von and oekentorp, k. 1974. Morphogenese und Bau der Tabulata unter besonderer Beriicksich-
tigung der Favositida. Palaeontograpbica, A145, 79 194.

SCRUTTON, c. T. 1979. Early fossil cnidarians. In house, m. r. (ed.). Tbe origin of major invertebrate groups,
161-207. Academic Press, London and New York.

492

PALAEONTOLOGY, VOLUME 30

SCRUTTON, c. T. 1984. Origin and early evolution of tabulate corals. Palaeontogr. am. 54, 1 10-1 18.

SOKOLOV, B. s. 1962. Gruppa Chaetetida. In orlov, u. a. (ed.). Osnovy paleontologii, 2. Guhki, Arkheotsiaty,
Kishecimopolostnye, Chervy, 169-176. IzdateFstvo Akademii Nauk, Moscow. [In Russian.]

STEL, j. H. 1978. Studies on the palaeohiology of favositids, viii + 247 pp. Stabo-All Round B. V., Groningen.

and DE coo, j. c. m. 1977. The Silurian Upper Burgsvik and Lower Hamra-Sundre Beds, Gotland. Scripta

Geol. 44, 1-43.

and OEKENTORP, K. 1981. Desmidopora and Nodidipora: misfits in the coral world. Acta palaeont. pol. 25,

SWANN, D. H. 1947. The Favosites alpenensis lineage in the Middle Devonian Traverse Group of Michigan.
Contr. Mus. Paleont. Univ. Mich. 6, 235-318.

VACELET, J. 1977. Eponges Pharetronides actuelles et Sclerosponges de Polynesie fran^aise, de Madagascar et
de la Reunion. Bull. Mus. natn. Hist. nat. Paris (3^ ser.), 444, Zool. 307, 345-368.

1985. Coralline sponges and the evolution of the Porifera. In conway morris, s., george, j. d., gibson,

R. and PLATT, H. M. (eds.). The origin and relationships of lower invertebrates, 113. Clarendon Press, Oxford.
WEST, R. R. and clark, g. r., ii. 1984. Palaeohiology and biological affinities of Palaeozoic chaetetids.
Palaeontogr. am. 54, 337-348.

403-417.

Typescript received 28 April 1986
Revised typescript received 22 August 1986

COLIN T. SCRUTTON

Department of Geology
The University

Newcastle upon Tyne NEl 7RU

A STAURIKOSAURID DINOSAUR FROM
THE UPPER TRIASSIC I SCH IGU AL ASTO
FORMATION OF ARGENTINA AND
THE RELATIONSHIPS OF THE
STAURIKOSAURIDAE

by D. B. BRINKMAN Ciud H.-D. SUES

Abstract. A partial skeleton of a staurikosaurid, cf. Staurikosaums sp. is described from the Ischigualasto
Formation of Argentina. This specimen provides additional anatomical data for an evaluation of the phylo-
genetic affinities of the Staurikosauridae. Colbert’s (1970) assignment of Stawikosaurus to the Dinosauria is
supported. No synapomorphies exist to support close relationships between Herrerasaurus and Stanrikosaurus.
It is concluded that Stawikosaurus and Herrerasaurus form successive sister taxa to an assemblage composed
of Ornithischia and Saurischia, with the latter only including Sauropodomorpha and Theropoda.

The early diversification of the Dinosauria is documented virtually exclusively by skeletal remains
from a sequence of Late Triassic continental strata in South America (Bonaparte 1978). The oldest
known dinosaurs occur in the Santa Maria Formation of Rio Grande do Sul, Brazil, which is most
likely Carnian in age (Colbert 1970). Slaurikosaurus Colbert, 1970 is the best known of these forms
(Colbert 1970; Galton 1977) and is documented by lower jaws and a fairly complete postcranial
skeleton (MCZ 1669). Spondylosoma Huene, 1942 from the same formation is very poorly known
but has also been classified as dinosaurian (Huene 1942; Bonaparte 1978). The holotype material
consists of vertebrae, parts of both scapulae, the proximal end of a left humerus, the proximal end
of a right femur, and a partial right pubis and is currently being redescribed by Galton (pers. comm.).
The Ischigualasto Formation of north-western Argentina, which is probably slightly younger than
the Santa Maria Formation, has yielded material referable to three genera of dinosaurs (Bonaparte
1978). Of these, Herrerasaurus and Ischisaurus, first described by Reig (1963), are generally classified
as primitive saurischians, and Pisanosaurus Casamiquela, 1967 is referred to the Ornithischia. Reig
(1963) also described 'Triassolestes' {Trialestes Bonaparte, 1982) as a podokesaurid theropod but
Bonaparte (1978, 1982) has reinterpreted this taxon as a crocodylomorph archosaur. The material
referable to Herrerasaurus and Ischisaurus has only been described in a most preliminary fashion.
Pisanosaurus is based on an extremely poorly preserved specimen, and Bonaparte’s (1976) assign-
ment of this genus to the Heterodontosauridae (which are otherwise only definitely known from
the Lower Jurassic of southern Africa) is indeed questionable.

In this paper we describe a single fragmentary skeleton of a primitive dinosaur, which we interpret
as the first record of a staurikosaurid from the Ischigualasto Formation. This specimen provides
much additional information on the structure of these dinosaurs, and, based on these new data, the
phylogenetic position of Stawikosaurus and related forms will be reconsidered.

The occurrence of a staurikosaurid in the Ischigualasto Formation is documented by a single
specimen in the collections of the Museum of Comparative Zoology at Harvard University, MCZ
7064. It consists of a partial postcranial skeleton including the atlas-axis complex, parts of at least
five dorsal vertebrae, fragments of both scapulocoracoids, proximal and distal ends of both humeri,
a partial left ilium, the proximal ends of both ischia, the distal end of a right femur, the proximal
and distal ends of a right tibia, the proximal end of a right fibula, and some pedal phalanges. The

(Palaeontology, Vol. 30, Part 3, 1987, pp. 493-503.|

© The Palaeontological Association

494

PALAEONTOLOGY, VOLUME 30

E

TEXT-FIG. 1. Cf. Staurikosaiirus sp., MCZ 7064. a, b, atlas centrum, atlas intercentrum, and axis, in a, right
lateral and b, dorsal view, c-e, centrum of a posterior dorsal vertebra, in c, anterior, d, lateral, and e, dorsal
view. Abbreviations: f.ic.l— facet for atlas intercentrum, n.s — neural spine of axis, od — odontoid process.

material was collected by A. S. Romer in 1958 from a site 1 km north-west of Arroyo de Agua, San
Juan province, Argentina. In the field-notes, the specimen (field-number 295-58M) is recorded as
much of an indeterminate skeleton. Some of the preserved pieces still show evidence of articular
context: the tibia and fibula are preserved in articulation and the left humerus is in contact with the
left scapulocoracoid. Thus the material probably represents the remains of a single skeleton. A
nearly complete skull of a large archosaur, MCZ 7063, was possibly originally part of the same
specimen but regrettably all direct information bearing on this appears to have been lost. The skull
is definitely not referable to the rauisuchid Saurosiichus from the same formation (Bonaparte 1978),
but determination of its affinities must await further preparation. Many of the bones were covered
by hematite, the removal of which is extremely laborious.

DESCRIPTION
Subdivision archosauria

DINOSAURIA

Family staurikosauridae
cf. Staurikosaurus sp.

Postcrcmial axial skeleton. The postcranial axial skeleton is documented by the atlas-axis complex
and parts of at least five dorsal vertebrae. The atlas centrum, axis intercentrum, and axis are
preserved in association (text-fig. 1) but it is uncertain whether they are fused or whether their
contacts are merely obscured by adhering hematite.

BRINKMAN AND SUES: STAURIKOSAURID DINOSAUR

495

The atlas centrum bears a prominent odontoid proeess (od, text-fig. 1a, b). Its height is about
half that of the axis centrum. Below the atlas centrum, the anterior face of the axis intercentrum
forms a crescent-shaped surface for the atlas intercentrum (f.ic.l). The axis is elongate, the length
of its centrum being about twice its height. The neural arch bears a large posteriorly directed spine
(n.s). The neural spine is bifurcated at its apex, with each branch extending to the posterior extremity
of a postzygapophysis. Transverse processes are absent. The sides of the axis centrum are pinched
in at a point just about mid-height; the ventral edge of the centrum is rounded. The posterior
articular face of the centrum is concave.

One dorsal centrum is nearly complete (text-fig. 1c-e). It is short, slightly more than half as long
as it is high (text. -fig. Id), and probably represents a posterior dorsal. Its articular ends are nearly
flat, although the posterior face shows a slight central depression. The floor of the neural canal on
broken centra narrows toward the eentre where it sinks deeply into the centrum (text-fig. 1e), rather
than extending continuously on level with the pedicles of the neural arch. A similar inward extension

TEXT-FIG. 2. Cf. Stawikosawus sp., MCZ 7064.
Conjoined right scapula and coracoid in lateral
view. Abbreviation: gl.c — glenoid cavity.

V

496

PALAEONTOLOGY, VOLUME 30

of the floor has elsewhere only been reported in dorsal vertebrae of the theropod Dilophosaurus
(Welles 1984); the distribution of this peculiar feature among other archosaurian groups remains to
be determined.

Appendicular skeleton. The scapula and coracoid form a single element but the line of fusion between
the two bones remains apparent. The scapulocoracoids are documented by the left scapula and

TEXT-FIG. 3. Cf. Stawikosaums sp., MCZ 7064. a, b, proximal end of right humerus in a,
medial and b, anterior views; c, d, distal end of left humerus in c, medial and d, anterior
views. Abbreviations: dp.c— delto-pectoral crest, h— proximal articular head.

BRINKMAN AND SUES: STAURIKOSAURID DINOSAUR

497

anterior half of the conjoined coracoid and a nearly complete right coracoid, and the base of the
attached scapula.

The scapulocoracoid is characterized by a very slender scapular blade and a large, plate-like
coracoid (text-fig. 2). The basal portion of the scapulocoracoid is rectangular in outline and bears
the glenoid posteriorly. The scapular blade is narrow anteroposteriorly and is oval in transverse
section. Although its dorsal margin is not preserved there are no indications for an expansion of
the dorsal (vertebral) end.

The humerus (text-fig. 3) is represented by the proximal and distal ends of both elements. Its
overall length cannot be determined. The articular ends are two-and-a-half times as wide as the
humeral shaft. The humeral head (h, text-fig. 3a) is hemispherical and is restricted to the centre of
the proximal end of the bone. The deltopectoral crest (dp.c) is prominent and arises slightly distal
to the proximal articular end. It extends more or less perpendicular to the long axis of the proximal
portion of the humerus (text-fig. 3b). The humeral shaft is circular in transverse section distal to
the deltopectoral crest. The distal articular end of the humerus bears distinct radial and ulnar
condyles. The radial condyle faces distinctly ventrally. The ulnar condyle is situated distal to it and
faces primarily distally. The ectepicondyle has a notch of uncertain significance in its lateral margin.

The partial left ilium (text-fig. 4c) includes the supra-acetabular rim and the pubic ramus. The

TEXT-FIG. 4. Cf. Stawikosaunis sp., MCZ 7064. a, b, partial right ischium in A, lateral and b, proximal end
view, c, partial left ilium in lateral view. Abbreviations: ac— acetabular wall, f.pu— articular facet for pubis.

498

PALAEONTOLOGY, VOLUME 30

supra-acetabular rim is prominent. It is concave ventrally, forming a socket for the reception of the
proximal head of the femur. The lateral aspect of the iliac blade lacks the strong vertical buttress
extending dorsally from the supraacetabular rim present in certain rauisuchians (Chatterjee 1985).
The well-developed medial wall of the acetabulum (ac) extends far ventrally and forms a knife-like
edge at about the level of the articular contact for the pubis (f.pu). While the acetabulum probably
was perforated, the opening was quite small. The pubic peduncle of the ilium is robust and large.

The proximal ends of both ischia are preserved. The ischium has a broad, triangular proximal

TEXT-FIG. 5. Cf. Staurikosaurus sp., MCZ 7064. Distal end of right femur in a, posterior, B, anterior, and c,
distal end view. Abbreviations: l.c— lateral condyle, m.c— medial condyle.

BRINKMAN AND SUES: ST AU R I KOS AU R I D DINOSAUR

499

TEXT-FIG. 6. Cf. Staurikosciunis sp., MCZ 7064. Partial right tibia and fibula. A, B, articulated proximal portions
of tibia and fibula in a, anterior and b, proximal end view, c, d, distal end of tibia in c, anteromedial and d,

distal end view.

portion and a rod-like distal region (text-fig. 4a, b). The articular surfaces for the ilium and pubis
are separated only by a slight narrowing of the bone, with that for the former being larger and
situated at a low angle relative to the latter.

The femur is documented by the distal end of the right bone (text-fig. 5). The diameter of the
femoral shaft increases toward the distal articular end. At the proximal end of the preserved
fragment, the diameter of the shaft is slightly more than half the width across the distal end. This
increase in width is developed asymmetrically so that the medial portion of the distal end is more
prominent than the lateral one (text-fig. 5a, b). The distal end of the femur is rounded in outline. It
bears a central depression, which is surrounded in front and on either side by a low, broad ridge
(text-fig. 5c). A smaller depression is developed anterior to the central one. The distal condyles (l.c,
m.c, text-fig. 5a) occupy subterminal positions on the posterior (ventral) aspect of the femur.

Both the proximal and distal ends of the right tibia are preserved. It has an expanded proximal
portion that rapidly becomes narrow distally (text-fig. 6a). A distinct cnemial crest, laterally bor-
dered by a groove, is developed. The proximal portion of the right fibula (text-fig. 6a, b) is preserved
in its original articular context. It has a broad semilunate proximal end, which fits tightly against
the lateral aspect of the tibia. Together the two bones form a rounded articular surface. The distal
end of the tibia (text-fig. 6c, d) is rounded in articular view and is much smaller in diameter than
the proximal end. Its articular surface has a helical shape with the two ends of the spiral joined by

500

PALAEONTOLOGY, VOLUME 30

a flat surface. The difference in the relative position of the two ends of the spiral produces a notch
for the reception of the ascending process of the astragalus.

No tarsal bones are preserved. The pedal digits are only documented by a few non-diagnostic
articular ends of phalanges.

TAXONOMIC AFFINITIES OF MCZ 7064

MCZ 7064 is identified as a staurikosaurid, rather than as a herrerasaurid, based on the structure
of the ischium and of the distal end of the tibia. The triangular proximal end of the ischium matches
that of Staurikosaurus pricei (MCZ 1669) closely and differs from that of Herrerasaurus, which
shows a distinct angulation between the posterior margin of the ischiadic shaft and the posterior
edge of the acetabular portion (Reig 1963, fig. 2). The outline of the distal end of the tibia is again
closely similar to that of S. pricei (Galton 1977, fig. 2m) and different from that in Herrerasaurus
where the distal end is more expanded (Reig 1963, fig. 3b, c). The distal ends of both tibiae in the
holotype of S. pricei have a deep groove extending proximally from the notch formed by the helical
articular surface but this feature has been exaggerated by overpreparation; the distal end of the
right tibia in MCZ 7064 has no comparable distal groove.

Direct comparison of MCZ 7064 with the type specimens of Spondylosoma ahscomUtum and /.
cattoi was not possible. The former differs from Staurikosaurus pricei in the development of the
pubic apron (Galton, pers. comm.) as well as in the structure of the vertebrae (Colbert 1970).
Staurikosaurus differs from I. cattoi, currently being restudied by F. Novas, in the presence of a
low lesser trochanter on the proximolateral aspect of the femur (Galton 1977). In both Ischisaurus
and Herrerasaurus the lesser trochanter forms a small but very prominent ridge in a more distal
position on the proximolateral aspect of the femur (Novas, pers. comm.).

Considering its slightly later occurrence in time and the possible structural differences to the
holotype of S. pricei, specimen MCZ 7064 is tentatively identified as cf. Staurikosaurus sp.

DISCUSSION

The new data on skeletal structure provided by MCZ 7064 invite examination of the phylogenetic
position of the Staurikosauridae relative to dinosaurs and other archosaurs. Outgroups used in
determining the polarity of character states displayed by Staurikosaurus were Lagosuchidae (Bon-
aparte 1975r/), Ornithosuchidae (Walker 1964; Bonaparte 19756), and Rauisuchia ( = Rauisuchidae
-f Poposauridae; Chatterjee 1985). These extensive comparisons with non-dinosaurian archosaurs
were undertaken because of the current debate about which group of archosaurs is most closely
related to the Dinosauria. We regard Dinosauria as a monophyletic assemblage, following the
recent discussions by Benton (1984) and Gauthier and Padian (1985), rather than as an artificial
grouping comprising two distinct orders Ornithischia and Saurischia, which supposedly have inde-
pendent, possibly multiple origins among Thecodontian’ archosaurs (Charig 1982).

The first problem to be considered is the placement of Staurikosaurus in the taxon Dinosauria.
Colbert (1970) listed seven characters in support of his assignment of this genus to the Dinosauria:

( 1 ) transverse processes supported by a pair of strong ventral buttresses;

(2) acetabulum perforated;

(3) ischium rod-like;

(4) ilium as deep as long and truncated posteriorly;

(5) femur shorter than tibia;

(6) proximal head of femur set off from shaft;

(7) fourth trochanter on femur strongly developed.

Since the publication of Colbert’s original description of Staurikosaurus much new material of early
Mesozoic non-dinosaurian archosaurs has been described, including forms that share some of the
character-states enumerated by Colbert. Characters ( 1 ) and (3) are also developed in the poposaurid
rauisuchian Postosuchus (Chatterjee 1985). Lagosuchus shares characters (5) and (7) with dinosaurs

BRINKMAN AND SUES: STAU RI KOS A U RI D DINOSAUR

501

(Bonaparte 1975o). Characters (5) and (6) are present in pterosaurs (Wellnhofer 1978; Padian 1983).
The size of the acetabular opening is uncertain in MCZ 7064, and the opening (2) may not have
been much larger than that in certain Ornithosuchidae (Walker 1964; Bonaparte 19756) and in
Postosuchus (Chatterjee 1985). Character (4) is possibly autapomorphous for Staurikosciurus, par-
ticularly the marked posterior truncation of the iliac blade. None of the above characters are unique
to dinosaurs.

In the asymmetrical development of the distal articular end of the femur, Staurikosaurus resembles
both other dinosaurs and Lagosiicinis. In most more primitive archosaurs the distal end is also
asymmetrical but the lateral portion is more prominent. This structural difference can be related to
a difference in the function of the femur; the femur extends laterally during femoral retraction in
primitive archosaurs (Brinkman 1980).

Gauthier and Padian (1985, p. 189) hypothesized the following set of synapomorphies in a
common ancestor of Dinosauria (= Sauropodomorpha + Theropoda and Ornithischia):

(1) Manus with phalangeal formula 2-3-4-3-2 (reduction in outer digits);

(2) semiperforate acetabulum;

(3) prominent supraacetabular buttress {supraacetabular rim in our usage);

(4) fossa on ventral margin of postacetabular portion of ilium (for origin of M. caudifemoralis
brevis);

(5) prominent anterior (or lesser) trochanter of femur;

(6) prominent cnemial crest on tibia, projecting beyond femoral condyles and curving anterolater-
ally;

(7) tibia in which proximal end is expanded anteroposteriorly and in which distal end is broadened
transversely (‘twisted’ tibia), with notch in distal end for reception of ascending process.

In addition, Gauthier and Padian note the existence of ‘several other synapomorphies’ but these
were not specified in their paper and cannot be critically evaluated. Of these features, Staurikosaurus
shares (2) to (4) and (6). The presence of character (I) cannot be determined. Character (5) is
found in Herrerasaurus but the feature is apparently developed in a rather different fashion in
Staurikosaurus. Staurikosaurus lacks tibial twisting, representing the plesiomorphous condition, but
has a notch in the distal end of the tibia for the reception of the ascending process of the astragalus.
Thus character (7) actually consists of two independently acquired features and should be modified
accordingly.

Benton (1984, pp. 13-14) has presented a list of possible synapomorphies for a monophyletic
assemblage Dinosauria, which, in addition to characters (5) and (7) listed by Gauthier and Padian,
includes the following features:

(1) Absence of postfrontal;

(2) deltopectoral crest extending far down along humeral shaft;

(3) forelimb about half as long as hindlimb;

(4) reduced contact between pubis and ischium;

(5) fourth trochanter prominent and positioned low on femur;

(6) proximal head of femur set off from shaft;

(7) reduced, roller-like astragalus with ascending process;

(8) calcaneum reduced or absent;

(9) advanced mesotarsal ankle joint;

(10) pedal digits II to IV bundled, elongate, and subequal in length;

(11) pedal digits I and V reduced and divergent;

(12) foot with digitigrade pose.

Characters (8), (10), and (1 1) of this list are also found in Lagosuchus (Bonaparte 1975a); metatarsal
II in Lagosuchus is but slightly shorter than are metatarsals III and IV. characters (1), (3), (6), part
of (7), (8), (9), part of (10) and (12) are shared by pterosaurs (Wellnhofer 1978; Padian 1983).
Characters (1), (3), and (7) to (12) cannot be determined in the presently available staurikosaurid
material but characters (2), (4), (5), and (6) are developed in Staurikosaurus.

502

PALAEONTOLOGY, VOLUME 30

We support Colbert’s assignment of Staurikosaunis to the Dinosauria. Like Herrerasaurus,
Staurikosaurus occupies a basal position within this group as defined by recent authors and is more
primitive than other dinosaurs in the outline of the distal end of the tibia. Herrerasaunis also has a
semiperforate acetabulum with a strongly developed medial wall but is more derived in the transverse
expansion of the distal end of the tibia (Reig 1963; Benedetto 1973). The markedly anteroposteriorly
expanded distal end of the pubis is probably an autapomorphy for this genus. This feature is also
developed, to a lesser extent, in S. pricek in the podokesaurid theropod Coelophysis the pubis
terminates distally in a knob-like thickening (Colbert 1970).

Staurikosaurus and Herrerasaurus were placed in a single family Herrerasauridae by Benedetto
(1973) who noted numerous similarities between them. Galton (1977), emphasizing certain differ-
ences between these two genera, proposed a separate family Staurikosauridae for the reception of
Staurikosaurus but left the question of their interrelationships unresolved. With the possible excep-
tion of the anteroposterior expansion of the distal end of the pubis, we find no synapomorphies in
support of a sister-group relationship between Herrerasaurus and Staurikosaurus and regard the
similarities between the two taxa as plesiomorphous. Only one feature of Staurikosaurus, the
presence of a narrow scapular blade, is an apparent autapomorphy for this genus. Judging from
outgroup comparisons (particularly with Lagosuchus; Bonaparte 1975u) and the condition in most
early dinosaurs including prosauropods and Coelophysis, a wide scapular blade is primitive for
Dinosauria. According to Reig’s (1963, pp. 6-8) list of skeletal material for Herrerasaurus, the
scapula in this form is unknown. We hypothesize Staurikosaurus and Herrerasaurus as successive

of

TEXT-FIG. 7. Hypothesis of interrelationships for Staurikosaurus, Herrerasaurus, and other Dino-
sauria (Saurischia only including Theropoda and Sauropodomorpha). Lagosuchus is included in
the cladogram as an outgroup, following Bonaparte (1975n) and Gauthier and Padian (1985).
Selected synapomorphies are: (1) neck sigmoidally curved. Three-regionalized’ vertebral column,
femur with moderately developed lateral condyle, astragalus with ascending process, mesotarsal
joint (Gauthier and Padian 1985); (2) semiperforate acetabulum with prominent supraacetabular
rim, distinct lesser (anterior) trochanter on femur, distal end of tibia with fossa for reception of
ascending process of astragalus (see text); (3) distal end of tibia transversely expanded (Twisted’
tibia); (4) medial wall to acetabulum less well developed, pedal digit V small.

BRINKMAN AND SUES; STAURIKOSAU R1 D DINOSAUR

503

sister taxa to a clade Saurischia + Ornithischia as defined by Gauthier and Padian (1985) (text-fig.
7). Staiirikosaurus is clearly the most primitive known representative of the Dinosauria. Galton
(1977) classified both Herrerasaunis and Staiirikosaurus as Saurischia iucertae seeks. We accept
Gauthier and Padian's more restrictive use of the term ‘Saurischia’ to include only Theropoda and
Sauropodomorpha (which, in fact, agrees with the traditional usage of that name) and suggest
placement of the two primitive South American dinosaurs in separate and distinct higher taxa to
reflect their respective phylogenetic positions.

Acknowledgements. We thank P. M. Galton, University of Bridgeport, and F. Novas, Museo Argentino de
Ciencias Naturales, for sharing unpublished information with us. K. Padian, University of California, and
Greg Paul provided an insightful review of the manuscript. D. Sloan, Tyrrell Museum of Palaeontology,
prepared text-figs. 2 6.

REFERENCES

BENEDETTO, J. L. 1973. Herrerasauridae, nueva familia de saurisquios Triasicos. Ameginniana, 10, 89-102.

BENTON, M. J. 1984. Eossil reptiles from the German Late Triassic and the origin of dinosaurs. In reif, w.-e.
and westphal, f. (eds.). Third Symposium on Mesozoic Terrestrial Ecosystems. Short Papers, 13 18. At-
tempto Verlag, Tubingen.

BONAPARTE, J. F. 1975fl. Nuevos materiales de Lagosuchus talampayensis Romer (Thecodontia— Pseudosuchia)
y su significado en el origen de los Saurischia. Chanares inferior, Triasico medio de Argentina. Acta geol.
lilloana, 13, 5-90.

1975/). The family Ornithosuchidae (Archosauria: Thecodontia). Ccdl. Internatn. C.N.R.S. 218, 485-502.

1976. Pisanosauriis mertii Casamiquela and the origin of the Ornithischia. J. Paleont. 50, 808 820.

- 1978. El Mesozoico de America del Sur y sus tetrapodos. Op. lilloana, 26, 5-596.

1982. Faunal replacement in the Triassic of South America. J. vert. Paleont. 2, 362-371.

BRINKMAN, D. 1980. The hind limb step cycle of Caiman sclerops and the mechanics of the crocodile tarsus
and metatarsus. Can. J. Zool. 58, 2187 2200.

CHARiG, A. J. 1982. Problems in dinosaur phylogeny: a reasoned approach to their attempted resolution.
GOihios, Mhn. Spec. 6, 113- 126.

CHATTERJEE, s. 1985. Postosuclius, d new thecodontian reptile from the Triassic of Texas and the origin of
tyrannosaurs. Phil. Trans. R. Soc. B309, 395-460.

COLBERT, E. H. 1970. A sauriscliian dinosaur from the Triassic of Brazil. Am. Mus. Novit. 2405, I 39.

GALTON, p. M. 1977. On Stawiko.saunis pricei, an early saurischian dinosaur from the Triassic of Brazil, with
notes on the Herrerasauridae and Poposauridae. Paldont. Z. 51, 234-245.

GAUTHIER, J. and PADIAN, K. 1985. Phylogenetic, functional, and aerodynamic analyses of the origin of birds
and their flight. In hecht, m. k., ostrom, j. h., viohl, g. and wellnhofer, p. (eds.). The beginnings of birds,
185 197. Freunde des Jura-Museums Eichstatt, Eichstatt.

huene, f. V. 1935-1942. Die fossilen Reptilien des sikiamerikanischen Gondwanalandes. Ergebnisse der Saurier-
grabungen in Siidbrasilien I928I29, viii -h 332 pp. C. H. Beck, Munich. (Archosauria are described in
Lieferung 3, which was published in 1942.)

PADIAN, K. 1983. Description and reconstruction of new material of Dimorphodon macronyx (Buckland)
(Pterosauria, Rhamphorhynchoidea) in the Yale Peabody Museum. Postilla, 189, I 44.

REIG, o. A. 1963. La presencia de dinosaurios saurisquios en los ‘Estratos de Ischigualasto’ (Mesotriasico
Superior) de las provincias de San Juan y La Rioja (Republica Argentina). Ameginniana, 3, 3-20.

WALKER, A. D. 1964. Triassic reptiles from the Elgin area; Ornithosuchus and the origin of carnosaurs. Phil.
Trans. R. Soc. B248, 53- 1 34.

WELLES, s. p. 1984. Dilophosaurus wetherilli (Dinosauria, Theropoda). Osteology and comparisons. Palaeonto-
graphica, A185, 85 180.

WELLNHOFER, p. 1978. Pteiosauria. In p. wellnhofer (ed.). Handbiich der Paldoherpetologie, Teil 19, x-t-82
pp. Gustav Eischer Verlag, Stuttgart.

DONALD B. BRINKMAN

Tyrrell Museum of Palaeontology, Box 7500
Drumheller, Alberta TOJ OYO, Canada

HANS-DIETER SUES

Typescript received 10 June 1986 Museum of Comparative Zoology, Harvard University

Revised typescript received 13 October 1986 Cambridge, MA 02138, USA

iif’t

k’;:' .J.'’/i|^/;

•.. ■ 'V, .■.'■VJI^tm'tr'' M': ■ I ' ' ’'^

•• •■"■i’fd ■! "• ‘■■'i'¥®*'^i

«> • ^ ■•■ f

2 ^

A

v; » AV '-lfi*'"'* •
.. V <>;■■ '■ ■^,i' .

t

.■('■ t; Lfj

.u; , j" ;

M i’ •*{iii'.

'.. li .i.'.-. ' '■ " ' ‘■■' i

•'" '■ :

.;, V-

. )■, rv "

V

t,.* ' ■ ".'

»'■ .. .ft* in'-w -■•(■■■« *®„'^ ■■

harfi’^'i i‘<' * j)

A CARIBBEAN RUDIST BIVALVE IN OMAN:
ISLAND-HOPPING ACROSS THE PACIFIC IN
THE LATE CRETACEOUS

by p. w. SKELTON Cindy . p. wright

Abstract. The hippuritid rudist bivalve Torreites is described from the Maastrichtian of Oman and the
United Arab Emirates. Together with the single fragment from the region on which T. milovanovici Grubic,
1979 was based, the specimens are placed in a newly recognized geographical subspecies of T. sanchezi
(Douville, 1927), a species previously considered endemic to the Caribbean Province. The Arabian T. s.
milovanovici Grubic differs from its Caribbean parent, T. s. sanchezi (Douville) (incorporating all previously
recognized T. sanchezi, as well as T. coxi Grubic, 1979) only in the angle (a) between the arete cardinale and
the ventralmost pillar: a is 12°-75° in the former and 50°-126° in the latter.

The loss of the normal hippuritid pore and canal system in Torreites is conhrmed and shown to have been
associated with exposure of the mantle margins, which may have contained symbiotic algae, as in the living
Tridacna.

Homeomorphy and plate tectonic drifting are rejected as explanations for the apparent disjunct endemism
of Torreites. Rather, larval dispersal along a corridor of shallow staging posts is favoured. A Mediterranean
Tethys/Atlantic route is considered unlikely, because of barriers. There is good evidence, in contrast, for such
staging posts across the Pacihc and eastern Tethys in Campanian Maastrichtian times.

Although the distinctive Late Cretaceous rudist bivalve Torreites has generally been considered
endemic to the Caribbean Province (Kauffman 1973), a single worn fragment of a right valve from
Sheikdom Sharjah, in the United Arab Emirates (UAE), was assigned to the genus by Grubic, in
1979. Such a strikingly disjunct distribution for a sessile benthic inhabitant of shallow equatorial
seas demands a palaeobiogeographical explanation. Yet this and other purported cases of apparent
disjunct endemism between the eastern Tethyan and Caribbean regions (e.g. Chubb 1956; Elliott
1981) have attracted surprisingly little comment in the literature. Possibly there has been a tacit
(though uninformed) assumption that such distantly separated forms are more likely to have been
mere homeomorphs than true congeners, or that the currently known localities are but preser-
vational relics of an originally cosmopolitan range.

During geological reconnaissance work in Oman (by V.P.W. in 1983) and in the UAE (by P.W.S.
in 1984), we found several further, well-preserved specimens of Torreites. The purpose of this
paper is to establish recognition of the genus in the region beyond any doubt, on the basis of a
suite of characters, the replication of which by homeomorphy would have been exceedingly
unlikely. The quality of our material also prompts a systematic re-evaluation of the new species
of the genus erected for the Middle Eastern form by Grubic, and allows the reconstruction of
some previously unrecognized aspects of the original soft part anatomy and life habits of the
animal.

Grubic (1979) inferred that the disjunct distribution of the genus implied closer proximity of
Oman to the Caribbean in Late Cretaceous times. However, recent Deep Sea Drilling Project work
in the Pacific (Schlanger et al. 1981) has shown that, in the Late Cretaceous, shallow-water benthos
with Caribbean affinities extended far across the Pacific, exploiting shallow-water ‘stepping stones’
formed by numerous volcanic seamounts and islands. We argue here that the disjunct distribution
of Torreites was more probably the product of larval dispersal via such staging posts than of the
drifting apart of Oman and the Caribbean implied by the hypothesis of Grubic.

{Palaeontology, Vol. 30, Part 3, 1987, pp. 505-529, pis. 61-62.| © The Palaeontological Association

506

PALAEONTOLOGY, VOLUME 30

GEOLOGICAL SETTING

The main suite of specimens described here was collected from the foot of a small escarpment of
Cretaceous strata east of the main Tertiary escarpment, some 30 km south-south-west of the Saiwan
airfield, in eastern central Oman, at latitude 20° 39' N. and longitude 57° 31' E. (text-fig. I). The
area is remote, with no previous published account of its Mesozoic geology, although a generalized
geological map of the region is given by Gorin et al. (1982). The scarcity of landmarks and
settlements makes it impossible to provide more precise locality details. The rudists were loose
specimens from the talus at the foot of the escarpment. They are associated with a yellow-weathering
bioclastic calcarenite. The Upper Cretaceous of the area comprises a basal clastic member overlain
by a limestone member, which yields the rudists. Both members fall within the Aruma Group of
Glennie (1977), of Santonian-Maastrichtian age. It is unclear as to whether the rudist-bearing
limestone is high in the Fiqa Formation (a predominantly marly formation of Santonian-Maastrich-
tian age) or if it corresponds to the Simsima Formation (Late Maastrichtian, according to Harris
et ai 1984, though work in progress on this formation in the UAE by P.W.S. with S. C. Nolan,
indicates, rather, an early to medial Maastrichtian age). The age of the Torreites-htdLx'xng units,
however, must be considered tentative because of the lack of earlier work in the area. One further
specimen was collected in Oman, at Filim in northern Masirah Bay (lat. 20° 36' N. and long. 58°
12' E.) (text-fig. 1 ).

The specimens from the United Arab Emirates were found at two localities (text-fig. 1). A small
hillock, Qarn Murrah, projecting through the desert sands some 8 km west of Jebel Faiyah, in
eastern Sharjah (lat. 25° 07' N. and long. 55° 46' E.), has several specimens in life position, in a
sequence of reddish bioclastic packstones to grainstones. To the south the basal chert conglomerate
exposed around the core of an anticline at Jebel Huwayyah, some 10 km north-east of Buraimi/Al
Ayn (lat. 24° 16' N. and long. 55° 48' E.), yielded a single worn right valve. In both cases, a

TEXT-FIG. 1. Localities in Oman and the United
Arab Emirates where the specimens of Torreites
described in this paper have been collected and/or
observed.

SKELTON AND WRIGHT: ISLAND-HOPPING RUDIST

507

Maastrichtian age is indicated, by the accompanying large foraminifera, Orhitoides media (d’Arch-
iac) and Omphalocyclus macroporus Lamarck, as well as by the hippuritid rudist Pironaea praesla-
vonica Milovanovic, Sladic and Grubic, at Qarn Murrah, and by the large foraminifer, Lofiusia sp.
in conformably overlying beds at Jebel Huwayyah. Work in progress by P.W.S., with S. C. Nolan,
suggests that both occurrences may indeed be confined to the Lower Maastrichtian.

Grubic (1979, p. 85) stated that his specimen came from ‘Guru Mileih, Sheikdom Sharjah’, and
he assigned it a Maastrichtian age from its association with Orhitoides media and Omphalocyclus
macroporus. The lithology of the matrix on this specimen is very similar to that observed at Qarn
Murrah, however, and the specimen may indeed come from there (enquiries about the names of
small hills in deserts often provokes confusing responses).

SYSTEMATIC PALAEONTOLOGY

Superfamily HIPPURITACEA Gray, 1848
Family HIPPURITIDAE Gray, 1848
Genus TORREITES Palmer, 1933

Type Species. Hippurites (Vacciuites) Sanchezi Douville, 1927.

Emended diagnosis. Medium to large-sized hippuritid, with a conical right valve (RV), and an
operculiform left valve (LV) with a ventrally biased apex. The outer shell layer of the LV is externally
smooth and devoid of pores and canals. It is also much thinner than that of the RV, the internal
margin of which is thus exposed. Apically blind radial canals penetrate the inner shell layer of the
LV. The RV exterior is radially ribbed, with three solid radial infoldings (text-fig. 2) comprising a
dorsal ''arete cardinale' (Po) and two posterior pillars (Pi and Pi). The former has a rounded inner
tip and extends much further inwards than either of the subequal pillars. The three infoldings tend
to be uniformly thick, though the arOe cardinale may taper inwards. Their crests project through
sinuses indented from the LV margin. They are positioned around an angle of arc of between 12
and 126°. The two teeth and posterior myophore of the LV are strongly projecting, and are arrayed
at between 25 and 50° across the inner end of the arete cardinale.

Remarks. The original description of the type species (Douville 1927) drew attention to the unusually
elongate arete cardinale and to the even thicknesses of the pillars. Palmer’s (1933) original diagnosis
for the genus noted most of the other key features: the imperforate outer layer of the LV; the
radiating canals in the inner shell layer of the LV; and the sinuses in the LV for the RV infoldings.
Rutten (1936) demonstrated that the canals in the inner shell layer of the LV open directly over the
body cavity (in contrast to those in the outer shell layer of the LV in other hippuritids: see Skelton,
1976). He, MacGillavry (1937), Jung (1970), and Van Dommelen (1971 ) provided many quantitative
data, including a higher range of values (75-120°) for the angle of arc formed by the RV infoldings
in T. sanchezi than that observed (45-70°) in a smaller species, T. tschoppi MacGillavry. Grubic
(1979) erected three new species, two based on previously described Caribbean specimens (T. coxi,
based on specimen G. 14066 of Jung, 1970, and T. chuhbi, based on that described by Chubb, 1971 )
and the third, T. milovanovici, on the single RV fragment from Sharjah. This latter was distinguished
by the very low angle of arc between Po and P2 — only ‘about 10°’. Our specimens from Oman and
the UAE share the latter feature, but are otherwise so similar to T. sanchezi that we would judge
them merely to represent a subspecific variant of that species (see the synonymy below).

Stratigraphical ranges. In the Caribbean region the genus Torreites ranges from the Sanlonian or lowest
Campanian to the Lower Maastrichtian. T. sanchezi is characteristically found in the "Barrettia Beds’, in
Cuba, Puerto Rico, and Jamaica, with either or both of Barrettia monilifera Woodward and B. gigas Chubb
(Van Dommelen, 1971, p. 34; though N. F. Sohl, pers. comm, of August 1983, refers more specifically to a
consistent co-occurrence with the latter species). MacGillavry and his co-workers (summarized in MacGillavry
1937) favoured a Maastrichtian age for the Barrettia Beds, based largely on orbitoid foraminifera in those of
the Cuban Habana Formation. In contrast, Hawkins (1924), working on Jamaican echinoids, and Muellerried

508

PALAEONTOLOGY, VOLUME 30

(1936), extrapolating from ammonite-bearing sequences in Chiapas, Mexico, postulated a Turonian age.
Subsequent work has refuted the conclusions of Hawkins and Muellerried (N. F. Sohl, pers. comm.), and
largely vindicated those of MacGillavry and his co-workers. B. monilifera is now judged to be of Late
Campanian, and B. gigas, of latest Campanian/earliest Maastrichtian age (Sohl and Kollmann 1985, fig. 19).
Dc la Torre et al. (1978) referred the Barrettia fauna of Cuba to the upper part of the Lower Maastrichtian.
So the inferred range of T. sanchezi in the Caribbean is best bracketed within the Upper Campanian to Lower
Maastrichtian interval.

The specimen described by Chubb (1971) as "T. cf. sanchezi' (= T. chubbi Grubic, 1979), however, comes
from the Peter’s Hill Limestone of Jamaica, now considered of latest Santonian to earliest Campanian age
(N. F. Sohl, pers. comm.).

T. tschoppi was considered ‘probably Upper Campanian’ by MacGillavry (1937, p. 20), though Van Dom-
melen (1971, text-fig. 18) extended its range into the Santonian, albeit with uncertainty. It is certainly older
than T. sanchezi (Van Dommelen, 1971; N. F. Sohl, pers. comm.).

The specimens from the UAE are all considered Maastrichtian, while those from Oman can probably be
assigned to that stage, too, though with less certainty (see. p. 506).

Characters of systematic value. Text-fig. 2 shows the characters measured and the abbreviations used for them
in the description of specimens that follows.

The five main characters of systematic value are:

1. the overall size of the shell (L, Di and D2);

2. the angle of arc between the arete cardinale and the ventralmost pillar (a);

3. the lengths (I) and widths (w) of the arete cardinale (Pq) and pillars (Pi and P2);

4. the angle made by the myocardinal array with the inner tip of the arete cardinale (jiy,

5. the character of the RV outer shell layer (its thickness, rib-width, and depth of intervening grooves
between ribs).

The symbols used alongside the synonymy list are as explained in Matthews (1973).

Torreites sanchezi (Douville)

Plate 61, figs. 1-5; Plate 62, figs. 1 -4

*1927 Hippurites ( Vaccinites) Sanchezi Douville, p. 54, pi. 4, fig. 1.

1933 Torreites sanchezi Douville; Palmer, p. 100, pi. 7, figs. 1 and 2; pi. 8, figs. 1 and 2.

1936 Torreites sanchezi (Douville); Rutten, p. 135, text-fig. 4g.

1937 Torreites sanchezi (Douville); Vermunt, p. 269 (description only).

1937 Torreites sanchezi (Douville); MacGillavry, p. 128, pi. 5, fig. 4e-h.

1970 Torreites sanchezi (H. Douville); Jung, p. 5, pis. 1-3, text-figs. 1 and 2.

1971 Torreites sanchezi (Douville); Van Dommelen, p. 34, table 5.

V. 1979 Torreites milovanovici Grubic, p. 84, pis. 1 and 2; text-fig. 4.

1979 Torreites coxi Grubic, p. 86, text-figs. 5 and 6 [text-fig. 5=7’. sanchezi (H. Douville); Jung,
1970, pi. 3].

Emended diagnosis. Large species of Torreites: D may exceed 100 mm and L, 200 mm. Ventral flank
of RV flattened, even embayed, and separated from somewhat flattened anterior part by a blunt
Carina. Outer shell layer of RV very thick (5-10 mm), and marked externally by deeply indented
radial grooves separated by broad rounded ribs usually over 5 mm wide. These indentations
correspond to salient ridges on the inner margin of the valve, and are present along the three
infoldings of the RV wall, as well as around its periphery. The outer shell layer of the LV is only
about 1 mm thick, and its marginal growth laminations are recurved over at least the peripheral
parts of the valve’s outer surface. The arete cardinale and pillars are spaced around an arc (a) of
12-126°. Arete cardinale much longer than the subequal pillars (of which Pi tends to be the longer).

Holotype. Hippurites ( Vaccinites) Sanchezi Douville, 1927, p. 54, pi. 4, fig. 1.

Material studied. Five full-sized specimens were collected by V.P.W. some 30 km south-south-west of the
Saiwan airfield, eastern central Oman, in 1983 (see p. 506), and have been deposited in the British Museum
(Natural History) (nos. BMNH LL 28000-28004). One further specimen (P.W.S. Collection, no. 84/x.l) was
collected by V.P.W. from Filim (north Masirah Bay), in 1983. All are well preserved, though with some patchy

SKELTON AND WRIGHT; ISLAND-HOPPING RUDIST

509

TEXT-FIG. 2. Diagrams showing the morphology and measured characters of Torreites sandiezi. A, RV interior.
Abbreviations are: am, anterior adductor muscle scar; ats, anterior tooth socket; be, body cavity; Pq arete
cardinale\ Pi and P2, the two pillars (infoldings of the RV outer shell layer); pms, socket for the LV posterior
myophore; pts, posterior tooth socket, b, as in a, showing measured characters. Abbreviations are: a, angle of
arc described by lines drawn through the middles of Pq and P2; P, angle of arc described by lines drawn
through Po and the myocardinal array (using the black dots centred on the sockets shown in a); Di, commissu-
ral diameter parallel to Pq; D2, commissural diameter normal to Po; IP, pillar length; wP, pillar width, c,
outline of both valves. Abbreviations are: L, RV length, measured along the external trace of Po; Iv, left valve;

rv, right valve.

silicification of the outer (calcitic) shell layer, and with the inner shell layers (originally aragonitic) now
replaced by calcite spar. Four the the specimens have all or part of both valves preserved, while the remaining
two (BMNH LL 28003 and 28004) are RVs only, though the latter also has a fragment of another, juvenile
RV attached to its flank. A slightly worn RV was recovered by P.W.S., in 1984, from the basal chert pebble
and shell rubble conglomerate underlying the Simsima Formation at Jebel Huwayyah, near A1 Ayn, UAE
(see p. 506) (P.W.S. Collection, no. 84/32.1). Other specimens, in life position, were studied in the field (but
not collected) by P.W.S., in 1984, at Qarn Murrah, Sharjah (see p. 506).

Description. Measurements of the specimens are shown in Table 1, and plots of their overall dimensions, and
of changes in shape of the arete cardinale and pillars, as well as in their arrangment (a), with respect to
commissural diameter (DJ, are shown in text-figs. 3, 4, and 5, respectively.

The RVs in BMNFI LL 28000 28004 are of gently curved elongate conical form, reaching over 230 mm in
length in LL 28001. The other two specimens have RVs of more obtusely conical form. The TVs are operculi-
form and only very gently convex towards the strongly ventrally biased apex. The commissure is of rounded
trigonal form, with the anterior and ventral margins slightly flattened, and the posterodorsal margin, with the
infoldings, somewhat bulging (PI. 61, fig. 1; text-fig. 6e g).

510

PALAEONTOLOGY, VOLUME 30

TABLE 1. Measurements on specimens of Torreiles sanchezi milovanovici Grubic from the Maastrichtian of
Oman and the United Arab Emirates, described in this paper. Specimen numbers beginning with LL are
housed in the British Museum (Natural History), and those with P.W.S., in the collection of the senior author.
Locality data are given in the text (p. 506), and the abbreviations for the measurements explained in text-fig. 2.

Specimen

Shell size
(mm)

Hinge

Angle

n

p

Pillar

Angle

n

a

Pillar dimensions (mm)

L

D.

Da

IPo

IPi

IP2

wPo

wPi

wPa

(a) LL 28000

190

103

89

—

39

32

29

26

8

10

9

(h) LL 28001

234

96

no

—

31

46

40

34

6

13

10

80

91

37

31

21

25

6

10

9

> 48

56

49

18

13

14

7

6

6

35

42

70

11

> 9

> 7

5

5

6

(c) LL 28002

226

104

113

—

28

36

29

42

14

11

7

(d) LL 28003

164

93

89

—

42

36

29

23

8

10

10

(c) LL 28004

> 127

100

98

25

15

45

36

33

10

12

11

70

78

31

27

19

19

8

10

8

44

57

61

15

12

11

8

7

7

(/) LL 28004

—

> 48

> 52

—

50

16

> 11

> 10

5

7

7

(small RV)

(g) P.W.S. 84/x.l

160

108

113

—

40

45

33

23

8

~10

~10

(h) P.W.S. 84/32.1

> 42

116

> 93

—

20

> 50

> 35

20

16

12

9

38

45

75

22

20

16

6

6

5

The pale brown, calcitic outer shell layer of the RV is more than 5 mm thick in all the adult specimens, with
a highly distinctive ornament (PI. 61, fig. 4): broad ribs, 5-8 mm wide, with coarse growth rugae, are separated
by deeply indented grooves. The latter are expressed on the inner valve margin as salient spurs (PI. 62, fig. 3).
The spurs continue along the infolded shell wall of the arete eardinale and pillars, giving them the distinctively
branched medial structure, when seen in section (PI. 61, figs. 2 and 3), noted by Grubic (1979).

The calcitic outer layer of the LV is only about I mm thick, and, apart from faint growth inductions, is
smooth (PI. 62, fig. 1); no vestiges of any pores or canals are seen in it. Sections across the margins of this
outer shell layer (PI. 62, fig. 4; text-fig. 7) show the growth lines to be recurved on to at least the periphery of
its upper surface. The subdued character of its growth undulations contrast markedly with the coarse rugae
of the RV.

The discrepancy in thickness between the outer shell layers of the two valves means that most of the inner
margin of the RV extends out beyond the rim of the LV (PI. 62, figs. 1 and 3).

Although the inner shell layers, which would originally have been aragonitic (Skelton, 1976), are now
replaced by calcite spar, evidence for the blind canals that penetrate the LV inner shell can be seen in one of
the specimens (PI. 61, fig. 5).

EXPLANATION OF PLATE 61

Figs. 15. Torreites sanchezi milovanovici Grubic, 1979. 1 -3, BMNH LL 28004, unnamed limestone member
in Aruma Group (probably Maastrichtian), 30 km south-south-west of the Saiwan airfield (lat. 20° 39' N.
and long. 57° 31' E.), eastern central Oman. 1, RV interior (see text-fig. 2 for explanation); 2, 3, sections
across successively younger ontogenetic stages of the RV, which is attached to the RV of another juvenile
individual {lower right), all x 1. 4, BMNH LL 28002, locality and age details as in 1-3; ventral flank of
RV, xO-5. 5, P.W.S. 84/x.l, unnamed limestone member of Aruma Group (probably Maastrichtian),
Filim, northern Masirah Bay (lat. 20° 36' N. and long. 58° 12' E.), Oman; detail of broken section across
posteroventral part of LV, showing the (dark) thin calcitic outer shell layer above, and the canaliferous
recrystallized inner layer (originally aragonite), below, forming a descending prominence, x 3.

PLATE 61

SKELTON and WRIGHT, Torreites

512

PALAEONTOLOGY, VOLUME 30

140|-

Ei20

E

Q100

cT

CO

QC 80

LU

I—

LU

2

<

I 60

_J

<

oc

w

CO

o 20

□

1 1

10^

J I L

8 fg

eg □

I

I-

D.

Key to specimens
□ New world ■ I
o Old world • I

a

J I I

20 40 60 80 100 120 140 160 180 200 220 240

LENGTH OF RV (L) in mm.

TEXT-FIG. 3. Graph of Di and Di against L for the specimens of Toireites
sanchezi listed in Tables 1 and 2. The measurements are explained in text-fig. 2.

EXPLANATION OF PLATE 62

Figs. 1-4. Toireites sanchezi milovanovici Grubic, 1979. 1, BMNH LL 28002, unnamed limestone member in
Aruma Group (probably Maastrichtian), 30 km south-south-west of the Saiwan airfield (lat. 20° 39' N. and
long. 57° 31' E.), eastern central Oman; bivalved specimen, viewed from the LV side, showing the (incom-
plete) operculiform LV fitting in the RV, the inner margins of which are thus exposed around the LV
periphery, x 0.75. 2, BMNH LL 28001, locality and age details as in 1; detail showing blade-like posterior
myophore of LV (centre), seen in section from dorsal side, projecting down between the arete cardinale (Po,
right) and Pi (left) of the RV, x 1 -25. 3, 4, BMNH LL 28003, locality and age details as in 1. 3, posterodorsal
region of bivalved shell, viewed from the LV side, showing the somewhat crushed operculiform LV, with
subdued external growth undulations, as well as the exposed RV inner margin, with salient ridges, running
around the LV periphery and projecting up through sinuses in the latter, above the arete cardinale (Po, left)
and pillars (Pi and P2), x 2-5; 4, close-up of broken section across a chip of the LV margin, indicated by an
arrow in 3, showing the recurved growth lines in the outer shell layer (see text-fig. 7 for explanation), x 10.

PLATE 62

SKELTON and WRIGHT, Torreites

514

PALAEONTOLOGY, VOLUME 30

Key to specimens

COMMISSURAL DIAMETER (D, ) in mm

TEXT-FIG. 4. Graphs of IP/wP against Di for Pq, Pi, and P2 of the specimens of Torreites
sancliezi listed in Tables 1 and 2. The measurements are explained in text-fig. 2.

Three deep radial grooves on the posterodorsal flank of the RV correspond with the infolded arete cardinale
and pillars. The crests of these project through wide sinuses embayed from the margin of the LV (PI. 62,
fig. I).

The elongate arete cardinale and the shorter, subequal pillars are wide and finger-like in section. Pi is
usually, but not invariably longer than P2 (see Table 1). In some specimens the arete cardinale tapers inwards
slightly, so becoming narrower than the pillars (PI. 62, fig. 1). The strikingly low values of a are achieved
ontogenetically: sections across the RV show the angle to reach 75° in the juvenile shell (PI. 61, figs. 1 3; text-
figs. 5 and 6e-g).

SKELTON AND WRIGHT; ISLAND-HOPPING RUDIST

515

0 20 40 60 80 100 120 140

COMMISSURAL DIAMETER (D, ) in mm.

TEXT-FIG. 5. Graph of a against Di for the specimens of Torreites sanchezi listed in Tables 1
and 2. The measurements are explained in text-fig. 2.

The dentition (PI. 61 , fig. I ), shows ji to be about 25°. The narrow sockets for the TV teeth are separated by
a pinched, wall-like RV tooth lying posteriorly to the inner tip of the arete cardinale. The latter has a rounded
inner termination indicating complete absence of a ligament. The tooth-like posterior adductor myophore of
the LV is situated dorsally behind the teeth, projecting down between Pi and the arete cardinale (PI. 62, fig.
2), where it is received in the RV by a large socket (PI. 61, fig. 1; text-fig. 2a). The anterior adductor myophore
of the LV forms an arcuate ledge extending some way around the anterior and ventral margins, where it faces
on to a broad inclined shelf in the RV, supporting a distinctively reticulate muscle scar (PI. 61, fig. 1; text-
fig. 2a).

The body cavity is very shallow, and its volume further reduced by the large, downwardly projecting
myocardinal elements of the LV. Much of the apical 'limb' of the RV is filled by tabulae, though these are
largely obscured by recrystallization.

Palaeoecology. Though found close together, specimens BMNH LL 28000-28004 were recovered as loose
blocks, and so their original life position remains uncertain. Evidence for attachment of one individual to
another is shown by specimens LL 28001 and LL 28004 (PI. 61. fig. 3). Otherwise they represent solitary
individuals. The elopgated and curved horn shape of specimens LL 28000, LL 28001, and LL 28002 suggests
a reclining, boat-like habit, with the convex flank of the RV shallowly embedded in the sediment and the
commissure raised up, away from it (text-fig. 8). The evidence of associated borings and epibionts lends some

516

PALAEONTOLOGY, VOLUME 30

TABLE 2. Measurements on specimens of Torreites sanchezi sanchezi (Douville) (1 12) from the Upper
Campanian to Lower Maastrichtian of the Caribbean, 'T. cf. sanchezi' (13) from the topmost Santonian
to basal Campanian of Jamaica, and T. s. milovanovici Grubic (14) from the Maastrichtian of Sharjah
(UAE), taken from the literature. The original species designations and references thereto (where
locality details may be found) are recorded for each specimen cited. The abbreviations for the measure-
ments are explained in text-fig. 2.

Shell size
(mm)

Hinge

Angle

(-)

Pillar
Angle
( )

Pillar dimensions (mm)

Specimen

L

D,

D^

P

a

IPo

IPi

IP2

wPo

wPi

WP2

T. sanchezi
1. Douville (1927)

83

>83

34

81

40

>32

>14

7

8

7

(holotype)

2. Palmer ( 1933),

130

> 120

126

90

20

52

7

1 1

6

pi. 7, fig. 2
3. Palmer ( 1933),

124

120

50

105

70

49

27

6

8

8

pi. 8, fig. 2
4, Ruttcn(1936)

20-30

(description)
5. Vermunt ( 1937),

230

115

80

(description)
6. MacGillavry

‘90-100’

‘75-120’

(1937)

7. MacGillavry

50

_

_

(1937),
exceptional
specimen
8. Jung (1970),

135

99

122

95

54

40

25

8

7

9

G. 14065 on
pi. 2

9. Jung (1970),

75

>57

>40

98

98

>72

>43

104

34

102

94

118

105

39
27
> 16
65

33

>21

>11

45

>20

>13

>9

26

7

5

4

5

7

7

4

7

8

7

5

5

G. 14066 on
pi. 3

10. Van Dommelen

74

(50)

111

27

15

5

5

6

5

(1971), J.3702n
(table 5)

1 1 . Van Dommelen

76

(90)

49

121

>30

26

14

7

9

7

(1971), J.37026
(table 5)

12. Van Dommelen

>160

86

78

32

118

33

24

14

6

6

8

(1971), J.3676
(table 5)

'T. cf. sanchezi'

13. Chubb (1971)

>110

27

65

45

28

II

5

7

'T. milovanovici'
14. Grubic (1979)

>130

100

12

>60

50

60

10

11

10

SKELTON AND WRIGHT: ISLAND-HOPPING RLIDIST

517

TEXT-FIG. 6. Drawings of sections across the RVs of Torreites sauchezi showing
the ontogenetic divergence in pillar arrangement between the Caribbean (a d)
and the Arabian (e g) subspecies. The drawings show the outlines of the outer
shell layer, as well as some of the internal features in a, d, and G, and arc all the
same scale (scale bar = 1 cm), a is from the holotypc (Douvillc 1927, pi. 4, fig. 1 ).
Abbreviations arc as in text-fig. 2a. b d are from a specimen (G. 14065) illustrated
by Jung (1970, pi. 2, fig. 16, c, and d respectively), e-g are from specimen BMNH
LL 28004, illustrated here in Plate 61, figs. 3 1 respectively.

518

PALAEONTOLOGY, VOLUME 30

growth lines on LV exterior

TEXT-FIG. 7. Explanatory drawing of
the marginal fragment of LV outer
shell layer of Torreites sanchezi (speci-
men BMNH LL 28003), shown in
Plate 62, fig. 4. A radial section across
the valve rim, adjacent to the inner tip
of P], is shown facing the observer.

support. In LL 28001 there is a slight preponderance of borings on the concave dorsal flank of the RV, where
the fragment of another individual is also attached: the convex ventral flank would here seem likely to have
been lowermost, and such a position is further substantiated by the orientation of some fine geopetal sediment
inside the shell. LL 28002 is of similar shape, but has borings and rare encrusters scattered around all its
flanks: possibly it suflfered post-mortem displacement. However, it clearly shows xenomorphic overgrowth on
to another rudist fragment high up on its convex ventral flank, again hinting at this surface having lain against
the seafloor in life. In LL 28000 the anterodorsal face is convex and the posteroventral face is concave. A
slight preponderance of borings on the latter face, as well as a couple of compactional indentations on to that
surface hint at it having been uppermost.

LL 28003 and LL 28004 have more or less straight RVs, with borings scattered around all their flanks.
These would seem to have lived in an essentially upright position (text-fig. 8). Burial in such a position is
indicated in specimen LL 28003 by compaction along the central axis of the RV.

The matrix associated with these specimens is a pale yellow (weathered), medium-grained bioclastic pack-
stone with subangular to subrounded bioclasts: it is presumably of shallow marine origin, as suggested by the
grain-supported texture and rounding of the bioclasts.

Specimens PWS 84/x.I and PWS 84/32.1 have obtusely conical RVs, and are likely to have been shallowly
embedded, barrel-like elevators (text-fig. 8). Specimens of similar form were observed in situ at Qarn Murrah,
Sharjah, with such an upright habit (text-fig. 9). Again, these are associated with medium-grained to fine-
grained biomicrite packstones of presumed shallow marine origin.

TEXT-FIG. 8. Diagram showing the two kinds of life position apparently exhibited by Torreites sanchezi
milovanovici—-d reclining, boat-like habit (left) and an upright habit (right). Note that elevation of the
commissure from the substratum (dotted ornament) is achieved in both cases. The difference between them
probably only reflects variability in the orientations established during larval settlement and early growth.

SKELTON AND WRIGHT; ISLAND-HOPPING RUDIST

519

TEXT-FIG. 9. Two RVs of Toneites
sanchezi milovanovici shown in sec-
tion in upright life position (cf.
text-fig. 8, right), on the upper
surface of a bedding plane of
Maastrichtian limestone (Simsima
Formation) at Qarn Murrah,
Sharjah (UAE) (see text-fig. 1 and
p. 506 for locality details). The lens
cap is 5-5 cm across.

Discussion. The possibility of our specimens merely being Old World homeomorphs of Terreites
can be immediately rejected: they share with the New World forms too many constructionally
independent, specialized features for coincidental convergence to be plausible. The features in
question comprise: ( 1 ) the smooth upper surface of the left valve; (2) the apically blind canals in the
LV inner shell; (3) the unusual relative sizes and shapes of the RV pillars and arete caniinale, and
their exposure through broad sinuses in the LV; and (4) the distinctive radially indented ornament
of the RV outer shell.

Plots of the overall dimensions (text-fig. 3), and of the shapes (text-fig. 4), and relative positions
of the pillars and the arete cardinale (text-fig. 5) of our specimens, drawn from Table 1, form
compact, well-defined clusters; they clearly all belong to one species population. The fragmentary
RV holotype of ‘T. milovanovici Grubic’ represents a slightly larger individual than any of ours (see
Table 2, specimen 14), but all its measured features plot comfortably as extrapolations from the
cluster of data points for our specimens. In fact the obliquity of the section across the valve figured
by Grubic (1979, pi. 1) is slightly misleading in making the pillars look more extended than they
are. Our measurements are taken directly from the specimen itself, with allowances being made for
this ‘cut effect’.

All the Old World specimens, then, may satisfactorily be grouped in a single species. It is the
relation of this to the New World species that is more problematical.

Prior to Grubic (1979) only two species of Torreites had been recognized, the type species T.
sanchezi (Douville) and T. tschoppi MacGillavry. The latter is typically much smaller than the Old
World form (rarely > 40 mm in commissural diameter), with simple costulate ribs about 1-2 mm
wide, and is found in dense clusters of individuals, rather than in ones or twos (MacGillavry 1937,
p. 129). Besides, it also appears to be older than the Old World specimens (see p. 506). It is thus
clearly distinct from the Old World species. However, it approaches the latter in having a relatively
low a, of about 45-70°, according to MacGillavry (1937).

Measurements reported for, or taken directly from the published figures of all described specimens
referred to T. sanchezi are shown in Table 2, and are also plotted on text-figs. 3-5. Grubic (1979)
erected two new species from among these, T. co.xi and T. chuhhi. The former was founded on two

520

PALAEONTOLOGY, VOLUME 30

specimens, the holotype being one of those described by Jung (1970) (specimen 9, Table 2). It has
the same size, ornament, and basic pillar arrangements as in the other T. sanchezi, but was considered
to dilfer in possessing a relatively longer and thinner arete cardinale, and pillars with slightly bulging
inner tips. With such small samples available, these subtle distinctions do not carry conviction. The
holotype simply lies at the edge of a rather broad cluster of points for arete cardinale shape in the
New World T. sanchezi (text-fig. 4). Moreover, the swelling of the pillar tips shown in Jung’s Plate
3 is only very slight. The species cannot be upheld and we here suppress it as a junior subjective
synonym of T. sanchezi.

T. chuhhi is based on a single poorly preserved specimen described as T. cf. sanchezi by Chubb
(1971) (specimen 13, Table 2). Again, its independent status is dubious, though the specimen is
intriguing in possessing a relatively low value of a (27°, rather than the 40° reported by Grubic). It
is also older than the other specimens considered here (probably latest Santonian to earliest Cam-
panian; see p. 506). Further material is needed to clarify its position, particularly in relation to the
coeval T. tschoppi.

The final comparison left to be made, then, is with T. sanchezi itself. Plots of the shapes and
relative positions of the pillars and arete cardinale (text-figs. 4 and 5) for all the specimens included
here in the New World T. sanchezi (specimens 1 -12 in Table 2) form coherent, if somewhat broadly
spread clusters of data points, supporting their inclusion within one species population. Their
commissural diameters (text-fig. 3) also plot together reasonably well, with the exception of the
small specimen 10 on Table 2; the variation in shell length can readily be attributed to ecophenotypic
variation in relation to life position (text-fig. 8).

The Old World specimens share with the New World T. sanchezi the same adult size range (text-
fig. 3) as well as precisely the same distinctive external ornament: the broad radial ribs of the Old
World specimens have widths varying from about 5 to 8 mm, exactly as in those of the New World
specimens (e.g. Jung 1970, pi. 1, figs. 2 and 3). The one value of fi from the Old World specimens
(from BMNH LL 28004) is at the lower end of the range exhibited by those from the New World
(compare Tables 1 and 2). Pillar form in the two populations (text-fig. 4) is closely similar in smaller
(younger) individuals, though in the largest individuals Pi tends to become relatively longer and
more slender in the New World specimens. A similar trend in the arete cardinale (Po) is even more
pronounced. There is thus some slight ontogenetic divergence between the two populations from a
more or less similar juvenile condition. This divergence may be linked constructionally with the
rather more marked separation of the two populations on the basis of a values (text-fig. 5): those in
the Old World specimens are notably smaller than those of the others. Again, there appears to have
been ontogenetic divergence, involving reduction in a throughout growth in the Old World forms,
but with little change in the New World forms. Yet the existence of some specimens among the
New World population with low values of a (e.g. specimen 7 on Table 2, cited by MacGillavry
1937, p. 129) gives support for close linkage between the two populations.

Although these few differences between the adults of the Old and New World populations are
sulficient to render each distinctly recognizable, it is the close similarity of the two populations in
all other respects that is the more striking. Given the small samples involved and our relative
ignorance of many aspects of rudist functional morphology, it is really a matter of subjective
judgement as to whether the Old World population should still be recognized as a distinct
species CT. milovanovicf), or whether it should be treated, as we propose here, as a geographical
subspecies of T. sanchezi, which we label T. s. milovanovici Grubic, 1979, in contrast with the
New World stock of T. s. sanchezi (Douville, 1927), in which we include all the New World
records of T. sanchezi recognized here. We favour this latter option because of the existence
of exceptional specimens in the New World with a values close to those of specimens from the
Old World. These suggest that the one significant diagnostic feature of the Old World popula-
tion may already have existed within the range of morphological variability of the New World
population.

The paradox of the extraordinarily wide geographical separation of the two subspecies will be
considered in a later section (p. 522).

SKELTON AND WRIGHT: ISLAND-HOPPING RUDIST

521

PALAEOBIOLOGY AND EVOLUTIONARY RELATIONSHIPS OF TO RREITES

Torreites is rare in the Caribbean Province; Grubic (1979, p. 94) noted that the then published
record of the genus was founded on only about twenty specimens (which have nevertheless fuelled
at least nine systematic papers alone!). Our material, as well as expanding the recorded total of
specimens by at least a third, is also significant for its good state of preservation. Placement of the
genus in the Hippuritidae can be confirmed, while the reason for its aberrant lack of pores in the
LV can now be explained from investigation of the growth lines in that valve.

Though generally considered a hippuritid because of its pillars (Douville, 1927), Torreites diflTers
from all other hippuritids in its lack of pores. Van Dommelen (1971, p. 65) commented: ‘In reality,
however, all we know about the VS [= LV] of Torreites is, that the outer shell layer is imperforate
in the parts between L, S and E [— Po, Pi and Pij. But what about the slit-like openings over these
infoldings of the available eroded specimens? Have they been covered or not covered, and if covered,
was this cover a reticulum or something different?’ MacGillavry (1937, p. 128) too, queried whether
or not the pillars really had been exposed through open sinuses in the LV. It is unambiguously
clear from our specimens (e.g. PI. 61, fig. 5 and PI. 62, figs. 1, 3, 4) both that, no roof of any kind
existed over the crests of the arete cardinale and the pillars, and that the LV outer shell layer is
indeed imperforate. Is Torreites thus a hippuritid that has lost its pores, or one descended from a
primitive form that never had them, or not a hippuritid at all? The latter option is the least likely.
Not only does Torreites possess the pillars so characteristic of the family, albeit of unusual form, it
also has a typically hippuritid myocardinal apparatus with its markedly tooth-like posterior my-
ophore in the LV, received in a socket in the RV. Its overall shell form, with an elongate-conical
RV and operculiform LV, is also typical of, though not exclusive to the family.

If Torreites is accepted as a hippuritid, the lack of pores is unlikely to be a primitive trait. Pores
are already well established in the earliest hippuritids, which predate the oldest Torreites by about
two stages. Torreites is further characterized by such specializations as the extreme elongation of
the arete cardinale, despite loss of the ligament itself, and the blind canals in the LV interior.

Two features in our specimens together suggest that the lack of pores is secondary: ( 1 ) the thinness
of the outer shell layer in the LV relative to that in the RV, with the margin of the latter projecting
well beyond that of the former (PI. 62, fig. 3); and (2) the recurvature of the growth lines in the
outer shell layer of the LV on to its outer surface (PI. 62, fig. 4; text-fig. 7), indicating that mantle
tissue reached out on to at least the peripheral parts of the outer surface of the valve. Mantle tissue
would thus have been freely exposed both around the inner margin of the RV and around the
periphery, if not more, of the outer surface of the LV. In normal hippuritids, no mantle tissue was
directly exposed in this manner. Rather, extensions of it lined the pore and canal system in the LV
outer shell layer, where its cilia are interpreted as having driven water currents, drawn from above
the shell, over the covered inner rim of the RV, enabling the entrapment there of food particles,
without the need of valve gaping (Skelton 1976). The free exposure of mantle tissue in Torreites
would have made redundant the canal and pore system of its ancestors. Its atrophy would then
have been hastened by the lowering and retreat of the LV outer shell layer to allow maximum
exposure of the RV inner rim. A reconstruction of Torreites in life, with its corona of exposed
mantle tissue is shown in text-fig. 10.

What was the adaptive significance of this exposure of mantle tissue? One possibility is that the
externalization of food entrapment (Skelton 1976) was taken a stage further than in the ancestral
hippuritids, with direct trapping of food particles on the exposed mantle projections, as has been
proposed for radiolitids (Skelton \919a). But such a trivial modification of function seems an
unsatisfactory explanation for so considerable a morphological change; it is hard to conceive of
any obvious benefit to feeding efficiency that might have been won thereby, and the increased
exposure would seem in any case to have carried an increased risk of physical or biological damage
to the mantle rims. One significant effect of the exposure, however, would have been the emergence
of mantle tissue from the darkened confines of the ancestral pore/canal and covered rim system,
into the light. This immediately calls to mind the possibility of symbiotic zooxanthellae, by analogy

522

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 10. Reconstruction of the appearance in life
of Torreites sanchezi milovanovici. The upper part of
the shell is shown, with thick extensions of mantle
margin projecting out between the valve rims, in the
manner of the living Giant Clam, Tridacna. It may be
surmised that, as in the latter, these mantle extensions
were vividly and variably coloured.

with the Giant Clam, Tridacna. The possibility of algal symbiosis in rudists has frequently been
raised in the literature (e.g. Kiihn 1937; Philip 1972; Coates 1973; Kauffman and Sohl 1974; Vogel
1975; and Skelton 1979r?) and has recently been ably reviewed by Cowen (1983). The consensus of
most of these works is that such symbiosis was widespread in the group— a conclusion supported
by Cowen himself. Skelton (1979a), in contrast, argued that positive evidence for possession of
zooxanthellate tissue, in the form of clear adaptation of the shell to allow maximum exposure of
mantle tissue to the light, is only seen in certain broad-rimmed radiolitid genera. Their symbiosis is
considered to have arisen as a secondary adaptation of the enlarged mantle margins associated
originally with food particle entrapment. The layout of the pore and canal system in normal
hippuritids, in particular, was well suited for supplying the RV mantle rim with feeding currents,
and is most satisfactorily explained thus. Cowen nevertheless felt that the tissue within the canal
and pore system could have been exposed to light through the shelly cover over the radiating canals.
Yet in some hippuritids this cover is actually quite thick (often over 3 mm), and, moreover, the
canals show no tendency to flatten out in such a way as to maximize exposure to light. Many
hippuritids seem, in any case, to have lived in somewhat turbid, poorly lit waters (Skelton \919h).
It thus remains unlikely that any of the normal hippuritids, at least, possessed zooxanthellae. With
Torreites, however, we may have an exception; its dear morphological adaptation for free exposure
of mantle tissue suggests a hippuritid ‘redesigned’ as a photophile. The scarce palaeoecological
information on Torreites — implying a basically upward growth form and occupation of shallow,
clear waters— is consistent with the algal symbiosis hypothesis.

SIGNIFICANCE OF THE PRESENCE OF T. SANCHEZim OMAN AND THE UAE

Torreites is not alone among Campanian-Maastrichtian shallow marine benthos in showing disjunct
endemism between the Caribbean Province of Kauffman (1973) and various sites in the eastern part
of the Tethyan Realm, ranging from the Middle East to the East Indies. Chubb (1956), recognized
such a distribution in the radiolitid rudist genus, Thyrastylon, remarking (p. 39): ‘It is indeed
interesting that in the same epoch, the Maestrichtian [^/c], a form closely resembling T. coryi [a
Caribbean species] was living in Persian seas, so that the geographic range of Thyrastylon extended
from Guatemala to Persia, a distance of nearly 10,000 miles.’ Kollmann and Sohl (1980) stated that
the itieriid gastropod, Vernedia friesi Kollmann and Sohl, from the Upper Cenomanian or Eower
Turonian of Colima Province, Mexico, had a closer affinity with a southern Indian form, V.

SKELTON AND WRIGHT: ISLAND-HOPPING RUDIST

523

glohoides (Stoliczka), from the Campanian to Maastrichtian Arrialoor Group of the Trichinopoly
District, than with the European and Transcaucasian species of the genus. Another gastropod,
Actaeonella borneensis Nuttall and Leong, is known both from an uncertain Cenomanian to
Campanian level in Borneo and from Campanian to Maastrichtian strata in Mexico and Cuba
(Sohl and Kollmann 1985). The codiacean alga Ovulites occurs in the Upper Cretaceous of the
Caribbean and in northern Iraq, Afghanistan, and Tibet, though not from the Mediterranean
Tethys (Elliott 1981). These are but a few examples that complement the clear-cut case of
Torreiles to establish this form of disjunct endemism as a palaeobiogeographical problem in need
of a solution.

There are three possible explanations for such a highly disjunct distribution of shallow-water
benthic taxa: ( 1 ) false synonymy of coeval homeomorphs; (2) plate tectonic drifting apart of formerly
united shallow-water provinces; and (3) temporary range extension between the two regions brought
about by the development of a continuous intervening chain of shallow-water ‘staging posts’ for
planktonic larval dispersal.

The first option, of coeval evolution of not one, but several pairs of homeomorphs in the two
regions, over the same limited time interval, is in itself improbable, and becomes yet more so with
every new example of the disjunct distribution that is added to the list. The importance of Torreites
is that its several constructionally independent diagnostic features allow us expressly to reject
homeomorphy as a reasonable hypothesis in its case; any argument for homeomorphy in the other
examples cited must now be relegated to special pleading for particular examples.

The second explanation, based on drifting apart of the two regions was evidently that favoured
by Grubic (1979, p. 94), who concluded: ‘The presence of a specific Caribbean Upper Cretaceous
rudist from in the eastern Mediterranean can by no means be interpreted other than by assuming
that both Americas, Antilles and the Mediterranean were much closer in the Upper Cretaceous.’
Such an explanation would be reasonable if there were matching geotectonic evidence. However,
there is none for Oman having been anywhere near the Caribbean in Late Cretaceous times. The
strata containing Torreites in Oman and the UAE represent the first marine autochthonous deposits
upon the Semail Nappe, itself already obducted on to the Arabian foreland (see Glennie 1977 and
Murris 1980 for details). So the area in question was then, as it is today, firmly part of the Arabian
continent. Likewise, the Caribbean sites with Torreites (Cuba, Jamaica, and Puerto Rico) were
unambiguously associated with the central American region in the Late Cretaceous (Mattson and
Lewis 1980).

We are then left with the third possibility, of planktonic larval dispersal along a chain of staging
posts. In the absence of recorded fossils of rudist prodissoconchs, the character of rudist larvae can
only be surmised from circumstantial evidence. The geologically almost simultaneous appearance
in the Late Turonian of the primitive hippuritid species, Hippiirites resect us Defrance in Europe
and North Africa (Douville 1890-1897, 1910) and of the almost certainly synonymous ‘//. mexicamts
Barcena’ in Mexico (Muellerried 1930), for example, favours a readily dispersed planktonic larval
stage (incidentally, this is against the notion of a larval brood pouch speculated upon by Skelton
(1976)). So it is not unreasonable to suppose that such a larval stage existed in Torreites.

The two possible oceanic routes for dispersal between the Old and New World sites of Torreites
are; (1) via the Mediterranean Tethys and Atlantic; and (2) across the Pacific and eastern Tethys
(text-fig. 11).

The Mediterranean/Atlantic route is considered unlikely for two reasons. Lirst, the bivalve
endemism data of Kauffman (1973) suggest that the ‘north Indian Ocean sub-province’ (incorporat-
ing Oman and the UAE) and neighbouring ‘eastern Mediterranean sub-province’ became signifi-
cantly distinct from the ‘western Mediterranean sub-province’ in Campanian-Maastrichtian times,
implying the presence between them of barriers to dispersal. In a recent study of Campanian/
Maastrichtian rudist distributions in the Mediterranean region, Philip (1985) similarly establishes a
clear distinction between an ‘Aquitano-Pyrenean palaeobiogeographic unit’ and an ‘eastern and
central Mediterranean’ one, with continental barriers lying between them. A few central Mediter-
ranean forms are preserved in southern Spain, howeVer, presumably derived via a North African

524

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. II. Inferred Campanian Maastrichtian palaeogeography of the World, showing oceanic surface
currents (arrows). The Americas are duplicated at each end of the map to allow direct comparison between
the Mediterranean Tethys/Atlantic and Pacific marine connections between the two sites from which Torreites
has been recovered (indicated by T). It is argued in the text that a chain of shallow staging posts, spread
across the Pacific and eastern Tethys, was the most likely means by which the range of Torreites was extended,
by larval dispersal, from the Caribbean, across to the Arabian region. Continental positions derived from
Smith and Briden (1977), land (dotted ornament) and sea (white) distributions from Zeigler et cil. (1983), and

ocean currents from various sources cited in the text.

route (Philip 1983). In any case, there are no records, as yet, either of Torreites or Thyrastyloii—
both very distinctive fossils — anywhere in the Mediterranean region.

The second objection to the Mediterranean/Atlantic route is the increasingly impassable width
of the Atlantic in Campanian-Maastrichtian times, as reflected in the rising generic dissimilarity
between its two sides with respect to bivalves (Kauffman 1973) and, in particular, to rudists (Coates
1973).

The alternative explanation is that Torreites, Thyrastylon, and the other disjunct endemics dis-
cussed earlier were somehow able to cross the Late Cretaceous Pacific Ocean. In which direction
might such dispersal have taken place? Many authors have argued for a circum-global East to West
equatorial current during the Cretaceous, passing through Tethys, and across the Pacific Ocean
(text-fig. 11; Luyendyk et al. 1972; Gordon 1973; Berggren and Hollister 1974, 1977; and Lloyd
1982). The Tethyan Realm was thus extended for several thousand kilometres into the eastern
Pacific (Gordon 1973) and many Pacific seamounts have now been found to have been capped by
rudist-bearing atolls in Aptian to Cenomanian times (recently reviewed, with references therein, by
Winterer and Metzler 1984 and Konishi 1985). From mid-Aptian times onwards, the opening
Atlantic promoted the growth of endemism in the Caribbean, with respect to the rest of Tethys
(Coates 1973; Skelton 1982), to provincial levels by Late Santonian times (Kauffman 1973). This
suggests that the Late Cretaceous Torreites is most likely to have originated in the Caribbean, and
then dispersed westwards towards eastern Tethys (text-fig. 11). Such a model is consistent with
MacGillavry’s (1937) identification of the older Caribbean species T. tschoppi as an ancestral form.

SKELTON AND WRIGHT: ISLAND-HOPPING RUDIST

525

Evidence for staging posts that would have facilitated the trans-Pacific dispersal of Torreites and
other forms is now well documented from DSDP work. Beckmann (1976), for example, has de-
scribed redeposited Campanian-Maastrichtian shallow-water foraminifera from the Line Islands
seamount chain. Two of the genera, Asterorhis and Sulcoperculina, had previously only been
recorded from the Caribbean and surrounding areas. A third, typical Caribbean form, Pseudorbito-
ides israelskyi Vaughan and Cole, was also known from eastern Tethyan sites (details in Dilley
1973). Other findings, reviewed by Schlanger et al. (1981), have extended the evidence for such
staging posts as far across the ocean as the Nauru Basin (Marshall Islands). These records indicate
that the Pacific Plate was studded with islands and/or seamounts which served as stepping stones
for some Caribbean shallow-water benthos in Campanian-Maastrichtian times. Evidence that the
Farallon Plate, which at that time separated the Pacific Plate from the Americas by some 6000 km,
was similarly endowed was also provided by Schlanger et al. (1981). This survives in the form of
ophiolite complexes with exotic limestones, plastered on to the western flanks of the Americas
during subduction of the plate.

There is thus direct evidence for former staging posts, carrying a Caribbean-derived shallow
marine benthic fauna of Campanian-Maastrichtian age, for over half the distance between the
Caribbean and Oman. Documentation of suitable staging posts along the remaining (eastern
Tethyan) part of the route is very much more difficult, because the available evidence is now caught
up in the various Tethyan suture zones running from the East Indies to the Middle East (Audley-
Charles et al., 1980). It seems, however, that some shallow or even emergent physiographical
prominences, such as island arcs and small continental blocks may then still have lain in the oceanic
gap between India and Eurasia, although most of these had already been accreted on to Eurasia by
mid-Cretaceous times (Tapponnier et al. 1981). Moreover, Maastrichtian orbitoid and rudist-
bearing facies are known from the southern edge of the Lhasa Block, in Tibet (Herm et al. 1985
and pers. comm.), which was by then accreted on to the southern flank of Eurasia.

From the evidence given above, it is therefore not unreasonable to postulate that Torreites spread
from the Caribbean to Oman, exploiting a continuous chain of shallow-water staging posts that
stretched across the Pacific and eastern Tethys during Campanian-Maastrichtian times.

As a test for this hypothesis, we predict that Torreites, or other forms with the same disjunct
distribution, such as Thyrastyloii, will eventually be recovered either from DSDP material from the
Pacific, or from shallow-water Tethyan carbonates of Campanian-Maastrichtian age caught up in
Himalayan and other eastern Tethyan suture zones. Conversely, the discovery of such faunal
elements in the western Mediterranean (e.g. north-western Africa or southern Spain) would militate
against our hypothesis.

Two palaeobiogeographical corollaries to this model are worth noting. First, the repeated popu-
lation-sampling effect that would have accompanied the successive westward dispersals of spat from
staging post to staging post ought, surely, to have resulted in a pronounced ‘founder’ effect in the
population that eventually became established in Oman and the UAE (see Mayr 1970, for expla-
nation of the founder principle). It is thus remarkable that the only significant modal deviation of
the Old World population from the Caribbean population is in the ontogenetic reduction of a
(p. 520), which, in our view, only merits distinction at the subspecific level.

Our model for the historical biogeography of Torreites and its ‘fellow travellers’ also bears on
the current debate about the relative roles of vicariance and dispersal in explanations for the
geographical distributions of taxa (concisely reviewed by Forey 1981). In identifying the Caribbean
as the centre of origin for Torreites, and treating its appearance in eastern Tethys as a result of
dispersal along staging posts between the two areas, we have clearly adopted a ‘dispersalist’ expla-
nation for the disjunct distribution of the genus. In our view the appearance of the staging posts
caused the decline of the Pacific Ocean as a barrier to dispersal of Caribbean shallow-water benthos
during the Campanian-Maastrichtian. The Pacific, then, became a ‘filter’ for biogeographical range
expansion (Simpson 1962) at that time, because selected taxa appear to have made the full crossing.
Forey’s (1981) characterization of such accounts as emphasizing processes supposed from ad hoc
considerations to have brought about distributional patterns, and as lacking in any general principles

526

PALAEONTOLOGY, VOLUME 30

of pattern analysis — an advantage reserved for vicariance studies— is a misleading criticism, suggest-
ing, as it does, that the accounts are difficult, if not impossible, to test. The hypothesis of the
appearance of a trans-Pacific filter for the westward dispersal of Caribbean shallow marine benthos
in the Late Cretaceous, could readily be tested by an analysis of fossil distributions. If our model is
correct, then among the taxa showing apparent disjunct endemism on either side of the Pacific there
should be an overwhelming preponderance of stratigraphically older records, for each of the taxa
considered, on the Caribbean side. The test should be workable, because the purported sequence of
events was spread over a ‘geological’ time scale of millions of years, and so ought to have left a
realistically detectable imprint on the fossil record. In the case of Torreites, for example, there is a
generous stratigraphical spread between (1) the first record of Torreites in the Santonian or earliest
Campanian, (2) the establishment of the full chain of staging posts in the Campanian-Maastrichtian,
and (3) the first record of Torreites in the Old World (Early Maastrichtian). It should be noted,
however, that what is being discussed here is a biogeographical range expansion provoked by
changes in the geological context for dispersal. The process of larval dispersal itself is admittedly
beyond geological analysis, because such events on the ‘ecological’ time scale would be, to all intents
and purposes, geologically instantaneous.

Since we are dealing with the effect on the distribution of taxa of the disappearance of a former
barrier (the Pacific Ocean prior to the completion of the chain of staging posts), we could not,
indeed, employ any vicariance method of analysis, simply because these are all irrelevant to such a
phenomenon. As Forey (1981) makes clear, the aim of vicariance analysis is to unravel the history
of progressive fragmentation of an already widespread ancestral biota. Were we investigating, say,
the partitioning of the Tethyan Realm in the Late Cretaceous as a result of the appearance of
barriers, then some form of vicariance analysis might be both appropriate and effective. Our point,
then, is that barriers come and go on a geological time scale, and so the two methods of analysis
both have roles to play in palaeobiogeography: vicariance analysis where developing barriers
fragmented an ancestral biota; and ‘dispersal’ analysis where the demise of barriers has allowed
range extensions to take place. Both methods, we believe, can yield hypotheses about historical
biogeography that can be further tested from fossil distributions.

SUMMARY OF CONCLUSIONS

Systematics. The hippuritid rudist bivalve species, T. sanchezi (Douville, 1927), is here considered
to comprise two geographical subspecies. The nominotypical subspecies, T. s. sanchezi (Douville) is
known from the Upper Campanian to Lower Maastrichtian of the Caribbean Province of Kauffman
(1973), and is considered, by subjective synonymy, to include T. coxi Grubic, 1979, as well as the
other unquestionable records of T. sanchezi in the faunal province. An Arabian subspecies, T. s.
milovanovici Grubic, 1979, is recognized from new specimens collected by ourselves from the
Maastrichtian of eastern central Oman and the United Arab Emirates together with the single
specimen from Sharjah Emirate, which constitutes the holotype for the former ‘T. milovanovici
Grubic’. The only essential morphological difference between the two subspecies is in the angle (a)
between the arete cardinale and the ventralrnost pillar in adult shells: that in T. s. sanchezi ranges
from 50 to 126°, while, in T. s. milovanovici, this angle is from 12 to 75°.

Palaeobiology. The extension of the inner margin of the RV beyond that of the LV and the
recurvature of the shell growth laminations in the LV on to its smooth upper surface, in Torreites,
indicates that marginal mantle tissue was freely exposed in life (text-fig. 10). It is considered likely
that the mantle tissue, thus exposed, contained symbiotic zooxanthellae, as in the living Giant Clam,
Tridacna. The externalization of the mantle tissue accounts for the atrophy of the (redundant) canal
and pore system in the LV, diagnostic of other hippuritids. The upward growth tendency, and
preference for shallow, well-lit marine settings, of Torreites, as indicated by sedimentological data,
are consistent with the zooxanthellate hypothesis.

SKELTON AND WRIGHT: ISLAND-HOPPING RUDIST

527

Palaeohiogeography. Three possible explanations for the apparent disjunct endemism of Torreites
(along with certain other shallow marine benthic taxa) in the Caribbean and in Oman and the UAE
are considered: (1) homeomorphy is rejected due to the negligible probability of several distinctive
and constructionally independent diagnostic features arising in two coeval populations; (2) the plate
tectonic drifting apart of originally neighbouring areas (as favoured by Grubic 1979) is rejected
since all the available geotectonic evidence is against any possibility of the two regions having been
conjoined in the Late Cretaceous; and (3) an episode of range extension between the regions,
brought about by larval drift along a continuous but temporary chain of staging posts between
them, is favoured by us. Of the two possible routes thus implicated, we prefer the trans-Pacific
option to that via the Mediterranean Tethys and across the Atlantic. First, palaeobiogeographical
data from the Mediterranean suggest that significant barriers to dispersal existed there by
Campanian/Maastrichtian times (Kauffman 1973; Philip 1985). Secondly, the Atlantic is also known
to have become a major barrier by this time (Kauffman 1973). In contrast, the equatorial Pacific
Ocean floor is known to have been studded with shallow seamounts and islands, stocked by shallow
marine benthos derived from the Caribbean, at this time (Schlanger et al. 1981). These, combined
with the shores and/or other shallow marine promontories of eastern Tethys could have served as
staging posts for the trans-Pacific dispersal of Torreites and the other taxa sharing its strikingly
disjunct distribution.

Acknowledgements. This paper is published with the permission of Amoco Production Co., USA, who funded
the fieldwork, and the Ministry of Petroleum and Minerals of the Sultanate of Oman. We are very grateful to
the staff of the Earth Resources Institute at the University College of Swansea, UK, who organized the
fieldwork and provided helpful and genial company in the field. Valuable advice and criticism were provided
by Dr N. F. Sohl (US Geological Survey) and Dr R. W. Scott (Amoco Production Co., Tulsa, USA). Any
remaining errors or misinterpretations are, of course, the responsibility of the authors. Janet Dryden, John
Taylor, and Helen Boxall (Open University, UK) assisted with the material preparation of the manuscript.

REFERENCES

AUDLEY-CHARLES, M. G., LATREiLLE, M. and NAYOAN, G. A. s. 1980. Comparison of orogenic features of the
Tethyan mountain belt from the Himalayas to the East Asian arcs. Mem. Bur. Rech. geol. minieres., (Fr.),
115, 155-165.

BECKMANN, j. p. 1976. Shallow-water foraminifera and associated microfossils from Sites 315, 316 and 318,
DSDP leg 33. In schlanger, s. o. et al. (eds.). Initial Reports of the Deep Sea Drilling Project. 33, 467-490.
US Government Printing Office, Washington, DC.

berggren, w. a. and hollister, c. d. 1974. Paleogeography, paleobiogeography and the history of circulation
in the Atlantic Ocean. In hay, w. w. (ed.). Studies in Paleo-oceanography. Spec. Pubis.. Soc. econ. Palaeont.
and Miner.. Tulsa. 20, 126 186. Tulsa, Oklahoma.

1977. Plate tectonics and paleocirculation— commotion in the ocean. Tectonophysics. 38, 1 1 48.

CHUBB, L. j. 1956. Thyrastylon. a new rudist genus from the Upper Cretaceous of Guatemala, the Antilles,
and Persia, with a discussion of the functions of rudist osculcs and pillars. Palaeontogr. Am. 4, 31 48.

1971. Rudists of Jamaica. Ibid., 7, 162-257.

COATES, A. G. 1973. Cretaceous Tethyan coral-rudist biogeography related to the evolution of the Atlantic
Ocean. Spec. Pap. Palaeontol. 12, 169-174.

cowEN, R. 1983. Algal symbiosis and its recognition in the fossil record. In tevesz, m. j. s. and mccall, p. l.
(eds.). Biotic interactions in Recent and fossil benthic communities. Topics in Geobiology. 3, 431 479.
Plenum Publishing, New York.

DE LA TORRE, A., JAKUS, p. and DE ALBEAR, J. F. 1978. Nuevos datos sobre las asociaciones de rudistas en Cuba.
In MACGiLLAVRY, H. J. and BEETS, D. J. (eds.). The 8th Caribbean Geological conference (Willemstad, 1977).
Geologic. Mijnb. 57, 143-150.

DiLLEY, F. c. 1973. Cretaceous larger foraminifera. In hallam, a. (ed.). Atlas of palaeobiogeograpliy. 403-419.
Elsevier, Amsterdam.

DOUVILLE, H. 1890-1897. Etude sur les Rudistes— Revision des principales especes d’Hippurites. Mem. Soc.
geol. Fr. Paleont. 6, 1 -236.

528

PALAEONTOLOGY, VOLUME 30

DOUViLLE, H. 1910. Etudes sur les Rudistes: Rudistes de Sidle, d’Algerie, d’Egypte, du Liban et de la Perse.
Mav. Soc. geol. Fr. Paleont. 41, 1 84.

1927. Nouveaux rudistes du Cretace de Cuba. Bull. Soc. gk)l. Fr. (4), 27 (4), 49 56.

ELLIOTT, G. F. 1981. The Tethyan dispersal of some chlorophyte algae subsequent to the Palaeozoic. Palaeo-
geogr., PalaeoclimatoL, Palaeoecol. 32, 341-358.

FOREY, p. L. 1981. Biogeography— Introduction. In forey, p. l. (ed.). The evolving biosphere, 241-245. British
Museum (Natural History) and Cambridge University Press.

GLENNIE, K. w. 1977. Outline of the geology of Oman. Mem. h. ser. Soc. geol. Fr. 8, 25-31.

GORDON, w. A. 1973. Marine life and ocean surface currents in the Cretaceous. J. Geol. 81, 269-284.

GORIN, G. E., RACZ, L. G. and WALTER, M. R. 1982. Late Precambrian-Cambrian sediments of Huqf Group,
Sultanate of Oman. Bull. Am. Ass. Petrol. Geol. 66, 2609-2627.

GRUBic, A. 1979. Torreites milovanovici sp. nov. iz Omana, T. coxi sp. nov. i T. chubhi sp. nov. sa Jamajke,
nov prikaz roda Torreites Palmer i osvrt na znacajnjegovog paleogeografskog rasprostranjenja. Vesn. zavod
Geol. i Geqfiz. istraz. 37 (A), 81-95. [In Serbo-Croat with an English summary.]

HARRIS, p. M., FROST, s. H., SEiGLiE, G. A. and SCHNEIDERMANN, N. 1984. Regional unconformities and depo-
sitional cycles, Cretaceous of the Arabian Peninsula. In schlee, j. s. (ed.). Interregional unconformities and
hydrocarbon accumulation. Mem. Am. Ass. Petrol. Geol. 36, 67-80.

HAWKINS, H. L. 1924. Notes on a new collection of fossil Echinoidea from Jamaica. Geol. Mag. 61, 312-316.

herm, d., WEiDiCH, K. F., WILLEMS. H., CHUNHUi, Y., BiNGGAO, z. and SHiLAN, z. 1985. Marine sediments and
faunas of Cretaceous and Lower Tertiary in southern Xizang (Tibet, China) — 1. Terra Cognita, 5, 264.
[Abstract.]

JUNG, P. 1970. Torreites sanchezi (Douville) from Jamaica. Palaeontogr. Am. 7, 1-13.

KAUFFMAN, E. G. 1973. Crctaccous Bivalvia. In hallam, a. (ed.). Atlas of palaeobiogeography, 353-383. Elsevier,
Amsterdam.

and SOHL, n. f. 1974. Structure and evolution of Antillean Cretaceous rudist frameworks. Verb, naturf.

Ges. Basel. 84, 399 467.

KOLLMANN H. A. and SOHL, N. F. 1980. Western Hemisphere Cretaceous Itieriidae gastropods. Prof. Pap. U.S.
geol. Surv. 1125-A, 115.

KONiSHi, K. 1985. Cretaceous reefal fossils dredged from two seamounts of the Ogasawara Plateau. In kobay-
ASHi, K. (ed.). Preliminary report of the Hakuho Mai n Cruise KH 84-1, 169 179. Ocean Research Institute,
University of Tokyo, Tokyo.

KUHN, o. 1937. Morphologisch-anatomische Untersuchungen an Rudisten. I. Die Siphonen der Hippuriten.
Zentbl. Miner. Geol. Paldeont. Abt. B, 5, 229-240.

LLOYD, c. R. 1982. The mid-Cretaceous Earth: paleogeography; ocean circulation and temperature; atmo-
spheric circulation. J. Geol. 90, 393-413.

LUYENDYK, K. p., FORSYTH, D. and PHILLIPS, J. D. 1972. Experimental approach to the paleocirculation of the
oceanic surface waters. Bull. geol. Soc. Am. 83, 2649 2664.

MACGILLAVRY, H. J. 1937. Geology of the Province of Camaguey, Cuba with revisional studies in rudist paleonto-
logy. Utrecht Rijks-Univ., Geogr. Geol. Meded., Phys. Geol. Reeks Diss. No. 14, 168 pp. Amsterdam.

MATTHEWS, s. c. 1973. Notes on open nomenclature and on synonymy lists. Palaeontology, 16, 713 719.

MATTSON, p. H. and LEWIS, J. F. 1980. Structural trends and tectonic history of the northern Caribbean. Man.
Bur. Rech. geol. minieres., (Fr.), 115, 30-37.

MAYR, E. 1970. Populations, Species, and Evolution, 453 pp. Belknap Press of Harvard University Press,
Cambridge, Mass.

MUELLERRIED, F. K. G. 1930. El FUppurittes [iic] calamitiformis Barcena. An. Inst. Biol. Univ. Mex. 1 (2), 169-
174.

1936. Estratigrafia Preterciaria preliminar del Estado de Chiapas. Bol. Soc. geol. Mex. 9(1), 31 -41 .

MURRis, R. J. 1980. Middle East: stratigraphic evolution and oil habitat. Bull. Am. Ass. Petrol. Geol. 64, 597-
618.

PALMER, R. H. 1933. Nuevos rudistas de Cuba. Revta. Agric. Habana, 14 (15-16), 95 125.

PHILIP, J. 1972. Paleoecologie des formations a rudistes du Cretace Superieur — L’exemple du Sud-Est de la
France. Pcdaeogeogr., PalaeoclimatoL, Palaeoecol. 12, 205-222.

1983. Le Campanien et le Maastrichtien a rudistes et grands foraminiferes de Quatretonda (Province de

Valence, Espagne): une clef pour la biozonation et les correlations stratigraphiques dans le domaine meso-
geen. Geol. Mediterr. 10 (3-4), 87-98.

1985. Sur les relations des marges tethysiennes au Campanien et au Maastrichtien deduites de la distri-
bution des rudistes. Bull. Soc. geol. Fr. (8), 1 (5), 723 731.

SKELTON AND WRIGHT: ISLAND-HOPPING RUDIST

529

RUTTEN, M. G. 1936. Rudistids from the Cretaceous of northern Santa Clara Province, Cuba. J. Paleont. 10,
134 142.

SCHLANGER, s. o., JENKYNS, H. c. and PREMOLi-siLVA, I. 1981. Volcanism and vertical tectonics in the Pacihc
Basin related to global Cretaceous transgressions. Earth Planet. Sci. Lett. 52, 435-449.

SIMPSON, G. G. 1962. Evolution and geography, 64 pp. Oregon State System of Higher Education, Eugene.

SKELTON, p. w. 1976. Functional morphology of the Hippuritidac. Lethaia, 9, 83 100.

1979(7. Preserved ligament in a radiolitid rudist bivalve and its implication of mantle marginal feeding in

the group. Paleobiology, 5, 90 106.

1979/). Gregariousness and proto-cooperation in rudists (Bivalvia). In larwood, g. and rosen, b. r.

(cds.). Biology and Systeniatics of Colonial Organisms. Syst. Assoc, spec. Vol. 11, 257-279. Academic Press,
Eondon and New York.

1982. Aptian and Barremian rudist bivalves of the New World: some Old World similarities. Cretaceous

Research, 3, ]45 153.

SMITH, A. G. and BRIDEN, J. c. 1977. Mesozoic and Cenozoic pcdeocontinental maps, 63 pp. Cambridge University
Press, Cambridge.

SOHL, N. F. and kollmann, h. a. 1985. Cretaceous actaeonellid gastropods from the Western Hemisphere.
Prof. Pap., U.S. geol. Surv. 1304, 1-108.

TAPPONNiER, p., MATTAUER, M., PROUST, F. and CASSAiGNEAU, c. 1981. Mcsozoic opliiolites, sutures, and large-
scale tectonic movements in Afghanistan. Earth Planet. Sci. Lett. 52, 355 371 .

VAN DOMMELEN, H. 1971. Ontogcnctic, phylogenetic and taxonomic studies of the American species of Pseudo-
vaccinites and of Torreites and the multiple-fold hippuritids. Ph.D. thesis (unpublished), 125 pp. University
of Amsterdam.

VOGEL, K. 1975. Endosymbiotic algae in rudists? Palaeogeogr., PalaeoclimatoL, Palaeoecol. 17, 327-332.

WINTERER, E. L. and METZLER, c. V. 1984. Origin and subsidence of guyots in Mid-Pacific Mountains. J.
geophys. Res. 89 (B12), 9969 9979.

ZIEGLER, A. M., SCOTESE, c. R. and BARRETT, s. F. 1983. Mesozoic and Cenozoic palcogeographic maps. In
BROSCHE, p. and sundermann, j. Tided friction and the Earth's rotation II, 240-252. Springer-Verlag, Berlin,
Heidelberg.

PETER W. SKELTON
Department of Earth Sciences
Open University, Walton Hall
Milton Keynes MK7 6AA, UK

Typescript received I March 1986

Revised typescript received 12 September 1986

V. PAUL WRIGHT

Department of Geology
University of Bristol
Wills Memorial Building
Queen’s Road
Bristol BS8 1 RJ, UK

, ;A-

f ' ■■■ ...:/' ■; ,.

■■■■ , •■■■ If-.

r^( •

I

>

A REINTERPRETATION OF ICHTHYOSAUR
SWIMMING AND BUOYANCY

by MICHAEL ALAN TAYLOR

Abstract. A new functional analysis of the reversed heterocercal caudal fin of ichthyosaurs suggests that its
function, other than propulsion, was not control of buoyancy, but to produce powerful downwards pitching
moments. These moments were used to initiate manoeuvres, to dive after breathing at the surface, and, in one
form, to feed. This model is of potential value in analysing the palaeobiology and evolution of ichthyosaurs
and other marine reptiles with similar caudal fins.

The caudal fin of ichthyosaurs is usually assumed to have had the primary function of propelling
the animal, but this does not explain why many ichthyosaurs had a caudal fin of the reversed
heterocercal type, with a fleshy dorsal lobe, and a ventral lobe containing the terminal vertebral
column. Previous studies have inverted existing analyses of the unreversed heterocercal caudal fin
of sharks (e.g. Alexander 1974) to conclude that the secondary role of the ichthyosaurian caudal
fin was to neutralize positive buoyancy (McGowan 1973). I here apply a new analysis of the shark
caudal fin by Thomson (1976) and Thomson and Simanek (1977) to conclude that the secondary
role of the ichthyosaurian caudal fin was, rather, to initiate manoeuvres. Furthermore, this new
analysis indicates potential new evidence for the palaeobiology of different ichthyosaurs and for
the reasons behind the evolution of the reversed heterocercal caudal fin in ichthyosaurs and other
marine reptiles.

FUNCTIONAL ANALYSIS

Previous analyses of the swimming and buoyancy of ichthyosaurs (McGowan 1973; Wade 1984)
assume that each ichthyosaur was lighter than water and reconstruct the caudal fin as producing a
forwards and slightly downwards directed thrust (text-fig. 1a). The downwards component of this
thrust has the function of neutralizing part of the upthrust from the negative buoyancy; the
remainder of this upthrust is neutralized by lift forces produced by the pectoral fins. The pectoral
fins are assumed to be anterior to the centre of balance so that the moment which they produce
about the centre of balance is in the opposite sense to, and therefore balances, that produced by the
caudal fin. However, by inversion, this analysis is subject to some of the criticisms directed at the
original analysis of the shark caudal fin (Alexander 1974; Thomson 1976; Thomson and Simanek
1977). The lift and drag forces vary with speed while the weight and buoyancy remain constant,
leading to shifts in the overall balance of forces. In many sharks, and possibly in ichthyosaurs, the
pectoral fins are so close to the centre of balance that they have disproportionately short lever arms
about the centre of balance and so have to produce large lift and correspondingly large drag forces
to counter the moment produced by the caudal fin, with its much longer lever arm (although this
could have been alleviated by the use of the pelvic fins). Most importantly, ichthyosaurs were not
necessarily lighter than water, and their buoyancy varied with factors such as fatness, repletion or
starvation, pregnancy, and above all, the depth of water above the immersed animal (Wade 1984).
During a dive, the increase of pressure with depth would compress the air in the lungs and lead to
rapid increases in overall density and loss of positive buoyancy and gain of negative buoyancy, as
in modern reptiles (Seymour 1982). The ichthyosaur would have had to cope with rapid changes in
the magnitude and polarity of its buoyancy.

IPalaeontology, Vol. 30, Part 3, 1987, pp. 531-535.|

© The Palaeontological Association

532

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 1 . Old and new models of forces acting on a swimming ichthyosaur. B. upthrust or downthrust
due to buoyancy; CB, centre of balance; D, drag; P, hydrodynamic downthrust from pectoral and
perhaps pelvic fins; T, propulsive thrust from caudal fin. a, old model, assuming positive buoyancy.
The caudal fin’s thrust is directed forwards and downwards and has the role of partially counteracting
the positive buoyancy. The lift from the pectoral fins counteracts the remaining buoyancy and its
moment about the centre of balance counters that from the caudal fin. b, new model, showing how the
caudal fin’s thrust is forwardly and upwardly directed so as to pass close to or through the centre of
balance in normal swimming. The lift from the pectoral (and perhaps pelvic) fins serves only to neutralize
upwards or downwards forces remaining from the addition of the upwards component of the caudal
fin’s thrust to any positive or negative buoyancy. This is the worst case, when the animal is at the
surface and positive buoyancy is greatest, and the pectoral and pelvic fins produce lift and therefore
drag, c, new model, when the animal has dived to just below neutral depth and it has slight negative
buoyancy. The upwards component of the caudal fin’s thrust cancels out the negative buoyancy, and
the pectoral and pelvic fins need to produce no lift. This is the most efficient situation. Hunting or
cruising ichthyosaurs would probably swim in this efficient manner.

The new analysis of the shark caudal fin by Thomson (1976) and Thomson and Simanek (1977)
can be inverted to reinterpret the ichthyosaurian caudal fin (text-fig. 1b). The propulsive force from
the caudal fin is directed forwards and slightly upwards. Its angle with the horizontal can be varied
by controlling the beat of the dorsal and ventral lobes, as in sharks. In an ichthyosaur swimming
straight and level, the line of thrust passes through the centre of balance, and the pectoral fins need

TAYLOR: ICHTHYOSAUR SWIMMING

533

produce only enough lift to compensate for residual up- or down-thrust remaining when the tail’s
downthrust is added to any positive or negative buoyancy. The lines of thrust from the caudal and
pectoral fins pass through, or close to, the centre of balance, so that little or no moments are
produced about it and correspondingly little lift and drag are wasted on balancing these moments.
The ichthyosaur can now control its buoyancy of whatever magnitude or direction.

produces a strong downwards pitching action which can be used (as here) when starting diving during
breathing at the surface, or can be turned into any other manceuvre by use of the pectoral and

pelvic fins.

The secondary role of the caudal fin is in manoeuvring, as in sharks (Thomson 1976; Thomson
and Simanek 1977). The ichthyosaur would initiate a turn by raising the line of thrust of the tail
above the centre of balance, producing a strong downwards pitching moment which could be
converted by the pectoral and pelvic fins into a turn in any required direction (text-fig. 2). Flexion
of the body and tail would contribute to this pitching effect (Appleby 1979). In sharks the positively
heterocercal caudal fin produces an upwards pitching action which brings the ventrally located but
protrusible mouth into action against prey. By contrast, ichthyosaurs had terminally located narrow
rostra. The inverted heterocercal caudal fin would, however, have allowed these air-breathing
animals to breathe at the surface. Swimming at, or just below, the water surface is energetically
costly because of drag caused by surface turbulence and the production of bow waves (Goldspink
1977). The ichthyosaur could swim up to the surface, start pitching downwards at the surface, and
already be diving while it breathed through the nostrils placed high on the sides of the snout just in
front of the eyes (text-fig. 2). A strong diving action is particularly important since the animal is
most buoyant at the surface.

On the face of it the new analysis of ichthyosaur swimming incorporates an apparent inefficiency
(text-fig. fu). Any positive buoyancy adds to the upwards component of the caudal fin thrust to
produce an upthrust which must be neutralized by the production of lift, and therefore drag, by the
pectoral and pelvic fins. However, this would be worst at the surface or at shallow depths, when
swimming is in any case energetically costly. When the animal dived again, and especially if it
exhaled, it would reach neutral depth and then, below that, a point where it became slightly negatively
buoyant. At this point the buoyancy and the upthrust from the caudal fin would balance and there

534

PALAEONTOLOGY, VOLUME 30

TEXT-HG. 3. The head of Eurhinosaums showing the greatly undershot lower jaw (after McGowan 1979,

pi. 5, fig. 2). Original c. 110 cm long.

would be no need for the pectoral and pelvic fins to produce any lift, and associated drag (text-
fig. Ic). The relative magnitude of the upwards component of the caudal fin thrust would depend
upon the trade-off between efficient swimming and manoeuvrability, as has been suggested for sharks
(Thomson 1976; Thomson and Simanek 1977). A relatively small upwards component, and thus a
nearly horizontal line of thrust, would promote minimal drag and therefore high speed, or efficient
cruising; a more sharply inclined line of thrust would produce greater manoeuvrability at the expense
of efficiency.

PALAEOBIOLOGICAL INFERENCES

The aberrant ichthyosaur Eurhinosaums may provide evidence for this hypothesis. It had a long,
tooth-armed upper jaw overhanging a much shorter lower jaw (text-fig. 3), and is reconstructed as
having slashed downwards through shoals of small fish and cephalopods (McGowan 1979). The
ability to pitch downwards strongly is implied by this habit.

Further testing of the hypothesis may come from analysis of variation within ichthyosaurs. In
sharks the variation of the caudal fin, especially in the degree of asymmetry about the horizontal
axis, the angle of the terminal vertebral column with the horizontal, and the aspect ratio, has been
correlated with the ecology of different forms, as this variation controls the angle with the horizontal
made by the caudal fin thrust and therefore the balance between manoeuvrability and the energetic
efficiency in terms of drag (Thomson 1976; Thomson and Simanek 1977). The existence of excep-
tional cases of soft part preservation of ichthyosaurs (McGowan 1973, 1979; Martin et al., 1986)
provides evidence for the outline of the caudal fin (so long as it is authentic, Riess 1985). This is an
opportunity to correlate palaeobiological inferences from caudal fin form with independent evidence
from overall body form, paired fin structure, and feeding adaptations, so as to reconstruct the
palaeobiology of different ichthyosaurs and different age-classes within species, and test the present
hypothesis of caudal fin function.

The hypothesis of caudal fin function will also be relevant to studies of the origin and evolution
of the ichthyosaurian reversed heterocercal caudal fin during evolution from terrestrial ancestors,
and of the independently evolved reversed heterocercal caudal fin of other marine reptiles such as
the thalattosuchian crocodilians (Buffetaut 1979).

Acknowledgements. I thank Mr. J. Martin for access to work in press and Professor R. McNeill Alexander for
comments.

REFERENCES

ALEXANDER, R. MCN. 1974. Functional design in fishes. 3rd edn. 160 pp. Hutchinson, London.
APPLEBY, R. M. 1979. The affinities of Liassic and later ichthyosaurs. Palaeontology, 22, 921 946.

TAYLOR: ICHTHYOSAUR SWIMMING

535

BUFFETAUT, E. 1 979. The evolution of the crocodilians. Sc7F«L 237 ( 1 0), 124 132.

GOLDSPiNK, G. 1977. Energy cost of locomotion. In Alexander, r. mcn. and goldspink, g. (eds.). Mechanics
and energetics of animal locomotion, 1 53 167. Chapman and Hall, London.

MCGOWAN, c. 1973. Differential growth in three ichthyosaurs: Ichthyosaurus communis, I. hreviceps and Steno-
pterygius cpiadriscissus (Rept., Ichthyosauria). Contr. R. Ont. Mus. Life Sci. 93, 1-21.

1979. A revision of the Lower Jurassic ichthyosaurs of Germany, with descriptions of two new species.

Palaeontographica, A166, 93 135.

MARTIN, J., FREY, E. and RiESS, J. 1986. Soft tissue preservation in ichthyosaurs and a stratigraphic review of
the Lower Hettangian, of Barrow-upon-Soar, Leicestershire. Trans. Leicester lit. phil. Soc. 80, 58-72.

RIESS, J. 1985. Biomechanics of ichthyosaurs. In riess, j. and frey, e. (eds.). Konstruktionsprinzipien lehender
und ausgestorhener Reptilien. Principles of construction in fossil and recent reptiles, 199 205. Sondcrfor-
schungbereich 230, Universities of Stuttgart and Tubingen, Stuttgart.

SEYMOUR, R. s. 1982. Physiological adaptations to aquatic life. In gans, c. and pough, f. h. (eds.). Biology of
the Reptilia. Volume 13, Physiology D. Physiological Ecology, 1 51. Academic Press, London.

THOMSON, K. s. 1976. On the heterocercal tail in sharks. Paleobiology, 2, 19 38.

and siMANEK, D. 1977. Body form and locomotion in sharks. Am. Zool. 17, 343-354.

WADE, M. 1984. Platypterygius australis, an Australian Cretaceous ichthyosaur. Lethaia, 17, 99 1 1 3.

MICHAEL ALAN TAYLOR
Earth Sciences Section
Leicestershire Museums
Art Galleries and Records Services
96 New Walk

Typescript received 8 August 1986 Leicester LEI 6TD

Xj

- { . ■

'> ■ •
f '■ 's;

*

-

i*'

§ 1

A NEW SILURIAN XIPHOSURAN FROM
PODOLIA, UKRAINE, USSR

by PAUL A. SELDEN and DANIEL M. DRYGANT

Abstract. A single incomplete specimen of a xiphosuran, Pasfenuikevia podolica gen. et sp. nov., from the
Ludlow Series of Podolia, Ukraine, USSR, is described. It has a smooth, spatulate carapace and rounded
genal cornua. The opisthosoma bears nine free tergites (second to tenth); the first tergite is reduced and
hidden beneath the carapace. The tergites have a broad axial region and small pleurae; the second tergite is
hypertrophic. Telson and appendages are not preserved. P. podolica resembles Pseudonisats Nieszkowski,
1859 and Cyamocephaliis Currie, 1927; it is thus placed in the infraorder Pseudoniscina Eldredge, 1974, but
certain characters are shared with the synziphosurines. It comes from the lagoonal deposits of the upper part
of the Ustye Suite (Bagovytsa Horizon) where it occurred together with Baltoewyptems tetragonophthalmus
(Fischer, 1839).

AhoTAUUI. OmicyeTbCM e/riinnii HcnoBHin'i CK3CMnA«p MCMoxBocxa Pastenuikevia podolica gen. et sp. nov. is
Ay^AOBChKoro flpycy IIomAAS (YPCP) . npc/rcTaBAemifi Bin i Aa/i,KiiM AoriaTOBii/(HiiM Kapanai<coM is sanpyi AerniMii
iiiiBHiiMii iiimiaMU i onicTOCOMOro. OnicTocoMa CKAa4aeTboi s acb’sitii Bii/riiMiix TCprixiB; iicpiinin Teprir pc/;yKO-
BaHiiii, cxoBamiii n'lA KapanaKCOM. Teprirn luaioTb iimpOKi ocbOBi MacTiiini ra viaAi n.-xespii; /ipyriiii reprir riiicp-
Tpo(j)OBaHm'i. I’cAbcoH i npn/iarKu ne sbcpci Aiicfl. McHoxnicT no4,i6Hiiii 40 Pseudoniscus Nieszkowski, 1859 ra
Cyamocephaliis Currie, 1927, y SB’asKy s bum Bi4HeccHiiii 40 in(|)pa3aroHy Pseudoniscina Eldredge, 1974, xon
Mae xaKorK i 4C«Ki osnaKii cnMu,ii(|)03yp. IIoxo4iiTb bIh is Aaryiimix Bi4i<Aa4iB BcpxHboi nacTimn ycxiBCbKoi csixii
baroBHUbKoro ropnsonxy, 4c 3Haii4eHiifi pasoM is Baltoeurypterus tetragoiiophtlialmiis (Fischer, 1 839).

The late Silurian saw the climax of the first phase of evolution of the Xiphosura. In the Treatise
(Stormer 1955) these middle Palaeozoic xiphosurans were united in the suborder Synziphosurina
Packard, 1886, but some are now considered to belong to the sister group, the suborder Limulina
Richter, 1929 (Eldredge 1974; Bergstrom 1975), and the infraorder Pseudoniscina Eldredge, 1974
was erected for them and some primitive bellinuroids. Xiphosuran phytogeny is founded to a large
extent on scattered records of genera based on few, or commonly single, specimens; new finds often
seem to confound rather than confirm established conceptions. The new genus described below
follows this pattern since, while it undoubtedly lies within the Pseudoniscina, it also shares at
least one character with some synziphosurines. For this and other reasons outlined below — and
particularly if some bellinuroids are included in the Pseudoniscina (Eldredge 1974; cf. Fisher 1982,
fig 1; 1984, fig. 2)— the monophyletic status of the infraorder must be considered suspect.

Eldredge (1974) convincingly argued that the most anterior opisthosomal tergite in synziphosu-
rines and pseudoniscines belongs to the second opisthosomal somite. The tergite corresponding to
the first opisthosomal somite is reduced to the form of an articulating half-ring and can only be
seen in dorsally flexed specimens. Our identification of the opisthosomal tergites follows that of
Eldredge (1974), so that the nine visible tergites are numbered second to tenth (II-X).

Stratigraphy and geological setting. The described specimen comes from dolomite marl (domerite)
of the upper part (c. 22-23 m above the base) of the Ustye Suite of the Bagovytsa Horizon, which
crops out on the left bank of the Dniester River c. 1-5 km downstream of the village of Velyka
Slobidka (Podolia, Ukraine, USSR; text-fig. 1). Subsequent extensive searches failed to reveal any
more specimens of the genus.

The section consists of light-grey, vesicular, granular and platy, pelitomorphic, and rarely stroma-
tolitic dolomites alternating with fine, platy domerites (text-fig. 2). Bed thicknesses are 015-

[Paiaeontologv, Vol. 30, Part 3, 1987, pp. 537-542.|

© The Palaeontological Association

538

PALAEONTOLOGY, VOLUME 30

1 -50 m, and suite thickness as a whole is 30 m. These rocks are lagoonal in origin and contain only
rare remains of chelicerates, of which the best known is Baltoeurypterus tetragonophthalmus (Fischer,
1 839); shrinkage cracks commonly occur on the surfaces of domerite beds (Nikiforova et al. 1 972).

In the lower part of the suite, rare, thick beds of limestones occur with abundant faunal remains
which allow their correlation with Eltonian to lowermost Leintwardinian (Ludlow) strata of Britain
(Tsegelnjuk et al. 1983; Drygant 1984); the layer with Pastemakevia correlates with the lowermost
Leintwardinian.

Preservation. The fossil consists of a single piece; the cast and internal mould of the whole carapace
and slightly damaged opisthosoma of nine tergites. Since the animal is not flexed dorsally, the half-
ring belonging to the first opisthosomal somite cannot be seen and is presumed to be hidden beneath
the carapace. The cuticle is rather thin, consisting of dark-grey matter which is clearly distinguished
on the pale rock background. Though originally chitinous, the cuticle has almost certainly been
replaced by material whose nature has not been determined.

SYSTEMATIC PALAEONTOLOGY

Phylum CHELiCERATA Heymons, 1901
Class xiPHOSURA Latreille, 1802
Order xiphosurida Latreille, 1802
Suborder limulina Richter and Richter, 1929
Infraorder PSEUDONisciNA Lldredge, 1974 (emended)

Genus pasternakevia gen. nov.

Type and only known species. Pasternakevia podolica sp. nov.

Derivation of name. In honour of Professor S. I. Pasternak, a prominent researcher of the Cretaceous fauna
of the Ukraine.

Diagnosis. Carapace spatulate, nearly as long as opisthosoma (excluding telson); cardiac and oph-
thalmic morphology obscure; genal cornua broad and rounded, lacking anterior median projection.

SELDEN AND DRYGANT: SILURIAN XIPHOSURAN

539

TEXT-FIG. 2. Columnar section of the outcrop located on text-fig. 1 showing occurrence of Pasternakevia

gen. nov.

Nine opisthosomal tergites with broad axial region and small pleurae; tergite of first opisthosomal
somite greatly reduced, that of second hypertrophied. No fused tergites.

Pasternakevia podolica sp. nov.

Text-fig. 3a, h, d

Holotype. Lviv State Natural Museum of the Ukrainian Academy of Sciences (Monographical Funds),
No. 35611; single dorsal piece (only known specimen) consisting of the carapace and opisthosomal tergites;
from the upper part (c. 22-23 m above the base) of the Ustye Suite (Bagovytsa Horizon, middle Ludlow)
on the left bank of the Dniester River c. L5 km downstream of the village of Velyka Slobidka (Podolia,
Ukraine, USSR).

Derivation of name. After Podolia, the region in which the locality lies.

Diagnosis. As for the genus.

Description. Pasternakevia is nearly half as wide as it is long (excluding the telson which is not preserved),
with the carapace occupying 0-44 of its length. The three largest opisthosomal tergites are nearly as wide as
the carapace and thus the body appears parallel-sided for much of its length.

The carapace is spatulate, nearly as long sagitally (sag.) as it is wide; the sides are parallel in the posterior
half, and the anterior rim of the carapace forms a semicircle. The posterior border of the carapace is gently
procurved, meeting the lateral borders in blunt genal cornua. A long, crescentic feature on the carapace,
widest anteriorly and tapering gradually to merge with the lateral borders near the genal angles, may represent
the impression of the prosomal doublure or ventral marginal plate(s) adpressed against the carapace. The
carapace is convex dorsally, the highest part being a broad circular area forming the posterior two-thirds; the
surface slopes steeply away from this area to the crescentic feature and the genal areas. The posterior border is

540

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 3. a, h, (i, Pasternakevia podolica gen. et sp. nov.. No. 35611, holotype and only known specimen;
upper part of Ustye Suite, Bagovytsa Horizon (middle Ludlow), near Velyka Slobidka, Lfkraine, USSR;
a, dorsal view in low-angle light, x 3 0; b, dorsal view in high-angle light, x 1-5; d, outline drawing of
specimen showing main features and tergite numbers, x 3 0. c, Bcdloeuryptems tetragonophthalmns (Fischer,
1839), No. 35612, from same horizon and locality as Pasternakevia, x 1.

gently arched. Typical xiphosuran carapace features cannot be discerned, except for a faint parabolic ridge
situated centrally at the anterior side of the circular area; this could represent the anterior part of the cardiac
lobe. Though obscured by a general wrinkling of the carapace surface, any original features must have been
faintly expressed in life.

Nine dorsal tergites are readily apparent on the opisthosoma (belonging to the second to tenth opisthosomal
somites); the first tergite is presumed to be concealed beneath the carapace (see above). Since the telson is not
preserved attached, we cannot be certain that the most posterior tergite is the last, but its small size suggests
that it is. The second tergite is the largest and is obviously hypertrophied. The third is only half the length
(sag.) of the second, but is as wide. The fourth is three-quarters the length (sag.) of the third, and is also as
wide as the second and third. Thereafter the tergites are roughly the same length but become increasingly
narrower (exsag.). Each tergite consists of a wide, raised axial part, occupying about two-thirds of the total
width, and narrower (exsag.) pleurae. The axial part of the second tergite is greatly swollen. On the third to
tenth tergites the axial part has straight anterior and posterior borders. The anterior and posterior borders of
the axial region of the third and more posterior tergites are depressed to accommodate adjacent tergites during
flexure of the opisthosoma; that anteriorly on the third tergite is recurved on its posterior side to accommodate
the hypertrophied second tergite. The pleurae are separated from the axis by dark coloured depressions,
possibly indicating the presence of muscle attachments beneath. Together, these depressions line up as a pair
of axial furrows which run nearly straight and converge from the genal angles of the carapace to the presumed
anterolateral corners of the telson. As the outline of the opisthosoma is broadly curved, the pleurae are widest
on the middle tergites. The pleurae curve gently backwards as spatulate lobes, with the posterior more strongly

SELDEN AND DRYGANT: SILURIAN XIPHOSURAN

541

curved than the anterior. At least the third to the eighth pleurae bear furrows running from the anteromedial
to posterolateral corners, becoming shallower posterolaterally. The opisthosoma lacks ornamentation.

Dimensions (in mm). Lengths (sag.); total (excluding telson), 30-7; carapace, 13-5; opisthosoma (excluding
telson), 17-2; tergites II, 4-8; III, 2-4; IV, 1-8; V, 1-6; VI, 11; VII, 1-4; VIII, 10; IX, 1-6; X, 1-5. Widths:
carapace, 14 0; tergites II, 13-2; III, 13-2; IV, 13-2; V, 12-5; VI, 11-5; VII, 10-6; VIII, 81; IX, 7 0; X, 4-7.

Discussion. Pasternakevia resembles Pseudonisciis, known from Saaremaa, Estonia (P. clarkei
Rtiedemann, 1916), Lesmahagow, Scotland, and Ludlow, England (Pseudonisciis spp., Eldredge
1974), all of which are Ludlow in age. Similarities include: overall dimensions, spatulate carapace,
number of opisthosomal tergites, no fused tergites, and lack of opisthosomal tagmosis (see Eldredge
1 974, table 3). Pasternakevia differs from Pseudonisciis in lacking an anterior median carapace projec-
tion, its broad opisthosomal axial region, and hypertrophied second tergite. Another pseudoniscine,
Cyamocephalus Currie, 1927, resembles Pasternakevia. Cyamocephalus is known from rocks of
Ludlow age at Lesmahagow, Scotland (C. loganensis Currie, 1927), and Ludlow, England (C. cf.
/ogfluc/wA Eldredge and Plotnick, 1974).

Both Pasternakevia and Cyamocephalus lack the median anterior carapace projection found in
Pseudonisciis. Cyamocephalus has a long opisthosoma with fused sixth and seventh opisthosomal
tergites and the seventh hypertrophied, features absent from Pseudonisciis and Pasternakevia. These
dissimilarities warrant the separation of all three animals at the generic level. To include Paster-
nakevia within the infraorder Pseudoniscina Eldredge, 1974, this taxon requires emendation to
remove the character ‘second segment not hypertrophic’. At present, it seems appropriate to include
the genus within Pseudoniscina with this emendation.

The presence of a hypertrophic second opisthosomal tergite is a character which Pasternakevia
shares with the synziphosurines Biinodes Eichwald, 1854 and Limuloides Salter in Woodward, 1865
(Eldredge 1974). The question arises: is this character homoplasous, i.e. derived independently in
two separate clades? Eldredge (1974), Bergstrom (1975), and Stiirmer and Bergstrbm (1981) agree
that the synziphosurines (Biinodes, Weinhergina, Legrandella, and Limuloides) are separate from
other Xiphosura at high taxonomic rank. However, the distinguishing characters (carapace mor-
phology and opisthosomal tagmosis) are not always strictly dehnable. In the Pseudoniscidae, for
example, the carapace morphology is typically obscure, and in a reconstruction of Weinhergina by
Stiirmer and Bergstrom (1981, e.g. fig. 5) the separation of pre- and postabdomen is indistinct.
The Ordovician genus Lemoneites Flower, 1968 also shares characters in common with both
synziphosurines and limulines (Eldredge 1974). Consequently, the current phytogeny of these lower
and middle Palaeozoic Xiphosura must be considered speculative; as more new taxa are described,
character matrix analyses (e.g. Eldredge 1974) will help to answer the homoplasy question and
produce new phylogenetic schemes.

As mentioned above, in the dolomite rocks of Ustye Suite of Podolia a normal marine fauna is
absent. Pasternakevia is accompanied only by relatively few fossils of the eurypterid B. tetragon-
ophthahnus (Fischer, 1839) (mainly as isolated parts but occasionally as almost whole animals; text-
fig. 3c), fragments of Pterygotus sp., and some unidentifiable arthropods. No Baltoeiirypterus
specimen exceeds 10 cm in length; the fragments of Pterygotus indicate that the complete animals
were much bigger. These chelicerates lived in a shallow lagoon, separated from an open basin to the
west by a chain of bioherms (Drygant 1984). Sedimentary conditions in the lagoon were not stable,
hence the deposition of thin, rhythmical, magnesial sediments which were periodically enriched with
clayey material. Common desiccation cracks indicate frequent subaerial exposure in some places. The
bed with Pasternakevia is a thin-bedded, clayey dolomite formed during a regressive phase, but before
its maximum.

The mode of life of Pasternakevia cannot be determined with certainty. Sphaeroidal enrolment
was almost certainly possible, since it is known in Pseudonisciis (Bergstrom 1975) which has a
similar gross morphology. Flexure of the opisthosoma into a dorsally concave shape was also
possible, as evidenced by the topography of the axial region of the tergites. Enrollment was

542

PALAEONTOLOGY, VOLUME 30

undoubtedly a defence strategy, while flexure in the opposite direction was used in righting the
overturned animal and may also have been a help in burrowing. The spatulate carapace, effaced
features, and broad axial region of the opisthosoma give Pasternakevia a streamlined shape similar
to that of burrowing illaenid trilobites; it might also have been advantageous to a swimming form,
but at small size and slow swimming speeds (low Reynolds numbers) effacement confers little
advantage, so burrowing seems a more likely explanation. The hypertrophic second tergite suggests
enlarged ventral organs: genitalia or, more likely, gills. It could be that enhanced gas- or ion-
exchange abilities enabled Pasternakevia to inhabit a hypersaline environment. Bunodes lunula
Eichwald, 1854 also has a hypertrophic second tergite and is found in similar dolomitic limestones
in Estonia, but it is there accompanied by Pseudoniscus aculeatus Nieszkowski, 1 859, which has a
normal second tergite, so this hypothesis remains speculative.

REFERENCES

BERGSTROM, j. 1975. Functional morphology and evolution of xiphosurids. Fossils Strata, 4, 291-305.

CLARKE, J. M. 1902. Notes on Paleozoic crustaceans. Rep. N.Y. St. Mas. nat. Hist, (for 1900), 54, appendix 3,
83 1 10.

CURRIE, L. D. 1927. On Cvamoceplialus, a new synziphosuran from the Upper Silurian of Lesmahago, Lanark-
shire. Geol. Mag. 64, 153 1 57.

DRYGANT, D. M. 1984. Correlation and conodonts from the Silurian- Lower Devonian deposits of the Volyno-
Podolian, 192 pp. Kiev. [In Russian.]

EICHWALD, E. VON, 1854. Die Grauwackenschichten von Liev- und Esthland. Bull. Soc. imp. Nat. Moscoii, 27,
1 211.

ELDREDGE, N. 1974. Revision of the suborder Synziphosurina (Cheliccrata, Merostomata), with remarks on
merostome phylogeny. Am. Mus. Novit. 2543, 1 41.

and PLOTNiCK, R. E. 1974. Revision of the pseudoniscine merostome genus Cvamocephalus Currie. Ibid.

2557, 11 0.

FISCHER, G., DE WALDHEIM 1839. Noticc sur Lin crustacc fossile du genre Eurvpterus de Podolie. Bull. Soc. imp.
Nat. Moscou,\f\25- m.

FISHER, D. c. 1982. Phylogenetic and macroevolutionary patterns within the Xiphosurida. Proc. N. Am.
Palaeontol. Conv. 1, 175-180.

1984. The Xiphosurida: archetypes of bradytely? In eldredge, n. and Stanley, s. m. (eds.). Living fossils,

196-212. Springer-Verlag, Berlin, Heidelberg, and New York.

FLOWER, R. 1968. Merostomes from a Cotter Horizon of the El Paso Group. Mem. Inst. Min. Technol. New
Me.x. 22 (4), 35-44.

NIESZKOWSKI, J. 1859. Zusatze zur Monographie der Trilobiten der Ostseeprovinzen, nebst der Beschreibung
einiger neuen obersilurischen Crustaceen. Arch Naturk, Liv.-Est. u. Kurlands, (ser. 1), 1, 345-384.
NIKIFOROVA, o. L, PREDTECHENSKY, N. N. and ABUSHIK, A. F. 1972. Silurian and Lower Devonian key sections of
Podolia, 262 pp. Leningrad. [In Russian.]

RUEDEMANN, R. 1916. Account of some new or little-known species of fossils, mostly from Paleozoic rocks of
New York. Bull. N. Y. St. Mus. 189, 7-112.

ST0RMER, L. 1955. Merostomata. In moore, r. c. (ed.). Treatise on invertebrate paleontology. Part P. Arthro-
poda 2, P1-P41. Geological Society of America and Unversity of Kansas Press, Boulder, Colorado and
Lawrence, Kansas.

STURMER, w. and BERGSTROM, J. 1981. Weinhergina, a xiphosuran arthropod from the Devonian Hunsriick
Slate. Palaeont. Z. 55, 237 255.

TSEGELNJUK, P. D., GRITSENKO, V. P., KONSTANTINENKO, L. L, ISHCHENKO, A. A., ABUSHIK, A. F., BOGOYAVLEN-
SKAYA, O. V., DRYGANT, D. M., ZAIKA-KOVATSKY, V. S., KADLETS, N. M., KISELEV, G. N. and SYTOVA, V. A. 1983.
The Silurian of Podolia. The guide to excursion, 224 pp. Kiev. [In Russian and English.]

PAUL A. SELDEN

Department of Extra-Mural Studies, University of Manchester
Manchester M 1 3 9PL
and

DANIEL M. DRYGANT

Institute of Geology and Geochemistry of Fuel Minerals
Typescript received 9 July 1986 Ukrainian Academy of Sciences, Naukova 3a, Lviv 47

Revised typescript received 1 October 1986 290047 Ukraine, USSR

A TAPHONOMIC AND DIAGENETIC CASE STUDY
OF A PARTIALLY ARTICULATED ICHTHYOSAUR

by DAVID M. MARTILL

Abstract. A single carcase of the large Middle Jurassic ichthyosaurian Ophthalmosaums sp. was rapidly
decomposed in well-oxygenated bottom water of the Lower Oxford Clay Midlands basin. Parts of the soft
tissues lying within anoxic sediments were subjected to slower rates of decay and portions of the integument
are now preserved as replacements by bacterial and possibly fungal mats. Elements of the skeleton were
encrusted with epibionts on their upper surfaces. Burial diagenesis has signihcantly affected some skeleton
elements, with the infilling of voids with calcite, pyrite, and sphalerite. Rarely, bone phosphate has been
replaced by pyrite. Compaction of bones and septarian cracking of surrounding concretionary mudstone has
caused crushing and brecciation of trabecular bones. More massive bones with cross-sections capable of
transmitting overburden pressures have resisted compaction.

The partial skeletal remains of an ichthyosaur, Ophthalmosaurus sp. were discovered (by Mr Lez
Fitchett, an employee of French Kier Construction) in the autumn of 1982 at Milton Keynes,
Buckinghamshire, during the construction of Caldecotte Lake. The specimen has been examined in
detail whilst still in situ and has formed the basis of a case-study on the preservation of marine
vertebrates in bituminous shales. The skeleton is now a mounted specimen in Milton Keynes Public
Library, BCM 1983/1008.

A systematic excavation of the specimen was undertaken between 12 October 1982 and 15
October 1982. Samples of the surrounding sediment were collected with the specimen for micro-
palaeontological and sedimentological analysis.

The specimen is that of a mature adult, approximately 5 m long (text-fig. 1). Some skeletal
elements do not appear on the diagram as they were disturbed by the excavating machinery, and
cannot be accurately positioned. Misplaced elements include part of the coracoid, the right? ulna,
part of the rostrum, and numerous digits. Part of the right side of the rib cage was also slightly
disturbed.

Horizon. The stratigraphic distribution and preservation of fossil vertebrates in the Lower Oxford
Clay has been discussed by Martill (1985, 1986). The skeleton was found in the Lower Oxford Clay
(Middle Callovian, Middle Jurassic), lying partly within greenish bituminous shale and enclosed by
a large septarian concretion. Two thoracic vertebrae detatched from the main part of the skeleton
were enclosed in a pyrite concretion. There is no published lithological section for the site, but a
section for the nearby brick pits at Bletchley is given by Callomon (1968) and probably differs only
in minor details.

Beds 17 and 9 of Callomon’s section can be identified in the site at Milton Keynes, and from a
comparison of the concretion with those occurring at Bletchley it appears that this specimen is from
bed 9. Thus the skeleton is from the obductum Subzone of the coronatum Zone.

Locality. Caldecotte Lake is situated on the South side of Milton Keynes at National Grid Reference
SP 892 352. Excavation of the site is now complete and all exposures of the Oxford Clay have
disappeared due to flooding and landscaping of the site.

Taxonomy. The skeleton is that of an associated, partially articulated Ophthalmosaurus sp. Specific
identification cannot be determined as the diagnostic coracoids (Appleby 1956) are not sufficiently
well preserved to show the anterior and posterior notches. Two species of Ophthalmosaurus are

{Palaeontology, Vol. 30, Part 3, 1987, pp. 543-555, pi. 63.|

© The Palaeontological Association

544

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 1. Outline skeletal plan of Ophthalmosaurus sp. BCM 1983/1008 as found in situ at Caldecotte Lake,
Milton Keynes, Buckinghamshire. Certain elements of the skull and shoulder girdle are omitted from the plan

as they were disturbed by earth-moving machinery.

recorded from the Lower Oxford Clay; O. icenicus Seeley 1874, and O. monocharactus Appleby
1956. The differences between the two are small, and considering the degree of variability seen
within Ophthalmosaurus, the differences may be sexual (Andrews 1910).

MATERIAL AND METHODS

The specimen consists of a partially articulated skeleton wanting only a few elements. Due to intense breccia-
tion it is not clear which elements are present in the concretion. Those parts of the skeleton known to be
preserved include all of the posterior portion of the vertebral column from about the ?20th vertebrae to the
tip of the tail. A few of the neural arches, especially the more posterior ones are present. Of the skull there is a
right quadrate, left and right lachrymals, right coronoid, right dentary, angular and surangular, and portions
of the left and right premaxillae. There are also a few plates from the sclerotic ring. The rib cage is almost
complete, although the right side is very disarticulated. The fore limbs are represented by left and right radii,
left and right ulna, and numerous carpals and metacarpals. A highly brecciated humerous was found within
the concretion. The left ischio-pubis, left femur, and fused left tibia and fibula were found with numerous
digits.

The skeleton was found lying in clay, grey/green, rather tenacious and slightly bituminous, and was partly
enclosed by a large septarian concretion. Two detached vertebrae were found enclosed in a flat pyrite con-
cretion. The following excavation procedure was carried out so that the skeleton could be removed to the
laboratories of the Geology Department of the University of Leicester for cleaning, preparing, and subsequent
mounting.

I. The skeleton was completely exposed in situ by careful digging with knives, forks, spoons, and small
trowels. Where the overburden exceeded more than a few centimetres a spade was employed.

MARTILL: ICHTHYOSAUR TAPHONOMY

545

2. The exposed skeleton was overlain with clear acetate sheeting and the position of all the bones was
mapped out using an indelible ink marker pen. Each bone received a unique number on the plan and was
placed in a sealable polythene bag bearing the same number. Where the bone was broken into two or more
pieces, all of the pieces were placed in the same bag.

3. Fractured elongate bones of the lower jaws were collected in their entirety by excavating around them
until they were left lying on an elevated plinth of clay. The clay plinth was undercut using a ‘cheese wire’
technique and a length of square section guttering slid underneath. This was then lifted and individually
wrapped to protect the fragile bones during transit.

4. The specimen was removed to the laboratory for cleaning. During this stage a black and buff coloured
surface coating was discovered on the surfaces of some of the bones. Cleaning of the bones was achieved by
soaking in warm water using only mild detergents to assist break down of the clay.

5. Removal of hard rock (mainly fibrous calcite) adhering to bones found in the clay was achieved by use
of a pneumatic chisel ( Vihrotool) and steel needles.

6. Bones in the concretion were considered unsuitable for use in the reconstruction of the skeleton, but
were used for diagenetic studies. The large calcareous concretion was cut up with a rock saw. Thin sections
were made from the remaining fragments. Bone used for scanning electron microscopy was extracted from
the matrix using 10% acetic acid.

7. Prepared bone was examined by light and scanning electron microscopy. Bone coatings were examined
by both scanning and transmission electron microscopy. Mineral phases were identified using normal petro-
graphic thin sections.

TAPHONOMY

The remains constitute an associated skeleton of a single individual. Much of the skeleton is
disarticulated, but a few elements, notably the left side of the thoracic ribs are articulated, and show
true bone to bone relationships, as in the living animal. Some skull elements enclosed within the
concretion, and the anterior part of the rostrum disturbed by the excavator, are also articulated.
The remainder of the skeleton is disarticulated, but most of the individual bones have not been
moved from their original positions by more than a few centimetres. Thus the overall shape of an
ichthyosaur skeleton is retained.

The cause of death of the animal cannot be satisfactorily determined. Advanced ossification,
indicated by irregular expansions to the ends of the ribs, and by fusing of the ?left tibia and fibula
suggest it was an old animal. There is little disruption of the skeleton to suggest the animal was the
prey of a large carnivore such as a pliosaur, so it is likely that it died of old age or disease.

Apart from the right hind limb, all appendages are present, including the extreme tip of the tail.
This indicates that very little scavenging took place during post-mortem drifting, with only the right
rear paddle possibly missing due to scavenger activity. The carcass arrived on the sea-floor intact,
with most of the soft tissue present, serving to hold the skeleton together.

The position of the skeleton, lying on its left side suggests that it landed on the sea-floor ventrally,
and as the flesh decomposed the skeleton collapsed forwards and on to its left side.

The carcass descended to the sea-floor with a velocity sufficient to allow the tip of the rostrum to
penetrate the sediment. Penetration of the rostrum tip into the sediment indicates that the skull
arrived on the sea-floor first, and may have been suspended below the main body of the carcass.
Presumably the sediment was soft, possibly even soupy, and the sinking velocity of the carcass need
not have been great.

Decomposition of the soft tissue in the water column took place rapidly and left the right side of
the rib cage exposed to sea water. Soft tissues in contact with the sediment decomposed more
slowly, and in the case of some of the tissues which had sank into the sediment, decomposition was
not completed. Text-fig. 2 summarizes the taphonomic history of the specimen.

The bottom waters of the Lower Oxford Clay Midland basin were in general well oxygenated
and capable of supporting a diverse benthos (Duff 1975). The clays around the specimen, however.

546

PALAEONTOLOGY, VOLUME 30

1

2

4

TEXT-FIG. 2. Summary of taphonomic history of BCM 1983/1008 based on observations made from
position of skeleton and state of preservation of skeletal elements. I, carcass drifts in water column
with skull suspended below main body of carcass. 2, carcass descends to sea-floor and part of rostrum
penetrates sediment. Light scavenging may take place. 3, rapid decomposition of soft tissues in well
oxygenated bottom water. 4, collapse of skeleton and encrustation by epibionts.

MARTILL: ICHTHYOSAUR TAPHONOMY

547

yielded only a restricted benthos of nuculacean bivalves, scaphopods, and foraminifera; although
the latter cannot be dehnitely considered benthic. Oysters and serpulid worms were found encrusting
the skeleton but were not found in the surrounding sediment (see below). This suggests that the
skeleton acted as a benthic island. Sea water in contact with the sediment may have been slightly
depleted in oxygen, and the oysters and serpulids encrusting the skeleton survived due to their
elevated position in more oxygenated water. If this is the case then current activity must have been
at a minimum to prevent mixing of the oxygen-depleted water with the oxygen-rich water. However,
this is not the only mechanism for producing the restricted infauna; the soft substrate may also
have been a contributing factor. The restriction of the benthos limited scavenging of the carcass
while on the sea-floor, but the movement of a few bones, notably the two vertebrae in the pyrite
concretion, is not due to current activity, and can almost certainly be attributed to scavenging.
Current activity is ruled out on the grounds that the two vertebrae are large and a current strong
enough to move them would have also moved the smaller elements of the skeleton. If seaweed grew
on the surface of the bones, it is possible that added bouyancy might assist movement during storm
activity. It is, however, difficult to establish if the skeleton lay within the photic zone.

Epibionts. Many of the disarticulated elements of the skelton are pale buff in colour. These bones
are frequently encrusted with epibionts, including oysters and serpulid worms. The oysters are only
found encrusting the buff coloured bones, and are restricted to the upper surfaces (text-fig. 3). No
micro-epibionts have been found on the underside of the skeleton, or on the dark brown bones of
the articulated portions of the skeleton. The oysters are preserved in dark grey calcite. They
frequently reach a length of 4 cm, and on flat bones they remain attached continuously during

TEXT-FIG. 3. Portion of rostrum having penetrated soft sediment. Soft tissues are preserved below oxygen
minimum zone. Epibionts encrust bone in well oxygenated water.

548

PALAEONTOLOGY, VOLUME 30

ontogeny. Oysters encrusting bones with strongly curving surfaces, i.e. ribs, are only attached during
early ontogeny, later stages of shell growth migrate away from the bone surface and the oyster shell
becomes curved. On very smooth bone surfaces the oyster may not secrete shell material, but lie in
direct contact with the bone.

Serpulid worms are less common then oysters, and are usually small, being generally less than 1
cm long. They are preserved as white calcareous conical tubes, approximately 2-3 mm diameter
anteriorly. No geotropism or phototropism has been established, but the distribution pattern on
the skeleton follows that of the oysters.

Soft tissues. The undersides of the articulated parts of the skeleton are dark brown, and devoid of
epibionts. A black coating adheres to the underside of the articulated vertebrae that lie within the
shale, and also to the underside of some of the ribs from the left side of the rib cage. (In life this
would be the outer surface of the left side of the rib cage.) The black coating, overlain by a slightly
reflective white/bufT coating was also found on a portion of the premaxilla.

These coatings are restricted to the dark brown coloured bones. No oyster encrustations are
found on the dark brown bones which suggests that these bones were in contact with, and partly
buried in the sediment. The black and white coatings may be by-products of a decomposing
integument (PI. 63, figs. 1 and 2). The abundance of pyrite within the sediment and encrusting some
of the bones shows that reducing conditions were present within the sediment. If the oxic/anoxic
boundary was close to the sediment water interface, the penetration of the ichthyosaur carcass into
the sediment may have introduced some of the soft tissues to reducing conditions, thus reducing
the rate of decay.

Scanning electron microscopy of the black coating from the underside of the vertebral column
shows it to be composed of an amorphous mass of carbonaceous material, underneath which are
numerous ovoids approximately 1 /nn long, and about 0-5 j.im diameter (PI. 63, figs. 3 and 4). These
ovids are interpreted as lithified bacteria, similar to those reported from soft-part outlines of Eocene
frogs and bats (Wuttke 1983). The bacteria represent a replacement of some of the original soft
tissues. In the case of the ichthyosaur Steuopterygius spp. from the Posidonia Shale (Toarcian,
Lower Jurassic) of Holzmaden, West Germany, these black coatings may extend into the sediment
to produce an outline of the entire animal (McGowan 1979). In the Lower Oxford Clay the micro-
environment in which this process can occur was restricted to the undersides of bones lying within
anoxic sediment, and has prevented complete outlines from being preserved.

DIAGENESIS

Apart from compaction and mineralization effects the skeletal elements are preserved in two distinct
ways: as dark brown bones with a smooth surface; or, as light buff" bones with a soft powdery
surface. The powdery surface is considered to be an effect of prolonged exposure to sea water, but
might also be attributable to the encrusting of the bone surface by marine algae. Internally, the
light buff bones are indistinguishable from the dark brown bones. Parts of the rib cage of the right
side of the skeleton were freshly fractured due to the weight of the earth-moving machinery. It can
be seen that in some of the ribs no mineralization of the voids in the bone has taken place. This has

EXPLANATION OF PLATE 63
Preserved bacterial mats on surface of ichthyosaur bones.

Pig. 1. Rostrum of Ophthalmosaurus sp. BCM 1983/1008, showing black film overlain by buff coloured
reflective coating, x 1 .

Fig. 2. Underside of thoracic centrum of Ophthalmosaurus sp. BCM 1983/1008 showing black film only, x 1 .
Fig. 3. SEM of lithified bacteria composing black film, x 10 000.

Fig. 4. Transmission electron micrograph of ultra thin section from black film taken from below thoracic
vertebrae showing electron dense bacterial bodies. Osmium stained, x 30 000.

PLATE 63

MARTILL. ichthyosaur taphotwtny

550

PALAEONTOLOGY. VOLUME 30

TEXT-FIG. 4. Ultra structure of bone from Ophthalmosaurus sp. revealed by SEM after preparation in 10 %
acetic acid, a, portion of jaw showing lacunae, x 100. h, detail of single lacuna with lining of newly mineralized
collagen fibrils, x 400. c, tangled webs of mineralized collagen fibrils from vertebral centrum overgrown by
small euliedral pyrite crystallites, x 520. r/, bundles of collagen fibrils twisted in rope-like fashion from highly

trabecular vertebral centrum, x 900.

made the bones very fragile. In other bones, however, the void spaces have been filled with a variety
of mineral phases (see below), and are more robust. Scanning electron microscopy of both
the bone surface and fractures of bone trabeculae etched in acetic acid show that no alteration of
the phosphatic bone matrix has taken place during burial diagenesis.

Bone ultra-structure. Thin sections and acetic acid etched samples of the bone show that the
phosphatic matrix of the bones from this specimen have remained relatively unaltered since the
death of the animal, and that the structures observed in thin sections, and with the electron
microscope, are primary features. In thin section the trabecular bone of vertebrae and ribs is seen
to be rich in lacunae and canaliculae, most of which have remained as voids within the bone; only
in a few sections have these been filled with diagenetic minerals, notably pyrite. High power scanning
electron microscopy of the internal surface of lacunae (text-fig. 4a, b) shows an irregular mass of
phosphatized collagen fibres, which at ultra high-power show evidence of banding. The surface of
the bony trabeculae is likewise unaltered, even after the effects of mineralization. Parts of a vertebral
centrum filled with late ferroan calcite were etched in 10% acetic acid until all the carbonate phase
had been removed. An examination of the prepared surface showed bundles of phosphatized
collagen fibrils lying roughly parallel to the surface of the trabeculae. These are occasionally
transgressed by isolated fibrils of phosphatized collagen. In some parts of the bone the parallel

MARTILL: ICHTHYOSAUR TAPHONOMY

551

fibrils become tangled masses, and single bundles may bifurcate (text-fig. 4c, d). Each bundle of
phosphatized collagen consists of nine or more individual fibrils, all of which are entwined like
rope. The bundles are approximately 3 /.im diameter, with individual strands less than 200 nm
diameter. The longest individual strands are over 9 /nn long, but the bundles are several times
longer than this.

Good resolution work at high power is difficult to achieve, but when possible, the banding on the
phosphatized collagen fibres appeared to be due to spaces 15 nm across, between individual crystal-
lites of apatite approximately 1 00 nm across. The organic matrix of the bone was not present.

Diagenetic minerals

Pyrite. No alteration of the original mineral matrix of bone has taken place in those elements found
within the shales or the large calcareous concretion, but there has been some replacement of
phosphatic material in the two thoracic vertebral centra preserved in the pyrite concretion. Thin
sections of the pyrite concretion show that lacunae and canaliculae, and void spaces in the trabecular
parts of the bone are filled with pyrite. In some parts of the concretion the pyrite appears to
have spread outwards from lacunae and pyrite filled voids to replace the bone material itself
(text-fig. 5(7).

Pyrite is also abundant in the trabecular bone as aggregates of pyrite octahedra. The aggregates

TEXT-FIG. 5. Thin section through over pyritized trabecular bone, a, pyrite (black) filling lacunae, void spaces,
and partially replacing bone, x 20. b. chains of pyrite framboids on surface of bone. x20. r, blade-like
pyrite possibly pseudomorphing marcasite growing tangential to bone surface, x 20. cl, detail of blade-like

pyrite. x 20.

552

PALAEONTOLOGY, VOLUME 30

may completely fill voids in the bone, but frequently aggregates are less than 1 /im diameter and
pyrite crystallites may represent the activity of a single sulphate reducing bacterium. These aggre-
gates are seen in thin section as rounded bodies, isolated or in chains (text-fig. 5h). Individual
octahedra are less than 3 |(m diameter, but are well-formed crystallites, whereas the aggregates are
up to 20 //m diameter, but the crystallites are less distinct, giving the aggregates a granular appearance.

Massive pyrite
with burrows

Euhedral pyrite

Uncrushed bone

Crushed bone

Calcite

TEXT-FIG. 6. Disc-shaped pyrite concretion overgrowing thoracic centrum. Pyrite formation has preserved
burrows in the sediment and prevented the centrum from suffering severe compaction damage, but some
crushing has still taken place. Septarian cracking of the concretion has also affected the bone.

Concretionary pyrite is abundant in parts of the Lower Oxford Clay where it can be found
overgrowing ammonites and fossil wood. It commonly forms flat disc-like concretions in the plane
of the bedding. The concretionary pyrite on the two vertebral centra of this specimen is of a
similar habit (text-fig. 6). Each vertebra is totally overgrown, and extensions of the pyrite extend
horizontally, uniting the two centra. Two forms of pyrite are present, an amorphous pyrite mudstone
found on the outside of the centra, and a micro-crystalline pyrite found lining void spaces in the
bone, and lining later fractures in the amorphous pyrite.

The bone within the pyrite concretion has a few microfractures, and has been subjected to slight
compaction, but it is apparent that the bones in the pyrite concretion have resisted compaction
more than the bones in the shales.

Sphalerite. Small quantities of euhedral sphalerite are found in early fractures between bones, in
large voids in trabecular bone, and in the tooth groove of the premaxilla and dentary. Occasionally
sphalerite is found on the surface of the bone, but it appears to be most common in areas that have
not been subjected to compaction. Sphalerite post dates in part the main compaction phase, as it
can be found filling cracks in brecciated bone and surrounding bone shards (text-fig. lb), but it is
not found in fractures in the large calcareous concretion. It is considered to be of a later origin
than the pyrite, post compaction of the bone, but pre-brecciation of the concretion.

Calcite. Late ferroan calcite is abundant, and found filling voids in bones, small fractures in the
pyrite concretion, and large fractures in the calcareous concretion. In uncompacted trabecular bones
it is found as coarse crystals completely filling cavities (text-fig. la). Crystal boundaries are irregular,
suggesting some pressure solution at the boundaries. No fringing cements have been observed.

MARTILL: ICHTHYOSAUR TAPHONOMY

553

TEXT-FIG. 7. Void filling minerals in trabecular bone, n, trabecular bone with coarse-grained ferroan calcite
filling void spaces. Crossed nicols. x 20. euhedral sphalerite (dark grey) filling void and fracture in trabecular
bone (light grey) with later ferroan calcite (mottled). Transmitted light, x 20.

Many of the bones in the bituminous mudstone are coated with fibrous ferroan calcite (beef) up
to 2 mm thick, with a thin film of clay sandwiched in between. This fibrous coating causes problems
for the preparator as it requires removal with a vibrotool.

COMPACTION

Compaction has had a deleterious affect on the specimen, and has resulted in differential preser-
vation of the skeleton. Many of the elements of the skeleton that were lying within the bituminous
mudstone have been uniaxially flattened by compaction. Failure of the bones is of a brittle nature,
often with the complete shattering of all the inner trabeculae. The more solid margins of bones
have resisted compaction, as have those bones with shapes that can transmit overburden pressures
around their surface.

Compaction of some of the vertebrae has been greater than 50 % and has been unaffected by the
position of the bone in the sediment. Vertebrae lying flat on the bedding planes have been flattened
anteroposteriorly, while those lying vertically are flattened dorsoventrally. Confining pressure of
the sediment has kept the brecciated bone together, and later cementation by ferroan calcite has
allowed individual bones to be collected entire, although severely crushed. During acetic acid
preparation the compacted elements of the skeleton fall into thousands of bone shards most of
which are less than 1 mm long.

Early formation of the calcareous concretion has prevented bones enclosed by the concretion
from being compacted, but compaction has caused the concretion to brecciate (Hudson 1978). Wide
cracks have developed in the concretion which have penetrated the bones. Geopetal fabrics can be
observed where shards of bone have fallen to the bottom of the cracks (text-fig. 8). Uncompacted
clay has been squeezed into the cracks, and has penetrated cavities in the bone. Differential move-
ment of the brecciated concretion and re-cementation of the fragments by coarsely crystalline
ferroan calcite has resulted in a bone breccia.

Formation of the pyrite concretion appears to have post-dated the formation of the calcareous
concretion, as the two vertebral centra preserved within the concretion show a slight degree of
compaction, but this is not as severe as that which affected the bones in the shale.

Compaction has not affected the microstructure of the most highly compacted vertebrae, where
lacunae and canaliculae can still be observed. Thus failure due to compaction is entirely of a brittle
nature, with no observable alteration due to pressure solution or recrystallization. All fractures are
clean and sharp and some appear to be controlled by the cleavage of void filling calcites.

554

PALAEONTOLOGY, VOLUME 30

Pyrite /calcareous
mudstone

">

Calcareous mudstone

m

Voids

:::x| BOHe

Calcite

TEXT-FIG. 8. A composite section through mudstone concretion illustrating diagenetic and compactional
features affecting the preservation of enclosed skeletal elements of Ophthalmosaunis sp. Part of BCM

1983/1008.

CONCLUSION

The Milton Keynes specimen of Ophthalmosaurns died as a large adult, with disease possibly being
a contributing factor to its death. The carcass sank rapidly to the sea-floor and was almost unaffected
by scavengers. It lay partly buried in the sediment, where parts of the soft tissue underwent a long
slow period of incomplete degredation by sulphate-reducing bacteria. The upper part of the carcass
underwent rapid decomposition, and the exposed skeleton was encrusted by a restricted, but
abundant epifauna. Parts of the skeleton became detached due to a combination of benthic scaveng-
ing, drifting due to adhering seaweed and perhaps storm activity. After burial, early formation of a
calcareous concretion occurred around the anterior part of the post cranial skeleton. Compaction
due to burial crushed many of the more trabecular, and flat bones, and also caused brecciation of
the concretion, but later compaction caused some septarian cracking of the pyrite concretion also.
Small quantities of sphalerite formed in lower pressure areas after an initial compaction phase. The
three broad preservational styles in which this specimen occurs; in compacted shale, calcareous
concretion, and pyrite concretion, are typical of vertebrates in the Lower Oxford Clay of the South
Midlands, but at Peterborough pyrite is less abundant and normally only found as thin films on
the surface of bones. At Peterborough septarian concretions occur around vertebrates in the Jason
Zone (Martin 1985), although septarian, brecciation is generally less severe than in concretions
from the Milton Keynes district.

Acknowledgements. Mr Rod Branson helped with the preparation of the skeleton plan and assisted with the
cleaning of the specimen. Bill Teasdale and Arthur Meadows gave valuable technical assistance with the
mounting of the specimen. Thanks to Anglia Waterboard for providing on site facilities, including the use of
a Landrover. Thinks to Jill Martill for patiently typing the manuscript.

REFERENCES

ANDREWS, c. w. 1910 1913. ^ descriptive catalogue of the marine reptiles of the Oxford Clay., vol. 1, 205 pp.,
10 pis. (1910); vol. 2, 206 pp., 13 pis. (1913). British Museum (Natural History), London.

APPLEBY, R. M. 1956. The osteology and taxonomy of the fossil reptile Ophthalmosaunis. Proc. zool. Soc. Land.
126, 403-447.

MARTILL: ICHTHYOSAUR TAPHONOMY

555

CALLOMON, j. H. 1968. The Kellaways Beds and the Oxford Clay. In sylvester-bradley, p. c. and ford, t. d.
(eds.). The geology oj the East Midlands, 246-290. University of Leicester Press, Leicester.

DUFF, K. L. 1975. Palaeoecology of a bituminous shale— the Lower Oxford Clay of Central England. Palaeon-
tology, 18,443-382.

FISHER, I. ST J. 1983. Studies on the Formation of pyrite in Jurassic shales. Ph.D. thesis (unpublished), Leicester
University.

HUDSON, J. D. 1978. Concretions, isotopes, and the diagenetic history of the Oxford Clay (Jurassic) of central
England. Sedimentology, 25, 339- 37().

MARTILL, D. M. 1985. The preservation of marine vertebrates in the Lower Oxford Clay (Jurassic) of central
England. Phil. Trans. R. Soc. B311, 155 165.

1986. The stratigraphic distribution and preservation of fossil vertebrates in the Oxford Clay of England.

Mercian Geologist, 10(3), 161-188.

MCGOWAN, c. 1979. A revision of the Lower Jurassic ichthyosaurs of Germany with descriptions of two new
species. Palaeontographica, A166, 93-236.

SEELEY, H. G. 1874. On the pectoral arch and fore limb of Ophthalnwsaurus, a new ichthyosaurian genus from
the Oxford Clay. Q. Jl. geol Soc. Fond. 30, 676-707, pis. 45-46.

WUTTKE, M. 1986. A ‘Weichteil-Erhaltung’ durch lithifizierte Mikroorganismen bei mittel-eozanen vertebraten
aus den Olschiefern der ‘Grube Messef bei Darmstadt. Senckenberg. leth. 64 (5/6), 509 527.

DAVID M. MARTILL

Department of Geology
Eield Museum of Natural History
Chicago, Illinois, 60605
USA

Typescript received 4 August 1986

Revised typescript received 1 1 November 1986

Present address

Department of Earth Sciences
Open University, Walton Hall
Milton Keynes MK7 6AA, UK

V.

,

' \

■A--

■n. ' ■■ •'.* / vt'ts*i!f»'€l^;*A;.'.v;

, ■>:• ■•• ■' •■ •'<■

•• • . ^ -V *•• ♦ . •

- ■''', ^ • ' J< “

i, ^ , , , } ."I ■■

V

(

J,

:oi

THE ARMOURED DINOSAUR POLACANTHUS FOX I
FROM THE LOWER CRETACEOUS OF
THE ISLE OF WIGHT

by WILLIAM T. BLOWS

Abstract. A new specimen of the Lower Cretaceous nodosaurid ankylosaur Polaccmthus foxi was found in
the Wealden formation of the Isle of Wight in 1979. Three types of presacral spine are now recognized and
new dermal elements of unknown position were found. Cranial and cervical fragments are described for the
first time, including an axis formerly identified by Seeley, 1876, as IJgiumodon. The sacrum shows a widening
of the neural canal, as in Stegosaurus and others, and the tail plates differ from those of the holotype. The
systematic status of P. foxi is reviewed, and comparisons are made with other contemporary nodosaurid taxa
as far as possible. The genus Polacanthus appears to be valid despite suggested synonymy with Hylaeosaurus,
it may, however, have closer affinities with comtemporary taxa of the New World, notably Hoplitosaurus.
Polacanthus appears to have a greater geological range than Hylaeosaurus.

Lee (1843) described and illustrated three pieces of saurian dermal plates which, at the time, were
unknown to science. These specimens can now be identified as fragments of sacral dermal plate of
Polaccmthus. The site of these pieces is recorded as the ‘Hastings Sands’ of Sandown Bay, Isle of
Wight. According to Rawson et al. (1978) the Hastings Beds do not outcrop on the Isle of Wight,
and the specimens probably came from the Wealden Marls (Wessex Formation), the usual horizon
for Polacanthus. Not only is this possibly one of the earliest descriptions of Polacanthus material
(without using the name), but Lee’s manuscript has a footnote telling of the loss of these specimens
in a hackney coach.

In 1865 William Fox excavated a partial skeleton of an armoured dinosaur from Barnes High,
near Atherfield, Isle of Wight (Blows 1983). Much of the pelvis, hind limbs, and verterbral column
were collected together with considerable dermal armour, which was made the type of P. foxi. A
journal (Anon. 1865) described this discovery in a brief notice accompanied by a drawing of a
collection of these bones. This account also included the name Polacanthus for the first time, and
attributed this name to Owen. Huxley (1867) mentioned the name during an account of Acantho-
pholis, and again attributed it to Owen. However, Huxley is the earliest known author of the name.
Hulke (1881, 1887) made a full description of the remains when they arrived in London in 1881.
Again he attributed the name to Owen.

Since this time, only one other important specimen of Polacanthus has been described (Lydekker
1891) also from the Isle of Wight. It is a portion of the pelvis with some overlying dermal armour.
This armour was thought to be smooth rather than patterned (as in P. foxi) and this lead to the
suggestion that it could be a new species (Seeley 1891), later called P. hecklesi Hennig 1924, after
the collector Mr Beckles. However, this species is here synonymized with P. foxi. Other genera of
contemporary age are Hylaeosaurus Mantell 1833, Polacanthoides Nopcsa 1928, and Hoplitosaurus
Lucas 1902, all based on inadequate material, and their affinities with Polacanthus are poorly
understood.

The object of this paper is to describe the most recent discovery of Polacanthus (Blows 1982)
which was collected by the author from Compton Bay beach. Isle of Wight from 1979 to 1984.

Abbreviations. AMNH, American Museum of Natural History; BMNH, British Museum (Natural History);
BYU, Brigham Young University; GM, Gosport Museum; KU, Kansas University; MIWG, Museum of Isle

IPalaeontology, Vol. 30, Part 3, 1987, pp. 557-580.|

© The Palaeontological Association

558 PALAEONTOLOGY, VOLUME 30

of Wight, Geology; CAMSM, Sedgwick Museum Cambridge; USNM, United States National Museum;
YPM, Yale Peabody Museum.

SYSTEMATIC PALAEONTOLOGY

Order ornithischia
Suborder ankylosauria
Family nodosauridae
Genus Polacantlnis Huxley 1867

Type species. F. /b.v/' Hulke 1881.

Holotype. BMNH R175, partial skeleton with derpial armour (Eox Collection).

Type locality. Barnes High, West of Cowleaze Chine, Atherfield, Isle of Wight (NGR SZ 443 806).

Type horizon. Wealden Marls (Wessex Formation), Lower Cretaceous (Barremian).

Range. Wealden Marls (Wessex Formation) to Ferruginous Sands (?) (Aptian).

Diagnosis. Moderate to large size nodosaurid (Coombs 1978); presacral series of five fused vertebrae;
sacrum of five fused vertebrae; maximum sacral canal expansion at S2 level; five sacral ribs; long
posterior dorsal ribs supporting overlying dermal armour and adjoining anterior ilium; ribs flat
dorsally, supported by a ventral ridge giving a T-shaped cross-section; anterior caudal vertebrae
with long lateral processes, thickened neural process with a supraspinous notch; caudal series
terminating in a vertebral-dermal mass with ossified tendons (?primitive club); presacral dorsal
spines attain maximum heights over the shoulder region, reducing in height both anteriorly and
posteriorly; the spines have flattened bases, the dorsal keel in large specimens twist through nearly
90° from base to apex, presacral spines mounted in a double row laterally to the spinal column;
presacral ‘lumbar’, sacrum, and both ilia covered by a large, flat dermal plate of armour, approxi-
mately 1 m square, ornamented on the dorsal surface by ossifications; caudal armour of tall upright
or short roof-like plates in double row, descending in height posteriorly and having narrow, hollow
bases; rounded, oval, or subtriangular ossicles of sizes up to 110 mm across.

Referred specimens. BMNH R9293, three skull fragments (?nasal, supraoccipital, angular); left neural arch of
atlas; CAMSM B5337I, axis vertebra; BMNH R9293, (?)fourth cervical vertebra, four dorsal vertebrae and
vertebral processes; sacrum with part neural arch; two caudal vertebrae (?first and second); two fragments of
ilium; ?portion of ischium; two ribs complete; parts of three more ribs; rib head and fragments; terminal
phalanx; ossified tendons; numerous endoskeletal fragments; five dermal plates (caudal series); four dermal
spines (presacral series), two dermal spines (presacral series), one spiked plate (?caudal); ten large pieces of
sacroiliac shield; fifteen smaller pieces of sacro-iliac shield; thirty-six dermal ossicles; CAMSM B53353,
dermal ossicle; CAMSM B53354-53358, five dermal ossicles; CAMSM B53588-53591, four ossicles; CAMSM
B53594-53597, four dermal ossicles; BMNH R9293, numerous dermal fragments; small complete bone (one
side of bilateral pair) of unknown origin; CAMSM B53372, small complete bone (the other side of bilateral
pair) of unknown origin.

Locality. The remains lay scattered within a confined pocket exposed at very low tide on Compton Beach, Isle
of Wight (NGR SZ 374 845).

Horizon. The site occurs in the lowest beds of the Wealden Shales (Vectis Formation) and represents the first
recorded find of Polacanthus from this stratigraphic unit (Dr A. Insole, pers. comm.). The bed is a pale grey,
non-fissile, massive clay generally devoid of fossils. The strata dip strongly to the west and can be traced in
the corresponding cliff section.

DESCRIPTIVE ANATOMY

The skeleton was disarticulated with bones at different depths. They may have been eroding out for
some time. The axis vertebra (described Seeley 1876) and fourteen pieces of dermal armour were

BLOWS: ARMOURED DINOSAUR POLACANTHUS

559

registered at Cambridge Sedgwick Museum between 1860 and the 1940s, and are considered to be
the same skeleton as R9293. The (?)fourth cervical vertebra was registered with the British Museum
(Natural History) (R9293) in the 1960s, and the main pocket of bones was found and excavated by
the author from 1979 to 1984. Other elements may have washed out over this time span to be
destroyed by sea action, or be held by private collectors.

The area containing the remains was about 4 square metres. The lack of duplicated bones, beyond
that found within one skeleton, plus the rarity of nodosaurid finds in the Weald, suggests the
presence of a single animal. Large bones were smashed (e.g. pelvis and overlying armour), and
scattered with some loss.

Skull. The holotype of Polacanthus is without a skull. A referred specimen, a basisphenoid (BMNH
R4951) from the Isle of Wight, is suggestive of nodosaurids but cannot be assigned to a genus with
any certainty. A ‘lower jaw’ (BMNH RI75X) has been identified as a fragment of an Igitanodon
ilium (D. Norman, pers. comm. 1976). The new skeleton has two skull fragments and one possible
jaw fragment.

Nasal. This is a flat bone with only one edge and one corner intact. One surface is rounded with
grooves, the other is gently depressed with flat dermal ossifications. The preserved edge appears to
be part of the medial suture across the skull roof.

Supraoccipital. This is a fragment of bone found loose on the site. It is massive compared to
Stegosaurus and Silvisaurus, with a deeply roughened sutural surface for articulation with the
paraoccipital bone, and a smoother articulation surface for the parietal bone. The ventral surface
is grooved, the roof of the brain case, and together with the opposite smooth dorsoposterior surface,
wedges towards the foramen magnum.

Lower jaw. The jaw fragment is probably the left angular (Gallon, pers. comm. 1983). It is a boat-
shaped bone with a deeply concave inner surface, and a convex outer surface bearing a dermal
ossification. One edge is smooth and sharp, the other is irregular and gently curved.

The dentition of Polacanthus is unknown.

Post-cranial skeleton

Cervical vertebrae. The entire cervical series is missing in the holotype. A single cervical vertebra,
considered as part of the type (BMNH R175) is referable to an ornithopod, probably Iguanodon.
In the new specimen, one cervical element (from the atlas vertebra) is known from the site. However,
an axis vertebra (SMES B53371) and a (?)fourth cervical (BMNH R9293) are from the same locality
and probably from the same skeleton (test-fig. 1). The atlas as preserved, consists of only the left
neural arch. The right neural arch and intercentrum are missing. The atlas and axis vertebrae were
not fused in Polacanthus. The atlas corresponds well with Stegosaurus as described by Gilmore
(1914) except the narrowing below the processes, being more constricted anteriorly and overlain by
a pronounced prezygapophysis. The neural arch was not fused to either the intercentrum or the
neural arch of the other side along the mid-line, which corresponds with a sub-adult status. The
posterior process is missing.

The axis vertebra. Seeley (1876) first described this bone (CAMSM B53371) and regarded it as
‘probably Iguanodon . The characteristics are those of general ornithischian dinosaurs and are well
described by Seeley (text-fig. 1a). The presence of both diapophysis and parapophysis indicate the
double-headed nature of the cervical rib.

OFourth cervical vertebra. Parts of the neural process, the left pre- and postzygapophysis and left
lateral process only are missing in this specimen (BMNH R9293) (text-fig. 1b, c). The centrum
is amphiplatyan, the articular surfaces being nearly a third greater in width than height and
centrally depressed. The large neural canal is ovoid, being greater in height than width. The
floor of the neural canal dips downwards to create a central cavity within the centrum. The

560

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 1. Polacanthus foxi. Lower Cretaceous, Isle of Wight, a, SMES B53371, axis vertebra, left
lateral view, b, c, BMNH R9293, (?)fourth cervical vertebra, b, right lateral view; c, anterior view, d, e,
BMNH R9293, dorsal vertebra, d, right lateral view; e, anterior view, f, g, BMNH R9293, dorsal
vertebra. F, right lateral view; G, anterior view, h-j, BMNH R9293, dorsal vertebra, h, right lateral

view; i, anterior view; J, posterior view.

pre- and post-zygapophyseal articular surfaces lay approximately on the same horizontal level, the
prezygapophysis only extending beyond the limits of the centrum. The neural process is inclined
posteriorly.

Dorsal vertebrae. Four free dorsal vertebrae exist in both the holotype and the new specimen. These
latter are better preserved than the type, but in both, the vertebrae consist of two almost complete
(R175 numbered Cl 2, C14 and R9293 numbered 3, 4); one with only half a centrum (R175
numbered Cl 6, R9293 numbered 2); one being a centrum only (R175 numbered Cl 9, R9293
numbered 1). Other isolated dorsal vertebrae (BMNH 2527; MIWG 5188; CAMSM B53587) are
worthy of description. Hulke (1881) and Nopcsa (1905) described the main features of the dorsal
vertebrae, but in the new specimen the floor of the neural canal dips into the centrum in a V-
shaped manner, with a maximum depth half-way along the vertebral length, similar to the cervicals
previously described (text-fig. 2a, b). It occurs in all four vertebrae but is obscured by matrix in the
holotype, although fractures across the centra indicate this feature.

An isolated dorsal vertebra (BMNH 2527) from Barnes High, Isle of Wight (Hulke Collection)
is much larger than the other Polacanthus Vertebrae (see table of measurements). Another dorsal
vertebra from Atherfield, Isle of Wight (SMES B53587), is almost complete. It has transverse
processes angled at 60° rather than the usual 90°. A dorsal centrum (MIWG 5188) is complete and
unworn. Sutural grooves on each side of the neural canal suggests a juvenile nature. The neural
canal floor dips as previously described although partly filled with matrix.

BLOWS: ARMOURED DINOSAUR POLACANTHUS

561

A

B

TEXT-FIG. 2. Polacanthus foxi. Lower Cretaceous, Isle of Wight, a-e, BMNH R9293, vertebrae showing neural
canal (nc). A, cervical vertebra; b, dorsal vertebra; c, anterior sacral vertebra; d, sacral vertebra at S2 level; e,

caudal vertebra.

Sacrum. The sacrum (text-fig. 3a, b) is well understood from the holotype which has five vertebrae,
with a further five presacral (‘lumbar’) vertebrae fused into a presacral rod. The holotype has part
of the overlying dermal plate and ossified tendons attached. The new specimen of five sacral centra
and one presacral centrum is free from attachments, and the neural canal is fully exposed. Nopcsa
( 1 905) stated ‘There does not seem to exist any especial widening out of the neural canal in the sacral
region, as recorded for Stegoceras (sic- Stegosaurus) and also visible in Dacentrurus (Omosaurus)'.
However, sacral neural canal expansions are reported in other nodosaurids: Sauropelta (Ostrom
1970), Nodosaurus (Lull 1921), Panoplosaurus (Sternberg 1921), and Silvisaurus (Eaton 1960), and
some ankylosaurids (Maryanska 1977). In BMNH R9293 an extensive broadening of the canal
reaches a maximum width and height at the level of the second and third sacral vertebrae, confirming
this feature in Polacanthus. In this respect Polacanthus resembles Stegosaurus, where a neural
expansion occurs over the anterior half of the sacral canal (Gilmore 1914). The holotype sacrum
bears five pairs of sacral ribs, the first pair are least robust, arising between the centre of the
posterior presacral and first sacral vertebrae. The following three pairs are the most stout, and the
posterior ribs arise from the centrum of sacral 5. The sacral ribs and posterior neural arch of
the new specimen are missing. A third sacrum from the beach near Whale Chine, Isle of Wight, in
a block of sandstone (BMNH no number) bears all the sacral ribs which unite distally prior to the
iliosacral joints, as in the holotype. The ventral surface only is exposed, the first two anterior
presacrals are missing and two sacral centra are eroded. This is the first recorded discovery of
Polacanthus in the (?)Ferruginous Sands (F. A. Middlemiss, pers. comm.).

Caudal vertebrae. Nopcsa (1905) noted a discrepancy in the holotype tail vertebral count; there are
in faet twenty-one preserved. A tail-end mass consisting of two verterbrae, dermal armour, and
ossified tendons are included in this number (text-fig. 3f-g). Also the first caudal vertebra is attached
to the posterior sacrum at an angle by matrix and may have been regarded as saerum, or sacrocaudal,
and remained uncounted. Both Hulke (1881) and Nopsca (1905) missed this vertebra in their text,
regarding the first free vertebra as caudal 1, although Hulke (1887) figures and labels this vertebra
(c.v., caudal vertebra). The tail-end mass has a rod-like bone tapering to almost a point representing
the termination of the vertebral column. This curves dorsally towards the tip and is hidden mostly
by overlying dermal ossifieations. It must be a fusion of terminal vertebrae. Two larger dermal plates
attached to the mass represent the termination of a tail long bilateral row of plates. Only one surface
of the right plate is exposed, the left plate shows two surfaces with a sharp keel-edge. Irregular
dermal ossieles are scattered over the vertebral rod between and beyond the plates up to the tail

562

PALAEONTOLOGY, VOLUME 30

F

TEXT-FIG. 3. Polacanthus foxi. Lower Cretaceous, Isle of Wight, a-e, BMNH R9293, vertebrae, a, sacrum and
anterior neural arch in left lateral view, ps— presacral, s — sacral vertebrae; b, sacrum in superior view showing
neural canal (nc); c, first caudal vertebra, left lateral view; d, first caudal vertebra, anterior view;
E, second caudal vertebra, posterior view. f-g. BMNH R175, caudal end mass, f, right lateral view, cv—
caudal vertebra, da~dermal armour, r— rod of bone, t — tendons; G, restored caudal end mass, right lateral

view (see f for abbreviations).

BLOWS: ARMOURED DINOSAUR POLACANTHUS

563

TEXT-FIG. 4. Polaccmthus foxi Lower Cretaceous, Isle of Wight. BMNH R9293 sacrum and ribs. Left'. Sacrum,

lateral and superior views. Right'. Ribs.

tip. A group of tendons run throughout the length of the mass along the ventral surface to end at
the tail tip. The two caudals found with R9293 are most likely to be the first and second. The
second caudal has lateral processes which are thin and long, and both have stegosaur-like expansions
of the neural process with a supraspinous notch. A small proximal vertebra (BMNH R4952) from
Grange Chine, Isle of Wight, may be the first caudal of a junior individual. A caudal vertebra
(MIWG 5144) from Brook Bay, Isle of Wight, is a heavily pyritized, water worn centrum.

Ribs. The holotype has ten free anterior dorsal ribs (labelled left: L4, 5, 7, 8, 9, 10; and right: R3, 6;
with two unlabelled) and ten posterior dorsal ribs, five each side attached to the overlying sacral
armour. The new specimen has one complete anterior dorsal rib (text-fig. 4), fragments of three
others, a rib head attached to a dorsal vertebra, one almost complete posterior rib with evidence of
sacral shield attachment, and a few rib fragments. Hulke (1881) describes the ribs sufficiently to
establish their major features, including those of the type that are found on the ventral sacral shield
(Hulke 1887). These arise from the presacral ‘lumbar’ vertebrae, although their attachment to the
vertebrae is now missing in the holotype. The broad dorsal surfaces of these ribs, in contact with
the sacral shield above, are supported by a narrow ventral ridge throughout much of their length.
The most anterior of these ribs are the longest, extending flat, bilaterally to beyond the medial edge
of the ilia, the posterior three appear to fuse with the ilia. A similar situation, without the sacral
shield, occurs in Euoplocephalus (Coombs 1978, text-fig. 13). These ribs would not have needed to
curve very much proximally in order to unite with the presacral vertebrae in the holotype. However,
ribs of a similar nature (new specimen R9293 and R4134) show considerable curvature, suggesting
that the vertebral column was deeper set than the holotype shows. This is further supported by the

564

PALAEONTOLOGY, VOLUME 30

Steep angle of the lateral process on the sacral neural arch of the new specimen. Nopcsa (1905)
draws attention to evidence that shows a space existed between the ribs and sacral shield. There is
evidence that suggest the pelvic region at least, in the holotype, has been crushed (see pelvis below).

It is possible that the dorsal ribs anterior to the sacro-lumbar unit was deep set below the
maximum height curvature of the ribs, and this is supported by the steep angles of the vertebral
lateral processes found in dorsal vertebrae (Nopcsa 1905). This feature is seen in Palaeoscincus
(Gilmore 1930, fig. 1 1) where the neural process hardly reaches the maximum rib height. Presacral
ribs differ from anterior dorsal ribs by a strong ‘T’ cross-section, the dorsal platform roughened by
fibrous tissue compared to the smooth ventral surfaces.

Pelvis. The pelvis is well understood from the holotype (Hulke 1 887). Two fragments of ilium found
with the new specimen are poorly preverved and offer little new information. An acetabulum with
proximal end of the ischium was badly worn before burial. An anterior fragment of ilium shows
the dorsal surface not seen in the holotype because of the overlying sacral shield. This surface is
generally fiat but roughened, especially along the lateral border. A portion of pelvis (BMNH R1926,
Heckles Collection) was considered to be a different species of Polacanthus (P. hecklesi Hennig
1924) because of smooth dermal armour and greater thickness of ilium than the holotype (Seeley
1891). It is the medial aspect of the acetabulum with attached eroded sacral ribs and proximal end
of ischium, apparently uncrushed but water worn. The fragment of ilium from R9293 is broken
but undistorted, and demonstrates varying thicknesses approaching that of the Heckles specimen.
Possibly, the holotype may have been subject to earth pressure causing an artificial thinning of the
ilia (Lydekker 1891). This is not a factor on which to found a new species (the question of smooth
dermal armour will be taken up later).

Seeley (1891) described a bone suggested by him to be the pubis of Polacanthus (removed by
Nopcsa 1905, to an Iguanodon-like genus), the pubis of nodosaurs being much reduced to a ridge
on the anterior border of the acetabulum (Coombs 1978).

Limbs. No elements of the limbs have been recovered from the new specimen site except a single
terminal phalanx. Hulke (1881) briefly described some bones, considered as ‘unguals’ by William
Fox, as ‘broad, depressed and blunt’. The number is not stated, and now I can only define one
terminal phalanx among the isolated parts of the holotype. This specimen measures 82 mm long by
63 mm at the widest point, but it is distorted by compression, and this may have caused the
articulating surface to appear on the ventral side. The bone has a rounded termination, and no
narrow ‘neck’ between the articulating surface and the body of the phalanx, possibly an adult
feature. The terminal phalanx from the site, considered part of the same skeleton, is smaller than
the holotype, more pointed and has a fiatter base. A slightly narrowed neck gives rise to expanded
lateral boarders over the anterior two-thirds of the bone. These are separated from the raised body
by grooves anteriorly. The neck of the bone is possibly a juvenile feature compared to the holotype
ungual, suggesting further the sub-adult status of the new specimen.

Bones of undefined origin. The new specimen includes a small, complete bone of unknown origin.
Its appearance is suggestive of a dermal scute, and this was the label attached to its counterpart in
Cambridge (SMES B53372). The two are a bilateral pair and originate from the same lacality. Each
has a sharp-edged ridge across the superior surface with a low peak set at one end. The base inclined
to the ridge at an angle of about 45°, and is gently concaved and roughened to suggest a sutural
surface with another bone. The shape of the base is oval, but broader at one end than the other.
The surface features are unlike other dermal elements, and the existence of an exact bilateral pair is
suggestive of possible endoskeletal origin. However, microscopic examination of the surface is
suggestive of dermal origin (D. Cooper, pers, comm.). The two do not in themselves unite, but
appear to form mirror images some distance apart. Until more complete material is available the
origin, orientation, and symmetrical designation of these bones will be in doubt.

Armour. The dermal elements of Polacanthus were interpreted by Nopcsa (1905) for the holotype,
and the new specimen falls broadly within this scheme. The animal had a bilateral row of dorsolateral

BLOWS: ARMOURED DINOSAUR POLACANTHUS

Pre~sacral sp
Type C

5 cm

I 1

Caudal plate
5 cm

565

P

a

TEXT-FIG. 5. Polaccmthus foxi. Lower Cretaceous, Isle of Wight. BMNH R9293, presacral and caudal spines.
1, posterior view; 2, medial view; 3, anterior view; 4, inferior (basal) view, a— anterior, p— posterior.

566

PALAEONTOLOGY, VOLUME 30

spines of varying height and broad bases along the neck and thorax. The caudal region had a
double row of narrow plates set dorsolaterally in decreasing size from tail root to tip. The pelvis
and sacrum are covered by a large sacral shield, and numerous smaller ossicles were distributed
across the trunk and tail between the larger elements. This arrangement is well understood from
description of the holotype, but each of the components are worthy of reassessment in the light of
the new specimen. In addition, new dermal elements, unknown in the holotype (Blows 1982), are
described.

Presacral dorsal spines. The holotype has seven preserved spines (Type A, text-fig. 5), five more are
plaster replicas for display purposes. Nopcsa quotes eight spines and distributes them as three left
and five right. His illustration (Nopcsa 1905, pi. xii) of the skeleton indicates seven spines as
preserved, five right and two left. The new specimen has four recognizable spines (Type A) of
various sizes.

Each Type A spine has two distinct edges to the keel which unite at a sharp summit. In the largest
spine (text-fig. 8) the two keel edges twist through nearly 90° towards the summit. The bases are
solid, broad, flat, and generally asymmetrical; only the smallest spine showing some degree of
concavity and symmetry in the base. The base has a hook-like feature produced by an extension of
one keel edge below the base margin. Opposite this the larger spines have a groove just inside the
base margin. Following Nopcsa in the orientation of the larger spines, the lateral face of the keel is
smooth and slightly concave, the medial face is angled into three planes. All the faces of the keel
are irregularly grooved, often thought to be of vascular origin, and the base has a crossing fibrous
texture.

Several of these spines of the holotype are unusually flat and broad, possibly further evidence of
general post-mortem crushing. Nopcsa’s orientation of these spines shows a consistent pattern. The
anterior keel edge is curved outwards, the posterior edge being shorter. This makes the basal hook-
like feature anterior, the opposite groove in larger spines being posterior. Other spines are of a
slightly dilferent nature (Type B, text-fig. 5). They are MIWG 1 191a, a large dorsal spine missing
the top, from Sedmore Point, Isle of Wight; and MIWG 5307, a large complete dorsal spine from
grey marls several hundred metres east of Chilton Chine, Isle of Wight, about 1 1 ft. above beach
level. They are both presacral in origin because they have solid, broad bases, but they vary from
Type A by the following points: the dorsal keel is flat on both sides and the edges are straight, or
very gently curved, twisting slightly or not at all. The base has two distinct areas of dermal
attachment, a prominent medial anterior process and lateral posterior process. Both areas are
separated from the dorsal keel by a step in the bone. There is no hook-like feature in Type B spines.
These spines could reach the size of Type A (see MIWG 1 19k/). Two spines from the new specimen
are of the same general variety as Type B, one larger and one smaller, but differ by having narrower,
smooth and gently arched bases. These do not reach the full extent of the dorsal keel edges which
extend beyond the anterior limit of the base into a V-shape similar to caudal plates. Type C is
designated to these spines.

Postsacral caudal plates. The other free standing elements are tail plates extending along the tail in
descending size and shape (text-figs. 5 and 7). The holotype has fifteen tail plates set in two rows,
seven right and eight left. Nopcsa’s skeletal restoration illustrates twenty-two caudal plates in eleven
pairs (including the tail tip plates as a pair). The new specimen has five plates referable to the tail,
with a number of fragments of others. Three plates are left side and two right, with four fragments
referable to the right. The dorsal keel is flat on both sides and tapers to a round peak. The base is
narrow and hollowed throughout the posterior two-thirds, the cavity created extending as much as
40 mm into the plate. The base edges are asymmetrical and sinusoidal and are heavily roughened
with notches and grooves for dermal attachment. The caudal plates of the holotype differ consider-
ably, the above base shape appears in only one anterior plate (numbered BRl). The others have
symmetrical bases with a V-shaped cavity throughout the base length, and a similar hook-like
feature to that of Type A spines.

Two caudal plates show variation from this. The first is part of the new specimen (R9293;

BLOWS: ARMOURED DINOSAUR POLACANTHUS

567

TEXT-FIG. 6. Pokiccmthus foxi. Lower Cretaceous, Isle of Wight. BMNH R9293, ?anterior caudal dermal plate,
A, ?medial view, b, lateral view, c, Hoplitosaunis marslii. Lower Cretaceous, USA. USNM 4752, ?medial view

(redrawn from Gilmore 1914).

568

PALAEONTOLOGY, VOLUME 30

vh

TEXT-FIG. 7. Polacanthus foxi. Lower Cretaceous, Isle of Wight, a, b, GM 981.45, dermal plate; A, ?medial
view, B, posterior view, c, Hoplitosaunis marshi. Lower Cretaceous, USA. USNM 4752, dermal plate, posterior
view (redrawn from Gilmore 1914). d f, P. foxi. Lower Cretaceous, Isle of Wight. D, E, BMNH R9293, caudal
plate; d, ?medial view, E, posterior view. F, BMNH R175, caudal plate, ?medial view, g— groove, vh— ventral

hollow, vk— ventral keel.

BLOWS: ARMOURED DINOSAUR POLACANTHUS

569

text-figs. 6 and 8) and is distinguished by a long, solid base, the keel narrowing sharply to a peak
supporting a straight, narrow cylindrical spine. This spine is nearly half the overall height of the
plate, and is similar to the peaks of Type A spines. The length and asymmetry of the plate base
suggests early caudal positioning on the left side, possibly the first postsacral plate. The second
plate of unusual nature was described previously (Delair 1982) (GM 981.45; text-fig. 7) and is from
the Barnes High area of the Isle of Wight (Kemp Collection, 1974). The anterior and posterior
edges are straight and the peak is posteriorly inclined. The dorsal keel thickens from the anterior
edge to create two posterior edges separated by a groove from base to apex. The solid base supports
a rugosed, deep ventral keel for dermal attachment. These plates are discussed later.

Sacral shield. The dermal plate which covered the sacrum and ilium is well known and described by
Hulke (1887). This holotype shield is slightly flattened by crushing. It is a complete unit, averaging
8 mm thick, and is not fused to the underlying bone. The dorsal surface is patterned with round,
subtriangular or oval bosses bearing various degrees of raised peaks, separated by multiple tubercles
and covered with fine grooves. The ventral surface, nor seen in the holotype, has a surface pattern
of cross fibrous tissue, like coarse textile (e.g. sackcloth), and is pock-marked by multiple blood
vessel openings. Ten large fragments and many smaller pieces exist of this shield in the new specimen
(text-fig. 9). The central pieces are thicker, the shield thins towards the lateral and posterior borders.
The portion of Polacantlms ilium, referred to under ‘pelvis’ as P. becklesi (R1926) was thought to
have a fragment of smooth sacral shield attached; this diagnostic feature suggested the new species
(Hennig 1924). Close examination of the surface of the armour reveals the bases of at least three
bosses and surrounding tubercles; the armour was patterned. This specimen is referred to P. foxi,
the name P. becklesi is obsolete. The holotype shows a very shallow, almost flat patterning to the
armour over the same region of ilium as R1926, and the Beckles specimen is clearly badly water
worn, being impregnated with calcareous deposits from the sea.

Ossicles. The isolated elements of armour, which were not incorporated within the shield, and which
probably covered the dorsum, flanks and tail between the free standing armour were ossicles. They
range in size from 9 mm to 70 mm across and number thirty-six from the new skeleton. The larger
are subtriangular to nearly round and peaked to one edge, with flat or slightly convex bases. The
dorsal surface has vascular grooving. The smaller ossicles are rounded with flat bases. Two ossicles
are notched, they have a small rounded piece of one border missing; an incomplete border
development and not a fracture. In Sedgwick Museum, Cambridge, about thirteen ossicles exist, all
from the Isle of Wight and possibly from this new skeleton. In Sandown Museum, Isle of Wight, a
group of dermal pieces (MIWG 37) from Brook contains ossicles of larger size than the maximum
eollected with the new specimen. One ossicle measures 93 mm by 75 mm, another 60 mm by 100
mm. The British Museum (Natural History), London, has six ossicles plus thirty-nine of the holotype
(twenty-eight small, eleven large). Some of the holotype ossicles are the largest known for this genus
(e.g. 105 mm by 93 mm).

Summary of dermal armour in Polacantlius (text-fig. 10);

1. Presacral spines; Type A (both holotype and R9293), a bilateral row; Type B (isolated speci-
mens, MIWG), a bilateral row(?); Type C (R9293), a bilateral fringe (?).

2. Sacral shield (both holotype and R9293) covering pelvis and sacrum.

3. Caudal plates; roof-like (holotype) or tall (R9293) both descending in size in a bilateral row
down the tail; Hoplitosaur-like (R9293) (?)anterior plate only.

4. Ossicles; subtriangular or round between other elements (both specimens).

DISCUSSION

Preservation

The new speeimen is considerably less complete and more disarticulated than the holotype, yet
better preserved. The holotype is distorted by crushing throughout various sections, but crushing
in the new specimen is limited.

570

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 8. Polacanthus foxi. Lower Cretaceous, Isle of Wight. BMNH R9293, dermal spines and plates. Top
row: Type A presacral spines. Bottom row: Type C presacral spine, posterior spine and Hoplitosaums-Wke

caudal plate.

BLOWS: ARMOURED DINOSAUR POLACANTHUS

571

TEXT-FIG. 9. Polacantims foxi. Lower Cretaceous, Isle of Wight. BMNH R9293, dermal armour, two sacral

shield fragments, superior (dorsal) view.

Age

The sub-adult status of the new specimen is indicated by separated sutures within the skull frag-
ments, the elements of the atlas vertebra are separated and the characteristic shape of the terminal
phalanx.

Post-cranial skeleton

The skeleton of Polacantims is very similar to Stegosaurus in many detailed and overall structures,
suggesting a possible parallel evolution from a single source; the nodosaurs expanded more rapidly
after the decline of the stegosaurs. The osteology of stegosaurs is now well known (Gilmore 1914;
Ostrom and McIntosh 1966).

572

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 10. Polacanthus foxi. Lower Cretaceous, Isle of Wight. Restoration of skeleton, lateral and superior

views. Based on BMNH R175 and R9293.

Post-cranial armour

The structure of Lower Cretaceous nodosaur dermal anatomy is always difficult because the material
is usually isolated or disarticulated from endoskeletal remains. Nopcsa’s interpretation (1905) of
the arrangement of presacral spines remains valid. The angle between the spine base and dorsal
keel suggests a position on the animal approximately 40° from the mid-line; a dorsolateral position.
The presence of Type C spines in association with the new specimen possibly suggests a lateral
fringe over the shoulder region (as in Palaeoscincus Matthew 1922) or perhaps a similar arrangement
at the tail base. The lateral position of these spines is suggested by the unusual base shape and deep
dermal insertions; the weight of the spine would not be directly above the body, creating a need to
withstand gravitational forces.

The existence of Type B spines as purely isolated elements causes problems. They strongly suggest
a similar position as Type A spines, but their absence in skeletons remains unexplained. No Type B
spine can therefore be directly attributed to Polacanthus on the basis of known material. However,
with absence of other nodosaurs identified from the Isle of Wight Wealden strata, tentative assig-
nation of Type B spines to Polacanthus is made on the grounds of possible sexual diamorphism.
The discovery of Type B spines in association with other skeletal material will resolve this.

Variation within the caudal plates occurs between the holotype and the new specimen. In the
former, only the most anterior plate is of the type described in the new specimen, the others create
diminishing roof-like structures along the tail, and flat rounded elements at the tail end. The most
anterior plate of the new specimen is probably the Hoplitosaur-like plate with the central spine,
unlike any other seen in the holotype. The significance of these two tail armours may be sexual or
suggestive of species variation. If sexual diamorphism could be demonstrated in dermal structures.

BLOWS: ARMOURED DINOSAUR POLACANTHUS

573

the caudal plate variation and robust (R9293) or slender (R175) presacral spines could be suspected
as such in Polacanthus.

The Hoplitosaur-like plate may be indicative of closer affinities with the American Lakota genus
Hoplitosaurus (USNM 4752). The caudal plate described by Delair (1982) (GM 981.45) enhances
the Polacanthus-Hoplitosaurus link. The assignment of this specimen to Polacanthus is tentative on
the grounds that a wider range of armour existed within this genus than previously known, and
there is no evidence of another genus within the Lower Cretaceous of the Isle of Wight. The
structure of the plate is similar to those of an early caudal position and the deep ventral keel is
indicative of a lateral projection where deep dermal insertion is needed to overcome gravitational
forces.

Histological examination of Polacanthus dermal armour has been carried out by Dr Cooper of
Worthing. He sectioned both small presacral spines and sacral shield. He writes They show the
same features of ossified collagen bundles which pass at random in all directions. This armour has
clearly formed by transformation of dermal collagen directly into bone, i.e. by metaplasia of collagen
as in a tendinous insertion into a long bone. There has then been subsequent formation of marrow
spaces and a few Haversian systems in the armour. There is no evidence that any of the dermal
elements formed from cartilage or had any form of muscle attachments. The fine channels on the
surface of the armour could well be vascular but could also be to increase the surface area of the
bone-keratin interface to provide a very rigid attachment for the keratin horn’ (D. Cooper, pers.
comm.).

The accepted function of dermal elements has always been their protective value against pre-
dators. Recently, some work has been conducted on heat dispersal properties of dermal plates in
Stegosaurus which is well endowed with these elements (Farlow et al. 1976).

Dr Cooper writes T do not think now that one can compare Polacanthus armour with the plates
of Stegosaurus which appear to have a heat radiating function, though similar heat exchange might
incidentally occur in the larger lateral spines. However, their primary function is obviously armour
protection’ (D. Cooper, pers. comm.). A dual purpose may seem incompatible. Should such elements
become damaged in a predator confrontation a high vascularity necessary for heat dispersal could
predispose to severe blood loss. Selective and well-controlled vasoconstriction would be essential,
as postulated for Stegosaurus (Farlow et al. 1976).

THE RELATIONSHIPS OF POLACANTHUS
1 . Polacanthus with Hylaeosaurus

Coombs (1971) and others suggest that Polacanthus and Hylaeosaurus are possibly synonymous.
Hylaeosaurus appears limited to the South East English mainland geology, which occupies most of
the Wealden Series from Ryazanian to Barremian (Rawson et al. 1978). Hylaeosaurus has occurred
in the ‘Tilgate Forest’ (Mantell 1841; Owen 1856) which probably refers to the Tilgate Stone (Gallois
1965), extensively quarried in Mantell’s day, and the stata from which he obtained Iguanodon. It is
also known as Cuckfield Stone (Rawson et. al. 1978), and divides the Grinstead Clay into Upper
and Lower components (Hastings Beds). The palaeontological record is grossly incomplete, but the
current evidence restricts Hylaeosaurus to the Upper Valanginian (Tunbridge Wells Sand/Grinstead
Clay), which is unexposed on the Isle of Wight.

Polacanthus ranges from Wealden Marls to (?)Ferruginous Sands (Barremian to Lower Aptian)
mostly from the Isle of Wight with one specimen from the mainland. This is a slab of grey matrix
from the Greensands overlying the Lias at Charmouth, Dorset, and containing parts of four
disarticulated but associated dorsal vertebrae, a rib section, and portions of flat dermal armour
(sacral shield) (text-fig. 11). Between the ranges of the two genera a gap occurs in the Lower Weald
Clay (Hauterivian). This may well be a collecting anomaly, or may have evolutionary significance.
Currently it suggests a geological separation between the two genera, with Hylaeosaurus being
considerably the older. The Polacanthus range correlation on the mainland (Upper Weald Clay to

574

PALAEONTOLOGY, VOLUME 30

A

TEXT-FIG. 1 1. Polacanthus foxi. Middle Cretaceous, Charmouth, Dorset. Dorsal vertebrae, rib and sacral shield

armour in matrix.

Hythe Beds) does not seem to yield this genus. It is important to realize that the rarity of nodosaurids
and the inadequacy of Wealden collecting (especially of the mainland) constitutes a vestige of
material from which positive conclusions are difficult.

In Hylaeosaurus the series of long pointed spines appear to take a lateral fringe position over the
shoulder region, similar to the arrangement in Palaeoscincus (Matthew 1922; Gilmore 1930). Post-
mortem displacement is limited, as indicated by the largely articulated nature of the endoskeleton.
The presacral spines of Polacanthus appear to be positioned more dorsolateral, and Types A or B
are not found in the holotype or referred specimens of Hylaeosaurus. The shield of armour in
Polacanthus is found as fragments on the Isle of Wight. No evidence exists of this shield in
Hylaeosaurus, and no fragments of sacral shield are known from the mainland Wealden formation
or referred to Hylaeosaurus. A similar situation occurs with ossicles, which are numerous, variable
in shape, both large and small in Polacanthus, but remain rare in Hylaeosaurus, being round, small

BLOWS: ARMOURED DINOSAUR POLACANTHUS

575

and button-like. Caudal plates also are scarce in Hylaeosaurus, a nearly complete tail (BMNH 3789)
has the bases only preserved of what could be two caudal plates, whilst both Polaccmthus discoveries
are dominated with caudal armour. The current evidence suggests that Hylaeosaurus was a primitive
nodosaurid existing earlier than Polacanthiis, and having a different armour arrangement. At
present, these two genera should be considered completely separate.

2. Polacanthus with Polacanthoides

Nopcsa’s (1928) Polacanthoides can be dismissed on the grounds that;

a, two bones (BMNH R1106 and R1107) are from the Isle of Wight and not from ‘Bolney,
Sussex’ as stated by Nopcsa (1928); h, they are casts, the originals returned to the Isle of Wight and
are now lost, and therefore should not have been designated as holotypes; c, the scapula (BMNH
2584) from Bolney, Sussex is clearly associated with a tibia (BMNH 2615) (Mantell 1841, p. 143;
Lydekker 1888), but Nopcsa (1928) does not mention this tibia.

The difference between the scapula of Hylaeosaurus and Polacanthoides is the acromion process;
large and flange-like in Polacanthoides, thumb-like in Hylaeosaurus (Nopcsa 1928; Ostrom 1970).
Ostrom upheld the distinction of Polacanthoides ^rom Hylaeosaurus, but Coombs ( 1978) synonymizes
the two genera, and finds the acromion size to be no grounds for the foundation of a genus. The name
Polacanthoides (and thus P. ponderosus) is nomen dubium. The tibia and humerus (BMNH R1 106,
R1 107) (Hulke 1874) are from the Isle of Wight, and may be Polacanthus. The tibia is very similar to
Polacanthus (Ostrom 1970, p. 135), but no humerus is known of Polacanthus for comparison. On
geological grounds these two bones may be referable to Polacanthus rather than Hylaeosaurus.

3. Polacanthus with Hoplitosaurus

The new Polacanthus has suggested close affinities with the American genus Hoplitosaurus niarshi
(Lucas 1902; USNM 4752). Both genera have flat and standing dermal elements. Hoplitosaurus
armour is described by Gilmore ( 1 9 1 4, pp. 118-121), and I refer to his numbering:

1, simple flattened, which correspond to variations within the ossicles of Polacanthus',

2, rounded ossicle-like, which correspond to ossicles of Polacanthus',

3, keeled, which correspond to ossicle variation within Polacanthus',

4, triangular, plate-like, which correspond to caudal plates of Polacanthus, in particular R9293;

5, spined, of which he recognizes the following sub-types:

a, ‘scutes’, which correspond with the Hoplitosaur-like plate of R9293 (Gilmore 1914, pi. 28). Gil-
more misinterpreted Nopcsa ( 1 905) by thinking such elements were placed anterior to the sacral shield .
In fact, no elements of this kind exist in the holotype (R175), the first European plate of this type
occurs in R9293. Some presacral spines of R175 have a similar central spine, but differ in having
broader, plain bases. It is easy to see how this plate type has developed from the presacral spines but
has caudal plate qualities. The position is therefore most likely to be the first plate, immediately behind
the sacral shield. It would seem that R175 never did possess such a plate, whilst R9293 and
USNM 4752 did. Whether this is an indication of sexual variation, along with the difference in
caudal plates between R175 and R9293, is unknown.

b, dermal elements with grooved posterior borders, of which only one English example exists,
GM 981.45 from the Weald of the Isle of Wight (Delair 1982). The presence of grooved plates in
Hoplitosaurus (Gilmore 1914) suggests the possible inclusion of the English plate within the genus
Polacanthus. The deep ventral keel is its only unique character, suggestive of deep dermal insertion
to overcome gravity.

c, compressed spines with heavy, massive, expanded bases, comparable to the presacral (Type A)
spines of Polacanthus.

Hoplitosaurus and Polacanthus dermal armour compares well. The absence of a sacral shield in
Hoplitosaurus may be the only objection, although reference is made to this under (2) above (in
Gilmore 1914, p. 118). The two could be found synonymous, in which case Polacanthus (Huxley

576

PALAEONTOLOGY, VOLUME 30

1867) takes priority over Hoplitosaurus (Lucas 1902). If the sacral shield was considered sufficient
grounds to allow species differentiation then Polacanthus niarshi, new combination, could be re-
served for USNM 4752. This synonymy would extend the geographical range of Polacanthus outside
England to the American Lakota Formation as suggested for other Isle of Wight genera, e.g.
Hypsilophodon wielandi (Galton and Jensen 1979).

CONCLUSIONS

Polacanthoides ponderosus is not valid. Polacanthus and Hylaeosaurus are separate genera. Hoplito-
saurus is probably a subjective junior synonym of Polacanthus, possibly extending the range of
Polacanthus into America. If species differentiation was shown between the two holotypes, P.
marshi, new combination is proposed for USNM 4752.

Acknowledgements. I thank the following for allowing me to examine specimens in their care; Dr Angela
Milner and Sandra Chapman, British Museum (Natural History); Mr Steven Hutt (Museum of Isle of Wight,
Geology); Dr David Price (Sedgwick Museum); Mr David Kemp (Gosport Museum). My gratitude is also
extended to those persons who have helped and advised on the new specimen; Dr Peter Galton (Bridgeport
University); Dr David Norman (Oxford University); Dr Beverly Halstead (Reading University); Dr Alan
Insole (Museum of Isle of Wight, Geology); Dr Dale Russell (National Museums of Canada); Mr Richard
Ford (Yarmouth, Isle of Wight); Dr David Cooper (Worthing, Sussex). Finally, my gratitude goes to all those
persons who have helped with the site excavation, preparation of the fossils, and production of this paper;
Mr Terence Hunter; Mrs Jessica Blows.

REFERENCES

ANON, 1865. A new wealden dragon. Land. Illiist. News, September 1865.

BLOWS, w. T. 1982. A preliminary account of a new specimen Polacanthus foxi (Ankylosauria, Reptilia) from
the wealden of the Isle of Wight. Proc. Isle Wight nat. Hist, archaeol. Soc. 7, 303-306.

1983. William Fox, a neglected dinosaur hunter of the Isle of Wight. Arch. nat. Hist. 11, 299-313.

COOMBS, w. p. 1971. The Ankylosauria. Ph.D. thesis (unpublished). University of Columbia, 487 pp.

1978. The families of the Ornithischian dinosaur Order Ankylosauria. Palaeontology, 21, 143 170.

DELAIR, j. B. 1982. Notes on an armoured dinosaur from Barnes High, Isle of Wight. Proc. Isle Wight nat.
Hist, archaeol. Soc. 7, 297-302.

EATON, T. H. 1960. A new armoured dinosaur from the Cretaceous of Kansas. Univ. Kans. Pubis. Mus. nat.
Hist. 25, 1 -24.

FARLOW, J. o., THOMPSON, c. V. and ROSNER, D. E. 1976. Plates of the dinosaur Stegosaurus: forced convection
heat loss fins? Science, N. Y. 192, 1 123-1 125.

GALLOis, R. w. 1965. The Wealden district (fourth edition). Brit. reg. Geol. 101 pp. London.

GALTON, p. M. and JENSEN, J. A. 1979. Remains of ornithopod dinosaurs from the Lower Cretaceous of North
America. Brigham Young Univ. geol. Stud. 25, 1 10.

GILMORE, c. w. 1914. Osteology of the armoured dinosauria in the United States National Museum, with
special reference to the genus Stegosaurus. Bull. U.S. natn. Mus. 89, 1-143.

1930. On dinosaurian reptiles from the Two Medicine Formation of Montana. Proc. U.S. natn. Mus. 77,

1 39.

HENNiG, E. 1924. Kentrurosaurus aethiopicus die Stegosaurier-Funde vom Tendaguru, Deutsch-Ostrafrika.
Palaeontographica, suppl. 7, 103-153.

HULKE, J. w. 1874. Note on a reptilian tibia and humerus (probably of Hylaeosaurus) from the Wealden
formation in the Isle of Wight. Q. Jl geol. Soc. Land. 30, 516-520.

1881. Polacanthus foxi, a large undescribed dinosaur from the Wealden formation in the Isle of Wight.

Phil. Trans. R. Soc. 172, 653-662.

1887. Supplemental note on Polacanthus foxi, describing the dorsal shield and some parts of the endo-

skeleton, imperfectly known in 1881. Ibid. (B) 178, 169-172.

HUXLEY, T. h.1867. On Acanthopholis horridus, a new reptile from the Chalk Marl. Geol. Mag. 4, 65-67.

LEE, J. E. 1843. Notice of saurian dermal plates from the Wealden of the Isle of Wight. Ann. Mag. nat. Hist. 2,
5-7.

BLOWS: ARMOURED DINOSAUR POLACANTHUS

577

LUCAS, F. A. 1902. A new generic name for Stegosaurus marshi. Science, N. Y. 16, 435.

LULL, R. s. 1921. The Cretaceous armoured dinosaur Nodosaurus textilis Marsh. Am. J. Sci. 1 (2), ser. 5, 97-
126.

LYDEKKER, R. 1888. Catalogue of the fossil reptilia and amphibia in the British Museum. I. Dinosauria. British
Museum (Natural History), London.

1891. On part of the pelvis of Polacanthus. Q. Jl geol. Soc. Lond. 48, 148 149.

MANTELL, G. A. 1833. Gcology of the South East of England. London.

1841. Memoir on a portion of the lower jaw of the Iguanodon, and on the remains of the Hylaeosaurus

and other saurians, discovered in the strata of the Tilgate Forest, in Sussex. Phil. Trans. R. Soc.
131-151.

MARYANSKA, T. 1977. Ankylosauridae (Dinosauria) from Mongolia. Palaeont. pol. 37, 85-151.

MATTHEW, w. D. 1922. A super-dreadnaught of the animal world, the armoured dinosaur Palaeoscincus. Nat.
//A?., TV. F. 22, 333-342.

NOPCSA, F. 1905. Notes on British dinosaurs. Part II; Polacanthus. Geol. Mag. 2, 241 -250.

1928. Palaeontological notes on reptiles. Geologica hung. 1, 1 -84.

OSTROM, j. H. 1970. Stratigraphy and palaeontology of the Cloverly formation (Lower Cretaceous) of the
Bighorn Basin area, Wyoming and Montana. Bull. Peabody Mus. nat. Hist. 35, 1 234.

and MCINTOSH, j. 1966. Marsh's Dinosaurs, the collections from Como Bluff. Yale Univ. Press, 388 pp.

OWEN, R. 1858. Monograph on the fossil reptilia of the Wealden formations, part IV (Hylaeosaurus). Palae-
ontogr. Soc. [Monogr.] 10, 8-26.

RAWSON, P. F., CURRY, D., DILLEY, F. C., HANCOCK, J. M., KENNEDY, J. M., NEAL, J. W., WOOD, C. J. and WORSAM,

B. c. 1978. A correlation of Cretaceous rocks in the British Isles. Spec. Rep. geol. Soc. Lond. 9, 70 pp.

SEELEY, H. G. 1876. On the axis of a dinosaur from the Wealden of Brook in the Isle of Wight, probably
referrable to the Iguanodon. Q. Jl geol. Soc. Lond. 31, 460-464.

1891. On the os pubis of Polacanthus foxi. Ibid. 48, 81 85.

STERNBERG, C. M. 1921. A Supplementary study of Panoplosaurus minis. Trans. R. Soc. Can. 15, 93-102.

WILLIAM T. BLOWS

Typescript received 4 August 1 986 39 Lane

Revised typescript received 16 November 1986 Dartford, Kent DA2 6PG

APPENDIX I

Specimens 0/ Polacanthus foxi Hulke 1881

Holotype: BMNH R175, partial skeleton; Referred specimens: MIWG 37, seven ossicles, one tail plate base;
MIWG 4222, one ossicle (Poole Collection); MIWG 1983, one ossicle, one sacral shield fragment; MIWG
5144, caudal vertebra; MIWG 5307, dorsal spine (collected April 1983); MIWG 5145, caudal plate fragment;
MIWG 5186, six plate fragments, one ossicle; MIWG 5187, two ossicles; MIWG 1 191a, dorsal spine; MIWG
5188, dorsal centrum; CAMSM B53595, dermal ossicle; CAMSM B53594, small dermal ossicle; CAMSM
B53590, dermal ossicle; CAMSM B53588, dermal ossicle; CAMSM B53589, dermal ossicle; CAMSM B53591,
dermal ossicle; CAMSM 53597, dermal ossicle; CAMSM B53371, axis vertebra; CAMSM B53372, unidentified
bone; CAMSM B53353, large dermal ossicle; CAMSM B53354, flat ossicle; CAMSM B53358, flat ossicle,
CAMSM B53357, ossicle; CAMSM B53355, ossicle; CAMSM B53356, ossicle; CAMSM B53596, dermal
ossicle; CAMSM B53587, dorsal vertebra; CAMSM B53593, dermal ossicle; BMNH R4952, caudal vertebra
(Gunyon Collection 1923); BMNH 36515-36517, two dermal spines (Mantell Collection); BMNH 40458,
dermal ossicle; BMNH 37713-37714, two dermal ossicles (Saul Collection 1863); BMNH R643, dermal ossicle
(Lee Collection 1885); BMNH R1876, dermal plate (Beckles Collection); BMNH 34533, dermal ossicle
(Backhouse Collection); BMNH R1875, dermal spine (Beckles Collection); BMNH R2527, dorsal vertebra
(Hulke Collection 1869); BMNH 39556, rib head; BMNH R202, two ossicles; BMNH R202-R202(A), four
dermal spines (Fox Collection); BMNH R203, two dermal spines (Fox Collection); BMNH R1926, ilium
fragment with sacral armour (type of P. becklesi) (Beckles Collection); BMNH R4134, vertebral process and
rib head; BMNH R9905, sacrum; BMNH R9293, partial skeleton with dermal armour (Blows Collection
1979) and cervical vertebra; GM 981.45, dermal plate (Kemp Collection); block containing dorsal vertebrae,
rib, and dermal armour (private collection); spine and caudal vertebra (Ford Collection).

578

PALAEONTOLOGY, VOLUME 30

APPENDIX II

Table of measurements in millimetres:

i. Cervical vertebrae

SMES B53371

BMNH R9293

Length of centrum

54

50

Width of anterior surface

64

65

Height of anterior surface

40

43

Total height of vertebra

80

107

Greatest width of neural canal

25

23

Greatest height of neural canal

25

30

ii. Dorsal vertebrae

Length of
centrum

Height of
centrum

Width of centrum
Anterior Posterior

Height of neural canal
Anterior Posterior

Overall

height

Holotype

C12

62

50

—

61

30

25

130 +

RI75

C14

77

51

53

52

23

25

110 +

C16

—

—

—

—

—

—

—

C18

—

—

—

58

—

48

—

C19

80

50

56

52

—

—

BMNH R9293

1

70

53

61

61

25

30

-

2

—

56

63

—

25

—

130 +

3

75

59

63

56 +

22

27

150 +

4

73

57

63

65

22

31

135 +

BMNH 2527

84

73

82

—

23

—

212 +

SMES B53587

78

60

68

69

23

24

144 +

MIWG 5188

70

60

71

70

—

—

--

Charmouth

specimen

70

65

—

—

—

—

170

iii. Sacral vertebrae

R175

R9293

BMNH R9905

Length of five sacral centra

350

365

360

Length of five presacral centra

410

—

—

Width of posterior articular surface (S5)

—

78

—

Height of posterior articular surface (S5)

—

50

—

Maximum width of neural canal (S2-S3 level)

—

78

—

Maximum height of neural canal (S2-S3 level)

—

60

—

Maximum width between (R) and (L) sacro-iliac
joints

418

425

BLOWS: ARMOURED DINOSAUR POLACANTHUS

579

iv. Caudal vertebrae

R175

1st

2nd

R9293

1st

2nd

R9252

MIWG 5144

Overall height

165

146

155

107 +

.

Length of centrum

60

50

60

55

40

45

Width of centrum

85

90

83

82

52

55

Height of centrum

40

60

58

60

52

60

V. Terminal phalanges

R175

R9293

Total length

82

60 +

Greatest width

60

35

Height of articular surface

20

19

Width of articular surface

50

29

Measurements of spines (Type A)

Total

Length

Width

height

of base

of base

Holotype, BMNH R127

R2 (f6), large spine

325 +

205

100

R4 (f4)

305 +

230

90

LI (g7), medium spine

322

198

115

L2 (g6), large spine

392 +

210

no

R3 (f5)

340 +

215

111

R5 (t3), medium spine

175 +

190

no

R4 (fl ), smallest spine

90

140

80

BMNH R9293

Largest spine, left side, shoulder

322 approx.

190

148

Large spine base, left side

—

197

95

Medium spine, right side

215 approx.

140

95

Smallest spine, anterior to pelvis

90

117

74

Measurements of spines (Type B)

MIWG 1191u

MIWG 5307

Overall height

—

260

Length of base (along keel)

—

170

Width of base (opposite keel)

—

130

viii. Measurements of spines (Type C)

BMNH R9293

Overall height 207+ 195 +

Length of base (along keel) 200 120

Width of base (opposite keel) 130 50

580

PALAEONTOLOGY, VOLUME 30

ix. Measurements of plates

Overall

height

Length
of base

Greatest width
of base

Holotype, R175

Early caudal, BRl

200 +

—

—

Middle caudal, BL3

110

192

—

Late caudal, BR6

70

126

—

BMNH R9293

Spined plate, left

280

238

58

Early caudal, left

245 +

218

42

Early caudal, left

230

—

45

Early caudal, left

210

160

40

Early caudal, right

200 +

200

45

Late caudal, right
GM 981.45

Early caudal(?)

330

160

60

Measurements of modified ossicles

SMES

MIWG

Maximum width across base

87

77

Length of base parallel to keel

70

67

Height of ossicle

38

37

Bilateral bones of unknown origin

B53372

R9293

Length (overall)

40

39

Height (overall)

30

29

Length of base

35

34

Width of base

20

20

ENGLISH EOCENE CRUSTACEA (LOBSTERS AND

STOMATOPOD)

by W. J. QUAYLE

Abstract. The Eocene lobsters from the London Clay, Bracklesham and Barton Beds are revised. Nine
species of lobster are represented, three new, Homarus morrisi, Hoploparia wardi, and //. victoriae, belonging
to six genera. Trachysoma scabrum Bell is placed in Glyphea. The remaining species, H. gammaroides M'Coy,
Archaeocarahus bowerhanki M‘Coy, Linuparus eocenicus Woods, L. scyllariformis (Bell), and Scyllarides tiiber-
culatus (Konig), are redescribed with further information. The stomatopod Squilla wetherelli Woodward is
redescribed and placed in Bathysqidlla.

In 1849 M‘Coy described three new species of lobster from the London Clay, Hoploparia gamma-
roides, H. belli, and Archaeocarahus bowerhanki. In 1858 Bell described Trachysoma scabrum, new
genus and species, and Thenops scyllariformis, new species, and redescribed Scyllaridia koenigii,
first described by Konig (1825), as Cancer {Scyllarusl) tuberculatus; he also made additions to
M‘Coy’s descriptions of H. gammaroides, H. belli, and A. bowerhanki. All of Bell’s specimens came
from the London Clay. Woods in his monograph of the Fossil Macrurous Crustacea of England
(1925-1931) described Linuparus eocenicus, a new species from the London Clay, and redescribed
H. gammaroides, A. bowerhanki, L. scyllariformis, and Scyllarides koenigii', H. belli was relegated to
synonymy. In a footnote on p. 94 he mentioned a large specimen of Homarus sp. from the
Barton Beds of Hampshire, first mentioned by Gardner et al. (1888). On p. 73 he mentioned but
did not describe the only known specimen of Trachysoma scabrum and figured it on pi. 22,
fig. 1. Cooper (1974) listed the known records of the English Palaeogene decapod Crustacea up to
that time.

Further collecting from some of the older sites, with some new ones, has added considerably to
our knowledge of the English Eocene macrurous Crustacea.

STRATIGRAPHY

Some of the localities mentioned by the earlier authors are no longer available, though on rare
occasions temporary exposures occur through road works and building. The following is a list of
localities from which the specimens used in this work were obtained, either by the old collectors
such as Bowerbank, Wetherell, Meyer, and Caleb Evans, or in recent years.

1. Isle of Sheppey, Kent; cliff and foreshore exposures, London Clay, Divisions C, D, and E
(King 1981, 1984). Grid reference TQ 955 738-TR 024 717.

2. Herne Bay, Kent; Beltinge cliff, London Clay, Divisions A2, A3, and lower half of B (King
1981), TR 195 685. Cliffs are now landscaped, leaving a foreshore exposure of A2. Thanet Beds,
TR 2051 6881 and TR 2046 6875, Units E and G (Ward 1979).

3. Burnham on Crouch, Essex; tidal river, with foreshore exposures, London Clay, Division D
(King 1981), TQ 920 968-922 966.

4. Maylandsea, Essex (George and Vincent 1982, pp. 39-41 ); tidal river with foreshore exposures,
London Clay, lower part of Division B (King 1981), TL 908 034-908 037.

5. Steeple, Essex (George and Vincent 1977, pp. 105-107); tidal river with foreshore exposures,
London Clay, lower part of Division B (King 1981 ), TL 916 043.

|Palaeontolog>, Vol. 30, Part 3, 1987, pp. 581-612, pis. 64-67.|

© The Palaeontological As.sociation

582

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 1. Localities in a, the London Basin and b, the Hampshire Basin (see text for key to numbers).

QUAYLE: EOCENE CRUSTACEA

583

6. Portsmouth Docks, Hampshire; former temporary exposure, ‘Dockyard Extension Works’.
The crustacean fauna (Woodward 1871, 1873) is similar to that of Whitecliff Bay. All the Crustacea
came from the London Clay, the Dentalium Bed of Meyer (1871), in the upper part of Division B2
(King 1981 ).

7. Alum Bay, Isle of Wight, Hampshire; cliff exposure, London Clay, Divisions A, B, C, and D
(King 1981), SZ 306 853.

8. Aveley, large quarries south-west of Aveley, Essex; London Clay, Divisions A, B, and C (King
1981), TQ 557 805 and 555 809. Working has now ceased. Material collected by R. J. Kirby is in
the Oxford University Museum.

9. Bognor Regis, Sussex; foreshore exposure, London Clay, Divisions A to C, SZ 942 990-
895 970.

10. Whitecliff Bay, Isle of Wight, Hampshire; cliff and foreshore exposures, Bracklesham Group,
SZ 640 861. Crustacea found in a line of iron stained phosphatic nodules in the Wittering Division
(Lishers Bed 1), approximately 7-5 m above the base of the Bracklesham Group. Some 75 % of the
nodules collected in situ and others collected loose from the beach (probably from this horizon)
had pieces or specimens of crustaceans. Limiparus is common, in association with H. morrisi sp.
nov. A nodule in the Sedgwick Museum, C69620, collected by H. Keeping in 1887, and labelled
Lower Bagshot Beds, is thought to have come from this horizon.

11. Various London Clay localities around London, now no longer available, i.e. Whetstone;
Chalk Larm, Divisions B and C; Highgate, Division E (King 1981), and various motorway construc-
tion sites around London.

12. Lee-on-the-Solent, Hampshire; foreshore exposure, Elmore Lormation (Kemp et al. 1979),
Huntingbridge Division, Bracklesham Group, SU 569 500.

13. Barton Beds, Christchurch Bay; cliff with occasional foreshore exposures; stratigraphic
horizons as in Burton (1929), SZ 199 928-261 923.

Repositories. Specimens prefixed BM are in the Palaeontological Department, British Museum (Natural
History); SM, Sedgwick Museum, Cambridge; OUM, Oxford University Museum, Oxford; PE, Passmore
Edwards Museum, Eondon; JSQ, W. J. and S. Quayle Collection; JC, John Cooper Collection.

SYSTEMATIC PAEAEONTOEOGY

Infraorder ASTACIDEA Latreille, 1803
Lamily nephropidae Dana, 1852
Subfamily homarinae Huxley, 1879
Genus HOMARUS Weber, 1795

Type species. Cancer gammanis Einne, 1758, Recent, European, by subsequent designation (Rathbun 1904,
p: 170).

Diagnosis. Rostrum rather short and spiny, carapace without ridges or spines behind the suborbital
spine, cervical groove clearly developed only below gastro-orbital groove, postcervical groove long
and smoothly curved, connecting lowest part of cervical with posterior part of branchiocardiac
groove; chelae stout, heterochelous.

Range. Cretaceous-Recent.

Discussion. Woods (1931) could find no generic distinction between the fossil forms of Hoploparia
and the living species of Homarus. Stenzel (1945, p. 427) stated that eleven species of Honiarus had
been described from the Cretaceous of the United States east of the Rockies. These eleven are
largely made up of specimens identified by Rathbun (19266, 1935) as Hoploparia. Stenzel also
described two new species, Homarus hrittonestris and H. davisi. These are similar to the living
species of Homarus in the nature of the carapace grooves.

584

PALAEONTOLOGY, VOLUME 30

Glaessner (1969) whilst noting that ‘the distinction between some of these species is difficult and
disputed’ maintained that a number of characters show that the genera are distinct. In Homarus:
cervical groove clearly developed only below the gastro-orbital groove; postcervical groove long
and curved, connecting the lowest part of the cervical groove with the posterior part of the branchio-
cardiac groove; carapace without ridges or spines behind the suborbital spine. In Hoploparia:
cervical groove developed above and below the gastro-orbital groove; postcervical distinct and
connected with the cervical groove through a semicircle; carapace can have ridges or spines behind
the suborbital spine.

There appear to be two distinct patterns of grooves. The first compares with the living forms of
Homarus as depicted by Glaessner (1969, R461, fig. 16) and Holthuis (1974, p. 816, fig. 24). The
second is as for the fossil species of Hoploparia (text-fig. 4a).

TEXT-FIG. 2. a, Homarus morrisi sp. nov., London Clay, Soft Rock, Bognor Regis, holotype, BM
In. 63392, x 0-8. 6, c, Hoploparia gammaroides M"Coy, London Clay, Isle of Sheppey, lectotype,
BM 46366. b, carapace, x 2. c, somites one to five, x 2. d, H. wardi sp. nov., London Clay, Isle of

Sheppey, holotype, SM Cl 93 14, x IT.

585

QUAYLE: EOCENE CRUSTACEA
Homarus morrisi sp. nov.

Plate 64, figs. 5 and 7; text-figs. 2a and 3

1849 Hoploparia gammaroides M'Coy, p. 177.

1850 Astacus belli (M'Coy); Dixon, pp. 1 14, 222, pi. 15, figs. 3 and 4.

1858 Hoploparia gamrnaroides M‘Coy; Bell, p. 38, pi. 8, figs. 5 and 6?; pi. 9.

1980 Hoploparia gamrnaroides M‘Coy; Morris, p. 9.

Derivation of name. After Mr S. F. Morris, Department of Palaeontology, British Museum (Natural History).

Types. Holotype: BM In. 63392 (text-fig. 2a), collected by B. A. Williams (Venables 1971), London Clay, Soft
Rock, Bognor Regis. Paratypes: BM In. 35299, In. 48223, In. 48226, and In. 48227 collected by E. M. Venables;
BM In. 63364 (PI. 64, fig. 5), In. 63365 (PI. 64, fig. 7), collected by D. Bone; all from Soft Rock, London Clay,
Bognor Regis, Sussex.

Diagnosis. A Homarus with fine grooves, with two rows of spines on the inner edge of the palm and
a single spine on the outer margin at the base of the dactylus; carapace with fine pits.

Horizon and locality. London Clay: Soft Rock and the Beetle Bed, Bognor Regis; Isle of Sheppey: near
Copenhagen House, London; Bracklesham Group, Bracklesham Bay, Selsey and Whitecliff Bay, Isle of Wight;
Barton Beds, Christchurch Bay.

Description. Cephalothorax. Rostrum dentate, slightly less than half the length of the carapace; carapace with
a postorbital spine and directly below this, a suborbital spine. The cervical groove starts almost level with the
antennal base and drops down, curving to join the antennal groove. This runs almost straight to the front
margin. The postcervical groove starts on the median line, slightly more than half the carapace length from
the base of the rostrum. It drops almost straight down, then curves towards the front at a slight downwards
angle and dies out. The postcervical is joined at the bend by the faint branchiocardiac groove running away
towards the rear, another fine groove drops away at this point and runs parallel with the cervical groove for a
short distance before dying out. Carapace with fine pitting. (Refer to text-fig. 4a for position of carapace
grooves.)

Abdomen. On the second abdominal somite the pleuron is quadrate, the anterior angle rounded, and the
posterior angle with a posterior pointing spine. Surfaces are smooth with fine pores; there is a large pit between
the lateral and posterior margins before they join. Somites three to five have a transverse groove on the
anterior third, extending to the boundary of the tergum at the anterior margin; pleura are triangular, falcate,
with an acute posterior pointed spine, their surfaces smooth with fine pits. The telson is approximately as
broad as long, its longitudinal margins with a fine beading; on the inner edge of the beading there is a smooth
groove, then a ridge.

Appendages. On the large first cheliped the width of the palm is from three quarters to equal to the length; the
inside face of the palm is flatly convex, the outer face half round; the inner edge is flattened along its length
and the margin is armed with two rows of alternating teeth, five on the outer, four on the inner, the last on
the joint margin; the outer margin is half round; the palm is smooth with fine pits of various sizes. A flat
surface develops on the outer edge, extending to the full width at the base of the fixed finger, continuing
parallel with the margin and becoming more noticeable on the outer face; it is about a third or more of the
width of the fixed finger. The inner margin of the fixed finger is wide, almost flat, and armed with depressed
oval teeth of varying size that lie close together. From the base of the fixed finger are three or four transverse

TEXT-FIG. 3. Homarus morrisi sp. nov., reconstruc-
tion of carapace ( x 1 -2).

586

PALAEONTOLOGY, VOLUME 30

oval teeth of approximately the same size though their height gradually increases. These are followed by a
larger oval raised tooth (greater length along the median line) and two much smaller oval teeth. The dactylus
is about as long as the palm, its outer margin half round with a spine on the outer edge by the base. The outer
surface is lightly rounded with a spine by the joint, paired to one on the palm. The inner surface is nearly flat
with a similar spine to that on the outer surface. The prehensile edge is armed with one large oval tooth
followed by three others diminishing in size towards the front, then a large rounded one followed by several
smaller rounded ones.

On the small claw the palm is nearly as wide as long, the top and bottom surfaces convex. The inner margin
has nine spines in two rows, the last on the margin of the joint with the dactylus. The outer margin is slightly
flattened. The surfaces are smooth with fine pores, except for a spine on each surface at the joint with the
dactylus. Only the base of the dactylus was preserved, oval in cross-section with a single spine on the outer
margin at the joint, armed with fine spiniform teeth on the prehensile edge.

Discussion. Bell (1858, p. 39) comparing BM 59136 (pi. 9) with H. belli remarked That it more
nearly approximates the common lobster and would probably be a distinct species’. It is now
considered that this specimen and a large majority of the lobsters from the Isle of Sheppey are
Homarus. The larger lobsters from this locality though well preserved are usually found in a crushed
or flattened state, leading to distortion of the grooves. This makes identification difficult; where
claws are preserved a spine can usually be found at the base, on the outer edge of the dactylus.

As well as the London Clay sites, the Barton Beds of Christchurch Bay have produced several
large specimens of Homarus since the first recorded specimen (SM C7742) was collected by H.
Keeping (Gardner el al. 1888). These specimens are thought to have originated from Horizons A3
and F. Due to the crushed state of the carapace where present, identification is difficult; so far only
small pieces of the prehensile edges of the claws have been preserved; the pieces that do exist,
though much larger, are apparently H. morrisi.

The Thanet Beds of Herne Bay have also produced specimens of Homarus, though only one has
the remains of the prehensile edge and parts of the claws. These, together with the specimens from
the Bracklesham Group of Bracklesham and Whitecliff Bay, are also apparently H. morrisi. It is
appropriate to include these specimens with H. morrisi until better preserved material can be found
to prove or disprove this statement. The Barton material known to date is: SM C7742; BM In. 60905;
piece of limb, JC Colleetion; BA/ 103- 1 06, JSQ Collection. Herne Bay material; speeimen collected
D. Kemp, JC Collection. Bracklesham Bay material: BM In. 29208 collected E. Williams. Whitecliff
Bay material: W88, 1 17, and 123, JSQ Collection.

H. morrisi is similar to H. gammarus (Linne) and H. americanus Milne-Edwards, living members
of the genus from Europe and America. It has a similar flattened edge on the outer margin of the
propodus; the rostrum is dentate and downturned and the carapace grooves and spines are in

EXPLANATION OF PLATE 64

Figs. 1 and 2. A small specimen of the Recent lobster Homarus gammarus (Linnaeus). 1, lateral view of the
carapace, x 1-5. 2, claw showing spines on the outer margin of the dactylus and palm, x 1-75.

Figs. 5 and 7. H. morrisi sp. nov., London Clay, Bognor Regis. 5, crusher claw, BM In. 63364, x 1-3. 7, fine
claw, showing arrangement of spines on the outer margin of the dactylus and palm, BM In. 63365, x 1 -2.

Figs. 3, 4, 6, 10. Hoploparia gammaroides M'Coy, London Clay. 3, crusher claw, OUM L558, Aveley, x 1.
4, third maxilliped, BM In. 63356, Maylandsea, x 2-2. 6, lateral view of carapace, BM In. 63354, Aveley,
X 1-5. 10, arrangement of spines (only one of the two rows is visible) on the palm; the base of the dactylus
is smooth, OUM L563, Aveley, x 1-7.

Figs. 8 and 9. H. victoriae sp. nov., Bracklesham Group, Lee-on-the-Solent. 8, lateral view of carapace, BM
In. 63357, x 1-8. 9, prehensile edge of fixed finger crusher claw, showing the arrangement of the teeth (base
only preserved), BM In. 63363, x 2-5.

Fig. 11. //. wardi sp. nov., London Clay, Chalk Farm, lateral view, SM Cl 93 18, x 1-25.

Figs. 12-14. Glyphea scabra (Bell), London Clay. 12, lateral view of the holotype, BM 59146, Chalk Farm,
X 2. 13 and i 4, left and right lateral views of BM In. 63366, Aveley, x2-3.

Specimens in figs. 3, 5-8, 11, 13, and 14 have been whitened with ammonium chloride.

PLATE 64

QUAYLE, Homariis, Hoploparia, Glyphea

588

PALAEONTOLOGY, VOLUME 30

similar positions. It differs in that the inner margin of the palm has a double row of spines with the
last at the joint margin. H. gammanis and H. americanus have a single row with the last spine before
the Joint margin.

H. davisi Stenzel, 1945 from the Cretaceous, Dallas Co., Texas, has a dentate carina running
back from the antennal spine and a line of tubercles on the posterior edge of the postcervical groove;
H. morrisi has neither of these characters. H. hrittonestris Stenzel, 1945, also from the Cretaceous,
Dallas Co., Texas, has a dentate rostrum and an antennal ridge followed by a small spine. The
carapace is covered with tubercles, the largest over the gastric region and on top of the ridges in the
vicinity of the rostrum. H. morrisi differs from H. hrittonestris in having fewer rostral spines, no
antennal ridge, and a finely pitted carapace.

H. hakelensis (Fraas) from the Cenomanian shale. Mount Lebanon, Syria, was provisionally
placed in this genus by Glaessner (1945, p. 702). Due to the crushed state of these specimens it is
difficult to make any valid comparison.

Genus hoploparia M‘Coy, 1849

Type species. Hoploparia longimana (Sowerby), 1826, Upper Greensand, Lyme Regis, by subsequent desig-
nation (Rathbun 1926a, p. 129).

Diagnosis. Rostrum dentate, cervical groove clearly developed above and below the gastro-orbital
groove, postcervical groove distinct, connecting with cervical groove through a semicircle; chelae
strong, long, heterochelous.

Range. Lower Cretaceous Middle Eocene.

Hoploparia gammaroides M‘Coy, 1849

Plate 64, figs. 3, 4, 6, 10; text-fig. 2b, c

1849 Hoploparia gammaroides M'Coy, p. 177.

1 849 Hoploparia belli M‘Coy, p. 1 78.

1858 Hoploparia gammaroides M'Coy; Bell, p. 38.

1858 Hoploparia belli M'Coy; Bell, p. 39, pi. 10, figs. 1-3, 5-8.

1929 Hoploparia gammaroides M'Coy; Glaessner, p. 219 (see for intermediate synonymy).

1931 Homarus gammaroides (M'Coy); Woods, p. 93, pi. 26, figs. 5 and 6; pi. 27, figs. 1, 2, 4.

1974 Hoploparia gammaroides M'Coy; Cooper, p. 85.

1980 Hoploparia gammaroides M‘Coy; Morris, p. 9.

Types. M'Coy based his description on specimens at the University of Cambridge and the Bowerbank
Collection, which should now be in the Sedgwick Museum and the British Museum (Natural History). No
figure was given or reference made to identifiable specimens. In order to select a type for H. gammaroides

TEXT-FIG. 4. Carapace grooves, a, Hoploparia sp. ( x 1-3). b, Glypliea scabra (Bell) ( x2 0). Identification of
grooves; a, branchiocardiac; bl, hepatic; b, antennal; c, postcervical; d, gastro-orbital; e, cervical; co, promi-
nence omega; i, inferior.

QUAYLE: EOCENE CRUSTACEA

589

both of these collections were carefully examined for a specimen that M'Coy could have used in his original
description. It is almost certain that this description was based on material of more than one genus, i.e.
Homanis and Hoploparia. No suitable specimen was found in the Sedgwick Museum and only one possible
specimen in the British Museum (Nat. Hist.). This specimen, BM 46366 (Bowerbank Collection), has the
characters of the species and was collected from the London Clay, Isle of Sheppey. Due to the uncertainty
that M‘Coy used BM 46366 (text-fig. 2b, c) for his original description of H. gammaroides, however, it is here
designated the neotype.

Other material. The following specimens have been used as a basis for the description. OUM L461, L558
(PI. 64, fig. 3), L559, L560, L563 (PI. 64, fig. 10), L564, all ex Kirby Collection; BM In. 63354 (PI. 64, fig. 6),
ex King Collection; BM In. 63355 In. 63356 (PI. 64, fig. 4), ex JSQ Collection.

Diagnosis. A Hoploparia with the cervical joining the antennal groove at an obtuse angle; the
antennal region prominent but without spines; posterior edge of the postcervical groove bordered
with a line of fine granules; prominence omega, small and obscure, surface with fine tubercles.

Horizon and locality. London Clay, Lower Eocene of Aveley, Chalk Farm, Copenhagen House, Isle of
Sheppey, Maylandsea, and Steeple.

Description. Cephalothorax. The median line on the rostrum is flanked by carinae, with a shallow depression
between. The postorbital spine is below but almost level transversely with the start of these carinae. The
rostrum is half the length of the carapace; its outer edges converge towards the tip. Approximately half-way
along the rostrum a large, anterodorsally pointing spine projects on either side of the margin. A similar but
larger spine projects on either side at about three quarters the length of the rostrum which then continues
almost straight and parallel sided. At the second pair of spines the shallow depression has almost died out.

The cervical groove is deep, starting almost level with the suborbital spine and dropping nearly straight
down with a very slight forward angle to join the antennal groove at an obtuse angle, turning towards the
front and continuing straight until below the antennal region where it turns sharply anteriorly. At the junction
of these grooves and the hepatic groove is a small obscure area known as the prominence omega. The
distinctive broad postcervical groove runs from the median line slightly more than half the distance from the
base of the rostrum to the posterior margin of the carapace. It slopes forward to the level of the antennal
groove to join the hepatic groove. This is semicircular and joins the cervical groove. The hepatic groove starts
strong, weakens towards mid length, then becomes stronger. At the top of the cervical groove a faint gastro-
orbital groove runs towards the front. The branchiocardiac groove is evident faintly on some specimens,
running backwards and upwards from the cervical groove at the level of the suborbital spine. A further faint
furrow extends downwards from the postcervical groove at the level of the base of the orbit. The frontal and
orbital regions are finely punctate; the antennal has fine tubercles and the odd pit; the branchial has regular
spaced, fine and larger tubercles with some fine pitting towards the posterior margin; the gastric has scale-like
pits on the top half, and tubercles on the bottom half; the cardiac is strongly pitted. (Refer to text-fig. 4 for
position of grooves.)

Abdomen. The first abdominal somite is the smallest, the tergum with a transverse smooth groove, the raised
portion pitted. The lateral margin of the pleuron is straight and finishes in a forward pointing, flat, blunt,
tooth which overlaps the posterior edge of the carapace. The second somite has a transverse groove on the
anterior third of the tergum, extending down to form a semicircle on the pleuron; the surfaces are smooth in
front of the groove on the tergum, variably pitted behind. The pleuron is broadly quadrate, the anterior angle
rounded, the posterior angle extended into a spine; between the lateral and the posterior margins is a large
deep pit; surfaces are ornamented with fine pits, with some areas nearly smooth. The third and fourth somites
have a transverse groove on the anterior third which extends to the boundary of the tergum at the anterior
margin; surfaces are smooth in front of the groove, pitted behind. The pleuron is triangular, falcate, with an
acute posteriorly directed spine. A weak semicircular groove starts and finishes at the pleuron boundary;
below this groove and on the median line is a large pit; surfaces are variably pitted. The fifth somite is slightly
smaller. The sixth is almost as deep as the second, with a groove crossing the tergum, which is pitted, near the
anterior edge. The pleuron is triangular with an acute spine. A weak semicircular groove, which is a continua-
tion of the tergum groove, finishes at a small spine on the posterior edge at the pleural boundary.

Telson. The telson is longer than broad, the outer margins converging slightly towards the rear. On each side
of the median line a rounded ridge extends posterolaterally from near the front to a posteriorly pointing spine
at slightly over three quarters distance along the margin. Beyond this the margin is semicircular, the surfaces

590

PALAEONTOLOGY, VOLUME 30

finely pitted. The endopodite is nearly as broad as long, the posterior margin well rounded, the sides converging
towards the front to form a well-rounded margin. A longitudinal ridge diverges slightly from the median line
on the inner side of the endopodite; there are fine tubercles on the ridge and the outer side; the inner side
away from the ridge is smooth; the posterior quarter is finely ribbed at right angles to the margin. The
exopodite is elliptical, nearly twice as long as broad and divided transversely by a suture at approximately
two thirds of its length. The posterior margin of the anterior section is armed with a series of spines, the
largest on the outer angle; the surface is finely punctate. The bottom third is finely ribbed, parallel with the
median line, its surface finely punctate.

Appendages. Antennae long, reaching nearly as far forward as the outstretched first pair of pereiopods.
Antennal peduncle slightly more than one third the carapace in length, the three segments round to oval in
cross-section and more or less equal in length to one another. On the first section of the peduncle the outer
edge is flattened to form a sharp ridge which develops into a forward pointing spine. Jutting out in front of
this first section and above the spine can be seen the remains of the antennal scale on some specimens. Pieces
of the antennule peduncle can be seen on BM In. 63354, length approximately to the end of the second section
of the antennal peduncle.

The third maxilliped is approximately two thirds the length of the carapace. The merus is triangular in
cross-section, its outer margin rounded; the lower margin has small evenly spaced pores; surfaces are finely
punctate. The propodus is slightly longer than the carpus, both more or less oval in cross-section. The dactylus
is spear shaped, flattened, approximately the same length as the carpus.

On the right-hand cheliped, the crusher, the outer margin of the propodus is rounded, with a faint suggestion
of a lateral groove immediately inside the margin on either surface. The margin is smooth with lines of variable
sized pits running along the fixed finger on the outer and inner surfaces. The fixed finger is oval in cross-
section near the tip, changing to half round by the base of the finger, the inner surface nearly flat; a median
triangular ridge runs along the length of the fixed finger. At the base of the ridge and running parallel to it is a
line of pits; the remainder of the surface is ornamented with fine pitting. The ridge of the finger is armed with
various sized teeth; from the base there are three oval blunt teeth of similar size, then a large blunt oval one
(greater diameter along the median line); these are followed by groups of fine teeth divided by a slightly larger
tooth. The palm is slightly shorter than the finger, and half as wide as long, nearly oval in cross-section; it
thins rapidly near the inner margin to form a lateral groove on either surface. The inner margin of the palm
has three large triangular spines on the outer edge, with a smaller spine behind the posterior tooth. Slightly
away from the lateral line of spines but parallel with it are two similar spines on the under surface. These
alternate with the previous spines. The dactylus is oval in cross-section, slightly flattened on both surfaces,
the outer margin rounded. The prehensile edge is armed with a large oval tooth (the greater diameter along
the median line), a smaller tooth separated by a gap from a large round tooth, then an assortment of large,
fine and slightly larger teeth.

On the fine or left-hand claw, the outer margin of the propodus is smooth and round, straight for most of
the length with a slight curve inwards at the tip, the surfaces with fine pitting. The fixed finger is nearly oval in
cross-section towards the tip, half round in cross-section at the base; a triangular ridge on the median line
runs along the prehensile edge, with fine spines or teeth in groups of three or four with a slightly larger spine
in between. This sequence is repeated until approximately one third the length from base where there is a
large, slightly offset spine followed by groups of fine and slightly larger spines. A line of pits runs parallel to
the ridge; surfaces have variable pits. The finger is at least one and a half times the length of the palm, which
is approximately half as wide as long. The outer surface of the palm is deeply convex, starting to flatten out
near the inner margin. The inner surface is not as strongly curved. The inner margin is armed with three
similar sized triangular spines with a smaller spine behind; parallel with these is another line of two or three
similar spines on the under surface, alternating with the previous spines. The dactylus is almost oval in cross-
section, flattened slightly on both surfaces, the outer margin rounded, the prehensile edge with a triangular
median ridge on the top of which are groups of three or four spiniform teeth, with a slightly larger one in
between. The carpus is nearly as broad as long with various anteriorly directed spines on the top and the
margins, the underside with an anteriorly directed spine at the front, the surfaces variably pitted. The merus
is half the length of the large claw, flattened to oval in cross-section, the front outer margin drawn out into a
long sharp point with a smaller spine on the inner margin. Only pieces of the ambulatory legs are preserved;
these are rounded to oval in cross-section, smooth with fine lines of pores or isolated pits; the second pair is
chelate.

Discussion. When M‘Coy (1849) established Hoploparia he included two new species from the
London Clay, H. gammaroides and H. belli. Bell (1858) maintained both species but Woods (1931)

QUAYLE: EOCENE CRUSTACEA

591

found that the smaller specimens (H. belli) differed in no significant respect from the larger (//.
gammaroides) and united the two. A study of the original material confirms the view of Woods.

Bell (1858, p. 40) noted that H. belli {= H. gammaroides) appears to resemble H. longimanus
(Sowerby) of the Lower Greensand. The latter has a line of tubercles behind and parallel with the
postcervical groove, which in H. gammaroides are finer (PI. 64, fig. 6). The junction of the antennal
and the cervical groove is at an obtuse angle in both. The ornamentation of the carapace is quite
different. H. longimanus has spines or tubercles on the carinae and a dentate carina runs back from
the antennal spine towards the cervical groove. In H. gammaroides this area is prominent but
without tubercles; it can vary in individuals as do some of the other surface decorations. H. stokesi
(Weller, 1903) from the Lower to Middle Campanian, James Ross Island, is similar to H. longimanus,
the main differences being in the carapace grooves, greater spinosity, and the proportions of the
cephalothorax (Ball 1960, p. 12). The American species H. tennesseensis Rathbun, 19266 from the
Upper Cretaceous, Ripley Formation, has a deep groove along the outer margin of the propodus
and a row of three spines on the inner side of the inner margin, with one spine at the proximal end
of the outer side of this margin. H. gammaroides has a double row of spines but arranged differently
and no groove along the outer margin of the propodus. H. gammaroides differs from the above
mentioned Cretaceous species in that the prehensile edge of the claw is armed differently.

Several Eocene species of Hoploparia are known, some only from pieces of claw. H. klebsi
Noetling, 1885 from Germany has a row of spines running back from the antennal spine; the
rostrum carinae running back on to the carapace are dentate. The inner margin of the palm of H.
klebsi is armed with forward pointing teeth as well as the outer margin of the dactylus (Forster and
Mundlos 1982, fig. 2). H. gammaroides differs from this species by having a prominent antennal
region without spines; the prehensile edge of the claw and the inner margin of the palm are armed
differently. H. groenlandica Ravn, 1903 from the Eocene of Kap Dalton, East Greenland has the
cervical joining the antennal groove at an obtuse angle as in H. gammaroides. The hepatic groove is
distinct throughout its length, and two dentate carinae run from the base of the rostrum nearly to
the cervical groove. The chelae of H. groenlandica appear to differ from those of H. gammaroides
in having a flat area on the inner and outer margins, but this may be preservational. Ravn (1903,
p. 119) noted that this species has a greater resemblance to H. gammaroides though he relied on
Bell’s (1858) description and figures to come to this conclusion. The specimens are now assigned to
Homarus and Hoploparia wardi sp. nov. The hepatic groove of H. gammaroides is not distinct
throughout its length; the rostral carinae are not armed and finish almost level transversely with
the postorbital spine. These characters make H. gammaroides different from H. groenlandica. H.
corneti van Straelen, 1921 from the Eocene of Belgium is too poorly illustrated (van Straelen 1921,
p. 136) for any detailed comparison with other species.

Hoploparia wardi sp. nov.

Plate 64, fig. 1 1; text-fig. Id

1849 Hoploparia gammaroides WCoy,^. 177.

1858 Hoploparia gammaroides M‘Coy; Bell, p. 38, pi. 8, fig. 4.

1858 Hoploparia belli M'Coy; Bell, p. 40, pi. 10, fig. 9.

1931 Homarus gammaroides M‘Coy; Woods, p. 93, pi. 27, fig. 3.

1980 Hoploparia gammaroides M'Coy; Morris, p. 9.

Derivation of name. Named after Mr David Ward.

Types. Holotype: SM C19314 (text-fig. 2d), Forbes Young Collection, London Clay, Isle of Sheppey. Para-
types: SM C19318 (PI. 64, fig. 11), Walton Collection, and BM 59132b, Wetherell Collection (Bell 1858, pi. 10,
fig. 9), London Clay, Chalk Farm; BM 46356, Bowerbank Collection (Bell 1858, pi. 8, fig. 4; Woods 1931,
pi. 27, fig. 3), London Clay, Isle of Sheppey.

Diagnosis. A Hoploparia with a prominent carina running back from the antennal spine with two,
possibly three tubercles and a further one between the nearly parallel postcervical and the cervical

592

PALAEONTOLOGY, VOLUME 30

grooves; distinct hepatic groove throughout its length; prominence omega triangular with fine
tubercles; carapace ornamented with well-spaced fine tubercles, more closely spaced ventrally.

Description. Cervical groove deep, runs anteroventrally to join the antennal groove which curves gently
anteriorly (rest of groove not seen). The distinct postcervical groove starts on the median line and runs
anteroventrally to join the semicircular hepatic groove which joins up with the cervical groove. The hepatic
groove is distinct throughout its length. Between the postcervical and the cervical grooves and in line with the
cheek, is a swollen area capped with a prominent tubercle. Antennal region elevated into a strong keel with
two, possibly three, tubercles. Carapace ornamented with variably sized fine tubercles, rear margin with a
smooth flange. First abdominal somite traversed by a smooth groove, surface anterior of groove smooth with
fine pitting behind. Somites two to four have a transverse groove on the anterior third which extends to the
boundary of the tergum at the anterior margin; surfaces are smooth in front of the groove, with fine pitting
behind which appears to increase in size and quantity from the tergum boundary downwards.

Discussion. In Bell’s (1858) redescription of H. gammaroides, he quoted a sentence from M’Coy’s
original description: ‘the cheek is elevated into a strong keel with about three large spinose tubercles;
cheeks prolonged as a semi-cylindrical sheath to the outer antennae half the length of the rostrum.’
Bell noted that he had not observed this structure, yet his figure 4 on plate 8 clearly shows a tubercle
above the hepatic groove. The specimen ( BM 46356, Bowerbank Collection) plainly shows a tubercle
above the hepatic groove, with the cheek elevated into a strong keel, with two, possibly three,
spinose tubercles. As regards the cheeks, it is quite possible that on the specimens available the
antennal peduncle was fragmentary; it could then be mistaken as a continuation of the cheek. Bell
also noted p. 41 that BM 59132b (pi. 10, fig. 9) differed in some respects from H. belli.

Rathbun (1935, p. 61) used four incomplete specimens to describe H. johnsoni from the Lower
Eocene, Sucarnoochee Bed, Midway, Alabama. It has at least one spine on the lateral ridge leading
to the rostrum and the antennal ridge is weakly defined (Rathbun 1935, pi. 14, figs. 27 and 28). H.
wardi, though similar to H . johnsoni, is different in that it has a pronounced antennal ridge, smooth
carinae from the rostrum, and a different arrangement of the larger carapace tubercles. H. wardi
differs from its near neighbour H. gammaroides by the arrangement of the spines leading towards
the antennal spine, the deep and clearly defined hepatic groove, and the postcervical and cervical
grooves which run parallel and closer together. The prominence omega is triangular and ornamented
with fine tubercles on both of these species. In H. klehsi the postcervical and cervical grooves appear
to be parallel similar to H. wardi (Forster and Mundlos 1982, pi. 33, fig. 3) but in H. wardi the
Carina from the rostrum is smooth and the large carapace tubercles are different, in H. groenlandica
the postcervical and cervical grooves are not parallel but diverge dorsally as for H. gammaroides,
whereas in H. wardi, the postcervical and cervical grooves run parallel and the carapace is orna-
mented with several large tubercles.

Hoploparia victoriae sp. nov.

Plate 64, figs. 8 and 9

1979 Hoploparia gammaroides M‘Coy; Quayle and Collins in Kemp et al.,p. 102.

1981 Hoploparia gammaroides M‘Coy; Quayle and Collins, p. 735.

Derivation of name. After my daughter. Miss Victoria Quayle.

Types. Holotype: BM In.63357 (PI. 64, fig. 8). Paratypes: BM In.63358-In.63363 (PI. 64, fig. 9), all e.v JSQ
Collection, from Unit 7, Elmore Formation, Bracklesham Group, Middle Eocene, Lee-on-the-Solent,
Hampshire.

Diagnosis. A Hoploparia with the cervical joining the antennal groove in a curve; prominence omega
well rounded and conspicuous, plain or pitted surface; antennal region rounded but not prominent;
posterior border of postcervical groove plain or pitted.

Description. Cephalothorax. The rostrum is half the carapace in length, with a median depression bounded on
each side by a carina; at approximately the mid length is the base of a spine and at the end a further spine (the

QUAYLE; EOCENE CRUSTACEA

593

arrangement of the spines is possibly as for H. gammaroides). The cervical groove is deep, extending ventrally
from the level of the suborbital spine, forming a smooth curve to join the deep antennal groove. The antennal
groove becomes finer towards the anterior margin of the carapace, where it almost disappears. At the Junction
of these grooves and the weak hepatic groove is a small, prominent well-rounded area, the prominence omega.
The hepatic groove alongside this area is smooth, wide, and deep, but elsewhere, even on well-preserved
specimens, is very indistinct. A weak furrow extends anteriorly from the top of the cervical groove, passing
under the suborbital spine. The postcervical groove is deep, situated at over half the carapace length from the
base of the rostrum. The dorsal part cuts the median line at right angles and runs anteroventrally approximately
30° to the vertical to join up with the hepatic groove. A further deep furrow extends ventrally from the
postcervical groove at the level of the base of the orbits. The branchiocardiac groove curves posterodorsally
from the cervical groove, at the level of the suborbital spine. The marginal furrow is well marked and the
posterior margin bears a smooth, finely punctate flange. Surface ornamentation on the carapace is as follows:
frontal and orbital area, smooth with well-spaced pits; antennal area, pitting with small tubercles; branchial
area, blunt tubercles with pits; gastric and cardiac areas, surface rough with pitting.

Abdomen. The tergum of the first abdominal somite is crossed by a deep groove two thirds the distance from
the front; raised portions are pitted. A groove crosses the tergum of the second somite in the anterior third
and extends down to form a semicircle on the pleuron. The pleuron is quadrate, the anterior angle rounded,
the posterior angle with a posterior pointing spine. Between the bottom of the semicircle and the posterior
spine is a large clear pit; the rest of the surface is ornamented with variable pitting. The third somite is likewise
traversed by a groove which extends to the pleuron, where it forms a weakly defined semicircle. The pleuron
is falcate, with an acute posteriorly pointed spine, its surface ornamented with variable pits, the area in front
of the tergum groove smooth. Somite four is similar to three; the groove on the pleuron is very weak, the
pleura having a solitary pit similar to that on somite two. The fifth somite is like three and four but narrower,
its surfaces variably pitted.

Ventral surfaces. There is a median tubercle on the abdominal sterna of the first to third somite. Mandibles
are almost rectangular, smooth and transversely convex with other surfaces and margins rounded.

Appendages. On the right-hand first cheliped, the crusher, the palm is approximately twice as long as wide,
oval in cross-section, thinning rapidly towards the inner margin which has two rows of alternating, anteriorly
directed spines (six in all?). The palm is coarsely pitted. The fixed finger is nearly oval in cross-section, its
outer margin rounded with a slight groove on the upper surface, its inner margin wide, almost flat. From the
base of the fixed finger the prehensile edge is armed with depressed oval teeth that he close together with their
greater length at right angles to the median line. The first tooth is the smallest with a marked increase in size
to the third; then the teeth remain almost the same size up to the tenth (rest not seen). The dactylus (of which
only fragments are known) is oval in cross-section, its outer margin wide and flat, the prehensile edge armed
with close, round to oval, flattened teeth, smallest at the joint. On the upper and lower surfaces a spine is
evident by the joint. On the left-hand first cheliped, the small claw, the palm is approximately twice as long as
wide and oval in cross-section. It thins rapidly towards the inner margin which is dentate with a double row
of alternating spines. The fixed finger has a rounded outer margin, the inner margin almost flat for its entire
width, the prehensile edge armed with small spiniform teeth for most of its length. The carpus is approximately
two thirds the length of the merus, triangular in cross-section, its inner and outer margins tapering towards
the merus. The lower surface has a row of three or four anteriorly directed spines. The upper surface is slightly
rounded with an anteriorly directed spine towards the rear. Ornamentation on the upper surface is scale-like.
There are small blunt tubercles on the lower surface. The merus is over twice as long as wide, oval in cross-
section, its inner and outer margins tapering from the front towards the ischium. There are one or two blunt
spines on the anteroventral margin; surfaces are smooth, finely punctate. The ischium is smooth and finely
punctate. On the second and third pereiopods the merus and ischium are round and smooth with surfaces
finely punctate; the merus is approximately three times the length of the ischium.

Discussion. H. victoriae differs from H. gammaroides in that the antennal and cervical grooves join
in a curve; the posterior edge of the postcervical groove is plain or with pits; the antennal region is
rounded on the surface and only prominent at the front; prominence omega rounded with plain or
pitted surface; the prehensile edge of the propodus of the crusher claw is armed with oval teeth
increasing in size away from the joint and transverse to the median line. H. victoriae differs from
H. wardi and H. klehsi in that the cervical and postcervical grooves are further apart and diverge

594

PALAEONTOLOGY, VOLUME 30

dorsally, and in the carapace ornamentation, small tubercles with fine pitting. It differs from H.
groenlandica in that the rostral carinae are smooth.

Infraorder palinura Latreille, 1803
Superfamily glypheoidea Winckler, 1883
Family GLYPHEIDAE Winckler, 1883
Genus glyphea von Meyer, 1835

Type species. Palimous regleyanus Desmarest, 1822, by original designation.

Diagnosis. Carapace with short pointed rostrum; tuberculate longitudinal carinae on anterior part.
Cervical groove deep and steeply inclined in lateral view. Postcervical and banchiocardiac grooves
very oblique, joined medially and laterally and in some species at additional points. Anterior portion
of carapace rectangular in dorsal and lateral views. Branchiostegite with long narrow anterior
extension. Abdominal terga smooth, telson rounded, exopods of uropods with diaeresis. Antennal
scale pointed; first pereiopod subchelate.

Range. Upper Trias-Lower Eocene.

Glyphea scahra (QqW, 1858)

Plate 64, figs. 12 14; text-fig. 56

1858 Trachysoma scahrum Bell, p. 41, pi. 10, fig. 1 1.

1929 Trachysoma scabnim Bell; Glaessner, p. 387.

1930 Trachysoma scahrum Bell; Woods, p. 73, pi. 22, fig. la, h.

1969 Trachysoma scabrum Bell; Glaessner, R464.

1974 Trachysoma scabrum Bell; Cooper, p. 85.

1980 Trachysoma scabrum ^e\\-,y[orns,p. 17.

Holotype: BM 59146 (PI. 64, fig. 12), by monotypy, London Clay, Chalk Farm, Camden Town, London.

Other material. BM In. 63366 (PI. 64, figs. 13 and 14), collected by D. Ward, London Clay, Aveley, Essex;
PE 80/454, collected S. W. Vincent, A127 (roadworks) at Folkes Lane, Cranham (TQ 580 884).

Diagnosis. A Glyphea with the carapace covered in different size sharp pointed tubercles.

Description. Cervical groove deep, starting at a point about two fifths the distance from the front to the
posterior margin of the carapace. It runs nearly straight, until half-way across the carapace where it bends
slightly forward and continues to the antennal groove. In front of the cervical and above the antennal groove
are three longitudinal carinae armed on the top with sharp anteriorly directed tubercles; the surfaces between
the carinae are ornamented with further sharp tubercles. The antennal groove is deep, nearly straight with a
slight upwards angle. The branchiocardiac groove is clearly marked. It starts at approximately four-fifths the
distance from the front with a slight forward curve, then runs almost straight across the carapace at an angle
of 45° to meet the small lobe above the hepatic lobe. The indistinct postcervical groove forms an acute angle
with the branchiocardiac groove near the dorsal margin; anteriorly the grooves are joined by a semicircle and
a triangular area is formed between them. The hepatic groove starts at the junction of the branchiocardiac

TEXT-FIG. 5. a, Glyphea sp. ( x 1 -5). 6, Reconstruction of Glyphea scabrum (Bell) ( x 2-2).

QUAYLE: EOCENE CRUSTACEA

595

and inferior grooves, runs slightly anterodorsally from this junction, then ventrally before joining the cervical
groove above the antennal groove. The area bounded by the hepatic, inferior groove and the lower margin is
roughly rectangular and decorated with sharp tubercles. The remainder of the surface of the carapace is also
ornamented with tubercles.

What may be part of the propodus is evident on BM In. 63366 (PI. 64, fig. 13). It is oval in cross-section, the
top margin with anteriorly directed equally spaced teeth, the outer surface with a smooth groove running
parallel with the top margin; all surfaces with tubercles. (Refer to text-fig. 4b for position of carapace grooves.)

Discussion. Bell (1858) described the new genus Trachysoma from one specimen collected by
Wetherell, to which he gave the trivial name scabrum. The holotype consists of an incomplete
carapace with limb fragments and part of the antennal peduncle. Bell (p. 42) was not certain of the
affinities of this species. Woods (1930) figured this unique specimen (pi. 22, fig. la, h) but gave no
description or remarks. Cooper (1974) mentioned a specimen collected by D. Ward, BM In. 63366.
This consists of an imperfect carapace with a limb fragment. The only other known specimen (PE
80/454), part of a carapace, was collected by S. Vincent. Glaessner’s (1969) diagnosis of Trachysoma
refers to a long, low and narrow carapace with deep straight cervical grooves, and straight
postcervical and branchiocardiac grooves. He also included Glypheopsis Beurlen in this genus, the
type of which is Orphea ornatci Quenstedt, 1858. This appears to have straight postcervical and
branchiocardiac grooves (R465, fig. 269.2a, b). BM 59146 and BM In. 63366 have had little or no
preparation since being collected. What appears to be the rounded ventral margin of the carapace
on BM 59146 is in fact a limb lying across the carapace. Likewise what appears to be a dentate
ventral margin forward of the cervical groove is in fact a line of sharp tubercles. Both of these
specimens were masked by matrix in a similar position which concealed the true shape of the
carapace, which preparation of BM In. 63366 has subsequently revealed, though details of the
rostrum and anterior are still not known. A new reconstruction based on these specimens (text-fig.
56) shows that the carapace compares in its proportions and grooves to Glyphea (text-fig. 5a) to
which it is here assigned.

Superfamily palinuroidea Latreille, 1803
Family palinuridae Latreille, 1802
Genus archaeocarabus M‘Coy, 1849

Type species. Archaeocarabus bowerbanki M‘Coy, 1 849, London Clay, Isle of Sheppey, by original designation.

Diagnosis. Rostrum of moderate size made up of three spines, centre one gripped by two processes
of the ophthalmic somite; supraorbital spines widely separated. Sternal plate with four pairs of
tubercles.

Range. Lower Eocene, England.

Discussion. Woods (1931) diagnosed Archaeocarabus as like Palinurus but with the rostrum similar
to that of Jasus Parker, 1883, particularly in that it is gripped by two processes of the ophthalmic
somite. He distinguished it from Jasus by the more widely separated supraorbital spines and the
presence of a row of tubercles on the plastron. The differences in the rostral area between Jasus
and Archaeocarabus can be seen by comparing text-fig. 6c with text-fig. 6d. In Archaeocarabus there
are three spines; in Jasus a single central spine. The rostrum or central spine in each case is gripped
by two processes of the ophthalmic somite. These two processes in Archaeocarabus tend to lie
alongside the spine with a slight upwards curve, rather than come straight up from the ventral side
as in Jasus.

Archaeocarabus bowerbanki M‘Coy, 1849

Plate 65, figs. 1-9; text-fig. 6

1849 Archaeocarabus bowerbanki y[^Coy,p. 174.

1854 Archaeocarabus bowerbanki M'Coy; Pictet, p. 443, pi. 42, fig. 3.

1858 Archaeocarabus bowerbanki M‘Coy; Bell, p. 42, pi. 1 1, figs. 1-5.

596

PALAEONTOLOGY, VOLUME 30

1925 Archaeocarabus bowerbanki M‘Coy; Woods, p. 36, pi. 8, fig. 5; pi. 9, fig. 6; pi. 10, figs. 1-3.

1929 Archaeocarabus bowerbanki M'Coy; Glaessner, p. 57.

1962 Archaeocarabus bowerbanki M'Coy; Roberts, p. 176.

1969 Archaeocarabus bowerbanki M'Coy; Glaessner, R473, fig. 277.2a, b.

1974 Archaeocarabus bowerbanki M'Coy; Cooper, p. 85.

1980 Archaeocarabus bowerbanki M'Coy; Morris, p. 1.

Types. M'Coy’s syntypes are SM C7737, C7738, Cl 9074, BM 46358, 46359, and 59764, all from the London
Clay, Isle of Sheppey. SM C7737 (Woods 1926, pi. 9, fig. 6) is here designated lectotype.

Other material. BM 38388 (PI. 65, fig. 6) and 59140 (PI. 65, fig. 8); BM In.63378 (PI. 65, fig. 5), In.63379
(PI. 65, fig. 4), In.63380 (PI. 65, figs. 1 and 3), In. 63381 (PI. 65, fig. 2), In. 63382 (PI. 65, fig. 9), In. 63383-
In. 63386 (PI. 65, fig. 7), ex JSQ Collection; all from the London Clay, Isle of Sheppey.

Diagnosis. As for genus.

Description. Cephalothorax. The front is straight and occupies nearly half the carapace width; it is armed with
several anteriorly directed spines. Length of the rostrum is two thirds of its width and occupies approximately
a quarter of the width of the frontal margin; it consists of a thin median spine with a slightly shorter stout
lateral spine either side. The median spine is armed dorsally with one anteriorly directed spine at approximately
half the length, with a further spine at the base; lateral spines, broad at the base, are flat and triangular; the
median spine is gripped on either side by two slightly upturned rounded cone-shaped processes of the ophthalmic
somite. The somite is slightly wider than the rostrum; its outer margins are angled inwards; the top surface is
concave and the sides drop steeply away; surfaces are smooth with a few small pits. The remainder of the
front is made up of three short spines on either side of the rostral spines, followed by the supraorbital spines.
These are large, anteriorly directed, laterally compressed and triangular in shape and lean slightly outwards;
they are approximately half the carapace width in length and are situated at the anterolateral angle. Behind
each runs the postorbital carina; this is cut by the cervical groove and continues behind with a slight inwards
angle, gradually dying out and ending three quarters of the length of the carapace from the front. On the
carina there are two smaller, anteriorly directed spines behind the supraorbital spine and behind these, large
spiny tubercles are interspersed with smaller ones. Between the supraorbital spines the carapace is almost flat
with a few small tubercles. Half the distance from the front to the postcervical groove there is a transverse
row of three small tubercles. Behind this and on either side of the mid-line are two longitudinal rows of three
larger, anteriorly directed spines; from the second tubercle to the postcervical groove the general ornamentation
of the carapace changes to a greater number of tubercles.

The postcervical groove is deep and broad, straight for half the width of the carapace, then joining the
cervical groove which curves in front of the fourth spine on the continuation of the postorbital carina; it
continues to run straight to cut the lateral margin at right angles, drops vertically down the side as a deep
groove, then develops a slight forward curve which continues as the antennal groove. Behind the postcervical
groove a slight median carina with large and small tubercles runs almost to the posterior margin where it is
cut by a deep transverse groove, parallel to and just inside the flanged edge of the posterior margin. The
lateral margins behind the cervical groove are parallel and start with a large forward pointing spiny tubercle,
followed by four or five smaller spiny tubercles to half-way along the margin; a more general ornamentation
then develops of small and large tubercles. The rest of the carapace is ornamented with a variety of medium
and small tubercles. The large and small tubercles have on their front sides rows or groups of fine pores at the
base. Below the supraorbital spine the orbital margin finishes in a large triangular infraorbital spine similar to
but not as large as the supraorbital. Here starts the antennal carina which runs parallel to the postorbital
carina and finishes at the antennal groove; it is armed with three flattened anteriorly directed triangular spines.
At a point half-way between the carinae and alongside the cervical groove is the start of another carina that

EXPLANATION OF PLATE 65

Figs. 1-9. Archaeocarabus bowerbanki M^Coy, London Clay, Isle of Sheppey. 1 and 3, BM In.63380. 1, lateral
view, xl T. 3, dorsal view, xl l. 2, eye and orbit, BM In. 63381, x2. 4, basal segments of antenna, BM
In.63379, x 1-7. 5, ophthalmic somite, BM In.63378, x4. 6, anterior of BM 38388, x 2. 7, lateral view,
BM In. 63386, xO-8. 8, sternum, BM In. 59140, xl. 9, claw, BM In.63382, xL8.

Specimens in figs. 3, 5, and 9 have been whitened with ammonium chloride.

PLATE 65

QUAYLE, Arcluicocarahus

598

PALAEONTOLOGY, VOLUME 30

TEXT-HG. 6. Archaeocarabiis howerbanki M'Coy, diagrammatic reconstruction. A, B, lateral and dorsal view
( X 11). c, details of the ophthalmic somite and rostrum ( x 3-6). D, Jasus lalandii (Milne-Edwards), Recent,
South Africa, frontal area (after Glaessner 1969) (x L25): a2, base of antenna; m, articulating membrane;

al, antennular base; oph, eye stalk.

QUAYLE: EOCENE CRUSTACEA

599

forms the continuation of the lateral margin posterior to the cervical groove, which dies out before the orbital
margin and is armed with two or three blunt spines. There are a further two, anteriorly directed triangular
spines between this carina and the postorbital carina at the forward edge of the cervical groove.

Abdomen. On the abdominal somites the tergum is more or less semicircular in cross-section, with a grooved
and flanged posterior margin. The pleura have one to three flat spines on each margin and end in a large
recurved central spine. On the first somite the flange develops a triangular flap on each side, which accommo-
dates the posterolateral angles of the carapace when the abdomen is rolled up. The surfaces are variably
pitted.

Telson. The calcified part of the telson is parallel sided, with the posterior margin concave; there are a few
tubercles towards the centre. Approximately the first half of the outer margin of the exopodite and endopodite
is calcified. This area is elliptical, approximately three times longer than wide, smooth on the outer edge,
strongly serrate on the inner and terminates in a ventrally directed spine. Only fragments of the remainder of
the tail are known.

Ventral structures. The anterior margin of the epistome is concave on each side of the median spine and
reached by the median furrow. The anterior part of the sternum is a small rectangular area, a fifth of the total
length; the outer margins then diverge to give the maximum width at the second sternite, which remains
almost constant as far as the concave rear margin. Along the median line and on the rear margin of each
sternite behind the rectangular area are pairs of tubercles. The first and last pair are smaller than the
intermediate pairs. The surface of the sternum is finely granulated and decorated with tubercles on the outer
margins.

Appendages. The long slender basal podomere of the antennule reaches as far forward as the end of the second
podomere of the antenna. The basal podomere expands distally to accommodate the second slender podomere,
only small pieces of which are preserved. The first three podomeres of the antenna are round to oval in cross-
section and armed with various sized spines. The dorsal and ventral margins of the second and third podomeres
are flattened slightly longitudinally, and armed with large flattened anteriorly directed triangular spines. The
flagellum is unknown but was probably similar to Recent Palinurus. The eye is very large and the peduncle
short, about one third the width of the eye.

The first pereiopod is chelate, the propodus three to four times the width of that on limbs two to five, but
approximately the same length; in cross-section it is round to oval, becoming broadly dilated towards the
dactylus; the width of this flattened end is two thirds the length of the rounded, slightly curving, outer margin
of the pointed dactylus. On the outer margin, opposite the joint, is a large posteriorly directed spine. There
are two or three longitudinal rows of uniform pits on the outer and inner surfaces, with the odd larger pit. On
the second, third, and fourth pereiopods the merus is slightly longer than the propodus. Both are oval to
triangular in cross-section. The merus is smooth and finely punctate; the propodus has longitudinal lines of
small evenly spaced pores; the rest of the surface is finely punctate. The dactylus is slightly less than half the
length of the propodus, round in cross-section, with a pointed end.

Discussion. Due to the very fragile nature of the carapace of Archaeocarahits, it is rarely found
complete; BM In. 63380 (PI. 65, figs. 1 and 3), though small, is possibly the best preserved. The
carapace of the larger specimens appears to be more rounded, with the carinae becoming indistinct.

Rathbun 1935 described two new species of Arcliaeocarabu.si and in 1945, a further one. Roberts
(1962, p. 176) was doubtful whether these three species should have been placed in Archaeocarahus.

Genus LiNUPARUS White, 1847

Type species. Palinurus trigonus von Siebold, 1824, by original designation.

Diagnosis. Cephalothorax depressed, carapace with three longitudinal keels, no rostrum; supra-
orbital spines close to the median line, fused to form plate or separated by indentation.

Range. Lower Cretaceous Recent.

Discussion. Three species of Recent Linupanis were recognized by Berry and George (1972): L.
trigonus (von Siebold), 1824, western Pacific, Japan, South China Sea, Philippines, and eastern
Australia, L. sordidus Bruce, 1965, from the South China Sea to north-western Australia, and L.

600

PALAEONTOLOGY, VOLUME 30

somniosiis Berry and George, 1972, from the western Indian Ocean, Mozambique, and Natal, living
in depths from 81 to 328 m. There were variations in the type series of each, small specimens having
better developed spines than large ones. As pointed out by Woods (1925) there are also variations
in the fossil species.

Linuparus eocenicus Woods, 1925
Plate 66, figs. 1- 5; text-fig. la, h

1858 Thenops scyllariformis Bell, p. 33, pi. 7, figs. 5?, 6, 7?, 8?

1925 Linuparus eocenicus Woods, p. 31, figs. 3 -5; pi. 7, figs. 4-6; pi. 8, fig. 1.

1929 Linuparus eocenicus Woods; Glaessner, p. 233.

1974 Linuparus eocenicus Woods; Cooper, p. 85.

1980 Linuparus eocenicus Woods; Morris, p. 10.

1980 Linuparus scyllariformis {BeWy, Morris, p. 10.

Types. Woods’s syntypes are SM C7732, C7733 (PI. 66, fig. 3)-C7736, Meyer and Carter collections, from the
London Clay, Portsmouth Docks, Hampshire; BM 59145, London Clay, Highgate tunnel, London. SM C7735
(PI. 66, figs. 1 and 2) is here designated lectotype.

Other material. SM X1357-1362; OUM L552-L554, L568 L570; BM ln.63370, ln.63371 (PI. 66, fig. 4)-
ln.63374 (PI. 66, fig. 5)-ln.63377; all c.x JSQ Collection, Whitecliff Bay, Isle of Wight. The following of Bell’s
syntypes of L. scyllariformis-. BM 59012, 59107 59109, 591 1 1, 59141 certainly, and 46365, 46367, 59765, 59114,
and 591106 probably, belong to L. eocenicus.

Horizon and locality. The London Clay of; Portsmouth Docks; Isle of Sheppey; Burnham on Crouch; Highgate
and Chalk Farm, London; Frinton on Sea; Base of Division E, Barwell Court, Surrey; and the Wittering
Division, Bracklesham Group, Whitecliff Bay.

Diagnosis. A Linuparus with the supraorbital spines united for half their length; carinae with spiny
tubercles, raised areas decorated with fine tubercles.

Description. Cephalothorax. The supraorbital spines take up approximately one quarter the width of the
frontal margin. They are ridged longitudinally, the outer margin sloping steeply away from the ridge. The
inner margins, from the tip of the spine to the centre line where they meet, form a shallow U. They are united
for approximately half their length. The frontal margin outside the supraorbital spines is concave, with well-
spaced tubercles running along the top surface, and terminates in an anteriorly directed anterolateral spine. A
short distance behind each supraorbital spine is a postorbital spine, which continues posteriorly as a ridge as
far as a median spine. Behind this the median line is flanked by carinae, which diverge slightly and then
converge meeting just anterior of the cervical groove; these carinae bear small anteriorly directed spines. In
between these carinae almost at the cervical groove is a pair of small, deep, round pores, flanking the median
line. From near the base of the supraorbital spines, rounded carinae extend outwards in a curve to the cervical
groove and bear several forward pointing spines; the area bounded by these is concave. The raised surfaces
are covered with fine tubercles; other areas are nearly smooth. External to the curved carina the carapace at
first slopes sharply ventrally and then becomes nearly horizontal; surfaces are tuberculate. The lateral margin
in front of the cervical groove is angular and bears several small anteriorly directed spines and usually one or
two larger spines along its length; it is nearly parallel to the axis of the body. The cervical groove is deep and
in the form of a shallow V. At the lateral margin the deep groove runs directly ventrally and then bends
sharply anteriorly. Behind the cervical groove are three nearly parallel longitudinal carinae, the median
prominent, the outer angular. Between the median and lateral carinae the carapace is flattened or concave
and ornamented with regularly spaced uniform tubercles. Carinae are ornamented with numerous spines,
usually decreasing in size towards the posterior. External to the lateral carina the carapace is vertical except
near the cervical groove, where the upper part is concave and the lower convex. Just in front of the posterior
margin of the carapace is a broad smooth groove which interrupts the median carina and curves backward on
each side, becoming narrower towards the lateral margins. This part of the carina is produced into an
anteriorly directed spine.

Abdomen. The abdomen is similar in length to the carapace, tapering slightly posteriorly. The carinae on the
carapace are continued as slight ridges on each tergum. The first somite bears one anteriorly directed spine on

QUAYLE; EOCENE CRUSTACEA

601

TEXT-FIG. 7. a, h, Linuparus eocenicus Woods, a, diagrammatic reconstruction ( x 0-7); h, supraorbital spines
( X 21). c, L. scyllariformis (Bell), supraorbital spines ( x 1-7).

602

PALAEONTOLOGY, VOLUME 30

the median ridge at the posterior margin. In front of this a smooth groove runs almost parallel to the posterior
margin for approximately half the width of the somite, where it divides. One arm runs to the posterior margin
at the outer edge, the other to the outer edge near the front of the somite; the two triangular areas enclosed
are ornamented with some fine tubercles and pits; the area in front is smooth. Somites two, three, and four
have a median carina, which is interrupted in the middle of each tergum by a broad transverse groove,
sloping posteriorly on each side, narrowing and nearly reaching the posterior margin. The tergum surface is
ornamented with a mixture of pits and fine tubercles. Each tergum has two anteriorly directed spines on the
median carina, one in front and one behind the groove; on the outer margins there is a suggestion of a
longitudinal ridge with a spine at the front. On the fifth somite the spines on the median ridge have disappeared,
but the outer spine is still evident with the transverse groove becoming indistinct. The sixth somite, which is
almost twice the length of the others, has a smooth median groove with a longitudinal row of tubercles on
each side; the rest of the surface is ornamented with fine tubercles. The pleura on somites four and five bear
two spines on the forward margin followed by a central ventrally directed strong spine. The posterior margin
is smooth laterally becoming serrate, the size of teeth increasing adaxially; a concave area flanks the central
spine. Only parts of the pleura of somites two and three are known but they are presumably as for four and
five.

Ventral structures. The epistome has a strong median groove between prominent ridges which diverge an-
teriorly. The anterior margin is concave with a tubercle either side of the median groove. The anterolateral
margins are directed at approximately 45° to where they turn posteriorly and become almost parallel. Between
the margins and the two median ridges are two large depressions, one alongside each ridge.

The sternum between the third maxilliped and the first pair of pereiopods is triangular and bears two flat,
rounded bosses anteriorly, one behind the other; surfaces are pitted. The coxae of the third maxillipeds fit
between these bosses. Each sternite has ridged sides tapering towards the front, decorated with pits and
tubercles. The lateral and posterior margins join in a triangular, laterally directed spine. The sternite for the
first pereiopod is ridged along the median line; the second, third, and fourth have a large tubercle either side
of this line. The fifth has two pairs of small tubercles, one pair either side of the median line; the outer ridge is
smooth except for some tubercles towards the front. Between the tubercles on the sternite for the fourth limb
is a groove along the median line. There is a single central tubercle on the first abdominal somite. On the
second there are three tubercles, the two anterior approximately half the size of the posterior one. On the
third somite the three tubercles are in a similar position to the second but are all the same size with a smaller
tubercle in between the pair. The fourth and fifth have a pair of tubercles, one either side of the median line.

Appendages. The antennular somite is longitudinally rectangular with a broad median furrow and it extends
forward between the basal podomere of the antennae. The basal podomere of the antennule is round and
thin, extending past the second podomere of the antenna. The first podomere of the antenna is broad and one
third wider than the second. The outer and inner margins are sharp and armed with two or three forward
pointing spines. The second podomere is similar in length to the first; the third is the longest, but also the
narrowest.

The coxa of the third maxilliped is triangular, its surfaces pitted. The merus is approximately half the size
of that on the second pereiopod; surfaces nearly smooth with very shallow pits. The first pereiopod is larger
than the rest, the coxa nearly as long as the ischium and basis, surfaces with shallow pitting. The merus, three
times as long as the carpus, is flattened and forked at the carpus articulation with sharp spines on the ventral
margins by the fork; the propodus is half the length of the merus. Other limbs are similar but the fifth appears
to be smaller.

Discussion. Of the three Recent species of this genus L. trigonus is the nearest to L. eocenicus. On
L. trigonus the supraorbital spines have a dentate inner margin; the frontal margin has a prominent
spine half-way between the supraorbital and the anterolateral spine. On L. eocenicus the supraorbital
spines have a smooth inner margin and there is no prominent spine on the frontal margin. Possibly
the arrangement of the spines on the pleura is different.

The fossil species are known from the Cretaceous and the Eocene. L. canadensis (Whiteaves,
1885), Northwest Territory, Highwood River, Alberta, has two spines on the tergum boundary
whereas L. eocenicus has one. L. vancouverensis (Whiteaves, 1895) from the Nanaimo Group,
Vancouver and Hornby Islands, British Columbia, has a double row of tubercles on the carapace
carinae (Rathbun 1935, pi. 10, figs. 1-3); the carapace carinae continue back on to the tergum as
ridges armed with a single row of tubercles. In L. adkinsi Rathbun, 1935 (pi. 10, figs. 4-10) from
the Denton Clay, Texas, the three carinae behind the cervical groove are granulated, the median

QUAYLE: EOCENE CRUSTACEA

603

one with a double row of tubercles. The tergum has a pronounced median carina with a series of
tubercles. L. eocenicus bears a single row of tubercles on the carapace carinae which continue back
on to the tergum as a ridge with a single or double spine.

The English Cretaceous has produced L. car/m (Reed, 1911) from the Lower Greensand, Ather-
field. Isle of Wight. The cervical groove starts as in other species of the genus but forms an obtuse
angle where it becomes transverse just before it crosses the median line; from this obtuse angle a
further groove runs backwards. In addition the arrangement of the tubercles on the carinae and
the general ornamentation of the carapace (Woods 1925, pi. 7, figs. 2 and 3) differentiate this species
from other members of the genus.

The Eocene members of the genus include L. texamis Rathbun, 1935 (pi. 16, figs. 9 and 10) from
the Midway of Dimmit County, Texas. It is possible that this species, which was described from
one specimen, is not distinct. Rathbun (1935) described the abdomen as lacking the first and second
segments: it is suggested that these are really segments one to five and not three to seven. The
median spine on all the specimens of species examined is lost on the fifth segment with a groove
appearing on the sixth, which does not appear to happen in this case. L. wilcoxensis Rathbun, 1935
(pi. 16, figs. 11-14) from the Sucarnoochee Beds, Wilcox County, Alabama, appears to be very
similar to L. eocenicus. The abdomen has two spines on the median carina on segments two to four
with possible differences on the outer carinae of these segments.

Woods (1925, p. 32, pi. 7, fig. 6h) described the antennular somite of L. eocenicus as triangular.
On closer examination the sides of the Triangle’ are seen to be chipped or broken and in recently
collected specimens it is evident that this somite is long and parallel sided, with a deep median
furrow (PI. 66, fig. 5).

Linuparus scyllciriformis (Bell, 1858)

Plate 66, figs. 6 8; text-fig. Ic

1858 Thenops scyllariformis Bell, p. 33, pi. 7, figs. 1 4.

1925 Linuparus scyllariformis (Bell); Woods, p. 29, pi. 8, fig. 2fl, h.

1929 Linuparus scyllariformis (Bell); Glaessner, p. 233.

1974 Linuparus scyllariformis (Bell); Cooper, p. 85.

1980 Linuparus scyllariformis (Bell); Morris, p. 10.

Types. Bell’s syntype BM 59106, figured Woods (1925, pi. 8, fig. 2), London Clay, Whetstone, London is here
designated lectotype. Of the paralectotypes BM In.43325, BM 591 lOn, 59113, 59142, 59143 certainly and
59144 probably belong to L. scyllariformis.

Other material. OUM L265, L460, L462 (PI. 66, fig. 8), L555 (PI. 66, fig. 6), Kirby Collection; BM In. 63367-
In. 63369, ex JSQ Collection; PE 82/395 collected C. King.

Horizon and locality. The London Clay of: Herne Bay, Aveley, Maylandsea, Steeple, Whetstone, Felixstowe,
and roadworks on the M25, Ockendon Road, approximate grid reference TQ 565 926.

Diagnosis. A Linuparus with large separate pyriform supraorbital spines, with prominent pits as
surface decoration on ridges and spines.

Description. Cephalothorax. Large separate pyriform supraorbital spines on either side of the median line are
divided by a deep U-shaped depression. The frontal margin runs slightly concave outwards from near the
base of these spines to end in a prominent anteriorly directed anterolateral spine. The dorsal surface along the
frontal margin bears evenly spaced blunt tubercles; the front edge of the margin is at right angles to the dorsal
surface and is smooth except for rare fine tubercles, which tend to increase in number towards the anterolateral
angle.

Directly behind and in line with the supraorbital spines are blunt postorbital spines, which continue back
in the form of rounded ridges; between and parallel to these is a small rounded median ridge, separated by a
small gap from a rounded tubercle at the front. Posterior to this, there is on each side a rounded carina. These
diverge slightly almost as far as the cervical groove where they converge enclosing a flat area. From near the
base of the supraorbital spines rounded carinae extend outwards in a curve to the cervical groove. These

604

PALAEONTOLOGY, VOLUME 30

carinae are ornamented with two or three blunt tubercles. External to these carinae the carapace slopes
vertically, then becomes nearly horizontal as far as the lateral margin where it slopes steeply again and begins
to curve inwards. The lateral margin is slightly concave for one third the length from the anterolateral spine
to where it erupts into another prominent forward pointing spine, beyond which it curves inwards to the
cervical groove. The carapace in front of the cervical groove is ornamented with numerous pits on the spines,
ridges, tubercles, and carinae, whilst the flat or concave areas are nearly smooth with fine tubercles. The
cervical groove is deep. Behind it are three parallel longitudinal carinae, of which the median one is prominent;
the carapace between the median and the lateral carinae slopes steeply away from the centre with a slight
concavity until it reaches the outer margin. Just in front of the posterior margin of the carapace is a broad,
smooth transverse groove, which interrupts the median carina and curves back on either side, becoming
narrower towards the outer margins. The carapace behind the cervical groove is ornamented with fine equal-
sized tubercles except for the carinae which are covered with pits.

Abdomen. The abdomen tapers slightly towards the tail. The carinae on the carapace continue as ridges on
each tergum. The first somite has one anteriorly directed spine on the central ridge at the posterior margin. In
front of this a groove runs almost parallel to the posterior margin for approximately half the width of the
somite where it forks, one branch to the posterior margin at the outer edge, the other forward to the outer
edge near the front of the somite. The two triangular areas enclosed by these branches are deeply pitted, whilst
the area in front is smooth with very fine punctae. Somites two, three and four have two anteriorly directed
spines on the central ridge of the tergum, separated midway by a deep transverse groove; this slopes back on
either side to cut the outer ridge just in front of the posterior margin. On each of the outer ridges are two
forward pointing spines, one in front, the other behind the groove. The tergum surface is deeply pitted, except
the grooves which are smooth. The fifth somite lacks spines, and the posterior groove has almost disappeared.
The sixth has a deep median groove with a prominent longitudinal line of pits on either side. The tergum of
each somite is triangular in transverse section with a slight flattening towards the outer ridge.

Ventral surfaces. The epistome has a deep median furrow; anteriorly running parallel to this furrow is a large
forward pointing blunt tubercle, the top surface of which has a line of several deep pits. The front edge of
the epistome is concave with steeply sloping sides, ornamented with variably sized fine tubercles with some
shallow pits.

Appendages. The antennular somite is longitudinally rectangular, with a deep, broad, median furrow extending
forward between the basal segments of the antennae. The first podomere of the antennal peduncle has two
forward-pointing spines on the outer edge; the top surface is coarsely pitted, the underside smooth, with fine
well-spaced pits. The second podomere is nearly half the width of the first and has small spines on the outer
edge. The third is smaller than the second with two spines on the inner margin. The flagellum is longitudinally
grooved on the upper and lower surfaces.

On the third maxilliped the merus is two thirds the length of that on the first pereiopod, its outer side flat,
inner edge flattened, margin tuberculate, surfaces deeply pitted. The first pereiopod is larger than the remain-
der. The merus is equal in length to the total length of the carpus, propodus, and dactylus, roughly oval in
cross-section, flattening on the underside towards the front, the surface becoming concave; the inner margin
is tuberculate, the surfaces decorated with small tubercles. The carpus is rectangular, flattened at the merus
and increasing in depth towards the front (i.e. triangular in side view) and decorated with various sized
tubercles. The propodus is slightly longer than the carpus, approximately oval in cross-section but slightly
flattened, the surfaces with tubercles and pits. The dactylus, one third the length of the merus, is pointed
distally; fragments of the inner margin show a line of large shallow pits with very fine pits inside these. Merus,
carpus, and propodus of the second pereiopod are similar to the first but smaller.

EXPLANATION OF PLATE 66

Figs. 1-5. Linuparus eocenicus Woods. 1-3, London Clay, Portsmouth Docks. 1 and 2, dorsal and ventral
view of lectotype, SM C7735, x 1. 3, lateral view, SM C7733, xO-65. 4 and 5, Bracklesham Group,
Whiteclifr Bay. 4, anterior, showing supraorbital spines, BM In. 63371, xO-7. 5, antennular somite, BM
In.63374, xO-6.

Figs. 6-8. Linuparus scyllariformis (Bell), London Clay. 6, dorsal view, OUM L555, Aveley, x 1. 7, ventral
view, BM 59106, Whetstone, xO-8. 8, supraorbital spines, OUM L462, Aveley, x 1.

Specimens in figs. 1, 2, and 6 have been whitened with ammonium chloride.

PLATE 66

QUAYLE, Linupanis

606

PALAEONTOLOGY, VOLUME 30

Discussion. On the other species of this genus the carinae and ridges are usually dentate or tubercu-
late. L. scyllariformis differs in that the area in front of the cervical groove is ornamented with
numerous pits on the spines, ridges, tubercles, and carinae, and behind this groove the carinae are
covered with pits.

Family scyllaridae Latreille, 1825
Genus SCYLLARIDES Gill, 1898

Type species. Scyllarus aequinoctialis Lund, 1793, by original designation. Recent.

Diagnosis. Eyes near anterolateral angles; lateral margins of carapace without deep fissures, rostrum
salient.

Range. Lower Cretaceous- Recent.

Discussion. Woods (1926, p. 41) says of S. koenigi ( = S. tuherculatus) ‘This species agrees so closely
with living forms of Scyllarides and differs from that of Scyllarus that there seems no reason for
retaining Bell’s genus Scyllaridia' (1858, p. 35). Holthuis (1954) proposed use of the Plenary Powers
to render the name ^ Scyllarides' Gill, 1898 the oldest available for this species; Scyllaridia was
suppressed by the International Commission (1954, Opinion 293, pp. 134-136). Glaessner (1969,
R475) stated that ‘if the fossil is not congeneric with the recent genus as claimed by Woods (1926),
it must be given a new name’. However, the present author agrees with Woods that the fossil
belongs to the Recent genus.

Scyllarides tuherculatus (Konig, 1825)

Plate 67, figs. 16

1825 Cancer (Scyllarusl) tuherculatus Konig, p. 3, pi. 4, fig. 54.

1 843 Cancer tuherculatus Konig; Morris, p. 72.

1854 Zanthopsis tuherculatus Konig; Morris, p. 116.

1858 Scyllaridia Koenigii Bell, p. 35, pi. 8, figs. 1-3.

1870 Scyllaridia Bellii Woodward, p. 493, pi. 22, figs. 1 and 2.

1925 Scyllarides koenigi (Bell); Woods, p. 39, pi. 10, figs. 7-10.

1929 Scyllarides koenigi (Bell); Glaessner, p. 376.

1969 Scyllaridesl Aoen/g; (Bell); Glaessner, R475, fig. 28 1 .3.

1974 Scyllarides koenigi (Bell); Cooper, p. 85.

1980 Scyllarides tuherculatus (ikonig), Morris, p. 16.

Types. The holotype, by monotypy, is BM 42228 (PI. 67, figs. 4 and 5), London Clay, Isle of Sheppey. Of
Bell’s figured specimens, syntypes of S. Koenigii, the original of pi. 8, fig. 1, is missing; pi. 8, fig. 2 is BM 59115
(PI. 67, fig. 6) and fig. 3 is BM 46364, both from the London Clay, Isle of Sheppey.

Other material. OUM L566 (PI. 67, figs. 2 and 3); BM In. 63387-In. 63389 (PI. 67, fig. 1), ex JSQ Collection; all
from the London Clay, Isle of Sheppey.

EXPLANATION OF PLATE 67

Figs. 1-6. Scylarides tuherculatus (Konig), London Clay, Isle of Sheppey, Kent. 1, ventral view, BM In. 63389,
X 2-5. 2 and 3, OUM L566. 2, somites four to six with calcified part of telson, x 1 -4. 3, lateral view, x 2-4.
4 and 5, dorsal and lateral view of the holotype, BM 42228, x 1-2 and IT. 6, dorsal view, BM 59115,
X 1-45.

Figs. 7-9. Bathysquilla wetlierelli (Woodward), London Clay, Isle of Sheppey, Kent. 7 and 8, BM In. 63390.
7, fragments of carapace, thoracic somites six to eight, abdominal somites one to four, x 1 . 8, part of the
raptorial claw, x 5. 9, abdominal somites one to five, BM In. 63391, x2-3.

Specimens in figs. 1 -6 and 9 have been whitened with ammonium chloride.

PLATE 67

QUAYLE, Seyllcirides, BathysquiUa

608

PALAEONTOLOGY, VOLUME 30

Horizon and locality. The London Clay, Isle of Sheppey and Whetstone, London; derived material from the
London Clay, Red Crag, Felixstowe, and Walton-on-the-Naze.

Diagnosis. A Scyllarides with carapace and ridges covered with various sized tubercles, a large spiny
tubercle on the median carina and another between this and the cardiac carina.

Description. The following is not a full description but consists of additions or changes to that of Woods
(1925).

Cephalothorax. The margin between the rostrum and the preorbital spine is concave and occupies one quarter
the anterior width. The strong anterodorsally directed preorbital spine is the continuation of a longitudinal
carina which runs parallel to the median line, broken by the branchiocardiac groove, and continuing nearly
to the posterior margin. Between the preorbital spine and the branchiocardiac groove lies a small blunt
tubercle. The orbit is well rounded and terminates on the anterior margin; its lower margin bears an anteriorly
directed spine on the outer side; the orbits occupy approximately one quarter of the anterior margin. The
remainder of the margin to the anterolateral angle is slightly concave. There is a strong forward pointing
lateral spine at the anterolateral angle reaching slightly further forward than the rostrum. The postcervical
groove cuts the median line at right angles at a point approximately half the distance from the rostrum to the
rear of the carapace; it forms an obtuse angle with the cervical groove, which is directed anterolaterally.

Some of the carapace tubercles have groups or lines of fine pores, possibly for setae. Larger pores lie in
front of the tubercles on the elevated gastric region and lateral margins. There are other small round holes or
pits at various places on the carapace. The outer margin runs slightly inwards from the anterolateral angle to
the niche for the cervical groove. There are several forward pointing small blunt spines on this margin. The
remainder of the lateral margin runs almost parallel, with several anteriorly directed blunt spines.

Abdomen. On the tergum of abdominal somites two to five a deep transverse groove cuts the median line at
one third the distance from the front and runs forward until it reaches the anterior margin, just above the
tergum boundary. A carina is formed on the median line behind this groove. Approximately half-way down
the tergum a slight groove starts at right angles to the main groove and continues roughly in a semicircle to
the posterior margin at the tergum boundary. This groove encloses a raised area which continues on to the
pleura. The anterior margin slopes from the median line vertically to the tergum boundary, where a hollow
boss accommodates the cone-shaped tubercle on the posterior margin of the previous somite. The margin
then curves gently towards the rear and forms a scythe-like ventral spine with the posterior margin. The
posterior margin ventral to the cone-shaped tubercle is slightly convex with a serrate edge of five or six
downward pointing blunt spines. At the last of these spines the margin becomes strongly concave to form the
posterior edge of the ventral spine. The posterior margins of the terga of somites five and six bear posteriorly
directed blunt tubercles or spines. On the tergum of the sixth somite there is an additional groove, with a rear
margin with blunt tubercles, instead of a median carina. On the pleuron of the sixth somite the posterior
margin is slightly more concave, almost semicircular, to accommodate the tail members. The surfaces of
somites one to six have irregular sized pits, both deep and shallow.

Telson. The calcified part of the telson has blunt tubercles of various sizes with groups or rows of fine pores
towards the rear; towards the lateral margins the ornament changes to small irregular pits.

Ventral surfaces. The sternum, which is triangular with a median groove, is evident between the first and fifth
pair of pereiopods. There is a pair of small tubercles at the front, flanking the median line, opposite the coxa
of the first pereiopod. This is followed by four pairs of large tubercles, flanking the median line and opposite
the coxae of the succeeding pereiopods. Between the last two pairs of tubercles on the median line is a round
cavity. The posterior margin of the sternum is straight, the surfaces tuberculate on raised portions, smooth at
the bottom of grooves.

Appendages. The coxae of the first to fifth pereiopods are triangular. The basis and ischium have a combined
length similar to that of the coxa. The merus is oval in cross-section with a longitudinal ridge and pitted
surfaces.

Discussion. Konig (1825) described and figured a unique specimen (BM 42228, PI. 67, figs. 4 and 5)
from the London Clay of the Isle of Sheppey, to which he gave the name Cancer (Scyllarusl)
tuherculatus. Bell (1858, p. 36) considered that ‘the whole surface of the carapace is fictitious, and
the very tubercles on which the name was found exist only in obedience to the skill and trickery of
the artist’, an opinion with which Woods (1925) largely agreed. As a result of his misinterpretation.

QUAYLE: EOCENE CRUSTACEA

609

Bell disregarded Konig’s name tuberculatus and substituted the honorific Koenigi. On examination
by the present author, however, it became clear that this specimen consists of a normally preserved
(i.e. with surface detail) abdomen of Scyllarides and an internal cast of the carapace. The general
shape of the two tubercles on either side of the body as depicted by Konig (1825, pi. 4, fig. 54) is
evident on the specimen. Bell’s argument for changing the name of this species rested on the fact
that he considered that the carapace surface and the tubercles were fictitious. Konig’s original name
is retained here, as advocated by Morris (1980).

Order STOMATOPODA Latreille, 1817
Family squillidae Latreille, 1803
Genus BATHYSQUiLLA Manning, 1963

Type species. Lysiosquilla microps Manning, 1961, by original designation. Recent.

Diagnosis. Telson with blunt median carina and all four pairs of marginal teeth with movable
apices.

Range. Lower Eocene- Recent.

Balhyscjuilla wetherelli (Woodv/ard, 1879)

Plate 67, figs. 7-9

1879 Squilla wetherelli Woodward, p. 549, pi. 26, fig. 1.

1969 Sqiiillal wetherelli Woodward; Holthuis and Manning, R541 .

1974 Squilla wetherelli Woodward; Cooper, p. 85.

1980 S’l/n/Z/r/ H’d/icre/// Woodward; Morris, p. 17.

Type. The holotype, by monotypy, is BM 59780 (Woodward 1879, pi. 26, fig. 1), London Clay, Highgate,
Wetherell Collection.

Other material. BM 38399 collected W. Griffith; BM In. 38262 collected D. J. Jenkins; In. 63390 (PI. 67, figs. 7
and 8) and In. 63391 (PI. 67, fig. 9), both ex JSQ Collection.

Horizon and locality. All from the London Clay; BM In. 38262, Beltinge, East Cliff, Herne Bay; BM 38399,
BM In. 63390, and In. 63391, Isle of Sheppey.

Diagnosis. A stomatopod with a marginal carina on the abdominal somites and a fine transverse
groove on the tergum of the second abdominal somite; the propodus of the raptorial claw is armed
with spiniform teeth.

Description. The posterolateral angle of the carapace is well rounded, and the posterior margin straight. The
surface of the thoracic tergites curves steeply ventrally on either side of the median line in a regular curve to
the lateral margin of the tergum where there is a slight ridge. The pleura of somites seven and eight have a
deep U-shaped indentation in the lateral margin. The width of the somites increases slightly towards the
telson, eight being the widest at approximately half the width of the first abdominal somite. Abdominal
somites one to four curve steeply ventrally on either side of the median line in a regular curve to the lateral
margin where there is a slight ridge which turns slightly inwards to form a semicircular depression on the
pleuron. This depression varies in width and depth; on the second somite it is nearly the full width and half
the depth of the pleuron but it decreases in size on successively posterior pleura. The remaining surface of the
pleuron is almost flat. The posterolateral angle of each pleuron is produced backwards into a small acute
tooth and the posterior margin is straight. Each somite (one to four) has an oblong punctum on either side
near the lateral margin of the terga; somites two to four bear two small sub-central puncta on the anterior
border; three and four have a single central punctum on the posterior border. A fine groove, approximately
two thirds the width of the tergum in length, runs parallel to the posterior margin and cuts the median line of
the second somite at mid-point. Part of the tergum of the fifth somite is preserved with a single central
punctum near the posterior margin. The sixth somite has a scabrous surface ornamentation.

The remains of the raptorial claw are evident on BM In. 63390 (PI. 67, fig. 8). Part of the inner margin of
the propodus is preserved showing four equally spaced spiniform teeth. The dactylus is almost complete; the
inner margin bears at least seven strong triangular teeth increasing in size to the terminal one.

610

PALAEONTOLOGY, VOLUME 30

Discussion. The Recent members of this group are abundant but are mainly restricted to tropical
and subtropical seas. The type species Lysioscjuilla microps was described by Manning (1961,
p. 693) from two Recent specimens taken from 732 and 916-952 m off the east and west coasts of
Florida.

Fossil stomatopods are rare and usually imperfectly preserved. The oldest known, Sculda Miins-
ter, 1840 (Holthuis and Manning 1969, R541), occurs in the Jurassic lithographic limestones of
Solenhofen, Bavaria. New material from the Eocene of the Isle of Sheppey shows that the species
described here cannot be included in the genus Squilla. Some of the diagnostic characters are not
preserved but the longitudinal carinae on the abdomen and the upper margin of the propodus, and
the even pectination of closely placed short blunt spinules are absent. This species shows similarities
to both Harpiosquilla Holthuis, 1964 and Bathysquilla. The first closely resembles Squilla and has
spaced spiniform teeth on the upper margin of the propodus of the raptorial claw but in all species
the abdomen is strongly carinate. Bathysquilla lacks carinae on the abdomen and the propodus of
the raptorial claw is armed with spiniform teeth. The affinities of this species therefore lie with
Bathysquilla.

ACKNOWLEDGEMENTS

1 thank the following for allowing access to specimens in their care; the Keeper of Palaeontology, British
Museum (Natural History); Dr D. Price, Sedgwick Museum; Dr W. J. Kennedy, Oxford University Museum;
Dr B. J. Pyrah, Yorkshire Museum; Dr P. R. Crowther, Leicester Museum; Dr T. H. Riley, Sheffield Museum;
Dr I. G. Robertson, Passmore Edwards Museum. The following assisted in locating specimens and references;
Mr S. E. Morris and Mr J. Cooper BM(NH); Mr P. Powell (OUM); Miss E. Talbot (PE). My special thanks
to C. King for passing on specimens and drawing my attention to horizons in Whitecliff Bay; to Dr R. Forster
for helpful advice; to Dr J. F. Wickins for small Recent casts of Hornarus; to Dr G. A. Bishop, Dr R. M.
Feldman, Dr R. B. Manning, and Dr E. Benamy for information on American fossil and Recent specimens
and to my wife and daughter who ably helped in collecting. The following helped in collecting or loaned
material; D. Bone. W. George, M. Harris, J. Hooker, D. Kemp, A. Lawson, E. Mintrum, D. Rodgers, M.
Simpson, S. Tracy, D. and A. Ward, and G. Ward. Lastly my thanks to J. S. H. Collins for reading and
suggesting helpful changes to the paper.

REFERENCES

BALL, H. w. 1960. Upper Cretaceous Decapoda and Serpulidae from James Ross Island, Graham Land. Scient.
Rep. Falkld Isl. Depend. Surv. 24, 1-30, pis. 1 7.

BELL, T. 1858. A monograph of the fossil malacostracous Crustacea of Great Britain. Part 1, Crustacea of the
London Clay. Palaeontogr. Soc. [Monogr.], 1 -44, pis. 1-11.

BERRY, p. F. and GEORGE, R. w. 1972. A new species of the genus Linuparus (Crustacea, Palinuridae) from
south-east Africa. Zool. Meded. Leiden, 46, 17-23, 4 pis.

BURTON, E. ST. J. 1929. The horizons of Bryozoa (Polyzoa) in the Upper Eocene Beds of Hampshire. Q. Jl
geol. Soc. Land. 85, 223-239.

COOPER, J. 1974. The stratigraphical distribution of the English Paleogene Decapod Crustacea. Tertiary Times,
2, 83-85.

DIXON, F. 1850. The geology and fossils of the Tertiary and Cretaceous formations of Sussex, xvi+422 pp., 50
pis. London.

FORSTER, R. and MUNDLOS, R. 1982. Decapod Crustacea from the Paleogene of Helmstedt and Handorf
(Northern Germany). Palaeontographica, A179, 148-184, pis. 33-35.

GARDNER, J. s., KEEPING, H. and MONKTON, H. w. 1888. The Upper Eocene comprising the Barton and Upper
Bagshot Formations. Q. Jl geol. Soc. Land. 44, 578-635.

GEORGE, w. and VINCENT, s. 1977. A foreshore exposure of London Clay at Steeple, Essex. Tertiary Research,
1,105-107.

1982. An exposure of London Clay at Maylandsea (Lawling Creek), Essex. Ibid. 4, 39-41.

GLAESSNER, M. F. 1929. Crustacea Decapoda. In pompeckj, f. j. (ed.). Fossilium catalogus. 1; Animalia, Pars
41, 1 464. Berlin.

1945. Cretaceous Crustacea from Mount Lebanon, Syria. Ann. Mag. nat. Hist. 12, 694-707, pi. 8.

QUAYLE: EOCENE CRUSTACEA

611

1969. Decapoda. In moore, r. c. (ed.). Treatise on Invertebrate Paleontology, Part R, Arthropoda 4 (2),

R399-R533, R626-R628. Geological Society of America and University of Kansas Press, New York and
Lawrence, Kansas.

HOLTHUis, L. B. 1954. Proposcd use of the Plenary Powers to render the generic name ‘Scyllarides’ Gill, 1898
(Class Crustacea, Order Decapoda) the oldest available name for the species currently referred thereto.
In HEMMiNGS, E. (ed.). Opinion 293, Validation, under the Plenary Powers, of the generic name Scyllarides
Gill, 1898 (Class Crustacea, Order Decapoda). Opin. Decl. int. Commn zool. Norn. 8 (10), 131-142.
London.

1974. The lobsters of the superfamily Nephropidea of the Atlantic Ocean (Crustacea and Decapoda).

Bull. mar. Sci. 24, 723-884, fig. 24.

and MANNING, R. B. 1969. Stomatopoda. In moore, r. C. (ed.). Treatise on Invertebrate Paleontology, Part

R, Arthropoda 4 (2), R535-R552. Geological Society of America and University of Kansas Press, New York
and Lawrence, Kansas.

KEMP, D. j., KING, A. D., KING, c. and QUAYLE, w. J. 1979. Stratigraphy and biota of the Elmore Formation
(Huntingbridge Division, Bracklesham Group) at Lee-on-the-Solent, Gosport, Hampshire. Tertiary Re-
search, 2, 93-103, figs. 1-5.

KING, c. 1981. The stratigraphy of the London Clay and Associated Deposits. Tertiary Research, Spec. Pap.
6,1-158.

1984. The stratigraphy of the London Clay Formation and Virginia Water Formation in the coastal

sections of the Isle of Sheppey (Kent, England). Tertiary Research, 5, 121-160, figs. 1-12, 2 pis.

KONiG, c. H. 1825. leones fossilium sectiles. Part 1, 4 pp., 8 pis. London.

MANNING, R. B. 1961. A ncw deep water species of Lysiosquilla (Crustacea, Stomatopod) from the Gulf of
Mexico. Ann. Mag. nat. Hist. 3(13), 693-697, pis. 1 and 2, text-fig.

m‘coy, f. 1849. On the Classification of some British Fossil Crustacea, with notices of New Forms in the
University Collection at Cambridge. Ibid. 4 (2), 161-179, 6 figs.

MEYER, c. 1871. On Lower Tertiary Deposits recently exposed at Portsmouth. Q. Jl geol. Soc. Land. 27, 402-
417.

MORRIS, J. 1843. A Catalogue of British Fossils, 222 pp. London.

1854. A Catalogue of British Fossils. 2nd edn., 372 pp. London.

MORRIS, S. F. 1980. Catalogue of the type and figured specimens of fossil Crustacea (excl. Ostracoda), Chelicerata,
Myriapoda and Pycnogonida in the British Museum {Natural History), 53 pp., 3 pis. BM(NH), London.

NOETLiNG, F. 1885. Die Fauna des samlandischen Tertiars. Abh. geol. Speckarte preuss. thiir. St. Band VI, Heft
3. I Teil. Lieferung II: Crustacea und Vermes.

PARKER, T. J. 1883. Notes from the Otago University Museum. IV. On the Structure of the Head in "Palinurus\
with special reference to the Classification of the Genus. Nature, Fond. 29, 189-190, 1 fig., 1 table.

1884. On the Structure of the Head in Palinurus, with especial Reference to the Classification of the

Genus. Trans. Proc. N.Z. Inst. 16 (for 1883), 297-307, 1 table, 1 plate.

PICTET, F. J. 1854. Traite de Paleontologie ou Histoire Naturelle des Animaux Fossiles considh'es dans lews
rapports Zoologiques et Geologiques. 2nd edn., 2, 727 + 77, 110 pis.

QUAYLE, w. J. and collins, j. s. h. 1981. New Eocene crabs from the Hampshire Basin. Palaeontology, 24,
733-758, pis. 104 and 105.

QUENSTEDT, F. A. 1856 1 858. Der Jura. VI, 842 pp., 103 pis. Haupp Siebeck, Tubingen.

RATHBUN, M. J. 1904. Some changes in crustacean nomenclature. Proc. biol. Soc. Wash. 17, 169 172.

1926a. The fossil stalk-eyed Crustacea of the Pacific slope of North America. Bull. U.S. natn. Mus. 138,

vii+ 155, 39 pis.

19266. Arthropoda. In wade, b. The fauna of the Ripley Formation on Coon Creek, Tennessee. Prof.

Pap. U.S. geol. Surv. 137, 184-191, pis. 63-70.

1935. Fossil Crustacea of the Atlantic and Gulf Coastal Plain. Spec. pap. geol. Soc. Am. New York, 2,

vii + 160, 26 pis.

1945. Decapod Crustacea. In ladd, h. s. and hoffmeister, j. e. Geology of Lau, Fiji. Bull Bernice P.

Bishop Mus. 181, 373-383, pis. 54-62.

RAVN, J. p. J. 1903. The Tertiary Fauna of Kap Dalton in Gronland. Meddr Gronland, 29, 93-140, pi. 4.

REED, f. r. c. 191 1. New Crustacea from the Lower Greensand of the Isle of Wight. Geol. Mag. 8, 1 1 5-120, pi. 7.

ROBERTS, H. B. 1962. The Upper Cretaceous Decapod Crustaceans of New Jersey and Delaware. In
RICHARDS, H. G., RAMSDELL, R. G., MILLER, A. K., GARNER, H. F., REESSIDE, J. B., JELETSKY, J. A., ROBERTS, H. B.
and MILLER, H. w. The Cretaceous Fossils of New Jersey. Bull. geol. Surv. New Jers. 61, 163-191, pis. 80-89,
4 text-figs.

612

PALAEONTOLOGY, VOLUME 30

STENZEL, H. B. 1945. Decapod Crustacea from the Cretaceous of Texas. Univ. Tex. Pubis. 4401, 400-454,
pis. 34-45, 16 text-figs.

VAN STRAELEN, V. 1921. Hoploparia corneti, crustace decapode nouveau de I’Ypresien superieur de Cuesmes.

Bull. Soc. Beige Geol. 30, 136-138, pl. 1, text-fig.

VENABLES, E. M. 1971 . Bognor Lobster sixty million year old. West Sussex Gazette, 6062, 1 -20.

WARD, D. j. 1979. The Lower London Tertiary (Palaeocene) succession of Herne Bay, Kent. Rep. Inst. geol.
Sc/. 78/10, 1-12, figs. 1 and 2.

WELLER, s. 1903. The Stokes collection of Antarctic Fossils./. Geol. 11,413-419, pis. 1-2.

WHiTEAVES, J. F. 1885. Notes on a decapod crustacean from the Upper Cretaceous of Highwood River, Alberta,
Northwest Territory. Trans. R. Soc. Can. 2, 237- 238.

1895. On some fossils from the Vancouver Cretaceous. Ibid. 1, 132-133.

WOODS, H. 1925 1931. A monograph of the Fossil Macrurous Crustacea of England. Parts 1 7. Palaeontogr.
Soc. [Monogr.], 122 pp., 27 pis.

WOODWARD, H. 1870. Contributions to British Fossil Crustacea. Geol. Mag. 7, 493-497, pl. 22.

1871 . Notes on some new Crustaceans from the Lower Eocene of Portsmouth. Q. Jl geol. Soc. Land. 27,

90-92, pl. 4.

1873. Further Notes on Eocene Crustacea from Portsmouth. Ibid. 29, 25-31, pis. 1 and 2.

1879. Contributions to the knowledge of Fossil Crustacea. Ibid. 35, 549-551, pl. 26.

Typescript received 15 November 1985
Revised typescript received 27 June 1986

W. J. QUAYLE

5 1 White’s Road
Bitterne

Southampton S02 7NR

ORDOVICIAN ACRITARCHS FROM THE MEITAN
FORMATION OF GUIZHOU PROVINCE,
SOUTH-WEST CHINA

by LI JUN

Abstract. An early Arenig acritarch assemblage is described from the Meitan Formation of Tongzi, Guizhou
Province, south-west China. Twenty-four species belonging to twenty genera are included. One new genus,
Tongzia, and four new species, Rhopaliophora memhrcma, Pireci sinensis, Schizodiacrodiwnl nndtiramifenon, and
Tongzia meitana are proposed. The assemblage belongs to the Mediterranean Province and shows that a
homogeneous Arenig assemblage extended from east Newfoundland through the Mediterranean to south-west
China.

Compared with other areas, notably Europe, Palaeozoic acritarchs have been neglected in China
apart from some pioneering work. Cambrian sphaeromorphs from south-west China were described
by Ouyang Shu et al. (1974). Xing Yusheng (1980, 1982) reported on seven genera from the
Dachengsi Formation (Arenig) of Sichuan and on Lower Cambrian acritarchs from near Kunming
of Yunnan, both in south-west China. Zhong Guofang (1981) described an assemblage from the
Dawan Formation (late Arenig) from Hubei Province recording six genera of non-sphaeromorph
acritarchs. Seven genera were recorded (Li Zaiping 1982) from the Machiakou Formation (late
Arenig/Llanvirn) of Hebei Province, northern China.

The present paper describes acritarchs from the Meitan Formation of Guizhou Province (text-
fig. 1). In Arenig times this locality lay towards the centre of the Upper Yangtse Paraplatform about
350 km south-east of Xing Yusheng’s locality on the Mount Emei Shan and about 550 km south-
west of Zhong Guofang’s locality in Yichang also in the same paraplatform. Li Zaiping’s material
came from north China about 1500 km to the north-east.

MATERIAL

Two samples of yellow-grey shale from the Meitan Formation from the Honghuayuan section at Tongzi were
provided by Dr Geng Liangyu of the Nanjing Institute of Geology and Palaeontology. The section is at
Honghuayuan (106° 51 ' E., 28° 4' N.), about 7 km south of Tongzi County, on the west slope of a hill: the beds
dip at 7° to the north-east (bearing about 50°). The Ordovician consists of, in descending order, the Wufeng,
Jiantsaokou, Pagoda, Shihtzupu, Meitan, Hunghuayuan, and Tungtzu formations. According to the descrip-
tion of the section by Zhang and Chen (1964), the Meitan Formation is 1 80 m thick and comprises a 92 m thick
lower part, consisting of yellow-grey shales, and an 88 m thick upper part, consisting of yellow-grey sandy shales,
both with thin-bedded limestone intercalations. Four graptolite biozones were set up within the lower part of the
Meitan Formation, i.e. from base to top, Didyinograptus filifonnis Biozone, D. protobifidus Biozone, D. deflexiis
Biozone, and Azygograptus suecicus Biozone. These represent 15-5 m, 48 m, 23-5 m and 5 0 m respectively in
terms of thickness. Sample MDZ, from the D. deflexiis Biozone, is located 70 m above the base of the formation;
sample MAZ from the A. suecicus Biozone is located 85 m above the base.

The material was prepared in the Department of Geology, University of Sheffield using standard
palynological techniques. Organic residues were then sieved through a 20 mesh screen and mounted on
slides. Acritarchs are well preserved, abundant, and yellow-brown in colour. Some dark-brown to black
Chitinozoa were also seen. Specimens on strew slides are located by sample, slide, and co-ordinate code, for
example, MAZ 44, T22/4 is sample MAZ, slide number 44, England Finder reference co-ordinate T22/4. The

(Palaeontology, Vol. 30, Part 3, 1987, pp. 613-634, pis. 68-72. |

© The Palaeontological Associatioh

614

PALAEONTOLOGY, VOLUME 30

types and illustrated material are housed in the collections of Nanjing Institute of Geology and Palaeontology,
Academia Sinica, Nanjing, China.

SYSTEMATIC PALAEONTOLOGY
For brevity, only name changes are listed in synonomies

Group ACRITARCHA Evitt, 1963

Genus acanthodiacrodium Timofeev, 1958 emend. Deflandre and Deflandre-Rigaud, 1962
Type species. Acanthodiacrodium dentiferwn Timofeev, 1958.

Acanthodiacrodium? tasse/ii Martin, 1969
Plate 71, figs. 5, 7,8

Description. The vesicle is hollow and elongate oval in outline; the wall is single-layered and bears eight to fifteen
processes at each pole. These may either be solid or hollow with their interior communicating with the inner
cavity. The surface of the vesicle is ornamented by longitudinal ribs. No opening was observed.

Dimensions (five specimens). Length of vesicle 20-42 /j,m, breadth of vesicle 1 7-26 ;uin, ratio of vesicle length to
breadth 1:0-8, rib spacing 1-3 juin, length of processes 3-8 jum.

Remarks. Actinotodissus Loeblich and Tappan (1978) differs from Acanthodiacrodium by having
processes that communicate freely with the vesicle interior. When Martin (1969) established the
species A. tasselii she did not mention whether or not the processes communicate with the vesicle

LI JUN: ORDOVICIAN ACRITARCHS

615

interior. The present specimens, which resemble those described by Martin (1969) and Vavrdova
(1972), are thus assigned to the genus Acanihodiacr odium with reservation.

Genus adorfia Burmann, 1970
Type species. Adorfia firma Burmann, 1970.

Adorfia cf. firma Burmann, 1970
Plate 70, fig. 7

Description. The vesicle is polygonal in outline; the wall is single-layered and carries irregularly distributed
processes. These are short, plump, and communicate freely with the inner cavity. Distally they are manate with
capitate pinnae mainly at the distal portion of the processes. No opening was observed.

Dimensions (two specimens). Diameter of vesicle 20-50 p,m, process length 5-12;um, process breadth 1-5-30 ;u,m,
process number peripherally about fourteen, length of pinnae 1-2 ;um.

Remarks. The specimens here differ from type material of Adorfia firma Burmann 1970 from later
Arenig of GDR by having more processes. They are distinguished from Vogtiandia multiradialis
Burmann, 1970 (Arenig, GDR) and Evittia flosmaris Deunff, 1977 (Llanvirn, Morocco) by the
presence of capitate pinnae.

Genus arbusculidium Deunff, 1968
Type species. Arhusculidium destombesii Deunff, 1968.

Arbusculidium filament osum (Vavrdova) Vavrdova, 1972
Plate 68, figs. 1, 3, 5

1965 Dasydiacrodium fUamentosum Vavrdova, pp. 355-356, pi. 3, fig. 3; pi. 4, fig. 1; text-fig. Aa-c.

1972 Arbusculidium filament osum (Vavrdova); Vavrdova, p. 81, pi. 1, fig. 3.

1980 Dasydiacrodium fUamentosum Vavrdova; Xing, p. 438.

Description. The vesicle is hollow, ellipsoidal to subcylindrical in outline. The processes are restricted to the polar
areas; one pole carries three to five simple processes which are broad-based, hollow, communicating freely with
the vesicle cavity. They taper gradually towards closed, pointed tips. The opposite pole bears a set of branching,
anastomosing filose processes which are arranged in a circle around the pole, forming a collar-shaped net. The
vesicle wall is about 0-5 /^tin and decorated by longitudinal ribs. No opening was observed.

Dimensions (eleven specimens). Length of vesicle 35-40 breadth 25-32 /xin, rib spacing 2-5 /xm, process
length 15-20 /xm, length of the collar 10-20 urn.

Previous records. Arenig, Czechoslovakia (Vavrdova 1965); Arenig, GDR (Burmann 1968); ?Upper Arenig or
lower Llanvirn, Morocco (Cramer, Allam et al. 1974); Wenlock (reworked?), Belgium (Martin 1969); Arenig,
France (Rauscher 1974); Arenig-Llanvirn, East Newfoundland (Martin, in Dean and Martin 1978); Arenig,
China (Xing 1980).

Genus baltisphaeridium (Eisenack 1958) emend. Eisenack, 1969

Type species. Baltisphaeridium (as Ovum hispidum) longispinosum Eisenack, 1931. Elolotype lost. Neotype: B. I.
(as filifera) longispinosum Eisenack, 1959.

Baltisphaeridium longispinosum longispinosum (Eisenack 1931) Staplin et al., 1965

Plate 72, fig. 2

1931 Ovum hispidum longispinosum Eisenack, p. 110, pi. 5, figs. 6-12, 14-17.

1938 Hystrichosphaeridium longispinosum (Eisenack) Eisenack, p. 12, pi. 1, figs. 4, 6, 7.

1959 Baltisphaeridium longispinosum filifera Eisenack, p. 195, pi. 15, fig. 1.

616

PALAEONTOLOGY, VOLUME 30

1965 Baltisphaeridium longispiiwsiiin Eisenack; Eisenack, p. 134, pi. 13, figs. 1 and 2.

1965 Baltisphaeridium longispinosum longispinosum (Eisenack) Staplin et al., p. 190, pi. 20, figs. 1 1 and
15; text-fig. 1 1.

Description. The vesicle is hollow, circular to subcircular in outline. The wall is single-layered, bearing regularly
arranged processes. These are hollow, shut off from the inner cavity, and closed distally. The length of processes
is variable, usually equal to or exceeding the vesicle diameter. No opening was observed.

Dimensions (five specimens). Vesicle diameter 48-56 length of process 35-60 fxm, breadth of process
1 -0-2-5 /xm.

Remarks. This taxon, although characteristic of the Baltic Province, appears to have a world-wide
distribution.

Previous records. Upper Arenig to upper Llandovery, Baltic (Eisenack 1931); Upper Llanvirn to Llandeilo,
Sweden (Eisenack 1965; Staplin et al. 1965); Upper Arenig and Caradoc, Poland (Gorka 1969); Caradoc to
Ashgill, Belgium (Martin 1974); Llanvirn, France (Rauscher 1974).

Genus coryphidium Vavrdova, 1972
Type species. Coryphidium hohemicum Vavrdova, 1972.

Coryphidium hohemicum Vavrdova, 1972

Plate 72, figs. 5 and 9

1972 Coryphidium hohemicum Vavrdova, pp. 84-85, pi. 1, figs. 1 and 2; text-fig. 4.

Description. The vesicle is hollow, subquadrate in outline, with round corners. The wall is single-layered, about
0-5 |um thick and is ornamented by striate ribs, densely spaced (approximately 1 -2 fxxn) and parallel to the edges.
Numerous processes decorate the surface; they are proximally opened, truncated or bifurcate distally. The
process distribution is more or less regular over the surface, with some concentration at the corners. On some
specimens, a round to oval opening occurs.

Dimensions (more than thirty specimens). Length of vesicle edge 24-34 /xin, rib spacing 0-5-1 0 /xin, process
length 3-9 ^uin, process number at each corner eight to fifteen.

Previous records. Arenig, Llanvirn, Czechoslovakia (Vavrdova 1972); Arenig, France (Rauscher 1974; Cocchio
1982); Upper Arenig to lower Llanvirn, Morocco (Cramer, Allam et al. 1974); Arenig to Llanvirn, Belgium
(Martin and Rickards 1979); Arenig, east Newfoundland (Martin, in Dean and Martin 1978); Arenig, China
(Xing 1980); Arenig to lower Llanvirn, Britain (Downie 1984); Arenig, Sardinia (Albani, Di Milia et al. 1985).

Genus cristallinium Vanguestaine, 1978
Type species. Cristallinium cambriense (Slavikova) Vanguestaine, 1978.

EXPLANATION OF PLATE 68

All figures x 1000.

Figs. 1, 3, 5. Arbusculidium filamentosum (Vavrdova) Vavrdova, 1972. 1, MAZ 46, M40/1; 3, MDZ 1, N27; 5,
MDZl,F22/2.

Fig. 2. Veryhachium trispinosum (Eisenack 1935) Deunff, 1954 ex. Downie, 1959. MAZ 46, M24/3.

Fig. 4. Leiosphaeridia tenuissima Eisenack, 1958. MDZ 1, K43/1.

Fig. 6. Cristallinium dentatum (Vavrdova) Martin, 1982. MDZ 35, P33/4.

PLATE 68

LI JUN, Ordovician acritanhs

618

PALAEONTOLOGY, VOLUME 30

Cristallinium dentatum (Vavrdova) Martin, 1982
Plate 68, fig. 6

1976 Suiplinia dentaUi Vavrdova, p. 58, pi. 2, figs. 5 and 6; text-fig. 5.

1978 Dictyotidiiiml dentatum (Vavrdova) Martin, in Dean and Martin, p. 8, pi. 3, figs. 14 and 28.

1982 Cristallinium dentatum (Vavrdova) Martin, p. 36.

Description (based on a single specimen). The vesicle is subcircular to polygonal in outline. The wall is thin and
divided into numerous polygonal fields. The edges of the polygonal fields bear several short, solid and capitate
processes.

Dimensions. Diameter of vesicle 40-45 /am, height of processes about 1 /im, process number at each edge six to
nine.

Previous records. Upper Arenig, Czechoslovakia (Vavrdova 1976); Arenig, east Newfoundland (Martin 1982).

Genus cymatiogalea (DeunlT 1961) Deunff, Gorka and Rauscher, 1974
Type species. Cymatiogalea margaritata DeunlT, 1961.

Cymatiogalea cf. cristata (Downie 1958) Deunff, Gorka and Rauscher, 1974

Plate 72, figs. 3 and 4

Description. The vesicle is hollow, circular or hemicircular in outline; wall is usually divided into polygonal areas
by low sutural ridges. Four to six processes are arranged along each side, they are decorated with lateral spines
which branch into two to four pinnae distally; in some cases second-order branching can be seen. Excystment is
by a large subpolygonal to round opening.

Dimensions (eighteen specimens). Diameter of vesicle 24-45 /am, size of opening 14-38 /xm, length of processes
5-12 /xm, length of branching 1-3 pm.

Remarks. The specimens here differ from the Tremadoc type material of Cymatiogalea cristata from
England by the presence of lateral processes and second-order branching and the lack of flanges on
the vesicle wall.

Cymatiogalea ciivillieri^ (Deunff 1961) Deunff, 1964
Plate 72, fig. 8

Description. The single specimen recorded has a vesicle with subcircular outline, the surface of the vesicle is
divided into polygonal fields, the boundaries of which are formed by short spines, 2-3 pm in height, and about
three to five spines can be seen on each boundary. The wall is simple and c. 0-5 pm thick, decorated by small
verrucae. At one pole of the vesicle an opening occurs, the diameter of which is c. 18 pin, approximately three-
quarters the size of the vesicle.

Remarks. The single poorly preserved specimen is questionably assigned to Cymatiogalea cuvillieri, a
species usually recorded from the Tremadoc.

Genus leiosphaeridia (Eisenack 1958) Downie and Sarjeant, 1963
Type species. Leiosphaeridia haltica Eisenack, I958n.

Leiosphaeridia tenuissima Eisenack, 19586
Plate 68, fig. 4

19586 Leiosphaeridia tenuissima Eisenack, pp. 391-392, pi. 2, figs. 1 and 2.

Description. The vesicle is hollow and circular in outline, the wall is single-layered, very thin (c. 0-2 pm thick),
without process and ornamentation, although some compressional folds occur. No opening was observed.

LI JUN: ORDOVICIAN ACRITARCHS

619

Remarks. Sphaeromorphs are not common and are represented in the present study only by this
species and Synsphaeridiiau cf. gotlandicum Eisenack, 1965.

Previous records. Tremadoc, USSR (Eisenack 1958u); undivided Palaeozoic, North Africa (Combaz 1966);
Caradoc, England (Turner 1984).

Genus multiplicisphaeridium (Staplin 1961) Lister, 1970
Type species. Multiplicisphaeridium ramispinosum Staplin, 1961.

Multiplicisphaeridium cf. irregulare Staplin, Jansonius and Pocock, 1965

Plate 70, figs. 9 and 10

Description. The vesicle is subcircular in outline; the wall is thin, about 0-5 ftm thick, and carrying numerous
processes. These are conical, hollow, and communicate freely with the inner cavity. Distally they are simple or
branching. Branching occurs in an irregular manner at any distance from the process base. No opening was
observed.

Dimensions (seven specimens). Diameter of vesicle 12 nm. process length 11-13 /^m, number of processes
peripherally fourteen to twenty-two.

Remarks. The specimens here are considerably smaller than the typical ones from Anticosti Islands
(vesicle diameter 25-35 jiun, Staplin et al. 1965). Jacobson and Aicha (1985) argued that the depth of
3005 ft from which Multiplicisphaeridium irregulare was recorded, was not the Middle Ordovician
Trenton Formation as originally reported, but the Upper Ordovician (lower-middle Ashgillian)
Vaureal Formation.

Genus petaloferidium Jacobson, 1978
Type species. Petaloferidium stigii Jacobson, 1978.

Petaloferidium florigerum (Vavrdova 1977) Jacobson, 1978
Plate 72, fig. 7

1977 Evittia florigera Vavrdova, p. 116, pi. 4, figs. 110; text-fig. 6a, h.

1978 Petaloferidium florigerum (Vavrdova) Jacobson, p. 296.

Description (based on single specimen). The vesicle is hollow and subcircular to polygonal in outline. The wall is
single-layered, laevigate, and carries processes which are open to the inner cavity but are closed distally, where
there is a pronounced swelling and a darker spot.

Dimensions. Vesicle diameter 20 f^m, process length 3-5 ;um, process number fourteen.

Remarks. This species may need to be subdivided since specimens from early Llanvirn of
Czechoslovakia (Vavrdova 1977, 1982) included both laevigate and striate forms, multi-process and
two-process forms. Single specimen here fits Vavrdova’s multi-horn but has no striations.

Previous records. Early Llanvirn, Czechoslovakia (Vavrdova 1977, 1982).

Genus peteinosphaeridium Staplin, Jansonius and Pocock, 1965
Type species. Peteinosphaeridium trifurcatum trifurcatum ex P. hergstromii SiapUn, Jansonius and Pocock, 1965.

Peteinosphaeridium trifurcatum intermedium Eisenack, 1976
Plate 72, fig. 9

1976 Peteinosphaeridium trifurcatum intermedium Eisenack, p. 195, pi. 4, figs. 8-11.

1976 Peteinosphaeridium trifurcatum Eisenack; Kjellstrom, p. 36, fig. 29.

1984 Peteinosphaeridium trifurcatum intermedium Eisenack; Turner, pp. 132-133. pi. 9, figs. 1, 4-6.

620

PALAEONTOLOGY, VOLUME 30

Description. The vesicle is circular in outline, the wall is firm, single-layered, c. 1 /xin thick, carrying numerous
processes; these are solid, distally trifurcate or quadrifurcate and bear delicate, more or less transparent
longitudinal membranes. The vesicle wall is granulate. Excystment is by means of a pylome.

Dimensions (eighteen specimens). Diameter of vesicle 32-46 ,um, process length 6 12 ;u.m, branch length 1 -4 /xin,
number of processes peripherally twenty-five to thirty-eight.

Reniarks. Eisenack (1976) attributed two infraspecific taxa to the species Petemosphaeridium
trifurcatimu i.e. P. trifiircatum intermedium and P. trifurcatiim hypertrophicwn. Only the former was
found in the present study.

Previous Records. Arenig, Baltic (Eisenack 1976); Llanvirn-Llandeilo, Sweden (Kjellstrom 1976); Caradoc,
Kentucky, USA (Jacobson 1978); Caradoc, England (Turner 1984).

Genus pirea Vavrdova, 1972
Type species. Pirea collifonnis (Burmann) Vavrdova, 1977.

Pireu sinensis sp. nov.

Plate 70, figs. 1-4

Derivation of name. Sinae (Eatin), Chinese.

Holotype. Plate 70, fig. 1 (MAZ 44, X49).

Lsotypes. Plate 70, fig. 2 (MAZ 44, Q23); fig. 3 (MAZ 46, D23); fig. 4 (MAZ 44, E23/4).

Locality and horizon. West slope of Honghuayuan Hill in Tongzi County, Guizhou, China; 85 m above the base
of Meitan Formation, Arenig, Ordovician.

Diagnosis. Vesicle pear-shaped with a single process drawn out from apical end, wall decorated by
solid spines, randomly distributed in the antapical part but apically forming longitudinal rows.
Excystment by a pylome at the antapical end.

Dimensions (eleven specimens). Vesicle length 62-78 /xm, vesicle breadth 36-45 (um, process length 11-15 /xm,
process breadth 7-10 ;uin, spine length 0-5-10 ;um, spine spacing c. 1 /xm, spacing of longitudinal rows 10-1-5
/xin. Diameter of pylome 8-12 ;ixm.

Description. The vesicle is hollow, thin walled, and ovate in outline. A single process is drawn out from the apical
end. This is hollow, communicating with the vesicle, and is distally closed and slightly expanded. The wall of the
test in the lower three-quarters is densely covered with uniform, small, solid, often capitate spines distributed
randomly. At the apical end they form longitudinal rows. Excystment is by a pylome at the antapical end, with or
without operculum.

Comparison. The new species resembles Pirea ornatissima Cramer and Diez, 1977, but can be
distinguished by the ornamentation. P. ornatissima is also decorated with solid sculptural elements
on the lower part of the vesicle; but apically the wall is smooth, with about a dozen longitudinal folds.
P. sinensis is ornamented by capitate spines randomly distributed in the lower part of the vesicle, and
the spines are arranged in longitudinal rows that extend onto the process at the apical end.

Genus polygonium Vavrdova, 1966
Type species. Polygonium gracilis V'dvrdova, 1966.

Polygonium gracile Vavrdova, 1966

Plate 72, fig. 1

1966 Polygonium gracilis Vavrdova, p. 413, pi. 1, fig. 3; pi. 3, fig. 1; text-fig. 4a, h.

1972 Polygonium gracile Vavrdova, p. 80.

LI JUN: ORDOVICIAN ACRITARCHS

621

Description. The vesicle is hollow, polygonal or subpolygonal in outline, with sides more or less equal in length
and slightly concave. The wall is single-layered, with long, simple processes drawn out at corners. They number
twelve to fifteen, communicate freely with inner cavity and are closed distally, with wide bases and pointed tips.
No opening was observed.

Dimensions (thirty-two specimens). Diameter of vesicle 24-31 fj.m, length of process 10-18 /xm.

Remarks. The commonest species in the present study, with related species of Tectitheca which may
integrate with Polygoniimu it makes up more than 30% of the total assemblage.

Previous records. Arenig, Czechoslovakia (Vavrdova 1966, 1972); Arenig, France (Rauscher 1974); Later Arenig
to early Llanvirn, Morocco (Cramer, Allam et al. 1974); Upper Cambrian to lower Llanvirn, Britain (Downie
1984).

Genus rhopaliophora Tappan and Loeblich, 1971
Type species. Rhopaliophora Tappan and Loeblich, 1971.

Rhopaliophora memhrami sp. nov.

Plate 71, figs. 1 and 3

Derivation of name. Referring to the membrane-like processes.

Holotype. Plate 71, fig- 1 (MAZ45, N32/1).

Isotype. Plate 71, fig. 3 (MDZ 3, P33/3).

Locality and horizon. West slope of Honghuayuan Flill in Tongzi County, Guizhou, China; 70 and 85 m above
the base of the Meitan Formation, Arenig, Ordovician.

Diagnosis. Vesicle circular to subcircular in outline, with membrane-like processes appearing as
irregular extension; excystment by means of pylome.

Description. The vesicle is c. 1 pm thick, with membrane-like processes of varying shape, width, number, and
extension; the vesicle surface is ornamented by microgranules. Excystment is by a round pylome with an elevated
rim. Double pylome occurs on some specimens, and free opercula have been seen in some cases.

Dimensions (twe\\e specimens). Vesicle diameter 28-46 /xin, process length 3-16 /xin, pylome diameter 13-16 ^m.

Remarks. This species differs from other species of Rhopaliophora by the membrane-like processes.

Rhopaliophora pahnata (Combaz and Peniguel) Playford and Martin, 1984

Plate 71, figs. 4 and 6

1972 Peteinosphaeridium pahnatum Combaz and Peniguel, p. 136, pi. 2, figs. 4, 9-12.

1984 Rhopaliophora pahnata (Combaz and Peniguel) Playford and Martin, pp. 210-212, fig. 9a-n.

Description. The vesicle is hollow and circular to subcircular in outline. The wall is single-layered, carrying
numerous short, stout processes. The processes may be variable in shape and length, separated from inner cavity.
The body surface is psilate. Excystment is by a circular pylome with elevated rim.

Dimensions (eight specimens). Vesicle diameter 27-32 ;um, length of process 2-5 pm, process spacing 5-8 /xm,
pylome diameter 5-9 /xm.

Previous records. Arenig to Llanvirn, Canning Basin, Australia (Combaz and Peniguel 1972; Playford and
Martin 1984).

Genus SCHIZODIACRODIUM Burmann, 1968
Type species. Schizodiacrodium ramiferum Burmann, 1968.

622

PALAEONTOLOGY, VOLUME 30

Schizodiacrodhmfl multiramifenim sp. nov.

Plate 69, figs. 4-6

Derivation of name, multi (Latin), numerous; ramiferum (Latin), branched. Referring to presence of numerous
processes.

Holotype. Plate 69, fig. 4 (MAZ 46, T22/4).

Isotype. Plate 69, fig. 6 (MAZ 47, F43).

Locality and horizon. West slope of Honghuayuan Hill in Tongzi County, Guizhou, China; 85 m above the base
of the Meitan Formation, Arenig, Ordovician.

Diagnosis. Vesicle polyhedral, slightly elongated, with numerous processes arranged on polar areas;
processes bifurcate, trifurcate, or multifurcate, all branches recurved.

Description. The vesicle is slightly elongate, polyhedral in outline and somewhat centrally constricted. The wall is
thin, c. 0-3 ;um thick and carries numerous processes on polar areas. These are arranged in a circle, are broadly
based, communicating freely with the inner cavity, and branched distally. The branching angle varies from 80 to
150'’. Branching is bifurcate, trifurcate, or multifurcate. All branches are slender, characteristically recurved,
with closed tips. Occasionally second-order branches occur. No opening was observed.

Dimensions (nine specimens). Vesicle length 32-38 ;um, breadth 20-25 /xm, process breadth 2-5 /xm, process
number at each pole nine to eleven, branch length 5-7 ^uin, branch breadth 0-2-0-5 /xin.

Remarks. In some cases polar differentiation is weak (see PI. 70, fig. 5) and therefore this taxon is
questionably assigned to genus Schizodiacrodium. The present species differs from others of
Schizodiacrodiiim in having strikingly recurved branches. Multiplicispluieridium marocpiense Cramer,
Allam et a!., 1974 and Vogtlandia fios Martin, 1978 also have similarly recurved branches, the former
may difler from the new species by having fewer processes (four to seven) and the latter by radial
symmetry instead of axial symmetry.

Genus striatotheca Burmann, 1970
Type species. Striatotheca principalis Burmann, 1970.

Striatotheca principalis parva Burmann, 1970
Plate 70, fig. 5

1970 Striatotheca principalis parva Burmann, p. 300, pi. 8, fig. 6.

Description. The vesicle is quadrate to rectangular in outline, with approximately straight sides, each corner
passes into a gradually tapering process. The processes communicate freely with the inner cavity and are closed
distally. The body surface is decorated by subparallel ribs, which extend on to the proximal quarter of the
processes. No opening was observed.

Dimensions (twenty-four specimens). Vesicle length 25-34 ;um, rib spacing 1 p.m, rib breadth 0-5-0-8 jum, process
length 16-20 /xm.

Previous records. Arenig, GDR (Burmann 1970); Arenig, France (Rauscher 1974); Upper Arenig, Italy
(Tongiorgi and Di Milia 1984); Upper Arenig, Hungary (Albani, Lelkes-Felvary et al. 1985).

EXPLANATION OF PLATE 69

All figures x 1000.

Figs. 1 -3. Tongzia meitana gen. et sp. nov. I, holotype, MAZ 45, L48; 2, MAZ 44, T52/4; 3, MAZ 45, X54.
Figs. 4-6. Schizodiacrodiunfl multiramifenim sp. nov. 4, holotype, MAZ 46, T22/4; 5, MAZ 46, M 1 2; 6, MAZ 47,
F43,

PLATE 69

LI JUN, Tongzia, Schizodiacrodiwn?

624

PALAEONTOLOGY, VOLUME 30

Striatotheca quieta (Maxim 1969) Rauscher, 1974
Plate 70, tig. 6

1969 Veryhachiwn quietum Martin, pp. 101 - 102, pi. 5, fig. 225; pi. 6, fig. 293; text-fig. 48.

1970 Striatotheca acutiuscula Burmann, p. 303, pi. 8, figs. 3 and 4.

1974 Riigulidium quietum (Martin) Cramer, Allam el al., p. 60, pi. 25, fig. 12; pi. 26, figs. 15-17.

1974 Striatotheca quieta (Martin) Rauscher; p. 76, pi. 3, fig. 10.

Description. The vesicle is quadrate on outline, with slightly bulging sides. The processes, which are short,
conical, hollow, and with obtuse tips, extend from each corner. The surface of the vesicle is ornamented by
subparallel, fine, closely spaced ribs.

Dimensions {iour specimens). Length of vesicle 26-30 ;um, rib breadth 0-5 ;um, rib spacing 1 /xm, length of process
2-5 (um.

Previous records. Late Arenig or early Llanvirn, Morocco (Cramer, Kanes et al. 1974); Arenig, France
(Rauscher 1974); Upper Arenig, Italy (Tongiorgi and Di Milia 1984); Late Llanvirn, GDR (Burmann 1970);
Caradoc (reworked), England (Turner 1982); Wenlock (reworked?), Belgium (Martin 1969).

Striatotheca c( Iransformata Burmann, 1970
Plate 70, fig. 8

Description. The vesicle is pentagonal in outline. The wall is single-layered, c. 0-3 /^m thick. Five processes are
drawn out at corners; they communicate with the inner cavity and are closed distally, with pointed tips. Surface
of vesicle is decorated by ribs, which cover the entire surface of the body and extend to the base of the processes.
They are subparallel to the edges, forming incomplete pentagons. No opening was observed.

Dimensions (two specimens). Height of the pentagon 20-30 /xm, breadth of rib 0-3 /luti, rib spacing I 0-1 -5 ;u,m,
length of edge 17-23 /xm, length of process 10-15 /x.m.

Remarks. Like Striatotheca transfonnata from GDR, which is also Arenig in age, the specimens here
have five processes, but differ in having pentagonal outlines instead of a quadrate one.

Genus synsphaeridium Eisenaek, 1965
Type species. Synsphaeridium gotlandicum Eisenaek, 1965.

Synsphaeridium cf gotlandicum Eisenaek, 1965
Plate 71, fig. 2

Description. The individual bodies are subcircular to ellipsoidal in outline, and form irregularly shaped clusters.
The wall is single-layered, decorated with hairs. No opening was observed.

Dimensions (two clusters). Individual body diameter 19-30 /<,m, number of individuals on each cluster fifteen to
thirty, cluster diameter 60-100 /xm, length of hairs 0-5- 1 0 /xm.

EXPLANATION OF PLATE 70
Unless otherwise specified, all figures x 1000.

Figs. 1 -4. Pirea sinensis sp. nov. 1, holotype, MAZ 44, X49; 2, MAZ 44, Q23; 3, MAZ 46, D23; 4, MAZ 44,
E23/4.

Fig. 5. Striatotheca principalis parva Burmann, 1970. MAZ 44, B34, x 900.

Fig. 6. S. quieta (Martin) Rauscher, 1974. MAZ 44, G44.

Fig. 7. Adorfia cL firma Burmann. MAZ 46, L26/3.

Fig. 8. S. cf transfonnata Burmann. MAZ 44, R39, x 900.

Figs. 9 and 10. Multiplieisphaeridium cf irregulare Staplin et al. 1965. 9, MAZ 44, Q43; 10, MAZ 45, C42/2.

PLATE 70

LI JUN, Ordovician acritarchs

626

PALAEONTOLOGY. VOLUME 30

Remarks. The individual bodies here correspond to those of Synsphaeridium gotlandicum in shape
and ornamentation but are smaller than the type material. The type material is Silurian, but the genus
has also been reported from the Arenig of Bohemia (Vavrdova 1973, p. 286).

Genus tectitheca Burmann, 1968
Type species. Tectitheca valida Burmann, 1968.

Tectitheca cf. additionalis Burmann, 1968
Plate 72, fig. 6

Description. The vesicle is hollow and subpentagonal in outline, comprising a shortened upper part extending
into a long, rapidly tapering apical process with a conically expanded base. In the middle part are four processes
with six to seven similar processes around the basal line. The processes are closed distally and communicate
freely with the inner cavity. Both vesicle and process surface are thin and faintly granulose. No opening was
present.

Dimensions (six specimens). Diameter of vesicle 23-27 ^m, length of process 20-25 /xm.

Comparison. The specimens here differ from the type material from upper Arenig or lower Llanvirn of
GDR in the more broadly based processes.

Genus tongzia gen. nov.

Derivation of name. From Tongzi County.

Type species. Tongzia meitana sp. nov.

Diagnosis. Vesicle circular to subcircular in outline, bearing radiating processes that are hollow, and
plugged at the base. All processes bifurcate in a similar fashion. No second-order branching was seen.
Surface of vesicle and processes decorated with granules.

Remarks. The method of opening is not known. This new genus is distinguished from Baltisphaeridium
by always having bifurcate processes. Ordovicidium Tappan and Loeblich, 1971 differs from the new
genus by having multifurcate processes. Skictgia Downie, 1982 has processes with funnel-shaped
ends.

Tongzia meitana sp. nov.

Plate 69, figs. I 3

Derivation of name. After the Meitan Formation.

Holotype. Plate 69, fig. 1 (MAZ 45, L48).

Isotypes. Plate 69, fig. 2 (MAZ 44, T52/4); fig. 3 (MAZ 45, X54).

Locality and horizon. West slope of Honghuayuan Hill in Tongzi County, Guizhou, China; 85 m above the base
of the Meitan Formation, Arenig, Ordovician.

EXPLANATION OF PLATE 71
Unless otherwise specified, all figures x 1000.

Figs. I and 3. Rhopaliophora memdrana sp. nov. 1, holotype, MAZ 45, N32/1; 3, MDZ 3, P33/3.

Fig. 2. Synsphaeridium cf. gotlandicum Eisenack, 1965. MDZ 1, Q31/4.

Figs. 4 and 6. R. palmata (Combaz and Peniguel) Playford and Martin, 1984. 4, scanning electron micrograph,
X 1800; 6, MAZ 45, C36.

Figs. 5, 7, 8. Acanthodiacrodium‘1 tasselii Martin, 1969. 5, MAZ 48, F32; 7, MAZ 46, H 1 1/2; 8, MAZ 44, H5 1 .
Fig. 9. Peteinosphaeridium trifurcation intermedium Eisenack, 1976. MAZ 44, Q35.

PLATE 71

LI JUN, Ordovician acritarchs

628

PALAEONTOLOGY, VOLUME 30

Diagnosis. As for genus.

Dimensions (twelve specimens). Diameter of vesicle 44-56 /.im, process spacing 3-8 /uin, granule spacing 0-5-1 0
/xm, process length 4-7 /xin, occasionally 14 /xm, branch length 4-10 ;um, ratio of process to vesicle diameter
c. 0- 1 : 1 .

Description. The vesicle is hollow, circular to subcircular in outline. The wall is firm, single-layered, and carries
numerous hollow, evenly distributed processes. These do not communicate with the cavity and bifurcate distally,
with pointed and closed tips. The branching angles varied from 60 to 160°, usually 110°. Surface of vesicle,
process, and branches are ornamented by densely arranged granules. No second-order branching nor opening
were observed.

Genus VERYHACHIUM (Deunff 1951) Deunff, 1954 ex. Downie, 1959
Type species. Veryhachiion frisiilcum (Deuntl") Deunff, 1954.

Veryhachium trispinosuni (Eisenack 1935) Deunff, 1954 ex. Downie, 1959

Plate 68, fig. 2

1938 Hystrichosphaeridiwn trispinoswn Eisenack, p. 16, fig. 2.

1959 Veryhachium trispinosuni (Eisenack) Downie, pp. 68 -69, pi. 12, fig. 7.

Description. The vesicle is triangular in outline, with three long, hollow processes drawn out at corners. The wall
is single-layered, 0-5 ^xm thick, without ornamentation. The processes are simple and closed distally with pointed
tips. Some specimens show excystment by epityche.

Dimensions (twenty-two specimens). Vesicle diameter 25-32 (um, process length 17-29 fxm.

Remarks. This is a common species from late Arenig to the Carboniferous. It first appears in the early
Arenig where it is rare; it is common in the present study.

DISCUSSION

Composition of assemblages. The overall composition of the two assemblages is shown in Table 1.
Although they are from different graptolite biozones Didymograptus defiexus (sample MDZ) and
Azygograptiis suecicus (sample MAZ), they are remarkably similar. The commonest species is
Polygonium gracile which, with the closely related Tectitlieca cf. additionalis, makes up 36 and 29 % of
the total. Also common are the genera Striatotheca (13 and 18%), Veryhachium (9 and 15%), and
Coryphidium (6 and 8%). Total number of identified species is 24, 23 from MAZ and 21 from MDZ;
20 are common to both.

Age. As mentioned above, the samples came from the D. defiexus and the A. suecicus Biozones
respectively. The D. defiexus Biozone occurs in Britain, and the A. suecicus Biozone, which is just
above the former, was correlated with the Isograptus gihherulus and D. nitidus Biozones of Britain

EXPLANATION OF PLATE 72

Unless otherwise specified, all figures x 1000.

Fig. 1. Polygonium gracile Wavrdova, 1966. MAZ 44, C49/2.

Fig. 2. Baltisphaeridium longispinosum longispinosum (Eisenack) Staplin et al. 1965. MAZ 44, B34, x 600.
Figs. 3 and 4. Cymatiogalea cf. cristata (Downie) Deunff et al., 1974. 3, MDZ 3, H21/4; 4, MAZ 44, H46/2.
Figs. 5 and 9. Coryphidium bohemicum Vavrdova, 1972. 5, MDZ 1, P49/3; 9, MDZ 1, R35/3.

Fig. 6. Tectitlieca cf. additionalis Burmann, 1968. MAZ 45, C52/4.

Fig. 7. Petaloferidium ftorigerum (Vavrdova) Jacobson, 1978. MAZ 44, E38/3.

Fig. 8. Cymatiogalea cuvillieril (Deunff) Deunff, 1964. MAZ 44, C51.

PLATE 72

LI JUN, Ordovician acritarchs

630

PALAEONTOLOGY. VOLUME 30

TABLE 1. Composition of assemblages from two horizons in the
Meitan Formation, Tongzi, Guizhou Province, south-west China.

Composition of assemblages (%)

Species present

MAZ

MDZ

Acanthodiacrodiuml tasselii

2

1

Adorfia cf. firma

5

—

Arhusculidium filamentosum

3

2

Baltisphaeridium longispinosum longispinosum

1

1

Corvphidium hohemicum

6

8

Cristallinium dentatum

—

1

Cymatiogalea cf. cristata

1

1

C. cuvilliern

4

—

Leiosphaeridia tenuissima

2

2

Multiplicisphaeridium cf. irregulare

1

2

Petaloferidium florigeriim

3

—

Peteinosphaeridium trifurcation intermedium

3

1

Pirea sinensis

3

2

Polygonium gracile

32

24

Rhopaliophora membrana

1

3

R. palmata

1

4

SchizodiacrodiunP. multiramiferum

1

7

Striatotheca principalis parva

8

11

S. cpdeta

2

2

S. cf. transformata

3

5

Synsphaeridium cf. gotlandicum

2

1

Tectitheca cf. additionalis

4

5

Tongzia meitana

3

2

Veryhachium trispino.mm

9

15

(Mu Enzhi et al. 1979; Zhang Wentang et al. 1982). Graptolites from the section indicate an early
Arenig age (Harland et al. 1982, p. 14). The acritarchs support the age determination. The
assemblages characterized by diagnostic acritarch genera such as Coryphidium, Striatotheca, and
Pirea are clearly from the Arenig/Llanvirn ( Burmann 1 968, 1 970; Cramer, Allam et al. 1 974; Cramer,
Kanes et al. 1974; Cramer and Diez 1977; Vavrdova 1972). Some species, e.g. Arhusculidium
filamentosum, Cn'stallinium dentation, and S. principalis parva, are not known from the Llanvirn thus
confirming an Arenig age. The abundance of V. trispinosum is unusual in the early Arenig since it is
generally very rare until the late Arenig (Downie, /;cr5. comm.).

Environment. This was marine and the richness and variety of the assemblage suggest a continental
shelf site of deposition.

Provincialism. Comparison within China is limited to three localities. The Tongzi locality is some 350
to 550 km from the nearest of these. The two localities, however, are still in the Upper Yangtse Para-
platform of deposition. Xing Yusheng (1980) listed Coryphidium boliemicum and A. filamentosum,
two very characteristic species in an assemblage apparently similar to that of Tongzi. Zhong Guofang
(1981) described an assemblage which is more difficult to compare due to poor preservation. Cory-
phidium and Arhusculidium were not recorded by Zhong. The assemblage described from northern
China consists largely of sphaeromorphs and Baltisphaeridium and Micrhystridium (Li Zaiping
1982). It is in the North China Platform and has none of the distinctive species characteristic of the
Tongzi assemblage.

LI JUN: ORDOVICIAN ACRITARCHS

631

In making comparison with other regions of the world, most of the data come from Europe
and North Africa. Vavrdova (1974) divided the European Arenig/Llanvirn assemblages into
a Baltic Province and a Mediterranean Province. According to her division, the Baltic Province
comprising the northern part of the Soviet Union, Sweden, Poland, northern Germany, and prob-
ably part of the British Isles is characterized by the dominance of acanthomorphs with many species
of Baltisphaeridium, Peteiiwsphaeridium, and Gotnosphaeridium. Representatives of diacromorphs
are apparently absent. The Mediterranean Province extending from Belgium, France, Spain,
northern Africa to southern Germany, central Bohemia, and Bulgaria is characterized by the
prevalence of diacrodians such as Arhusculidium, Coryphidium, and Striatotheca.

Elements of the Mediterranean flora have been described from GDR (Burmann 1968, 1970),
Czechoslovakia (Vavrdova 1965, 1966, 1972, 1977), Hungary (Albani, Lelkes-Felvary and Tongiorgi
1985), Sardinia (Albani, Di Milia et cd. 1985), Bulgaria (Kalvacheva 1982), Ireland (Smith 1981),
Britain (Turner 1982; Downie 1984), France (Rauscher 1974), Belgium (Martin 1977), and Spain
(Cramer and Diez 1976). That described from Czechoslovakia by Vavrdova is the richest, comprising
some forty-two Arenig species and fifty-eight from the Llanvirn. A rich Arenig to Llanvirn
assemblages of more than fifty species is reported from Morocco (Cramer, Allam et al. 1 974; Cramer,
Kanes et al. 1974; Cramer and Diez 1977). This includes many characteristic Mediterranean genera
listed by Vavrdova (1974) with a number of new species. The occurrence of Coryphidium species in
Libya and Saudi Arabia (Cramer and Diez 1976) indicates that the two localities should be included
in the Mediterranean Province. Martin (in Dean and Martin 1978) lists forty-one species belonging to
twenty-one genera from east Newfoundland. The presence of the genera Arbuscididium, Coryphidium,
and Striatotheca shows a clear affinity with the Mediterranean Province.

This assemblage from the Upper Yangtze Paraplatform clearly has a greater affinity with the
Mediterranean Province than with the Baltic. It has fewer species than Czechoslovakia and Morocco.
Of twenty genera recognized in Tongzi only two, Tongzia gen. nov. and Rhopaliophora, have not been
found in the Mediterranean Province. Comparison with the Baltic Arenig shows that only B.
longispinosum, P. trifurcatwn intermedium, and Leiosphaeridia tenuissima are in common and they
occur rarely. The assemblage from the North China Platform ( Li Zaiping 1 982) appears more similar
to the Baltic one but the data are not good enough to make a satisfactory comparison.

Little can be said about the rest of the world. Arenig to Llanvirn assemblages from the Canning
Basin, north-west Australia (Combaz and Peniguel 1972; Playford and Martin 1984) contain mostly
new species not known elsewhere. The acritarchs are thus difficult to compare with Baltic and
Mediterranean assemblages although similar assemblages are recorded from the western part of
North America (K. J. Doming, /7cr5. comm.). Of the species with widespread records, P. trifurcatum
and B. longispinosum are characteristic of the Baltic Province, but appear to have a world-wide
distribution. Veryhachium trispinosum and V. lairdii are widespread across the Mediterranean belt.
V. cf. oklahomense, similar to V. lairdii, is stated to have an ornament of fine ribs (Playford and
Martin 1984, p. 215) and so resembles Striatotheca species. Pirea, a species recorded previously only
in the Mediterranean Province, was also described. Rhopalioplwra, a genus originally described from
the late Ordovician of Indiana and not recorded in Europe and Africa, was recognized both in the
Canning Basin and in the present study.

With the exception of east Newfoundland there are few detailed published accounts of Arenig
assemblages from North America. However, Rhopalioplwra (Tappan and Loeblich 1971) and
Petaloferidium (Jacobson 1978) from the central United States are two genera also recorded in the
present study.

The significance of this assemblage from south-west China, in terms of palaeogeography, cannot
be definitely ascertained because of the absence of adequate knowledge of Arenig acritarchs from
large areas of the world, but it seems to indicate the presence of a fairly homogeneous Arenig
assemblage extending from east Newfoundland through the Mediterranean area and the Middle East
to south-west China (the Mediterranean Province of Vavrdova).

The primary factor controlling acritarch provincialism was thought to be palaeotemperature
which is mainly determined by palaeolatitude (Cramer and Diez 1974). The Mediterranean Province

632

PALAEONTOLOGY, VOLUME 30

was believed to represent a cold belt (Vavrdova 1982). Based on trilobites, a Lower Ordovician
Selenopeltis Province characteristic of cool waters is recognized (Whittington and Hughes 1972). The
extent of the Selenopeltis Province is roughly the same as that of the Mediterranean Province. The
Chinese trilobite fauna is classified as of ‘uncertain alTnity’ because it appears to belong in the
Selenopeltis Province at the family level but not so at the generic level (Whittington and Hughes
1972). Plotting the distribution of Neseuretus, a trilobite genus considered to be a good indicator of
epicontinental seas at relatively high latitudes, rather than comparing whole faunas at the generic
level, the existence of a broadly united Gondwanan continent in the early Ordovician, including
southern Europe attached to North Africa, and including also the southern part of China, has been
proposed (Fortey and Morris 1982). The present study supports this to some extent. Lu Yanhao et al.
(1976) argued that no Ordovician latitudinal and climatic difterentiation could be derived from the
distribution of Chinese faunas because they, both shelly and graptolitic, showed a mixture of
elements belonging to difl'erent provinces. The Ordovician graptolite fauna of south-west China is
believed to link up with those of Europe, North Africa, and South America, and the climate of the
Ningkuoan (equivalent of Arenig and Llanvirn) might have been warm and arid (Mu Enzhi et al.
1979). We should take these arguments into account and need more data for a better understanding
of bioprovincialism and differentiation of climatic zones.

Acknowledgements. The author is extremely grateful to Professor C. Downie for instructive discussion, reading,
and correcting the manuscript; to F. Martin, D. Edwards, and K. Doming for their advice; and to Professor
J. B. Dawson for permission to use the facilities of the Department of Geology, University of Sheffield. The
author thanks Mr S. J. Ellin, Mr M. Cooper, and Mrs D. Darwin for their willing technical assistance, and Miss
P. Mellor for her careful typing. The author is indebted to Dr Geng Liangyu for his kind provision of the samples
and stratigraphical information. Financial support, jointly from the British Council and Academia Sinica, made
the author’s visit to the Department of Geology, University of Sheffield possible.

REFERENCES

ALBANi, R., LEiKES-FELVARY, G. and TONGiORGi, M. 1985. First recoid of Ordovician (Upper Arenigian, acritarchs)
beds in Bakony Mts., Hungary. Geol. palaeont. Ahh. 170, 45-65.

Di MiLiA, A., MiNZONi, N. and TONGIORGI, M. 1985. Nuovi dati palinologici e considerazioni geologiche

suH’eta delle Arenarie di Solanas (Cambro-Ordoviciano-Sardegna Centrale). Atti Sac. tosc. Sci. nat.. Mem.,
Serie A, 92, 1 -20.

BURMANN, G. 1968. Diacrodien aus dem unteren Ordovizium. Palaeont. Ahh. B, Palaeohot. 2, 639-652.

1970. Weitere organische Mikrofossilien aus dem unteren Ordovizium. Ibid. 3, 289-349.

cocCHio, A. M. 1982. Donnees nouvelles sur les acritarches du Tremadoc et de 1’ Arenig dans le Massif du
Mouthoumet (Corbieres, France). Revue Micropaleont. 25, 26-39.

COMBAZ, A. 1966. Remarques sur les niveaux a TasmanaceesduPaleozoiqueSaharien. Palaeohotanist, 15,29-34.

and PENiGUEL, G. 1972. Etude palynostratigraphique de I’Ordovicien dans quelques sondages du Bassin de

Canning (Australie occidental). Bull. Cent. Rech. Pan, 6, 121-167.

CRAMER, E. H., ALLAM, B., KANES, w. H. and DiEZ, M. D. c. R. 1974. Upper Aieiiigian to Lower Llanvirnian
acritarchs from the subsurface of the Tadla Basin of Morocco. Palaeontographica, B145, 182-190.

and DIEZ, M. D. c. r. 1974. Early Paleozoic Palynomorph Provinces and Paleoclimate. In ross, c. a. (ed.).

Palaeogeographic Provinces and Provinciality. Spec. Puh. 21, Soc. Econ. Palaeontol. Mineral. 177-188.

1976. Seven new late Arenigian species of the acritarch genus Coryphidium Vavrdova, 1972. Paldont.

Z. 50, 201-208.

1977. Late Arenigian (Ordovician) acritarchs from Cis-Saharan Morocco. Micropaleontologv, 23,

339-360.

KANES, w. H., DIEZ, M. D. c. R. and CHRESTOPHER, R. A. 1974. Early Ordovician acritarchs from the Tadla

Basin of Morocco. Palaeontographica, B146, 57-64.

DEAN, w. T. and MARTIN, F. 1978. Lower Ordovician acritarchs and trilobites from Bell Island, eastern
Newfoundland. Bull. geol. Surv. Can. 284, 35 pp.

DEUNFF, J. 1954. Veryliaclnum, genre nouveau d’Hystrichospheres du Primaire. C. r. somm. Seanc. Soc. geol. Fr.
11, 305-307.

- — 1961. Un micropaleonton a Hystrichspheres dans le Tremadoc du Sahara. Revue Micropaleont. 4, 37-52.

LI JUN: ORDOVICIAN ACRITARCHS

633

1964. Systematique du microplancton fossile a Acritarches. Revision de deux genres de rOrdovicien

inferieiir. Ibid. 7, 1 19-124.

- 1968. Arhuscididium, genre nouveau d’Acritarche du Tremadocien marocain. C. r. somni. Seanc. Sac. geol.
Fr. 3, 101.

1977. Un microplancton a acritarches dans les schistes llanvirniens de I’Anti-Atlas (Zagora-Maroc). Notes

Serv. geol. Maroc, 38, 141-151.

GORKA, H. and RAUSCHER, R. 1 974. Observations nouvelles et precisions sur les Acritarches a large ouverture

polaire du Paleozoique inferieur. Geohios, 7, 5-18.

DOWNIE, c. 1958. An assemblage of microplankton from the Shineton Shales (Tremadocian). Proc. Yorks, geol.
Sue. 31, 331-349.

1959. Hystrichospheres from the Silurian Wenlock Shale of England. Palaeontology, 2, 56-71.

1982. Lower Cambrian acritarchs from Scotland, Norway, Greenland and Canada. Tran. R. Soc. Edinh.

Ear. Sc. 72, 257-285.

1984. Acritarchs in British stratigraphy. Geol. Soc. London, Spec. Rep. 17, 26 pp.

EISENACK, A. 1931. Neue Mikrofossilien des baltischen Silurs. I. Paldont. Z. 13, 74-1 18.

1938. Hystrichosphaerideen und verwandte Formen im baltischen Silur. Z. Geschiebeforsch. Flachlandged.

14, 1-30.

1958c/. Tasinanites Newton 1875 und Leiosphaeridia n. g. als Gattungen der Hystrichosphaeridea.

Palaeontographica, AllO, 1-19.

1958/x Mikrofossilien aus dem Ordovizium des Baltikums, 1. Markasitschicht, Z)/cnY///c/;/c/-Schiefer,

Glaukonitsand, Glaukonitkalk. Senckenherg. leth. 39, 389-405.

1959. Neotypen baltischer Silur-Hystrichospharen und neue Arten. Palaeontographica, A1 12, 193-21 1 .

1965. Die Mikrofauna der Ostseekalke. 1. Chitinozoen, Hystrichospharen. Neues Jh. Geol. Paldont. Ahli.

123,115-148.

1976. Mikrofossilien aus dem Vaginatenkalk von Hiilludden, Oland. Palaeontographica, A154,

181-203.

FORTEY, R. A. and MORRIS, s. F. 1982. The Ordovician trilobite Neseuretus from Saudi Arabia, and the
palaeogeography of the Neseuretus fauna related to Gondwanaland in the earlier Ordovician. Bull. Br. Mas.
Nat. Hist. (Geol.), 36, 2, 63-75.

HARLAND, W. B., COX, A. V., LLEWELLYN, P. G., PICKTON, C. A. G., SMITH, A. G. and WALTERS, R. 1982. A geologic
time scale, 131 pp. Cambridge University Press, Cambridge.

GORKA, G. 1969. Microorganismes de I’Ordovicien de Pologne. Acta palaeontol. pol. 22, 102 pp.

JACOBSON, s. R. 1978. Acritarchs from the Upper Ordovician Clays Ferry Formation, Kentucky, USA.
Palinologia num. extraord. 1, 293-301 .

and Ai'CHA, A. 1985. Acritarch biostratigraphy of the Dicellograptns coinplanatiis graptolite zone from the

Vaureal Formation (Ashgillian), Anticosti Island, Quebec, Canada. Palynology, 9, 165-198.

KALVACHHVA, R. 1982. Palynological evidence for the Early Ordovician (Arenigian) age of ophiolites in the
Botevgrad District (West Balkan Mountain) Bulgaria. C. r. Acad. hidg. Sci. 35, 1 101-1 104.

KJELLSTROM, G. 1976. Lower Viruan (Middle Ordovician) Microplankton from the Ekon borehole no. 1 in
Gotland, Sweden. Sver. geol. Unders. Aarsh. 65 (1), 75 pp.

1 1 ZAIPING. 1982. Acritarchs from the Upper Machiakou Formation in North China and their geologic age. Acta
palaeont. sin. 21, 6, 715-732. [In Chinese.]

LOEBLiCH, A. R. Jr. and TAPPAN, H. 1978. Some Middle and Late Ordovician microphytoplankton from central
North America. J. Paleontol. 52, 1233-1287.

Lu, Y., CHU, €., CHiEN, Y., ZHOU, z., CHEN, J., LIU, G., YU, w., CHEN, X. and XU, H. 1 976. Oidovician biostratigraphy
and Palaeozoogeography of China. Mem. Nanjing Inst. Geol. Palaeont. Ac. Sin. 7, 83 pp. [In Chinese.]
MARTIN, E. 1969. Les acritarches de I’Ordovicien et du Silurien beiges. Determination et valeur stratigraphique.
Mem. Inst. r. Sci. nat. Belg. 160 (for 1968), 175 pp.

— 1974. Ordovicien superieur et Silurien inferieur a Deerlijk (Belgique). Palynofacies et Microfacies. Ibid.
174, 71 pp.

1977. Acritarches du Cambro-Ordovicien du Massif du Brabant, Belgique. Bull. Inst. r. Sci. nat. Belg. 51

(1975), Sci. de la Terre, 1, 1-33.

1982. Some aspects of late Cambrian and early Ordovician acritarchs. In m. g. bassett and w. t. dean

(eds.). The Cambrian-Ordovician boundary: sections, fossil distributions, and correlations, 29-40. Nat. Mas.
Wales Geol. Ser. 3, Cardiff.

and RICKARDS, b. 1979. Acritarchs, Chitinozoaires et Graptolithes ordoviciens et siluriens de la vallee de la

Sennette (Massif du Brabant, Belgique). Annls. Soc. geol. Belg. 102, 189-197.

634

PALAEONTOLOGY. VOLUME 30

MU ENZHi, ZHU ZHAOLiN, CHEN JUN YUAN and RONG JiAYU. 1979. The Ordovician System in Southwest China. In
NANJING INST. GEOL. PALAEONT. (cd.). The Carbonate hiostratigraphy of southwest china, 108-154. Science
Press, Beijing. [In Chinese.]

OUYANG SHU, YIN LEiMiNG and LI ZAiPiNG. 1974. The Cambrian paleospores. In nanjing inst. geol. palaeont.
(ed.). A handbook of the stratigraphy and paleontology in southwest China, 1 14-123. Science Press, Beijing. [In
Chinese.]

PLAYFORD, G. and MARTIN, F. 1984. Ordovician acritarchs from the Canning Basin, Western Australia.
Alcheringa, 8, 187-223.

RAUSCHFR, R. 1974. Rechcrclies Micropaleontologiques et Stratigraphiques dans I’Ordovicien et le Silurien en
France. Etude des acritarches. des chitinozoaires et des spores. Mem. Sci. geol. 38 (1973), 224 pp.

SMITH. D. G. 1981. Progress in Irish Lower Palaeozoic palynology. Rev. Palaeobot. Palynol. 34, 137-148.
STAPLiN, F. L. 1961. Reef-controlled distribution of Devonian microplankton in Alberta. Palaeontologv, 4,
392-424.

jansonius, j. and pocock, a. a. j. 1965. Evaluation of some acritarchous hystrichosphere genera. Neues Jh.

Geol. Paldont. Abh. 123, 167-201.

tappan, h. and lofblich, a. r. jr. 1971. Surface sculpture of the wall in Lower Paleozoic acritarchs.
Micropaleontology, 17, 385-410.

TIMOFEEV, B. V. 1958. Uber das alter sachsischer Grauwacken mikropalaophytologische Untersuchungen von
Proben aus der weesensteiner und Lausitzer Grauwacke. Geologie, 7, 826-845.

TONGiORGi, M. R. and Di MiLiA, A. 1984. The Solannas Sandstones of Central Sardinia: new paleontological data
(acritarchs) and an attempt of geological interpretation. (A ‘post-Sardinian’ molasse?). Bull. Soc. geol. Fr. 26,
665-680.

TURNER, R. E. 1982. Reworked acritarchs from the type section of the Ordovician Caradoc Series, Shropshire.
Palaeontology, 25, 1 19-143.

1984. Acritarchs from the type area of the Ordovician Caradoc Series, Shropshire, England. Palaeonto-

graphica, B190, 87-157.

VANGUESTAiNE, M. 1978. Ciiteies palynostratigraphiques conduisant a la reconnaissance d’un pli couche
Revinien dans le sondage de Grand-Halleux. Annls. Soc. geol. Belg. 100, 249-276.

VAVRDOVA, M. 1965. Oidoviciaii acritarchs from Central Bohemia. Vest, list fed. Ust. geol. 40, 351-357.

1966. Palaeozoic microplankton from Central Bohemia. Cas. Miner. Geol. 11, 409-414.

1972. Acritarchs from Klabava Shales (Arenig). Vest, listfed. Ust. geol. 47, 79-86.

1973. New acritarchs from Bohemian Arenig (Ordovician). Ibid. 48, 285-289.

1974. Geographical differentiation of Ordovician acritarch assemblages in Europe. Rev. Palaeobot.

Palynol. 18, 171-175.

1976. Excystment mechanism of Early Paleozoic acritarchs. Cas. Miner. Geol. 21, 55-64.

1977. Acritarchs from the Sarka Formation (Llanvirnian). Vest, ustfed. Ust. geol. 52, 109 -1 18.

1982. Phytoplankton communities of Cambrian and Ordovician age of Central Bohemia. Ibid. 57,

145-155.

WHITTINGTON, H. B. and HUGHES, c. p. 1972. Ordovician geography and provinces deduced from trilobite
distribution. Phil. Trans. R. Soc. B263, 235-278.

XING YUSHENG, 1980. Microplaiits and chitinozoans from the Lower Ordovician Dachengsi Formation of
Emeishan, Sichuan. In 5 International Palynological Conference Abstracts, 438. Cambridge.

1982. Microflora of the Sinian System and Lower Cambrian near Kunming, Yunnan and its stratigraphical

significance. Acta geol. sin. 56, 42-49. [In Chinese.]

ZHANG WENTANG and CHEN XU. 1964. Ordovician rocks in Honghuayuan, Tongzi. In nanjing inst. geol.
PALAEONT. (ed.). An excursion guide to Paleozoic rocks in N. Guizhou, 15-24. Nanjing. [In Chinese.]

LI JiJiN, ge meiyu and chen junyuan. 1982. Classification and correlation of the Ordovician in China. In

NANJING INST. GEOL. PALAEONT. (ed.). Stratigraphic correlation chart in China with explanatory text, 55-72.
Science Press, Beijing. [In Chinese.]

ZHONG GUOFANG. 1981. Early Ordovician microflora from the Dawan Formation at Huanghuachang Yichang.
Bull. Yichang Inst. Geol. M. R., Chinese Acad. Geol. Sci. Sp. Issue S. P., 118-126. [In Chinese.]

LI JUN

Nanjing Institute of Geology and Palaeontology
Academia Sinica

Typescript received 5 November 1985 Nanjing

Revised typescript received 21 November 1986 China

THE BELEMNITE ACROTEUTHIS IN THE
HIBOLITES BEDS (H AUTERI VI AN-BARREMI AN)
OF NORTH-WEST EUROPE

by J. MUTTERLOSE, G. PINCKNEY, and P. F. RAWSON

Abstract. Hiholites, a Tethyan-derived genus, was the dominant belemnite in north-west Europe for most of
Hauterivian to earliest Barremian (Early Cretaceous) time, while the Boreal genus Acroteuthis continued to
thrive in more northerly latitudes. Rare Acroteuthis occur in the Hiholites beds and are easily confused with
the slightly younger Aulacoteiithis. Two new Acroteuthis species belonging to the subgenus Boreioteuthis are
described here by Pinckney: A. (B). rawsoni and A. (B.) stolleyi-

The latest Jurassic-early Cretaceous belemnite genus Acroteuthis is a member of the boreal sub-
family Cylindroteuthinae, occurring mainly in Siberia, the Russian Platform, north-west Europe,
East Greenland, and Arctic Canada. In north-west Europe it first appeared in mid-Volgian (late-
Tithonian) times and died out early in the Barremian; over that interval, seven discrete assemblages
are distinguished (Pinckney and Rawson 1974). However, while the genus was the dominant belem-
nite in late Volgian to earliest Hauterivian sediments, it was almost completely replaced early in the
Hauterivian by the immigration of Hiholites from Tethys (Rawson 1973; Mutterlose et ctl. 1983). In
England the faunal change was taken as one of the main boundaries in the Speeton Clay (D/C beds
boundary: Lamplugh 1889), though here the abruptness of the change is exaggerated by condensed
sedimentation. Hiholites was almost completely replaced by a new boreal immigrant, Praeoxyteuthis,
early in the Barremian (C/B beds boundary at Speeton) though H. mimitus continued through the
Barremian.

While Hiholites was flourishing in north-west Europe during Hauterivian and earliest Barremian
times, Acroteuthis continued to evolve in other boreal areas. A few examples have been found in
the Hiholites beds of England and north Germany (text-fig. 1 ) and their occurrence is reviewed
here. They include two new species (assigned to the subgenus Boreioteuthis) described here by G.
Pinckney. The described specimens are from the following collections: BGS, British Geological
Survey, Keyworth; BM, British Museum (Natural History), London; GM, Geological Museum,
Copenhagen; GPIG, Geologisches-Palaontologisches Institut, Gottingen University; GPIH, Geo-
logisches-Palaontologisches Institut, Hannover University; NLfB, Niedersachsisches Landesamt
fur Bodenforschung, Hannover; SC, Stiihmer collection, Heligoland; WC, C. W. and E. V. Wright
Collection, London; UC, University College London.

ACROTEUTHIS ASSEMBLAGES IN THE HIBOLITES BEDS

Acroteuthis assemblages 6 and 7 of Pinckney and Rawson (1974) are represented in the Hiholites
beds. Assemblage 6 includes four species, A. (A.) suhquadratus (Roemer), A. (A.) explanatoicles
(Pavlow), A. (A.) acmonoides Swinnerton, and A. (A.) paracmonoides (Swinnerton), all of which
also occur in the underlying assemblage 5. It is characterized by greater numbers of the more slender
species A. acmonoides and A. explanatoides. Assemblage 6 occurs in the highest part of the
Acroteuthis beds and, at Speeton, as a relic fauna in the lower part of the Hiholites beds (text-
fig. 2). The constituent species are well known from Swinnerton’s (1937, 1948) monograph and are
not redescribed here.

Assemblage 7 is completely distinct from the underlying ones. It is characterized by forms with
an apical groove. Three species occur, A. (A?) conoides Swinnerton and two previously undescribed

IPalaeontology, Vol. 30, Part 3, 1987, pp. 635-645.|

© The Palaeontological Association

636

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 1. Map showing localities at which Acroteuthis occurs in the Hiholites beds.

ones, A. (Boreioteuthis) rawsoni Pinckney sp. nov. and A. {B. stolleyi) Pinckney sp. nov. All three
species are described and their phylogenetic significance discussed below. The assemblage 7 fauna
occurs in the middle to upper part of the Hiholites beds in both England and north Germany.

LOCALITIES

Seven localities in the Hiholites beds have yielded Acroteuthis: Speeton and Lincolnshire in eastern England,
the North Sea island of Heligoland, and four section's in Lower Saxony (text-fig. 1).

Speeton. The lithostratigraphy of the C beds (Hiholites beds) has been described in detail by Fletcher (1969)
and a lithic log published by Rawson (1971). Acroteuthis representing assemblages 6 and 7 have been found
in the mid C beds (regale to speetonensis Zones) and bed Cl (variahilis Zone).

Lincolnshire. A single A. (B). rawsoni was collected from the Tealby Limestone of Normanby during the last
century by the Revd J. Lee. The limestone is of earliest Barremian (variahilis-rarocinctum Zones) age.

Heligoland. Lower Cretaceous rocks are exposed on the sea-floor east of Heligoland in the 'Skit Gatt’. In
recent years skin divers have collected rich cephalopod faunas from these outcrops. The ammonite faunas
indicate that the lower part of the Hauterivian (ainblygonium, noricum, and possibly regale Zones) is condensed
or missing but the remainder of the Hauterivian and Barremian are well represented (Kemper et al. 1974;
Rawson 1975). The belemnites include several hundred H . jaculoides and one specimen each of A. (B.) rawsoni
znA A. (B.) stolleyi.

MUTTERLOSE, PINCKNEY AND RAWSON: AC ROTEVTHIS

637

Ahlimi. This old section is no longer visible: it was situated south of Ahlum, about 4 km east of Wolfenbuttel
(TK 25 Wolfenbiittel, Nr. 3829, re: 36 44 048, h: 57 82 550). Stolley (1906) described the sequence; from the
lowest beds he recorded Simbirskites (Craspedodiscus) of the phiUipsi group which indicates the discofalcatus
Zone. Two large Belemnites aff. siihquadratus were noted and later Stolley ( 1 925) included them (with additional
material) in A. ahlianensis Stolley nom. mid. He also recorded some much more slender forms. Stolley’s records
apparently represent A. (B.) stoUeyi and A. (B.) rawsoni, several examples of which are still preserved in old
collections from Ahlum.

STAGE

BELEMNITE

BEDS

AMMONITE

ZONES

England I Germany

BELEMNITE
OCCURRENCES
■ Germany &■ England
□ England

CC ^ CO

< ^ S

CQ

z

<

>

QC

LU

1-

Z)

<

X

Praeoxy-
teuthis
beds (pars)

HiboUtes

beds

rarocinctum

variabilis

discofalcatus

■ ■

marginatus

gottschei

CD

speetonensis

staffi

0)

05

inversum

□ £
n

regale

□ W

<

Acroteu-

this

beds (pars)

noricum

conoides

rawsoni

StoUeyi

amblygonium

TEXT-FIG. 2. Stratigraphical occurrence of Acroteuthis in the Hiholites beds of north-west Europe.

Delligseii (Hits). During recent excavations for the foundations of a sports hall in Delligsen (TK 25 Alfeld, Nr
4024, re: 35 54 900, h: 57 56 460), dark clays were exposed from which a single A. (B.) stoUeyi and a fragment
of S. (C.) discofalcatus were collected.

Fhelingen. In 1979 a new clay pit, Ziegelei Oltman, was opened near Frielingen, about 20 km north-west of
Hannover (TK 25 Garbsen, Nr. 3523, re: 35 34 275, h: 58 17 125). It exposes about 14 m of clay, forming a
rhythmic sequence of pale and dark beds, occasionally intensively bioturbated. A rich discofalcatus Zone
ammonite fauna occurs: S. (C.) discofalcatus (Lahusen), S. (C.)Juddi Rawson, S. (C.) phiUipsi group, S. (S.)
toensbergensis (Weerth), and Paracrioceras spatlii Kemper, Rawson and Thieuloy. These occur with abundant
H . jacidoides and hence the total assemblage indicates the middle part of the discofalcatus Zone: it correlates
with the lower part of the English variabilis Zone of earliest Barremian age.

The top two metres of the section have yielded sixteen specimens of Acroteuthis (Boreioteuthis), an unusually
large number: both A.(B.) rawsoni and A.(B.) stoUeyi occur.

Sarstedt. The clay pit of Ziegelei Gott lies on the outskirts of Sarstcdt, about 30 km south-east of Hannover
(TK 25 Sarstedt, Nr. 3725, re: 35 60 400, h: 57 90 650). It is one of the key sections in north Germany, exposing
about 70 m of shallow-water Upper Hauterivian, Barremian and Upper Aptian clays deposited on the flank
of a salt stock. (For section details see Mutterlose 1983, fig. 4). The section has yielded two loose specimens of
A. (B.) StoUeyi, probably from the discofalcatus Zone.

PALAEONTOLOGY, VOLUME 30

638

TEXT-FIG. 3, a h, Acroteiithis (Boreioteiithis) rawsoni Pinckney sp. nov. a-d, Paratypes, discofalcatus
Zone, Frielingen (GPIH 1 985-1-1, 1985-1-2); e,f, Paratype, bed C6 base (speelonensis Zone), Speeton
(BM.C. 59522); g-h, Holotype, mid C Beds, Speeton (BGS 24454). /, y, A. (Ad) conoides Swinnerton.
Bed "Cl or above’, Speeton (WC 18350). k, /, A. (B.) stoUeyi Pinckney sp. nov., Paratype, discofalcatus
Zone, Frielingen (GPIFI 1985-1-1 1). a, c, e, g, i, k, ventral views; h, d, f, h,j, /, lateral views; x 1.

MUTTERLOSE. PINCKNEY AND RAWSON: ACROTEUTHIS

639

SYSTEMATIC PALAEONTOLOGY
(by G. Pinckney)

Order bf.lemnitida Zittel, 1895
Suborder belemnitina Zittel, 1895
Family belemnitidae d’Orbigny, 1845
Subfamily cylindroteuthinae Stolley, 1919

Genus acroteuthis Stolley, 1911

Diagnosis. Rostrum depressed, generally wedge-like in profile; apical line markedly displaced
towards the venter; alveolus excentric and moderately deep.

Subgenus acroteuthis Stolley, 1911
Type species. Belemnites subqitadratus Roemer, 1836.

Diagnosis. Acroteuthis with relatively large rostra that are wedge-like in profile, with only a weakly-
developed ventral groove. Juvenile rostra moderately slender, spindle-shaped.

Acroteuthis (Acroteuthis?) conoides Swinnerton, 1937
Text-figs. 3/, /, 4n, h

V* 1937 Swinnerton; p. 17, pi. 6, fig. 2

71964 Acroteuthis cf. conoides Swinnerton; Jeletzky, p. 56, pi. 14, fig. 3.

71964 Acroteuthis aff. conoides Swinnerton; Jeletzky, p. 58, pi. 1 5, fig. 3.

Type. Holotype, BGS 17298 (Danford Collection), Beds C7 C8, Speeton Clay, Speeton, Yorkshire.

Material. 6 specimens from the C beds of the Speeton Clay: BM.C. 59521 (Rawson Collection) from C7E;
WC 21284 from C8 and 18350 from ‘C7 or above’; BGS (Danford Collection) 17299, 17318, and 17319 from
‘mid C’.

Diagnosis. Rostrum slender and conical; ventral apical groove well developed; apical angle acute in
both outline and profile; transverse sections slightly depressed in alveolar and stem regions, becom-
ing compressed posteriorly. Alveolus occupies about half length of guard and is weakly excentric.

Discussion. A. (A.?) conoides was described in detail by Swinnerton (1937). It is only tentatively
referred to the subgenus Acroteuthis because it appears to be a transitional form towards Boreio-
teuthis. It resembles the latter in the weakly depressed condition of the rostrum and in the slight
excentricity of the alveolus, but differs in the more feeble development of the ventral groove. Its
stratigraphical horizon is also transitional between the ranges of the two subgenera in north-west
Europe.

A. (A.?) conoides most closely approaches conical variants of A. (A.) explanatoides (Pavlow), but
has a more compressed rostrum, better developed ventral groove and more acute apex in both
outline and profile.

Geographic distribution: A. (A.?) conoides is known only from England, though very similar, possibly identical,
forms have been figured from Northern Canada (Jeletzky 1964).

Subgenus boreioteuthis Saks and Nalnyaeva, 1966
Type species. Acroteuthis (Boreioteuthi.s) niiga Saks and Nalnyaeva, 1966. Lower Volgian. northern Siberia.

Diagnosis. Acroteuthis with relatively large, moderately elongate, weakly depressed rostra; with a
well-developed ventral groove extending into the stem region; and with markedly slender, clavate
juvenile rostra.

640 PALAEONTOLOGY. VOLUME 30

TEXT-FIG. 4. a, b, Acroteuthis (A.?) conoides Swinnerton, bed C8 {regale Zone), Speeton (WC 21284);
c, d, A. {Boreioteuthis) stoUeyi Pinckney sp. nov., Iiolotype, discofalcatus Zone, Ahlum (NLf B Kv56).

a, f, ventral views; h, d lateral views; x 1 .

Discussion. The subgenus Boreioteuthis is distinguished from Acroteuthis s.s. by its better-developed
median ventral groove, which extends beyond the apical region, and by its more slender juvenile
rostrum. It differs from the subgenus Microbelus in the possession of a larger ‘adult’ rostrum, and
in the better development of the median ventral groove.

Our concept of Boreioteuthis is broader than that of Saks and Nalnyaeva (1966). Here, all
Acroteuthis species with a median ventral groove that extends beyond the apical region are included,
whereas Saks and Nalnyaeva (1966) only placed species possessing a ventral groove that extended
into the alveolar region in this subgenus.

Stratigraphical distribution. According to Saks and Nalnyaeva (1966), Boreioteuthis ranges in age from Oxfor-
dian to Barremian. However, in western Europe it is known only from strata of late Volgian and late
Hauterivian to earliest Barremian age.

Acroteuthis (Boreioteuthis) rawsoni sp. nov.

Text-fig. 3a- h

71901 Beienmites speetonensis Pavlow and Lamplugh; Pavlow, p. 82 pi. 8, fig. 7.
vp. 1937 Acroteuthis festucaiis Swinnerton, pi. 9, fig. 2, non pi. 9, figs. 1 and 3.

MUTTERLOSE, PINCKNEY AND RAWSON: ACROTEUTHIS

641

V. J974 Acroteuthis (Boreioteuthis) sp. nov. b, Pinckney and Rawson, p. 202.

V. 1982 Acroteuthis (Boreioteuthis) ci. festucalis Swinnerton; Stuhmer, Spaeth and Schmid, pi. 20, fig. 2.

V. J983 Acroteuthis (Boreioteuthis) rawsoni Pinckney [MS]: Mutterlose, p. 15 (nomen nudum).

V. 1983 Acroteuthis (Boreioteuthis) rawsoni Pinckney [MS]: Mutterlose, Schmid and Spaeth, p. 299
(nomen nudum).

Type series. Holotype: BGS 24454 (Danford Collection), mid C beds, Speeton Clay, Speeton. Paratypes: 1
specimen from the Tealby Limestone of Lincolnshire (BM. C. 59519); 2 from the Speeton Clay of Speeton
(UC Pinckney Collection 366 from C7H, BM.C. 59522 [Rawson collection] from the base of C6); 8 from the
middle of the discofalcatus Zone at Frielingen (GPIH 198 5-1-1 -1985-1-8); 4 from Ahlum, probably discofalcatus
Zone(GM 1923.599 [I of 3 specimens], GPIG 3 uncatalogued specimens); I from Heligoland (SC 1503, figured
Stuhmer et al. 1 982).

Diagnosis. Rostrum slender and weakly depressed; apex very acute and only slightly depressed;
median ventral groove quite well developed, extending into stem region; alveolus shallow.

Description. Rostra quite short, subconical to subcylindrical, very slender; actual length of rostrum up to 6-5
times the maximum width. In outline the sides of the alveolar and stem regions are usually quite straight and
subparallel, but may converge slightly adapically. Rostral sides are weakly curved in the apical region and
converge with moderate rapidity to form a long, acute to very acute apex, varying from about 32° to 40°. In
profile the sides of the rostrum are almost straight in alveolar and stem regions. They are subparallel in front
but converge gently adapically, so that the point of maximum thickness of the rostrum is near the alveolar
rim. Convergence of rostral margins in apical region is quite gentle. Dorsal surface only slightly more curved
than ventral, producing a feebly depressed, weakly asymmetrical apex, ranging from about 30° to 36°.

Transverse sections are weakly depressed and strongly quadrate through most of length of rostrum, except
in apical region where they approach an oval form. Dorsal surface almost semicircular. Ventral surface, which
is about same width as dorsal, is also semicircular in the alveolar region but flattens adapically. Flattening is
accompanied by the development of a distinct, long, median ventral groove which is deep and narrow in the
apical region but shallow and often excavate in the stem region, where it disappears. Rostral flanks markedly
flattened.

Lateral lines are generally poorly visible. On well-preserved rostra, the upper main lateral line is represented
by a relatively broad, flattened belt, and the lower by an indistinct, narrow, flattened belt. The minor lateral
line has not been seen.

Alveolus shallow, occupying about one third of the actual length of the rostrum. It consists of a ventrally
curved, conical depression, the apex of which is weakly excentric. The apical line is also excentric and curves
gently towards the venter throughout its course.

No juvenile specimens are known. Evidence from dorsoventral longitudinal sections indicates that the
juvenile rostrum is clavate and very slender, the maximum thickness being up to eight times the total rostral
length.

Discussion. Swinnerton (1937) included specimens here regarded as A. (B.) rawsoni 'm his new species
A. festucalis. However, the holotype of the latter differs from A. (B.) rawsoni in its greater size,
more weakly developed median ventral groove, and in the possession of a feebly clavate rostrum.
Moreover, the species are well separated stratigraphically, A. {A.) festucalis being of Volgian age
and A. (B.) rawsoni of mid to late Hauterivian age.

A. (B.) rawsoni superficially resembles Aulacoteuthis descendens (Stolley) but differs principally in
its slightly stouter form, less strongly developed ventral groove, and in the occurrence of cylindroteu-
thid lateral lines.

Acroteuthis (Boreioteuthis) stolleyi sp. nov.

Text-figs. 3k, /, 4c, d, 5a, h

v 71906 Belemnites aff. pseudopanderi Sinzov; Danford, p.7, pi. 3, fig. 16; pi. 6, fig. 16.

? 1925 Acroteuthis ahlumensis Sio\\Qy,y>. \M (Nomen nudum).

V. 1974 Acroteuthis (Boreioteuthis) sp. nov. c, Pinckney and Rawson, p. 202

V. 1980 Acroteuthis (Boreioteuthis) stolleyi Pinckney [MS]; Mutterlose, pp. 239, 240 (nomen nudum).

V. 1983 Acroteuthis (Boreioteuthis) stolleyi Pinckney [MS]; Mutterlose, Schmid and Spaeth, p. 299 (nomen
nudum).

642

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 5. Acroteuthis (Boreioteuthis) stolleyi Pinckney
sp. nov., paratype, discofalcatus Zone, Frielingen (GPIH
1985-1-10). a, ventral view; b, lateral view; x 1.

b

Type series. Holotype: NLlTi Kv56, Simbirskiten Schichten, Ziegelei Ahlum, Wolfenbuttel, near Braunschweig.
Paratypes: 8 specimens from Frielingen (GPIFI 1985-I-9-1985-F16); 6 from Ahlum (GM 1923-599 [2 of 3
specimens], 1923.600 [2 specimens], GPIG 2 uncatalogued specimens); 2 from Sarstedt (GPIH); 1 from
Heligoland (SC), 1 from Delligsen (Strohmeyer private collection: cast in GPIH).

Diagnosis. Rostrum quite stout and weakly depressed; apex obtuse and inflated; median ventral
groove quite well developed, extending into stem region; alveolus quite deep.

Description. Rostra large, subcylindrical to subconical and moderately stout, actual length of rostrum being
about 3-5 times the maximum width. In outline the sides of alveolar and stem regions are almost straight and
subparallel, but they converge slightly adapically. Rostral sides converge more rapidly and curve quite strongly
in apical region to form a moderately obtuse apex, ranging from about 55° to 65°. In profile, sides of
rostrum almost straight and subparallel in alveolar region, but they curve gently and converge in stem region.
Convergence of rostral margins in apical region is strong. Ventral surface quite weakly curved, but dorsal
surface markedly inflated, producing an obtuse, asymmetrical apex, varying from about 50° to 60°.

Transverse sections are weakly depressed and quadrate in alveolar region, but become more compressed
and oval adapically. Dorsal surface is subsemicircular in front and closely approaches a semicircular condition

MUTTERLOSE, PINCKNEY AND RAWSON: ACROTEUTHIS

643

in the apical region. The ventral surface, which is about the same width as the dorsal, is subsemicircular in the
alveolar region but becomes flatter and weakly concave in apical and stem regions, where the median groove
is developed. The groove is quite deep in the apical region but shallows adorally, eventually disappearing in
the stem region. It does not extend quite to the apical tip, where the venter is feebly swollen. Rostral flanks
quite strongly flattened.

Evidence of lateral lines is poor. The upper main lateral line is best developed and forms a broad, indistinct
belt in the alveolar and stem regions that disappears adapically. The nature of the other lateral lines is unclear.

The alveolus is deep and occupies almost two thirds of the total length of the rostrum. It comprises a
ventrally curved conical depression, the apex of which is moderately excentric. The apical line is also excentric
and curves gently towards the venter throughout its course.

No juvenile specimens of A. (B.) stolleyi have been seen. However, smaller individuals are more slender
with a more acute, less inflated apex and a shallower alveolus.

Discussion. A. (B.) stolleyi very closely resembles A. (A.) p.seuciopamleri (Sinzov emend. Pavlov) but
has a better-developed median ventral groove and a shallower alveolus. It can be distinguished
from A. (A.) acrei Swinnerton, A. (A.) partneyi Swinnerton, A. (A.) bojarkae Saks and Nalnjaeva,
and A. (A.) chelae Saks and Nalnjaeva chiefly in the stronger development of the ventral groove,
which extends into the stem region.

A. (B.) stolleyi apparently embraces A. ahlitmensis Stolley. The latter was neither figured nor
described and is therefore a nomen nudum. However, specimens of A. {B.) stolleyi from Ahlum were
purchased from Stolley by the Mineralogisk Museum, Copenhagen (now Geologisk Museum), and
bears his labels ''A. ahlumensis'.

A. (B.) stolleyi probably occurs in bed Cl (variahilis Zone) at Speeton. Two specimens are known,
though both are too corroded for Arm identification. One, Danford’s (1906) flgured Belemnites aff.
pseudopanderi, came from ‘the lower of the two mottled beds at the upper limit of the C division’
{— Cl B), while a fragment with the apical region missing has been collected recently from bed CIA
(BM.C. 59523: Rawson Collection).

PHYLOGENETIC PROBLEMS

The Acroteuthis suhgenera

The origin and phylogeny of the Acroteuthis subgenera is relevant to our interpretation of the late
forms described here and is therefore discussed briefly.

Acroteuthis is divided into three subgenera: Microhelus (Callovian-Hauterivian in the USSR,
Middle and Upper Volgian in north-west Europe), Boreioteuthis (Oxfordian-Barremian in the
USSR, Upper Volgian-Early Barremian in north-west Europe), and Acroteuthis s.s. (Middle
Volgian-Upper Hauterivian in USSR and north-west Europe, ?Aptian in USSR).

Microhelus is characterized by a relatively small, quite slender, depressed guard, with a median
ventral groove confined to the apical region. According to Saks and Nalnyaeva (1966, p. 172) it
evolved from Pachyteuthis (P.) parens Saks and Nalnjaeva in the Callovian. Saks and Nalnyaeva
(1966, p. 174) also suggested that Boreioteuthis and Acroteuthis s.s. both evolved from A. (Microhe-
lus) pseudolateralis Gustomesov, the former in the Oxfordian and the latter during the Volgian.
Both Boreioteuthis and Acroteuthis s.s. are larger than Microhelus: Boreioteuthis has a well-
developed ventral groove extending on to the stem region while Acroteuthis s.s. has a weakly
developed one.

Saks and Nalnyaeva (1966, p. 173) suggested that from the Volgian onward each subgenus
represented an independently evolving lineage.

While we agree that Boreioteuthis is probably derived from M. pseudolateralis by an increase in
the length of the ventral groove, the evolution of Acroteuthis s.s. is more problematic for two reasons:

1. There is a long time gap between the extinction of M. pseudolateralis in the Late Oxfordian
and the first appearance of true Acroteuthis in the Middle Volgian, and no transitional forms are
known.

2. Many of the earliest Acroteuthis s.s. (e.g. A. (A.) lindseyen.sis Swinnerton) closely resemble
Pachyteuthis species. These resemblances may be homeomorphic, but if not then Acroteuthis s.s.

644

PALAEONTOLOGY, VOLUME 30

may have evolved directly from Pachyteuthis (Callovian-Kimmeridgian) and the genus Acroteuthis
as currently defined would be polyphyletic in origin.

Origin and evolution of the conoides-rawsoni-stolleyi group

A. (A.) explanatoides (Pavlow) is a long-ranging form (earliest Ryazanian to early Hauterivian)
which apparently gave rise to several shorter-lived species. The last of these was A. {A.l) conoides,
an interpretation on which we concur with Saks and Nalnyaeva (1966, p. 175). The species occur in
stratigraphic succession at Speeton (text-fig. 2) with only a small gap between them which may well
reflect simply their rarity in the Hiholites beds.

Stratigraphic occurrences (text-fig. 2) suggest that there may be an evolutionary lineage from A.
(A.?) conoides through A. (B.) rawsoni to A. (B.) stoUeyi. Early A. (B.) rawsoni slightly overlap with
A. (A.?) conoides in the inversum Zone while younger examples occur with A. {B.) stolleyi in the
discofalcatus Zone. Three main evolutionary trends are identified in this apparent lineage:

1. an increase in the length and depth of the ventral groove;

2. a decrease in the degree of depression of the rostrum in transverse section;

3. a progressive change from an essentially conical to a cylindrical rostrum.

If this lineage is confirmed by further finds, then the species rawsoni and stolleyi will need to be
placed in a new subgenus which bears only a homeomorphic relationship to Boreioteuthis.

HOMEOMORPHY WITH AU LACOTEUTH IS

The development of an apical groove characterizes not only Boreioteuthis and some Acroteuthis
but also the oxyteuthinid genus Aulacoteuthis. The subfamily Oxyteuthinae characterizes strata
immediately overlying the Hiholites beds in north-west Europe (Mutterlose 1983) and exhibits
an evolutionary lineage: Praeoxyteuthis, a slender, ungrooved genus gave rise to Aulacoteuthis,
characterized by an apical groove, which in turn evolved into the ungrooved Oxyteuthis.

Most Aulacoteuthis are readily distinguished from the north-west European Boreioteuthis by a
longer groove and a slimmer guard. However, A. descendens, the youngest (mid-Barremian) form,
has a short, stout guard with a groove that is sometimes short. Such forms closely resemble A. {B.)
rawsoni and can only be differentiated by the lateral lines. While the latter possesses cylindroteuthid
lateral lines, A. descendens has oxyteuthid ones (Mutterlose 1983, fig. 57). The close homeomorphy
of the two species has caused confusion in the literature. From Simbirsk (now Ulyanovsk) on the
Volga, Pavlow (1901, p. 82, pi. 8, fig. 7) described a short-grooved, stout belemnite from the mid-
Hauterivian versicolor Zone and identified it as Belemnites speetonensis (i.e. Aulacoteuthis). It is
here tentatively placed in A. (B.) rawsoni, though Swinnerton (1948, p. 48) identified the same
specimen as A. descendens. This may be one of the reasons why Swinnerton regarded Aulacoteuthis
and Oxyteuthis (including Praeoxyteuthis) as two forms derived from a common stock rather than
agreeing with Stolley (1925, 1927) that the former was a grooved stage interposed in the evolution
of the latter two genera. The proven stratigraphic separation of the late Hauterivian to earliest
Barremian Boreioteuthis from the mid-Barremian Aulacoteuthis supports the morphological evi-
dence that they are homeomorphs.

CONCLUSIONS

While the rare Acroteuthis in the lower part of the north-west European Hiholites beds are simply relics of an
earlier fauna, later examples of the Acroteuthis (A.?) conoides Boreioteuthis group are a new, more strongly
grooved group that can be confused with slightly younger Aulacoteuthis. Material is insufficient to prove
whether the conoides-rawsoni-stolleyi succession evolved in the area or whether the species migrated individu-
ally from other boreal areas, but in situ evolution is our preferred model. It is striking that while Hiholites is
abundant in both shallow and deep water clays, Boreioteuthis is unusually common in the one section
(Frielingen) where deeper water clays are exposed, so it may have adapted to an offshore environment.

MUTTERLOSE, PINCKNEY AND RAWSON: AC ROTEUTHIS

645

Acknowledgements. J.M. acknowledges financial help from the British Council and the Deutsche Forschungs-
gemeinschaft; G.P.’s research was during the tenure of an NERC Research Studentship at Queen Mary
College, London: he gratefully acknowledges the permission of the Mobil Oil Corporation, New York, to
publish this paper. We are grateful to the custodians of the various collections mentioned above for access to
specimens. The figures were drawn by Janet Baker and the photographs are by Mike Grey, both of University
College London.

REFERENCES

DANFORD, c. G. 1906. Notes on the belemnites of the Specton Clays. Trans. Hull geol. Soc. 5, 1 14.

FLETCHER, B. N. 1969. A lithological subdivision of the Speeton Clay C Beds (Hauterivian), East Yorkshire.
Proc. Yorks, geol. Soc. 37, 323- 327.

JELETZKY, j. A. 1964. Illustrations of Canadian fossils. Lower Cretaceous marine index fossils of the sedimentary
basins of Western and Arctic Canada. Geol. Siirv. Pap. Can. 64-11, 1 101.

KEMPER, E., RAWSON, p. F., SCHMID, F. and SPAETH, c. 1974. Die Megafauna der Kreide von Helgoland und ihre
biostratigraphische Deutung. Newsl. Strut. 3, 121 - 137.

LAMPLUGH, G. w. 1889. On the subdivisions of the Speeton Clay. Q. Jl geol. Soc. Loiul. 45, 575 618.

MUTTERLOSE, J. 1980. Zur Gliederung des Unter-Barreme in NW-Europa mit Hilfe der Unter-Familie Oxyten-
thinae Stolley (Belemnitida). Newsl. Strut. 8, 238-243.

1983. Phylogenie und Biostratigraphie der Unterfamilie Oxyteuthinae (Belemnitida) aus dem Barrcme

(Unter-Kreide) NW-Europa. Palaeontographica, A180, 1 90.

SCHMID, F. and spaeth, c. 1983. Zur Paliiobiogeographie von Belemniten der Unter-Kreide in NW-

Europa. Zitteliana, 10, 293-307.

PAVLOW, A. p. 1901 . LeCrctaceinferieur de la Russieet sa faune. Nouv. Mem. Soc. {imp.) Nat. Mosc. ns 16, I 87.

PINCKNEY, G. and RAWSON, p. F. 1974. Acroteuthis assemblages in the Upper Jurassic and Lower Cretaceous
of northwest Europe. Newsl. Strut. 3, 193-204.

RAWSON, p. F. 1971. The Hauterivian (Lower Cretaceous) biostratigraphy of the Speeton Clay of Yorkshire,
England. Ibid. I (4), 61 75.

1973. Lower Cretaceous (Ryazanian Barremian) marine connections and cephalopod migrations be-
tween the Tethyan and Boreal Realms. In casey, r. and rawson, p. e. (eds). The Boreal Lower Cretaceous,
131 144. Seel House Press, Liverpool.

1975. Hauterivian (Lower Cretaceous) ammonites from Helgoland. Geol. Jh. A25, 55-83.

SAKS, V. N. and nalnyaeva, t. i. 1966. Upper Jurassic and Lower Cretaceous belemnites of northern USSR.
Genera Pachyteuthis and Acroteuthis, 260 pp. Nauka, Moscow and Leningrad. [In Russian.]

STOLLEY, E. 1906. Uber alte und ncue Aufschlusse und Profile in der unteren Kreide Braunschweigs und
Hannovers. Jher. Ver. Nat. Braunschweig, 15, 1 44.

1925. Beitrage zur Kenntnis der Cephalopoden der norddeutschen unteren Kreide. 1. Die Belemniten

der norddeutschen unteren Kreide. 2. Die Oxyteuthinae des norddeutschen Neokoms. Geol. Paldont. Ahh.
NF 14, 177-212.

1927. Zur Systematik und stratigraphie median gefurchter Belemniten. Jher. Vols. Geol. Ver. 20, 1 1 1 1 36.

STUHMHR, H. H., SPAETH, c. and SCHMID, F. 1982. FossUien Helgolands, Ted I. Nicderelbe-Verlag, 184 pp.

swinnerton, h. h. 1937. A monograph of British Lower Cretaceous belemnites. Palaeontogr. Soc. [Monogr.],
Part 2, 17-30.

1948. A monograph of British Lower Cretaceous belemnites. Ibid., Part 3, 31-52.

J. MUTTERLOSE

Geologisches-Palaontologisches-Institut
Universitat Hannover, Callinstrasse 30
D 3000 Hannover 1, West Germany

G. PINCKNEY

Mobil Oil Corporation, Exploration Department
150 East 42nd Street
New York, NY 10017, USA

P. F. RAWSON

Department of Geological Sciences
University College London
Gower Street, London WCIE 6BT

Typescript received 17 September 1985
Revised typescript 20 September 1986

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 CFl IXL, 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 1987 is £33 (U.S. S58) for Institutional
Members, and £16-50 (U.S. S30) 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 ILB, England.

RECENT PALAEONTOLOGICAL ASSOCIATION

PUBLICATIONS

Special Papers in Palaeontology

Numbers 1 -19 are still in print and are available (post free) together with those listed below:

20. (for 1977): Fossil Priapulid Worms, by s. conway morris. 155 pp., 99 text-figs., 30 plates. Price £16 (U.S. S24).

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 j. p. bujak, c. downie, g. l.
EATON andG. L. WILLIAMS. 100 pp., 24 text-figs., 22 plates. Price £15 (U.S. $23).

25. (for 1980): Stereom Microstructure of the Echmoid Test, bv 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.
1% pp., \0 text-figs., \9 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 t. c. w. cope and p. w. skelton. 202 pp., SO text-
figs., 4 plates. Price £30 (U.S. $45).

34. (tor 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., \6 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).

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).

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 -1-241 pp. Price £13-95. Available in the USA from Halsted Press at U.S. $24.95.

© The Palaeontological Association, 1987

Palaeontology

VOLUME 30 • PART 3
CONTENTS

Pfotocystites menevensis — a stem-group chordate (Cornuta)
from the Middle Cambrian of South Wales

R. P. S. JEFFERIES, M. LEWIS and S. K. DONOVAN 429

A review of favositid affinities

C. T. SCRUTTON 485

A staurikosaurid dinosaur from the Upper Triassic Ischigua-
lasto Formation of Argentina and the relationships of the
Staurikosauridae

D. B. BRINKMAN and H.-D. SUES 493

A Caribbean rudist bivalve in Oman; island-hopping across
the Pacific in the late Cretaceous

P. W. SKELTON unJ V. P. WRIGHT 505

A reinterpretation of ichthyosaur swimming and buoyancy

M. A. TAYLOR 531

A new Silurian xiphosuran from Podolia, Ukraine, USSR

P. A. SELDEN and D. M. DRYGANT 537

A taphonomic and diagenetic case study of a partially arti-
culated ichthyosaur

D. M. MARTILL 543

The armoured dinosaur Polacanthus foxi, from the Lower
Cretaceous of the Isle of Wight

W. T. BLOWS 557

English Eocene Crustacea (lobsters and stomatopod)

W. J. QUAYLE 581

Ordovician acritarchs from the Meitan Formation of Guizhou
Province, south-west China

LI JUN 613

The belemnite Acroteuthis in the Hiholites Beds (Hauterivian-
Barremian) of north-west Europe

J. MUTTERLOSE, G. PINCKNEY and P. F. RAWSON 635

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 1987-1988

President: Dr. L. R. M. Cocks, Department of Palaeontology, British Museum (Natural History), Cromwell Road,

London SW7 5BD

Vice-Presidents: Dr. D. E. G. Briggs, Department of Geology, University of Bristol, Bristol BS8 IRJ
Dr. L. B. Halstead, Department of Geology, University of Reading, Reading RG6 2AB
Treasurer: Dr. M. Romano, Department of Geology, University of Sheffield, Sheffield SI 3JD
Membership Treasurer: Dr. A. T. Thomas, Department of Geological Sciences, University of Aston, Birmingham B4 7ET
Institutional Membership Treasurer: Dr. A. W. Owen, Department of Geology, The University, Dundee DDl 4HN
Secretary: Dr. P. W. Skelton, Department of Earth Sciences, The Open University, Milton Keynes MK7 6AA
Circular Reporter: Dr. D. J. Siveter, Department of Geology, University of Hull, Hull HU6 7RX
Marketing Manager: Dr. V. P. Wright, Department of Geology, University of Bristol, Bristol BS8 IRJ
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. P. R. Crowther, City of Bristol Museum and Art Gallery, Bristol BS8 1 RL
Dr. D. Edwards, Department of Plant Science, University College, Cardiff CFl IXL
Dr. T. J. Palmer, Department of Geology, University College of Wales, Aberystwyth SY23 2AX
Dr. C. R. C. Paul, Department of Geology, University of Liverpool, Liverpool L69 3BX
Dr. P. A. Selden, Department of Extra-Mural Studies, University of Manchester, Manchester M13 9PL

Other Members

Dr. H. A. Armstrong, Newcastle upon Tyne Dr. G. B. Curry, Glasgow

Dr. M. E. CoLLiNSON, London Professor B. M. Funnell, Norwich

Dr. J. A. Crame, Cambridge Dr. P. D. Taylor, 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 1987 are:

Institutional membership £50 00 (U.S. $79)

Ordinary membership . £2100 (U.S. $38)

Student membership £1 1 50 (U.S. $20)

Retired membership £10-50 (U.S. $19)

There is ho admission fee. Correspondence concerned with Institutional Membership should be addressed to Dr. A. W. Owen,
Department of Geology, I'he University, Dundee DDl 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. A. T. Thomas, Department of Geological Sciences, University of Aston, Aston Triangle,
Birmingham B4 7ET. Subscriptions cover one calendar year and are due each January; they should be sent to the Membership
Treasurer. All members who join for 1987 will receive Palaeontology, Volume 30, Parts 1-4. Back numbers still in print
may be ordered from Marston Book Services, P.O. Box 87, Oxford 0X4 ILB, England.

Cover: Pedicle valve of the brachiopod Stroplionella euglypha (Dalman, 1828) from the Wenlock Limestone of Dudley,
West Midlands; x 2. Photography by Harry Taylor of the British Museum (Natural History) Photographic Studio. One of
the specimens illustrated in the Atlas of Invertebrate Macrofossils published by the Association.

EDIACARAN BIOTA OF THE
WERNECKE MOUNTAINS, YUKON, CANADA

by GUY M. NARBONNE and HANS J. HOFMANN

Abstract. Soft-bodied metazoans, trace fossils, and metaphytes occur throughout several hundred metres of
post-tillite, pre-Cambrian strata in the Wernecke Mountains, east-central Yukon Territory. The fossils occur in
three unnamed units of siltstone deposited under shallow shelf and deeper water conditions. The oldest
fossiliferous unit, the ‘Goz siltstone’, contains Cluinuodisciisl and C yclomechisa sp. Siltstone unit 2 has yielded
macrofossils (Beltanellifoj-mis brunsae, Medusinites asteroides, Riigoconitesl sp.), trace fossils (Planolites
montamis), and carbonaceous compressions (Vendotaenidl sp.). Siltstone unit 1, the youngest of the three
siltstones, exhibits abundant soft-bodied macrofossils (Beltanella gilesi, Beltcmelliformis brunsae, Charniodiscus
cf. arboreits. Cyclomedusa plana, C. sp., Ediacaria flinders!, Kullingia? sp., Medusinites asteroides, Nadalina
yukonenis gen. et sp. nov., Spriggia annulata, S. wadeae, Tirasiana sp.) and trace fossils (Gordia marina,
Neonereitesl sp., Planolites montanus, and a backfilled burrow). Overlying Proterozoic carbonates contain only
simple trace fossils.

The Wernecke assemblage is similar to other occurrences of the Ediacaran fauna, but is most closely
comparable with the Valday assemblage of the Russian Platform and the type assemblage in South Australia.
Similarity of the faunal sequence in the Wernecke Mountains with that present in the type Ediacaran
and Vendian implies that evolutionary stages previously identified within these systems may be globally
significant.

The Ediacaran fauna is a distinctive assemblage of soft-bodied metazoans and simple trace fossils
that occurs above the highest Varangian tillites and below the lowest fossiliferous deposits of the
Cambrian (see Glaessner 1984 and references therein). This fossil assemblage has been reported from
every continent except Antarctica (Glaessner 1984, Fig. 1/8), and is particularly well developed in
Australia, Namibia, Newfoundland, and the Russian Platform. The widespread occurrence of this
distinctive assemblage has supported calls for recognition of a formal, sub-Cambrian geological
period, variously termed the Vendian (Sokolov 1952; Sokolov and Fedonkin 1984), Ediacaran
(Jenkins 1981), Ediacarian (Cloud and Glaessner 1982), or Sinian (Grabau 1922; Xing 1984). While all
have ‘the base of the Cambrian’ as their common upper limit, none of these proposed formal units has
the same lower limit. A discussion of the relative merits of each is beyond the scope of the present
paper. However, for the purposes of this paper we prefer to use Ediacaran for the shortest interval with
soft-bodied fossil assemblages, because the first diverse assemblage to be described and used for
comparison derives from the Ediacara Hills region in South Australia. The faunally characterized
Ediacaran is encompassed chronologically by the longer Vendian interval (and the Vendian by the
still longer Sinian).

Elements of the Ediacaran fauna were first discovered in the Wernecke Mountains (text-fig. 1)
during a reconnaissance geologic study by Fritz et ai (1983). Their collection included the trace fossils
Gordia sp. and Palaeophycus tubularis Hall (Fritz et al. 1983; Nowlan et al. 1985), and the macrofossils
Cyclomedusa davidil Sprigg and Beltanelliformis brunsae Menner (Hofmann et al. 1983); Hofmann
(1984) subsequently described microfossils from these samples. During the summer of 1984, the
present authors carried out more detailed geological studies (Narbonne et al. 1985), collecting
numerous fossil specimens from three unnamed formations. These specimens help to elucidate the
taxonomy, palaeoecology, and palaeogeography of early metazoans, and also aid in the regional and
global correlation of the fossil-bearing strata.

jPalaeontology, Vol. 30, Part 4, 1987, pp. 647-676, pis. 73-75.|

© The Palaeontological Association

648

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 1. Index map showing location of sections studied in the eastern Wernecke Mountains.

GEOLOGICAL SETTING AND AGE

The Wernecke Mountains lie on the southern edge of the Yukon Stable Block (Lenz 1972), a site of
predominantly shallow-water sedimentation throughout most of the late Proterozoic and early
Palaeozoic. To the south, the strata pass into deeper-water shales and turbiditic conglomerates
(Gordey 1980; Cecile 1982).

The strata considered in this study occur in equivalents to the upper part of the Windermere
Supergroup, a predominantly clastic succession that can be traced throughout the Canadian
Cordillera (Eisbacher 1981). The base of the Windermere Supergroup is younger than 770 Ma (Park
and Aitken 1986) and probably younger than 730 Ma (Evenchick et al. 1984); its top is located at the
base of the lowest Cambrian unit (Eisbacher 1981). Ediacaran macrofossils occur in the upper part of
the Windermere Supergroup in southern British Columbia (Hofmann et al. 1985), western Northwest
Territories (Hofmann 1981) and eastern Yukon Territory (Hofmann et al. 1983; this study). The
Windermere Supergroup in the study area is more than 2 km thick, and exhibits glacial and
glaciomarine deposits possibly equivalent to the Sturtian and Marinoan glaciations in its lower half
(Eisbacher 1985). The Windermere Supergroup is disconformably overlain by the Vampire
Formation (text-fig. 2), which contains, in its basal beds, small shelly fossils including Anabahtes
trisulcatus and Protohertzina anaharica\ Nowlan et al. (1985) regarded this fauna as correlative with
the Anaharites-Circotheca-Protohertzina Assemblage Zone of south-western China and with the
Nemakit-Daldyn Horizon of Siberia, both of which would be placed immediately below the base of
the Cambrian if the proposed base in Yunnan, China is accepted (Cowie 1985). Trace fossils in the
same beds include two genera of arthropod burrows (Rusophycus and Cruziana) generally regarded as
Cambrian or younger in age (Nowlan et al. 1985; Narbonne et al. 1987). The Vampire Formation is
eonformably overlain by the Sekwi Formation, a Lower Cambrian carbonate unit containing
trilobites of the Fallotaspis and Nevadella zones near its base (Fritz 1978). Thus, the units
bearing the fauna reported here occur stratigraphically between late Proterozoic glacial deposits and
fossiliferous Lower Cambrian strata, and can therefore be regarded as Ediacaran in age.

Ediacaran strata in the study area comprise a succession of alternating carbonate and fine
siliciclastic units, with soft-bodied macrofossils recognized only in the siliciclastic units (Narbonne
et al. 1985). Due to the reconnaissance level of geological study, formal stratigraphic names have not
yet been proposed for most of these units nor can the palaeoecology be discussed in detail. This paper

NARBONNE AND HOFMANN: EDIACARAN BIOTA FROM CANADA

649

A B ^ ^

TEXT-FIG. 2. Stratigraphic sections with collection localities (Geological Survey of Canada [GSC] numbers).
Sections A, D from Narbonne et al. (1985), Section B from Aitken (1984), Section C from Fritz et al.
(1983). Risky Formation (Aitken, in press) = map unit 11 of Blusson (1971).

will use the names suggested by Fritz et al. (1983), Fritz (1984), Aitken (1984), Narbonne et al. (1985),
and Aitken (in press).

The oldest unit containing macrofossils is the ‘Goz siltstone’, which crops out on Goz Creek at
locality ‘E’ (text-fig. 1). The ‘Goz siltstone’ occurs as a small, isolated fault block; it cannot be directly
correlated with any unit in sections A-D, but is lithologically similar to Siltstone unit 3 at locality ‘D’
and the Sheepbed Formation of the western Northwest Territories (Narbonne et al. 1985). The ‘Goz
siltstone’ consists predominantly of thin- to medium-bedded siltstone. Slump and load structures
are common; ripple-marks are very rare. Granule-filled channels averaging 0 5 m deep occur
sporadically. Most channels are filled with clast-supported, normally graded material, but some
contain matrix-supported, massive fill. These features, combined with the apparent absence of
structures typical of shallow-water conditions, suggest that deposition occurred on a slope in
a deeper-water setting. Macrofossils occur on bedding surfaces in the siltstone.

Siltstone unit 2 at locality ‘D’ (text-figs. 1 and 2) consists predominantly of recessive-weathering
shale and siltstone with sporadic interbeds of quartzose sandstone and minor dolostone. Individual
sandstone beds are very thin and graded, and most likely represent storm deposits. The upper part of
the unit has several laterally discontinuous, coarsening-upwards sandstone packages up to 8 m thick;
lower parts of the cycles are characterized by planar-tabular and hummocky cross-stratification, and
the upper parts of cycles exhibit sporadic desiccation cracks. Deposition probably occurred on a

650

PALAEONTOLOGY, VOLUME 30

wave- or storm-dominated shelf, with the coarsening-upwards sandstones representing sand bars.
Carbonaceous remains of metaphytes occur in dark-grey shale in the lower part of the unit (GSC loc.
1 0 1 540), whereas metazoans and trace fossils occur as positive features on the soles of very thin to thin,
storm-deposited sandstone beds near the top of the unit (text-fig. 2).

Siltstone unit 1 crops out at localities ‘A’ to ‘D’ (text-figs. 1 and 2). It is lithologically similar to
Siltstone unit 2, differing mainly in that the thick sandstone beds are more continuous laterally.
Synaeresis cracks are abundant, but desiccation cracks were not observed. Deposition probably
occurred on a wave- or storm-dominated shelf. Macrofossils are found sporadically throughout the
lower two-thirds of the unit; trace fossils occur rarely in the lower half of the unit but are common
throughout the upper half. Macrofossils and trace fossils are seen chiefly as positive features on the
soles of very thin to thin, storm-deposited sandstone beds; only very rarely are they preserved as
negative features on the tops of these beds. No slabs exhibiting both macrofossils and trace fossils were
observed.

These occurrences suggest that the Ediacaran organisms preserved in Siltstone units 1 and 2 lived in
a shallow sublittoral environment, whereas those present in the ‘Goz siltstone’ lived in a deeper-water
environment. This is consistent with other reports of the Ediacaran fauna, which are from both
shallow shelf (e.g. Goldring and Curnow 1967; Jenkins et al. 1983) and deeper slope (e.g. Anderson and
Conway Morris 1982; Gibson et al. 1984) settings.

BIOSTRATIGRAPHY

As now known, the Ediacaran macrobiota of the Wernecke Mountains comprises an assemblage of
14 species of metazoan fossils (one of which is new), 1 metaphyte, and 5 trace fossils. In addition,
3 dubiofossils are reported. The greatest taxonomic diversity is exhibited by soft-bodied discoidal
structures: 1 1 of the 14 species can be broadly referred to 9 ‘medusoid’ genera; one species is a pennate
coelenterate, and one is a possible organ of attachment to the substrate. The numerically dominant
soft-bodied organism in the assemblage is a gregarious species of compressed globular structures
assigned to BeltaneUiformis brunsae. In addition, 5 species of ichnofossils characterize the upper part
of the sequence studied. Of these ichnofossils, only Planolites montamis is common.

The Ediacaran fauna of the Wernecke Mountains is considerably more diverse than that reported
from map-unit lOB of the Mackenzie Mountains of the western Northwest Territories by Hofmann
(1981). However, the presence of B. brunsae and Gordia marina in both areas is consistent with
lithostratigraphic evidence (Fritz et al. 1983; Aitken, in press) that the strata are equivalent.

The Wernecke assemblage is broadly similar to Ediacaran assemblages reported from other areas,
particularly Australia and Eurasia (text-fig. 3). Similarity of the Wernecke assemblage to Ediacaran/
Vendian assemblages from Australia, the Russian Platform, Siberia, China, England, Newfoundland,
and several other localities supports recent calls (e.g. Jenkins 1981; Cloud and Glaessner 1982;
Sokolov and Fedonkin 1984) for recognition of a formal geological period based on this fauna.

Although many authors believe that several biostratigraphic zones can be recognized within the
Ediacaran/Vendian, intercontinental correlation of these macrofossil zones has proved difficult. The
Vendian of the Russian Platform can be divided into four zones which Sokolov and Fedonkin (1984)
considered to be stages. The global applicability of these ‘zones’ or ‘stages’ is currently uncertain (see
Jenkins 1 98 1 , p. 181 and references therein). Nevertheless, there would appear to be a close correlation
between the faunal sequences of the Wernecke Mountains and the Russian Platform (text-fig. 4).

The basal, Volhyn ‘Series’ contains deposits of the Varangian glaciation, but apparently
lacks macrofossils and trace fossils (Sokolov and Fedonkin 1984). This is most likely equivalent to
strata below the fossiliferous units discussed in this paper, which contain glacial deposits and
apparently lack macrofossils and trace fossils (Eisbacher 1981). Intercontinental correlation of glacial
deposits in the Windermere Supergroup has been discussed by Eisbacher (1985), who recognized
equivalents of the Sturtian and Marinoan (= Varangian) glaciations in these strata.

The overlying Redkino ‘Series’, represented by the Mogilev, Yaryshev, and Nagoriany Formations
in the Ukraine and by the Pletevev and Ust-Pinega Formations in the Baltic region, contains a diverse

NARBONNE AND HOFMANN: EDIACARAN BIOTA FROM CANADA

651

SYNOPSIS OF WERNECKE MOUNTAINS MACROBIOTA

OCCURRENCES

"(5

TAXA

UNITS

Map

unit

11

Siltstone unit 1

Siltstone
unit 2

Goz

siltstone

Vendian, Eurasic

Ediacaran. Austr

SECTIONS

D

A

B

c

D

E

GSC

localities

99132

99104

o

o o

101533

101532

101531

101537

99042

101536

101535

99039

101534

99036

101539

101538

99099

99098

101543

99095

101542

101541

101540

101546

101545

101544

Phylum COELENTERATA

Beltanella gilesi Spngg, 1947
^ Charniodiscus cf. C. arboreus (Gl. in Gl. & D., 1959)
(£) C. ? sp.

Cyclomedusa plana Gloessner & Wode, 1966

Ediacaria flindersi Spngg, 1947
Kullingia? sp.

® Medusinites asteroides (Spngg) Gl & W., 1966
(^) Nadalina yukonensis gen. et sp. nov
^ Rugoconites? sp

Spnggia annulata (Spngg) Southcotl, 1958
S- wadeae Sun, 1986
Tirasiana sp
Phylum uncertoin

® Beltanelliformis brunsae Menner, 1974
ICHNOFOSSILS

j»0 Gordia marina Emmons, 1844
•••• Neonereites? sp.

Palaeophycus tubularis Holl, 1847
— Planolites montanus Richter, 1937
® Bockfilled burrow

Group VENDOTAENIDES Gnilovskoyo, 197!

Vendotaenia? sp.

DUBIOFOSSILS
0 Dubiofossil A

© Dubiofossil B

& Dubiofossil C

HJH 86

0

o

0

o

o

o

o

o

O O 0

0

o o

o

o

0

o

0

o

0 o

O

□

O

□

o

□

o

□

□

o

O

□

□

0

□

o

o

o

□

o

o

□

0 o

o o

O O O 0

O 0 o o

o

□

0

o

0

O O 0
o

o o

o

O 0

o

o

□

o

o

0

□

o

o

o

TEXT-FIG. 3. Synopsis of Ediacaran macrobiota of Wernecke Mountains. Specimens from the present study
are shown by circles; triangles identify specimens described by Nowlan et al. (1985) from map-unit 11
( = Risky Formation). For comparison, same (rhombs) or similar (squares) forms from Australia and Eurasia

are also indicated.

assemblage of soft-bodied metazoans along with vendotaenid algae and simple trace fossils (Palij
et al. 1979; Sokolov and Fedonkin 1984). This is similar to the ‘Goz siltstone’, Siltstone unit 2, and
the lower two-thirds of Siltstone unit 1, which contain a similar assemblage. Indeed, most of the
specific forms found in these three units have also been described from the type Redkino ‘Series’
(text-fig. 3).

The overlying Kotlin has yielded vendotaenid algae and simple trace fossils, but only sparse
‘medusoids’ (Palij et al. 1979; Sokolov and Fedonkin 1984). This compares with the upper third
of Siltstone unit 1 and all of the Risky Formation (= map-unit 1 1 of Blusson 1971; see Aitken, in

652

PALAEONTOLOGY, VOLUME 30

STRAT?GRAPHY

UNIT

WERNECKE MACROBIOTA

1 VENDIAN
1 OR

, LOWER

1 CAMBRIAN

lower

Baltic

‘Series’

Basal

Vampire

Fm.

Small shelly fossils - Anabarites, Protohertzina
Arthropod trace fossils - Rusophycus , Cruziana

VENDIAN 1

Kotlin

‘Series’

Risky

Fm.

Simple trace fossils - Palaeophycus

Siltstone
unit 1

Simple trace fossils - Planolites, Gordia,Neonereites?
Megafossils - indef. medusoids

'(0

a>

Simple trace fossils -Gordia

Megafossils - Beltanella, Beltanelliformis, Charniodiscus,
Cyclomedusa, Ediacaria, Kullingia?, Medusinites,
Nadalina, Rugoconites?, Spriggia, Tirasiana

Carbonate

0)

unit 1

No biota

/A

QC

Simple trace fossils - Planolites

HI

o

c

Megafossils - Beltanelliformis, Medusinites,

Q.

■o

Siltstone

Rugoconites?

3

e

oc

unit 2

Vendotaenid algae - Vendotaenia?

9 9

TEXT-FIG. 4. Proposed correlation of the upper part of the Windermere Supergroup with the Vendian of the
Russian Platform. Similar correlations can be made with the Ediacaran of Australia.

press), which exhibit simple trace fossils and sparse, predominantly indeterminate ‘medusoids’. Both
in the Russian Platform and in the Wernecke Mountains, this zone contains relatively few diagnostic
taxa, and is recognized primarily by its position between the more distinctive fossil assemblages of the
overlying and underlying strata.

Overlying the Kotlin ‘Series’ on the Russian Platform is the Baltic ‘Series’. This ‘series’ traditionally
has been regarded as Lower Cambrian (e.g. Sokolov 1973; Palij et al. 1979), but recently some authors
(e.g. Sokolov and Fedonkin 1984) have included its basal ‘horizon’, the Rovno, in the upper Vendian.
The small shelly fossil assemblage of the basal Vampire Formation is broadly similar to that of the
Nemakit-Daldyn (Nowlan et al. 1985), which Sokolov and Fedonkin (1984) regarded as the Siberian
equivalent of the Rovno ‘Horizon’. However, the abundance of arthropod traces {Rusophyciis,
Cruziana) in the basal Vampire suggests that correlation with the Lontova ‘Horizon’, which
immediately overlies the Rovno on the Russian Platform, is more probable. The Lontova ‘Horizon’ is
also part of the lower Baltic ‘Series’, and is generally regarded as Lower Cambrian (e.g. Fedonkin
1985h).

Similar correlations can be made with the Ediacaran of South Australia (Jenkins 1981). Marinoan
( = Varangian) glaciomarine deposits occur at the top of the Umberatana Group (Coats 1981), and

NARBONNE AND HOFMANN: EDIACARAN BIOTA FROM CANADA

653

have been correlated with probable glacial deposits in the middle part of the Windermere Supergroup
by Eisbacher (1985). Problematic remains of Ediacaran aspect occur sporadically throughout the
lower part of the overlying Wilpena Group (Cloud and Glaessner 1982, fig. 2f; Dyson 1985), and a
diverse Ediacaran assemblage is present in the upper part of the Wilpena Group (Pound Subgroup).
The macrofossil assemblage of the Pound Subgroup is closely similar to that of the Wernecke
Mountains (text-fig. 3) and the Redkino ‘Series’ of the Russian Platform (Jenkins 1981; Cloud and
Glaessner 1982; Sokolov and Fedonkin 1984), thereby implying that the strata are equivalent. The
Pound Subgroup is disconformably overlain by the Uratanna Formation, which contains Cambrian
trace fossils near its top (Daily 1972, 1973).

Similarity of the faunal and floral sequence between the Wernecke Mountains, the Russian
Platform, and the Adelaide Geosyncline suggests that intercontinental correlation of stages and even
zones may be possible within the Ediacaran/Vendian.

SYSTEMATIC PALAEONTOLOGY

The taxonomic affinities of the Ediacaran fauna are currently controversial. Until recently, most
authors followed Sprigg (1949), Richter (1955), and Glaessner (1979, 1984) in regarding most discoid
and pennate forms as the impressions of fossil coelenterates. Seilacher (1984) and McMenamin ( 1 986)
have questioned this interpretation, suggesting that some of these fossils may represent radial or
circular burrows, internal sandy skeletons, or representatives of an extinct phylum. A full evaluation
of the affinities of the Ediacaran fauna is beyond the scope of this paper and we have tentatively
followed the Treatise (Glaessner 1979) in referring most of our taxa to the Coelenterata (text-fig. 3).

In the following section, synonymies include only putative pre-Cambrian occurrences. An
alphabetical list for each major group of the Ediacaran biota of the Wernecke Mountains, and their
stratigraphic and geographic distributions, is shown in text-fig. 3. Terminology regarding preserva-
tion follows Frey (1973, table 5).

All figured specimens are in the repository of the National Type Collection of Invertebrate and
Plant Fossils (GSC) in Ottawa.

Phylum COELENTERATA
Genus beltanella Sprigg, 1947

Type species. Beltanella gilesi Sprigg, 1947.

Beltanella gilesi
Plate 73, fig. 6

For synonymy up to 1966, see Glaessner and Wade (1966).

1972fl Planomedusites grandis Sokolov, pi. 2, fig. 1.

1973 Planomedusites grandis Sokolov, p. 210, fig. 3/1.

1979 Planomedusites grandis Glaessner, p. A96, fig. 10/3.

Description. Smooth disc with narrow raised rim; two specimens preserved in convex hyporelief, respectively 36
and 46 mm in diameter and 1 -5 mm and 1-7 mm in maximum relief; smaller specimen with central protruberance
3 mm across, partly enclosed by low concentric ridge 9 mm in diameter; larger specimen exfoliated and
incomplete; both specimens surrounded by a flange about 2-5 mm wide.

Remarks. Our structures strongly resemble the holotype of B. gilesi Sprigg, which, however, has
uniformly sized and equidistantly spaced circular markings at two-thirds the distance to the margin,
as well as radial grooves. Glaessner and Wade (1966) hypothesized that these features might be
accidental, a view supported by undescribed new specimens from Australia (R. J. F. Jenkins, pers.
comm. 1986). Beltanella is similar in size and morphology to Ediacaria Sprigg (Glaessner and Wade
1966; Glaessner 1979), but can be distinguished by the presence of an outer flange.

654

PALAEONTOLOGY, VOLUME 30

Genus charniodiscus Ford, 1958
Type species. Charniodiscus concentricus Ford, 1958.

Charniodiscus cf. arhoreus (Glaessner, 1959)

Text-tig. 5d

Description. Two incomplete specimens preserved as large, ovate convex hyporcliefs. Larger specimen 98 x
72 mm, with at least 5-6 mm of relief at margin; roughly bilaterally symmetrical impression of flattened, bag-like

TEXT-FIG. 5. Macrofossils from the ‘Goz siltstone’ and siltstone unit 1. u. Cyclomedusa sp., epirelief
GSC loc. 101546. GSC 83016, x 1. 6, C/iur/i/odiscus? sp., epirelief GSC loc. 101 545. GSC 83017, x l.c,
Charniodiscus cf arboreus Glaessner. Two incomplete specimens (middle, GSC 83019; upper right,
GSC 83020). Hyporelief GSC loc. 101529, x 1.

NARBONNE AND HOFMANN: EDIACARAN BIOTA FROM CANADA

655

body with parallel-sided, regular segments between 7 and 10 mm wide, extending outwards in apparently
alternating series, at angles of about 60° with a poorly defined axial region, opening towards the more pointed
end of the specimen (distal apex). Segments separated by straight to very slightly curved 1 0- 1 -5 mm wide furrows
and associated one or two juxtaposed low, parallel, levee-like ridges; suggestion of short, oblique secondary
furrows directed inward at angles near 40° from main furrows in wider (lower) part of specimen. Apical region
subtriangular, concavo-convex, marked by fine wrinkles parallel with concave side of apical region; region
15 mm across immediately below triangular apex marked by delicate concentric wrinkles. Surface of segmented
portion with randomly spaced (and probably accidental) subcircular tubercles 0-5-7 0 mm across and up to
10 mm of relief. Transverse section across lower end showing two irregular narrow vertical zones marked by
mineralogical and textural contrast.

Smaller specimen an ovate hyporelief 50x26 mm, 10 mm high; surface bearing transverse convex
segments of unequal width, from 6-12 mm wide, separated by pronounced furrows; surface marked by fine,
oblique wrinkles continuous across furrows; transverse section showing recurved nature of impression and
sedimentary lamination paralleling lower surface of hyporelief.

Remarks. The structures are indicative of the organization and highly flexible nature of genera such as
Charniodiscus and Inkrylovia. The larger specimen somewhat resembles the Rangea arhorea
illustrated by Glaessner in Glaessner and Daily (1959, pi. 43, figs. 2 and 3; pi. 44, fig. 3; later assigned
to Arhorea arhorea [Glaessner and Wade 1966, p. 619], and finally placed in synonymy with
Charniodiscus [Glaessner 1979, p. A99]), and may represent a partially decomposed and over-folded
specimen of this taxon. The smaller specimen has few distinguishing characteristics; its preservation is
similar to a specimen referred to Inkrylovia from the correlative map-unit lOB in the Mackenzie
Mountains 250 km east-south-east (Hofmann 1981, fig. 3a, b).

Charniodiscus? sp.

Text-figs. 56, c and 6

Description. Nineteen specimens occurring as convex epireliefs and concave hyporeliefs at the interface between
two very thin beds of siltstone. Bipartite structure, consisting of a rough-textured central disc surrounded by a
smooth, flat outer ring with slight ( < 0-3 mm) relief. External diameter 7-5-2 10 mm (mean = 14-3 mm); diameter

20n

E

E

c 15-
cc

LU

I—

N=19

<

Q

O

CD

Q

QC

LU

10-

5-

• •

• •

0 I 1 1 1 1

0 5 10 15 20 25

EXTERNAL DIAMETER in mm

TEXT-FIG. 6. Size variations in Charniodiscus? sp. from the ‘Goz Siltstone’.

656

PALAEONTOLOGY, VOLUME 30

of the inner disc 4-15 mm (mean = 91 mm); diameter of the inner disc 50-90% (average 68%) of the entire
fossil (text-fig. 6). Outer margin of the fossil circular to subcircular, with minor indentations (text-fig. 5c). Inner
disc centrally to slightly eccentrically located. Faint impression of a rod-like stem 2-5-6 0 mm wide attached to
the inner disc. Specimens preserved as cleavage reliefs, with no structure visible below or above the bedding
surface.

Remarks. The apparent absence of vertical tubes below or above the specimens implies that the
structures did not form as a result of water or gas escape. Small pyrite concretions approximately one
metre lower stratigraphically exhibit a central core of pyrite and an outer ring of darker (reduced)
sediment; these can readily be distinguished from Charniodiscusl sp. by the fact that the outer ring has
a different mineralogy and colour, but no difference in relief. The faint impression of a stem further
serves to distinguish Charniodiscusl sp. from inorganic sedimentary structures and from medusiform
genera with bipartite organization (e.g. Medusinites Glaessner and Wade, 1966; Nitnbia Fedonkin,
1980).

Ford (1958) originally described Charniodiscus solely on the basis of its bipartite disc-like structure,
but later (1963) figured the entire specimen with the bipartite disc attached by a stem to a frond-like
structure. Subsequent workers (e.g. Jenkins and Gehling 1978; Glaessner 1979) have regarded the
frond as the diagnostic portion of the organism. The Wernecke specimens exhibit an incomplete stem
but lack the attached frond, and hence can only tentatively be referred to Charniodiscus.

Genus cyclomedusa Sprigg, 1947
Type species. Cyclomediisa davidi Sprigg, 1947.

C yclomedusa plana Glaessner and Wade, 1966

Plate 73, fig. 3

1966 Cyclomediisa plana Glaessner and Wade, p. 607, pi. 98, figs. 1-3.

1968 Cyclomediisa plana Zaika-Novatskiy and Palij, pp. 133-134, fig. 1 (English translation, pp. 269-
270, fig. 1).

1973 Cyclomedusa plana Sokolov, p. 209, fig. 2/1.

1979 Cyclomediisa plana Glaessner, p. A94, fig. 9/2«.

? 1981 C yclomedusa cf. plana. Fedonkin, pp. 58-59, pi. 2, fig. 1; pi. 3, fig. 4.

? 1983 Cyclomedusa cf. plana. Fedonkin in Velikanov et ai, pi. 28, fig. 7.

? 1985u Cyclomediisa cf. plana Fedonkin, pp. 71-72, pi. 2, figs. 3 and 5; pi. 5, fig. 7.

Description. One partially preserved circular impression (hyporelief) with bipartite organization; inner disc,
24 mm in diameter, with four coarse, slightly eccentric folds; small, bud-like, concentric pattern superimposed

EXPLANATION OF PLATE 73

Fig. 1. C yclomedusa sp. Sprigg. GSC loc. 101537. GSC 83021, x 1.

Fig. 2. Nadalina yukonensis gen. et sp. nov., epirelief Holotype. GSC loc. 101535. GSC 83022, x 1.

Fig. 3. C. plana Glaessner and Wade, hyporelief. GSC loc. 101535. GSC 83023, x 1.

Fig. 4. Sprigyia wadeae Sun, hyporelief. GSC loc. 101534. GSC 83024, x 1.

Fig. 5. Kiillinyia? sp., hyporelief. GSC loc. 101534. GSC 83025, x 1.

Fig. 6. Beltanella gilesi Sprigg, hyporelief GSC loc. 101531. GSC 83026, x 1.

Figs. 7-9. Medusinites asteroides (Sprigg), hyporelief 7, GSC loc. 99095. GSC 83027, x 1; 8, GSC loc. 99042. GSC
83028, X 1; 9, GSC loc. 101536. GSC 83029, x 1.

Fig. 10. S. annulata (Sprigg). Specimen on shale pebble. GSC loc. 101530. GSC 83030, x 1.

Fig. 11. Rugoconitesl sp., next to small, unidentified circular structure. Hyporelief GSC loc. 101541. GSC 83031,
X 1.

PLATE 73

NARBONNE and HOFMANN, Ediacaran biota

658

PALAEONTOLOGY, VOLUME 30

on third fold. Outer zone smooth, with 1-2 mm wide annulus of low relief at margin. Diameter of fossil
52 mm; total relief 0-7 mm, maximum in inner disc.

Remarks. C. plana is distinguished by its relatively small, coarsely corrugated central disc and its large,
smooth outer ring without rugosities. Unless the small concentric structure on the third ring of the
central disc represents the impression of a separate, small specimen of Cyclomedusa, this structure
could be interpreted as a bud, much like one of the Australian paratypes in which the central cone
appears to be twinned. Otherwise, the fossil most closely resembling our specimen is one from the
Yaryshev Formation of the Ukraine (Zaika-Novatskiy and Palij 1968, fig. 1; see also re-illustrations in
Sokolov 1973, fig. 2/1, and Glaessner 1979, fig. 9/2a).

Cyclomedusa sp.

Plate 73, fig. 1; text-fig. 5a

1983 Cyclomedusa davidi? Hofmann e( ai, p. 455, fig. 2a, b.

Description. Circular to elliptical structures with numerous coarse, concentric rugae, slightly eccentric in some
specimens; preserved as epi- and hyporeliefs; distinct small central tubercle (hyporelief ) or pit (epirelief ) at or near
centre; indistinct radial markings in some specimens (e.g. Hofmann et al. 1983, fig. 2b); diameters 17-76 mm
(N = 23, mean 38 6 mm), 0-5-3 0 mm in relief.

Remarks. The collection constitutes a heterogeneous lot, with specimens diflfering widely in quality of
preservation. It includes large, flatfish specimens as well as small discs with proportionately higher
relief. All are characterized by moderately coarse, concentric rugae. The fifteen specimens from the
‘Goz siltstone’ referred to C. davidil by Hofmann et al. (1983, p. 455) have now been determined to be
epireliefs. The collection has been supplemented by several additional specimens, which include one
showing a more eccentric position of the central depression, somewhat like the C. serebrina from the
Ukraine (Palij et al. 1979, pi. 48, fig. 4); however, the poorly preserved marginal wrinkles do not
overlap as in the Ukrainian specimen.

Cyclomedusa is the most cosmopolitan of the Ediacaran macrofossils (see also Wade 1972, p. 205),
and most specimens have been referred to C. davidi Sprigg. Sun (1986) has reviewed the taxonomy of
Cyclomedusa, restricting C. davidi to forms with numerous fine radial grooves on the oral surface. The
scarcity of radial grooves in the specimens from the Wernecke Mountains may be taxonomically
significant, or may simply reflect preservation of the aboral surface of C. davidi.

Genus ediacaria Sprigg, 1947
Type species. Ediacaria flindersi Sprigg, 1947.

Ediacaria flindersi Sprigg, 1947
Text-fig. 7

For synonymy up to 1966, see Glaessner and Wade 1966.

1978 Tirasiana disciformis Fedonkin, fig. 3, no. 6.

1979 Ediacaria flindersi Glaessner, p. A95, fig. 9/1.

1981 Tirasiana disciformis Fedonkin, p. 57, pi. 2, fig. 4.

1985a Ediacaria flindersi Fedonkin, pp. 74-75, pi. 1, figs. 2 and 5; pi. 2, fig. 4.

Description. Very large circular structures with tripartite organization, composed of three superimposed
concentric discs, preserved as convex hyporelief. Two specimens; complete specimen (text-fig. 7) comprising
inner disc 36 mm in diameter, with about 2 mm of relief, superimposed on second disc 100 mm across and
about 3 mm in relief, attached to third disc 165 mm across, with about 5-5 mm of relief Surface of innermost
disc with one faint concentric circular furrow midway between its centre and its periphery; second disc with at
least five distinct, narrow, unevenly spaced concentric ridges which are developed most strongly in its peripheral

NARBONNE AND HOFMANN: EDIACARAN BIOTA FROM CANADA

659

TEXT-FIG. 7. Ediacaria fUndersi Sprigg, hyporelief. GSC loc. 101532. GSC 83014, x 0 5.

portion; largest disc with faint concentric markings midway between inner and outer limits. Very faint radial
markings also locally visible. Middle and outer discs with indentations that coincide with an arcuate groove
which has large radius and is tangent to central disc. Part of specimen traversed by TO mm-wide shrinkage
crack filling. Second, incomplete specimen (GSC 83053) exhibiting inner disc 74 mm in diameter, with about
1 mm of relief, superimposed on second disc 1 10 mm in diameter with about 1 mm of relief attached to third
disc 222 mm in diameter with about 2 mm of relief: middle disc with four concentric ridges near its
periphery.

Remarks. E. flinder si is the largest ‘medusoid’ known from the Ediacara assemblage (Glaessner and
Wade 1966), and the two specimens from the Wernecke Mountains are at the upper end of its known
size range. Sprigg’s ( 1 947, 1 949) original specimens are marked by strong radial grooves, but these are
not present on the ex-umbrellar surface (Glaessner and Wade 1966; Sun 1986, p. 336). The specimens
from the Wernecke Mountains are most similar to a large specimen from the Vendian of the White Sea
coast, originally referred to Tirasiana disciformis Palij, 1976 by Fedonkin (1978, fig. 3, no. 6; 1981, pi. 2,
fig. 4) but now referred to E. flinder si (Fedonkin 1985u, pi. 1, fig. 5).

Genus kullingia Glaessner in Foyn and Glaessner, 1979
Kullingial sp.

Plate 73, fig. 5

? 1979 Kullingia concentrica Glaessner in Foyn and Glaessner, pp. 39-40, fig. 8u.

? 1985 Kullingia concentrica Gureev, pp. 99-100, pi. 39, figs. 1-4.

Description. Two specimens from GSC locality 101543 preserved as convex discoidal hyporeliefs; figured
specimen 54 mm in diameter, with about 1 mm of relief; surface smooth, provided with closely spaced, faint,
regular concentric wrinkles that are most prominent near periphery; central disc 7 mm across barely noticeable;
no radial pattern distinguishable; second specimen juxtaposed to first, partially preserved, about 60 mm across,
with 1 mm of relief.

660

PALAEONTOLOGY, VOLUME 30

Remarks. The structures have some of the characteristics of KLiUiugia, but folds in the central portion
of the discs are poorly preserved making the specimens resemble those of Beltanelliformis. Their larger
size and more regular concentric ridges and furrows set them apart.

Genus medusinites Glaessner and Wade, 1966
Type species. Medusinites asteroides (Sprigg) Glaessner and Wade, 1966.

Medusinites asteroides (Sprigg), emend. Glaessner and Wade, 1966
Plate 73, figs. 7-9

1949 Medusina mawsoni Sprigg, p. 89, pi. 13, fig. 4; text-fig. 7b.

1949 Medusina asteroides Sprigg, p. 90, pi. 13, text-fig. 7c.

1956 Protolyella asteroides Harrington and Moore, p. F155, fig. 127/1.

1956 Protolyella mawsoni Harrington and Moore, p. F155, fig. 127/2.

1966 Medusinites asteroides Glaessner and Wade, pp. 605-607, pi. 94, figs. 1-5.

1968 Medusinites asteroides Wade, pp. 259, 260, figs. 22 and 24.

? 1973 Medusinites patellaris Sokolov, p. 210, fig. 3/2.

1979 Medusinites asteroides Glaessner, p. A94, fig. 10/1.

? 1980 Paliella patelliformis Fedonkin, p. 10, pi. 2, figs. 1-3.

? 1981 Paliella patelliformis Fedonkin, pp. 62-63, pi. 31, figs. 2 and 3; pi. 32, figs. 1 and 2.

? 1983 Paliella patelliformis Fedonkin, pi. 28, figs. 1, 2, 4-6.

1983 Medusinites asteroides Fedonkin, pi. 28, fig. 10.

? 1983 Medusinites sp. Fedonkin, pi. 34, fig. 1.

1985a Medusinites asteroides Fedonkin, pi. 8, fig. 2.

? 1985a Medusinites sp. Fedonkin, pi. 8, fig. 3.

? 1985a Paliella patelliformis Fedonkin, pp. 73-74, pi. 3, fig. 9; pi. 10, fig. 5.

Description. Subcircular convex hyporeliefs, composed of smooth central disc, separated by a subcircular groove
from a broad, smooth outer ring, itself surrounded by a groove; disc one-third to one-half of diameter of whole
structure. Outer diameters of three specimens 9 0, 20-5, and 25-6 mm; disc diameter respectively 4-3, 10-9, and
8-3 mm; relief respectively 0-8, L3, and L8 mm.

Remarks. Glaessner and Wade (1966) erected the genus Medusinites to include both Medusina
asteroides Sprigg and, questionably, M. mawsoni Sprigg. Our three specimens, none of which shows
any radial elements, most closely resemble the holotype of M. mawsoni, and two specimens of
Medusinites asteroides illustrated by Glaessner and Wade (1966, pi. 97, figs. 1 and 2). Similar
structures from the USSR were illustrated by Sokolov (1973, p. 210, fig. 3, no. 5) as M. patellaris, and as
M. asteroides by Fedonkin (1983, pi. 28, fig. 10).

The genus Pcdiella (Fedonkin 1980, p. 10) is very close in morphology to Medusinites, and is said to
be distinguishable from it by the presence of radial grooves in the outer zone. However, such grooves,
though not dominant, are present in the type material of Medusinites (e.g. Sprigg 1949, text-fig. 7b, c;
Glaessner and Wade 1966, pi. 97, figs. 3 and 5). Moreover, Fedonkin (1983, pi. 28, figs. 5 and 6)
illustrated under Paliella specimens in which radial elements are not distinct. Paliella may thus be a
junior synonym of Medusinites, or, if the radial pattern were considered characteristic and dominant,
the structures could be referred to Protolyella Torell, in which case Paliella may be a junior synonym
of Protolyella.

Genus nadalina gen. nov.

Type species. Nadalina yukonensis sp. nov.

Diagnosis. Discoidal structure of centimetric size, with large smooth inner disc separated from
surrounding annular field with narrow marginal rim of small relief by a ring of numerous.

NARBONNE AND HOFMANN: EDIACARAN BIOTA FROM CANADA

661

equally spaced millimetric pits (as seen in epirelief); width of annulus about one-half of radius
of inner disc.

Etvmologv. Named for its occurrence in the Nadaleen River map area (National Topographical Series of Canada
Map 106C, 1 : 250,000).

Nadalina yukonensis sp. nov.

Plate 73, hg. 2

Diagnosis. As for genus.

Holotype. GSC 83022.

Etymology. Named for its occurrence in the Yukon Territory of Canada.

Description. One whole specimen, preserved as elliptical epirelief on medium grey, medium-grained sandstone;
62 X 55 mm in size, with about 1 mm of relief. Disc differentiated into 2 zones, an inner flat disc without distinctive
markings, 38 x 33 mm across, and an outer ring about 6-10 mm wide with less smooth surface and a 1 -2 mm wide
raised rim at the outer margin; between the 2 zones a partially preserved ring marked on one side of specimen by
at least 9 pits, 1-3 mm wide and up to 1 mm deep, more or less regularly spaced, 5-9 mm apart. Poorly preserved
partial impression of a second specimen (GSC 8301 5) on the same slab about 75 x 90 mm in size, with outer ring
12 mm wide, marked on inside with at least 7 pits 1-3 mm across, spaced 5-10 mm apart.

Occurrence and type locality. Siltstone unit 1, Section D (GSC locality 101538).

Remarks. The structure is unlike any known to us; we regard it as a new genus and species, and
interpret it as the impression of a medusoid. The illustrated specimen is sufficiently well preserved and
distinct to serve as a basis for a new taxon.

Genus rugoconites Glaessner and Wade, 1966
Type species. Rugoconites enignialicus Glaessner and Wade, 1966.

Rugoconites? sp.

Plate 73, fig. 1 1

Description. Single specimen, poorly preserved as convex hyporelief; subcircular disc 29 x 32 mm in diameter,
with 2-7 mm of maximum relief; outer zone on one side of specimen with short radial furrows spaced about
3 mm apart, apparently emanating from points of bifurcation uniformly located about 5 mm from the
periphery. Possible presence of ring surrounding specimen beyond margin, suggested by a diffuse, irregularly
patterned, 7-9 mm wide zone.

Remarks. The pattern and size of the questionably bifurcating furrows are suggestive of the
morphology of Rugoconites, of which two species have been described. The coarse furrow pattern
would fit R. enigmaticus better than R. tenuirugosus, but our only specimen is so poorly preserved as to
make even the identification of the genus doubtful. The genus has not before been reported from
outside Australia.

Genus spriggia Southcott, 1958
Type species. Madigania annulata Sprigg, 1949.

Spriggia annulata (Sprigg, 1949) Southcott 1958
Plate 73, fig. 10

1949 Madigania annulata Sprigg [partim], pp. 93-94, pi. 16, fig. 1.

1956 Madigania annulata Harrington and Moore, p. F154, fig. 124.

662

PALAEONTOLOGY, VOLUME 30

1958 Spriggia annulata (Sprigg), Southcott, p. 59, fig. 3.

1979 Cyclomedusa davidi Feyn and Glaessner, p. 40, fig. 5c.

? 1981 Cyclomedusa delicata Fedonkin, pp. 59-60, pi. 2, fig. 2.

1984 Spriggia annulata Jenkins, p. 97, pi. 1, fig. 6.

? 1985a Cyclomedusa delicata Fedonkin, pp. 72-73, pi. 1, fig. 4.

1986 Spriggia atmulata Sun, pp. 337-339, fig. 2e, fig. 5.

Description: Single, bipartite disc 28 mm in diameter, preserved in convex relief on a shale clast. Inner disc 1 7 mm
in diameter, slightly convex (3 mm relief), with slightly eccentric papilla surrounded by annular rugae. Outer
flange flat. Both inner disc and outer flange sculpted with numerous, submillimetric, concentric wrinkles.

Remarks. Spriggia has recently been revised by Sun (1986), who discussed its complex nomenclatural
history and its distinction from similar genera. Sun regarded Spriggia as the impression of a fossil
chondrophore.

Spriggia wadeae Sun, 1986
Plate 73, fig. 4

? 1979 Cyclomedusa minuta Fedonkin, in Palij et al., pp. 63-64, pi. 58, fig. 4.

? 1981 Cyclomedusa minuta Fedonkin, p. 59, pi. 4, fig. 2.

1986 Spriggia wadeae Sun, pp. 339-346, figs. 6a-d and 8a-c.

Description. Single disc preserved as convex hyporelief, 20 mm across, 0 7 mm high, with sharp outer margin
and distinctly flat appearance, though sculptured by numerous annular ridges increasing in width outwards to a
maximum of 10 mm wide. Attached to one side of specimen along about one-half of periphery, is an irregular
crescentic marking 1-6 mm high, with maximum width of 10 mm, characterized by poorly defined
submillimetric irregularities; shale matrix from underlying layer covering parts of disc and crescent.

Remarks. This structure is broadly similar to C. minuta from the White Sea coast (see synonymy),
particularly because of the presence of the crescentic projection. However, C. minuta is considerably
smaller and contains fewer annular ridges. The distinction between S. wadeae and S. annuluta has been
discussed by Sun (1986).

Genus tirasiana Palij, 1976
Type species. Tirasiana disciformis Palij, 1976.

Tirasiana sp.

Plate 74, figs. 2 and 4

Description. Small discoidal hyporeliefs 10-18 mm in diameter (N = 4), 0-5- 10 mm high, with tripartite
organization: small central tubercle surrounded by inner disc that extends about halfway to periphery of whole
impression; prominent circular groove separating inner disc from broad outer disc. Outer disc of specimen at
bottom of Plate 74, fig. 4 bearing indistinct radial markings and narrow concentric circular groove; inner disc of
specimen in Plate 74, fig. 2 with at least five circular markings of uniform size equidistant from centre, and
incomplete subsidiary concentric wrinkles.

EXPLANATION OF PLATE 74

Figs. 1, 3, 5-7. Beltanelliformis brunsae Menner. 1, vertical section along top margin of 3. GSC loc. 101534. GSC
83032, X 1. 3, hyporelief, cluster of large specimens with smooth centres. 5, hyporelief; cluster of specimens
with concentric wrinkles in central parts of discs. GSC loc. 99094. GSC 83035, x 1. 6, hyporelief; cluster of
small specimens with high relief cast in underlying shale. GSC loc. 101541. GSC 83036, x 1. 7, epirelief; cluster
of large specimens with low relief on underlying sandstone. GSC loc. 101532. GSC 83037, x 1.

Figs. 2 and 4. Tirasiana sp., hyporeliefs. 2, GSC loc. 99042. GSC 83033, x 1. 4, GSC loc. 101531. GSC 83034, x 1.

PLATE 74

NARBONNE and HOFMANN, Ediacaran biota

664

PALAEONTOLOGY, VOLUME 30

Remarks. The specimens are very similar to a slightly larger specimen of Tirasiana sp. from the
Yaryshev Formation in the Ukraine illustrated by Palij et al. (1979, pi. 49, fig. 7), which also has
circular markings in the middle ring surrounding the central tubercle, like the specimen in Plate 74,
fig. 4. The specimen in Plate 74, fig. 2 resembles Protoniohia Sprigg, 1949 (pi. 9, fig. 1) but lacks
the marginal subcircular structures. Protoniohia was regarded as a concretion by Cloud (1968), and
was treated as a ‘rejected and unrecognizable’ taxon by Glaessner (1979, pp. A112-A113), but we
regard our specimen as organic.

? 1969 ‘minute fossils’ Wade, pi. 69, fig. 7.

1974 Bellanelliformis hrunsae Menner, in Keller et al., p. 132, pi. 1, fig. 10.

1976 Nemiana .simplex Palij, pp. 70-71, pi. 21, fig. 5; pi. 22, figs. 1-3.

1979 Nemiana .simplex Palij et al., p. 64, pi. 49, figs. 1, 5, 6.

1981 Beltanelliformis hrunsae Fedonkin, p. 58, pi. 1, figs. 1-6.

1981 Nemiana simplex Fedonkin, p. 57, pi. 3, figs. 2 and 9.

1981 ISekwia excentrica Hofmann, fig. 4h.

1983 Beltafielliformis hrunsae Hofmann et al., p. 455, fig. Ic.

1985 Beltanelloides simplex (Palij) Gureev, pp. 97-98, pis. 35, 36, 37, figs. 1-4, 7; pi. 38, fig. 1.
1985« Beltanelliformis hrunsae Fedonkin, pp. 70-71, pi. 5, fig. 2.

1985a Nemiana simplex Fedonkin, p. 70, pi. 5, fig. 3.

1985 Nemiana simplex Bekker, pi. 29, fig. 6.

Phylum uncertain

Genus beltanelliformis Menner, in Keller et al. 1974
Type species. Beltanelliformis hrunsae Menner, in Keller et al. 1974.

Beltanelliformis hrunsae Menner, in Keller et al. 1974
Plate 74, figs. 1, 3, 5-7; Plate 75, figs. 1-8; text-figs. 8 and 9

Beltanelliformis

10

SILTSTONE 1
SILTSTONE 2
N = 413

E

E

8 ■

c

6 -

Li.

1 1 1

LU

cr

0

10

20

30

DIAMETER in mm

TEXT-FIG. 8. Size variation of Beltanelliformis hrunsae Menner from the
Wernecke Mountains.

NARBONNE AND HOFMANN: EDIACARAN BIOTA FROM CANADA

665

TEXT-FIG. 9. Vertical sections of Beltanelliformis hrunsae. 1 corresponds to Plate 75, fig. 1. 2 to Plate 75, fig. 2. 3 is
a tracing from a thin section cut across the three aligned specimens at right margin of Plate 75, fig. 3. 4 is a
tracing from a thin section cut parallel to the large specimen illustrated in Plate 75, fig. 4.

Description. Flat to button-shaped, circular to subcircular convex hyporeliefs and concave epireliefs; less
commonly, concave hyporeliefs and full reliefs; 2-2-331 mm across; (mean diameter = 915 mm; s = 5-88 mm;
N = 413); 0-1-8-5 mm in relief (mean relief = F35 mm; s = 1-21 mm; N = 413); specimens in Siltstone unit 1
generally larger than those in Siltstone unit 2 (text-fig. 8). Individuals typically very closely crowded, with
pronounced unimodal size distribution for specimens on individual bedding planes. Specimens with high relief
(diameter/relief < 10) smooth, or provided with one or more curvilinear, irregularly linear, bifurcating, or
star-shaped furrows or folds. Specimens of low relief (diameter/relief > 10) with narrow concentric peripheral
wrinkles or folds, and smooth central field; some vertical sections showing collapsed sediment immediately above
disc (e.g. PI. 75, figs. 1-3; text-figs. 9.1; 9.2, specimen F); all gradations between low- and high-relief specimens
present (text-fig. 8). Vertical sections of complete, high-relief specimens exhibiting lenticular nature, with
semicircular bottom, deformed upper semicircle, and involuted sides; internal sediment fill laminated, graded,
massive, or exhibiting slump structures (text-figs. 9.2, specimen D; 9.3; 9.4). Rare concave hyporeliefs
hemispheroidal with central circular marking (e.g. PI. 75, fig. 5) to relatively flat and wrinkled. Concave epireliefs
shallow and relatively smooth, or with irregular, partly concentric wrinkles. Specimens preserved in full relief
subspherical, with circular marking on top (PI. 75, figs 6 and 7).

Remarks. The discs here assigned to B. brimsae are the most common fossils in the Wernecke
assemblage. They typically are closely crowded, and the size distribution of specimens on individual
bedding planes is strongly unimodal, indicating that each sample represents a population of
individuals at the same stage of ontogenetic development assembled on a mud substrate, before the
arrival of storm-deposited sand.

The Wernecke specimens are similar to forms described from the Russian Platform under a variety
of names (see synonymy). The size range of typical "Nemiana simplex' from the Ukraine is reported to
be 2-60 mm (Palij 1976, p. 70; Gureev 1985, p. 97). Our specimens are mainly at the lower end of the
size range for this form, and are more similar to the W. simplex' from the White Sea coast
(Fedonkin 1981, p. 57, pi. 3, figs. 2 and 9), for which diameters are between 10 and 30 mm and the
relief is 0-3 mm. Specimens from the Soviet Union also exhibit a single circular marking on their
upper surface which Fedonkin (1985a) interpreted as an oral opening. Taxonomic assignment and
interpretation of the discs from the Soviet Union is difficult because, despite an abundance of
photographic illustrations of the basal surface, little has been documented of their internal structure.

The Wernecke discs show considerable variation in preservation from one bedding plane to
another, ranging from smooth, flat discs with numerous delicate marginal wrinkles ( = B. brunsae
Menner) to more strongly convex forms with fewer and coarser, more irregular wrinkles, some of
which extend into the central parts of the disc (= N. simplex Palij). This morphological transition

666

PALAEONTOLOGY, VOLUME 30

indicates to us that Beltanelliformis and Nemiana are best explained as preservational variants of a
single globular biological taxon.

Any interpretation of Beltanelliformis must account for: (1) the concentric wrinkling; (2) the variety
of sedimentary fills and structures in and above the fossils (text-fig. 9; PI. 74, fig. 1; PI. 75, figs. 1, 2, 4);
and (3) the complete gradation between Beltanelliformis-type and Nemiana-type preservation
(text-fig. 8). We propose that Beltanelliformis-type forms were preserved where globular bodies resting
on the mud substrate, either as single spheroidal organisms, or, less likely, as spheroidal colonies
comparable to those figured by Glaessner ( 1969, fig. 3), were buried by storm-deposited sand. If, upon
burial, the bodies quickly collapsed, undisturbed lamination would be preserved above the locus of
these bodies (e.g. PI. 74, fig. 1). If the bodies maintained their integrity long enough to allow the
accumulation of sand around and above them subsequent decay of the spheroids would have
produced the collapse of the sand into the space previously occupied by the bodies, resulting in the
disturbed lamination now encountered above some of the specimens (text-figs. 9.1; 9.2, specimen F). In
contrast, specimens exhibiting, N emiana-type preservation probably were partially buried in the mud,
most likely as a result of slow accumulation of mud around the bases of the spheroidal organisms. As
a consequence of rapid, storm-induced burial, some specimens were filled with graded to parallel
laminated sand (text-figs. 9.2, specimen D; 9.3). Other specimens remained unfilled, and subsequent
collapse of the organism produced disturbed (slumped) lamination (text-fig. 9.4). Folding of the tough
outer wall of the organism during rapid burial probably produced the concentric to irregular wrinkles
visible on most specimens of Beltanelliformis and 'Nemiana'.

Lithology also appears to have played an important role in determining the mode of preservation.
Gureev (1985) has pointed out that, in the Ukraine, Beltanelloides sorichevae {Beltanelliformis-type
preservation) occurs predominantly in shale, whereas 'Beltanelloides simplex' {N emiana-type
preservation) occurs predominantly in siltstone and sandstone. Based on this, Gureev (1985)
suggested that the two forms might be synonymous. Our specimens further support this hypothesis.
In the Wernecke biota, the best examples of Beltanelliformis-type preservation (e.g. PI. 74, fig. 2) are
cast by argillaceous siltstone, whereas the best examples of N emiana-type preservation (e.g. PI. 75, figs.
5-8) are cast by sandstone.

Some specimens of Beltanelliformis exhibiting N emiana-type preservation superficially resemble
hemispherical anemone burrows such as Bergaueria Prantl, 1945. Palij et al. (1979) pointed out that
'Nemiana' can be distinguished from Bergaueria by the presence of numerous wrinkles and folds
resulting from deformation of a soft-bodied organism following burial, the consistent absence of an
overlying vertical cylinder, and by the fact that adjacent specimens deform but do not cross-cut each
other. The sporadic occurrence of specimens preserved in concave hyporelief (PI. 75, fig. 5) in the
Wernecke assemblage further suggests that Beltanelliformis represents the impression of a soft-bodied
organism rather than a hemispherical burrow-fill. Hemispherical specimens of Beltanelliformis also
superficially resemble the base of the Cambrian-Ordovician fossil Protolyella Torell 1870, which
Seilacher (1984) has interpreted as the internal sandy skeleton of an anemone. However, Protolyella
exhibits concentric hemispherical sediment fill reflecting active packing by the organism, whereas
hemispherical Beltanelliformis were passively filled with graded, laminated, or massive sediment.

EXPLANATION OF PLATE 75

Figs. 1-8. Beltanelliformis hrunsae Menner. 1, vertical section along inclined left margin of specimen in fig. 3,
showing sagged laminae above specimens of Beltanelliformis. 2, vertical section along upper right margin of
fig. 3, showing draping of laminae over specimens, x 1. 3, lower surface, showing close association of high
relief (GSC 83038) and low relief (GSC 83039) forms. GSC loc. 101532. 4, vertical section showing laminated
fill and slight draping of large specimen. GSC loc. 101531. GSC 83040, x 1. 5, cluster with specimens in both
convex and concave hyporelief. GSC loc. 99095. GSC 83041, x 0-7. 6, bedding plane view of specimen
preserved in full relief. GSC loc. 99036, GSC 83042, x 2. 7, side view of specimen in fig. 6, x 2. 8, largest
observed specimens in the Wernecke assemblage. GSC loc. 101531, GSC 83043, x 1.

PLATE 75

Jit//

NARBONNE and HOFMANN, Beltanelliformis

668

PALAEONTOLOGY, VOLUME 30

Nevertheless, Beltanelliformis, Bergaueria, and Protolyella can be closely similar in plan view, and can
only be distinguished through study of their three dimensional form and the nature of their internal
sediment.

BelTanellifonnis bnmsae was originally regarded as amedusoid (e.g. Sokolov 1972a, h; Menner 1974;
Palijcfa/. 1979; Fedonkin 1981). However, Sokolov (1976), Sokolov and Fedonkin (1984, p. 13), and
Glaessner ( 1 984, pp. 24-25) who use the designation Beltanelloides sorichevae for such structures, later
related them to Chuaria-Wke organisms, which are typically preserved as carbonaceous compressions.
A lack of associated carbonaceous material, the apparent presence of a circular aperture on the upper
surface, and the centimetric size of some specimens, makes the comparison of Beltai'ielliformis with
chuariamorphids tenuous. Neither Beltanelliformis nor Beltanelloides appear in the Treatise
(Glaessner 1979), and we have not had the opportunity to compare type material of these two taxa.
The former may be an objective synonym (ICZN 1964, Art. 616), however, because of the possible
distinctness of specimens referred to B. sorichevae Sokolov (1972a, pi. 4, figs. 4-8), which have no
circular impressions in the centre, and to Beltanelliformis brunsae Menner (Sokolov 1972a, pi. 4, fig. 2),
and the apparent questionable nomenclatural status of the former (no holotype designated for species
from among two ‘forms’; no diagnosis given; ICZN 1964, Art. 13a (i), 15, 456, c, 72a), we have assigned
our structures to the validly published taxon Beltanelliformis.

ICHNOFOSSILS
Ichnogenus gordia Emmons, 1844
Type ichnospecies. Gordia marina Emmons, 1844.

Gordia marina Emmons, 1844
Text-fig. 10a

? 1976 ‘crawling trails’, Palij, pi. 26, figs. 1 and 2.

? 1979 ‘crawling trails, first variety’, Palij et a/., p. 77, pi. 53, figs. 2 and 4.

? 1981 Gordia sp., Hofmann, p. 309, fig. 56. '

1983 Gordia sp., Fritz et al., pi. 44.1, fig. 3.

? 19856 Gordia sp., Fedonkin, pi. 23, fig. 1.

Description. Two specimens, preserved as concave epireliefs and convex hyporeliefs on very thin beds of
fine-grained sandstone. Burrows horizontal and irregularly meandering; true branching absent, but cross-overs
present. Burrows smooth with a diameter of approximately 1 mm. Burrow fill similar to host lithology.

Remarks. The presence of numerous cross-overs is commonly used to distinguish Gordia from slender
specimens of Planolites and Helminthopsis (Ksi^ziewicz 1977, p. 155). This criterion, based on
Phanerozoic specimens, appears to be less significant in the Ediacaran, where forms transitional
between Gordia and Planolites are common (e.g. text-fig. 106; Glaessner 1969, fig. 5b; Palij et al. 1979,
pi. 53, figs. 2 and 4). Ksi^ziewicz (1977) and Crimes and Anderson (1985) recognized three ichnospecies
of Gordia: G. marina Emmons, G. molassica (Reer 1865), and G. arcuata Ksi^ziewicz 1977, but we agree
with Pickerill (1981) that G. molassica is indistinguishable from G. marina.

Gordia occurs widely in Phanerozoic strata and is one of the most commonly reported Ediacaran
ichnofossils. G. marina occurs on the Russian Platform and in northwestern Canada, G. arcuata has

THXT-FiG. 10. a, Gordia marina Emmons, concave epirelief. GSC loc. 101533. GSC 83044, x 1. 6, Planolites
njonfanas Richter, hyporelief GSC loc. 101531. GSC 83045, x 1. c, Aconere/res? sp., hyporelief. GSC loc. 101533.
GSC 83046, X 1. d, P. montanus Richter, hyporelief. GSC loc. 101532. GSC 83047, x 1. e, horizontal
backfilled burrow, ?epirelief. GSC loc. 101532. GSC 83048, x 1. 6 Dubiofossil A, probably epirelief. GSC loc.
101536. GSC 83049, x 1. g, Vendotaenial sp. GSC loc. 101540. GSC 83050, x 2. 6, Dubiofossil B. GSC loc.
101532. GSC 83051, x 1. /, Dubiofossil C. GSC loc. 101541. GSC 83052, x 1.

NARBONNE AND HOFMANN; EDIACARAN BIOTA FROM CANADA

669

670

PALAEONTOLOGY, VOLUME 30

been reported from northern British Columbia (Fritz and Crimes 1985), and Gordia sp. occurs in the
Ediacaran of Australia (Glaessner 1969, fig. 5b), the Russian Platform (Palij et al. 1979), the Mackenzie
Mountains (Hofmann 1981), and Newfoundland (Crimes and Anderson 1985). Gordia probably
represents the crawling or feeding burrow of a worm-like organism (Chamberlain 1977).

Ichnogenus neonereites Seilacher, 1960
Type ichnospecies. Neonereites hiserialis Seilacher, 1960.

Neonereitesl sp.

Text-fig. 10c

Description. Four specimens occurring in convex hyporelief on thinly bedded, fine-grained sandstone. Burrows
slightly to moderately sinuous, up to 90 mm long, each consisting of a uniserial string of spherical to slightly
ellipsoidal pellets 3-5 mm in diameter; pellets locally in contact, but typically irregularly spaced up to 5 mm apart;
pellets composed of well-sorted sand.

Remarks. Specimens closely resemble those figured by Fritz and Crimes (1985, pi. 3/) from the
Ediacaran of northern British Columbia. Among described species of Neonereites, they most closely
resemble N. uniserialis Seilacher, but differ in that the pellets are irregularly spaced. N. nniserialis has
been reported from the Vendian of the Russian Platform (Fedonkin 1977; Palij et al. 1979) and is
common in Phanerozoic deposits (Hantzschel 1975). Neonereites probably represents the feeding
burrow of a worm-like organism, most likely an annelid (Hakes 1976).

Ichnogenus planolites Nicholson, 1873
Type ichnospecies. Planolites vulgaris Nicholson and Hinde, 1875.

Planolites montaniis Richter, 1937
Text-fig. 106, d

1970 ‘hypichnial and exichnial casts’. Banks, p. 26, pi. 16, d.

? 1970 ‘curved trails’, Webby, p. 87, fig. 3d.

1970 Planolites hallandus Webby, p. 95, fig. 14a-c.

1973 ‘hypichnial and endichnial burrows’. Banks, p. 4, fig. 4a.

1977 Planolites sp., Fedonkin, p. 184, pi. 2d.

1979 ‘crawling traces, third variety’, Palij et al., pp. 77-78, pi. 54, fig. 2.

1979 Planolites cf. serpens, Palij et al., p. 73, pi. 42, fig. 6.

1984 Planolites sp. Glaessner, p. 70, fig. 2/7.

19856 Planolites cf. serpens Fedonkin, pi. 28, figs. 3, 6.

Description. Specimens occurring in convex hyporelief and concave epirelief on very thin beds of siltstone and
fine-grained sandstone. Burrows highly sinuous and undulatory, occurring on bedding surfaces as small knobs
and discontinuous, curved, burrow segments. Burrows cylindrical, with diameters ranging from 0 4 to L2 mm
(N = 100); indistinct, irregularly spaced constrictions giving burrows a faint ‘pinch-and-swell’ appearance. True
branching and cross-overs of adjacent specimens rare. Burrow fill structureless, and differing from host lithology.

Remarks. The taxonomy of Planolites has recently been reviewed by Pemberton and Frey (1982), who
concluded that Planolites can be distinguished from the morphologically similar burrow Palaeo-
phycus Hall by the presence of processed burrow-fill and the absence of a burrow lining in Planolites.
Many specimens of P. montanus, including the ones in this study, exhibit a faint ‘pinch-and-swell’
appearance reminiscent of Torrowangea Webby, 1970. However, the swellings in Torrowangea are
well defined and evidently represent a back-fill structure (Webby 1970), whereas the swellings in P.
montanus are poorly defined and apparently reflect the undulose nature of the burrow. The typically
meandering pattern of Torrowangea also differs from the sinuous to undulose pattern of P. montanus.
P. montanus is very common in the Wernecke assemblage, a feature typical of many Ediacaran

NARBONNE AND HOFMANN: EDIACARAN BIOTA FROM CANADA

671

assemblages (see synonymy above). P. montanus also occurs commonly throughout the Phanerozoic
(Pemberton and Frey 1982). Plmwlites probably represents the feeding burrow of a vermiform
organism (Pemberton and Frey 1982).

BACK-FILLED BURROW
Text-fig. lOe

Description. Single, fragmentary specimen on the (?)upper surface of a very thin bed of argillaceous, fine-grained
sandstone. Specimen a gently curved, partially compressed cylinder 1 1 mm wide and at least 17 mm long; with
well-developed back-fill.

Remarks. Although its fragmentary nature precludes definite identification, the specimen exhibits a
back-fill structure similar to Muensteria von Sternberg 1833 or Beaconites Vialov 1962. Similar traces
occur in the Ediacaran of Australia (R. J. F. Jenkins, written comm. 1986).

METAPHYTES

Group VENDOTAENiDES Gnilovskaya, 1971
Genus vendotaenia Gnilovskaya, 1971

Type species. Vendotaenia antiqua Gnilovskaya, 1971.

Vendotaenia? sp.

Text-fig. lOg

Description. Isolated, smooth carbonaceous ribbons, curved and bent, 04-2 0 mm wide, largest fragment of nine
specimens seen 30 mm long; faint, submillimetric longitudinal striae present in portions of filament.

Remarks. The ribbons have an appearance intermediate between those referred to Vendotaenia and
Tyrasotaenia from the Russian Platform. The broad diameter, the curved nature, and the faint
microscopic longitudinal striae suggest affinity with Vendotaenia, whereas twisted specimens bear
more resemblance to Tyrasotaenia. However, longitudinal striae have also been reported for the latter
genus, though these have been ascribed to folding of the thallus.

DUBIOFOSSILS
Dubiofossil A
Text-fig. 10/’

Description. Almost complete, flat, gibbous disk, 40 x 29 mm, with marginal groove 0-8 mm deep; margin on
one side almost rectilinear for about 17 mm, remainder evenly curved. No further identifiable markings.

Remarks. The specimen has no diagnostic features which would allow it to be classified. Possibly, it
represents a severely distorted Cyclomedusa, though the absence of concentric wrinkling is against
such an interpretation.

Dubiofossil B
Text-fig. 10/;

Description. Single specimen preserved in convex hyporelief, 39 x 28 mm; subhexagonal, with outer rim
L2-4-5 mm wide and 0-5-L2 mm high. Trapezoidal ridge 15 x 12-19 mm and up to 2 0 mm high near centre
of specimen. Faint parallel ridges approximately 1-0 mm apart locally preserved on central trapezoid and outer
rim.

672

PALAEONTOLOGY, VOLUME 30

Remarks. The hexagonal outline and central trapezoid are both features that have not previously
been described from Ediacaran macrofossils. The specimen appears to have been flattened, and
consequently it is difficult to determine whether these represent primary features. Patterns such as
those exhibited by dubiofossil B occur commonly on much smaller compressed leiospherid
microfossils (e.g. Timofeev 1969), and it is possible that our specimen represents a large, compressed,
spheroidal or sac-shaped organism.

Dubiofossil C

Text-fig. 10/

Description. Large number of minute pits, preserved on the upper surface of an olive-grey weathering siltstone
lamina; diameters 0-5-L2 mm, relief about OT mm; specimens scattered, generally not contiguous. Similar, but
smaller depressions exposed in one corner of the slab 10 mm below the first layer.

Remarks. The pits do not appear to be moulds of sand grains, inasmuch as the overlying sediment of
coarse siltstone/very fine sandstone is, within some pits as well as outside them, still attached to the
epirelief surface and does not contain coarse sand grains. Because of their small size and the inferred
subtidal setting of the sediment, they do not appear to be rainprints. We thus consider it possible that
they are fossils; they may represent an assemblage of juvenile forms of Beltauelliformis or Bergaueria.
Alternatively, they resemble small pits associated with some specimens attributed to Arumberia (e.g.
Bland 1984, figs, lb and 2b, c), which differ, however, by the presence of superimposed fine, parallel to
subparallel narrow ridges and wider grooves.

Acknowledgements. We gratefully acknowledge financial support from the Natural Science and Engineering
Research Council of Canada (Grants A2648 and N0108 to Narbonne and A7484 to Hofmann), the Canadian
National Committee of the International Geologic Correlation Programme (I.G.C.P. Project 29), and the
Queen's University Advisory Research Committee. J. D. Aitken and W. H. Fritz (Geological Survey of Canada)
provided logistic support and critically read the manuscript; D. T. Osborne and J. Carrick assisted us in collecting
the material. C. Peck prepared some of the illustrations, H. Helmstaedt and J. Kuska translated critical articles,
and Ann MacDonald, Larke Zarichny, and Michael Watson typed the manuscript. We thank W. Hofmann for
modification of the depth gauge used to obtain more precise measurements of bedding relief, and J. P. Bourque
and M. Kachaami for technical assistance. We are also grateful to M. M. Anderson, M. A. Fedonkin, R. J. F.
Jenkins, and R. K. Pickerill for critical review of the manuscript.

REFERENCES

AITKEN, J. D. 1984. Strata and trace fossils near the Precambrian-Cambrian boundary, Mackenzie, Selwyn and
Wernecke Mountains, Yukon and Northwest Territories: Discussion. Geol. Siirv. Pap. Can. 84-1 B, 401-407.

In press. Uppermost Proterozoic formations in the central Mackenzie Mountains, Northwest Territories,

Canada. Bull. geol. Siirv. Can. 368.

ANDERSON, M. M. and CONWAY MORRIS, s. c. 1982. A review, with descriptions of four unusual forms, of the
soft-bodied fauna of the Conception and St. John’s Groups (Late Precambrian), Avalon Peninsula, Newfound-
land. Proc. third N. Am. paleont. Conv. I, 1-8.

BANKS, N. L. 1970. Trace fossils from the late Precambrian and Lower Cambrian of Finnmark, Norway. Geol. J.
Spec. Issue, 3, 19-34.

1973. Trace fossils in the Halkhavarre section of the Dividal Group (?Late Precambrian-Lower Cambrian),

Finnmark. Norsk geol. Unders. 288, 1-6.

BEKKER, YU. R. 1985. Metazoa from the Vendian of the Urals. In sokolov, b. s. and Ivanovskiy, m. n. (eds.).
The Vendian System; Historic Geological and Pcdaeontological Basis, 1, 107-111. Nauka, Moscow. [In
Russian.]

bland, b. h. 1984. Arumberia Glaessner & Walter, a review of its potential for correlation in the region of the
Precambrian-Cambrian boundary. Geol. Mag. 121, 625-633.

BLUSSON, s. L. 1971. Sekwi Mountain map-area, Yukon Territory and District of Mackenzie. Geol. Surv. Pap.
Can. 71-22, 17 pp.

NARBONNE AND HOFMANN: EDIACARAN BIOTA FROM CANADA

673

CECiLE, M. p. 1982. The Lower Paleozoic Misty Creek Embayment, Selwyn Basin, Yukon and Northwest Terri-
tories. Bull. geol. Surv. Can. 335.

CHAMBERLAIN, c. K. 1977. Ordovician and Devonian trace fossils from Nevada. Bull. Nev. Bur. Mines Geol. 90,
24 pp.

CLOUD, p. E. 1968. Pre-metazoan evolution and the origins of the Metazoa. In: drake, t. (ed.). Evolution and
Environment, 1-72. Yale University Press, New Haven.

and GLAESSNER, M. F. 1982. The Ediacarian Period and System: Metazoa inherit the earth. Science, N .Y. 217,

783-792.

COATS, R. p. 1981. Late Proterozoic (Adelaidean) tillite of the Adelaide Geosyncline. In hambrey, m. j. and
HARLAND, w. B. (cds.). Earth's Pre-Pleistocene Glacial Record, 537-548. Cambridge University Press,
Cambridge.

cowiE, j. w. 1985. Continuing work on the Precambrian-Cambrian boundary. Episodes, 8, 93-97.

CRIMES, T. p. and ANDERSON, M. M. 1985. Trace fossils from Late Precambrian-Early Cambrian strata of
southeastern Newfoundland (Canada): Temporal and environmental implications. J. Paleont. 59, 310-343.

DAILY, B. 1972. The base of the Cambrian and the first Cambrian faunas. Spec. Pap. Centre Precambrian Res.,
Univ. Adelaide, 1, 13-41.

1973. Discovery and significance of basal Cambrian Uratanna Formation, Mt. Scott Range, Flinders

Ranges, South Australia. Search, 4, 202-205.

DYSON, I. A. 1985. Frond-like fossils from the base of the late Precambrian Wilpena Group, South Australia.
Nature, Land. 318, 283-285.

EiSBACHER, G. H. 1981. Sedimentary tectonics and glacial record in the Windermere Supergroup, Mackenzie
Mountains, northwestern Canada. Geol. Surv. Pap. Can. 80-27, 40 pp.

1985. Late Proterozoic rifting, glacial sedimentation, and sedimentary cycles in the light of Windermere

deposition, western Canada. Palaeogeog. Palaeoclimat. Palaeoecol. 51, 231-254.

EMMONS, M. A. 1844. The Taconic System, 68 pp. Albany, N.Y.

EVENCHiCK, c. A., PARRISH, R. R. and GABRiELSE, H. 1984. Precambrian gneiss and Late Proterozoic sedimentation
in north-central British Columbia. Geology, 12, 233-237.

FEDONKiN, M. A. 1977. Precambrian Cambrian ichnocoenoses of the East European Platform. Geol. J. Spec.
Issue, 9, 1 83-194.

1978. A new discovery of soft-bodied Metazoa in the Vendian of the Winter Coast. Dokl. Akad. Nauk.

SSSR, 239, 1423-1425. [In Russian.]

1980. New representatives of Precambrian coelenterates on the north of the Russian Platform. Paleont. Zh.

2, 7-15. [In Russian.]

1981. The Vendian White Sea biota. Trudy geol. Inst. Akad. Nauk SSSR, 342, 1-100. [In Russian.]

1983. Nonskeletal fauna of the Podolian Pridnyestrov'ya. In velikanov, v. a., aseeva, e. a. and fedonkin,

M. A. (eds.). The Vendian of the Ukraine, 128-139. Naukova dumka, Kiev. [In Russian.]

1985n. Systematic descriptions of Vendian metazoa. In sokolov, b. s. and Ivanovskiy, m. a. (eds.).

The Vendian System: Historic-Geological and Paleontological Ba.sis, 1, 70-107. Nauka, Moscow. [In
Russian.]

19856. Paleoichnology of Vendian metazoa. In sokolov, b. s. and Ivanovskiy, m. a. (eds.). The

Vendian System: Historic-Geological and Palaeontological Basis, 1, 112-116. Nauka, Moscow. [In
Russian.]

FORD, T. D. 1958. Precambrian fossils from Charnwood Forest. Proc. Yorks, geol. Soc. 31, 211-217.

1963. The Pre-Cambrian fossils of Charnwood Forest. Trans. Leicester lit. phil. Soc. 57, 57-62.

F0YN, s. and GLAESSNER, M. F. 1979. Platysolenites, other animal fossils and the Precambrian-Cambrian
transition in Norway. Norsk geol. Tidsskr. 59, 25-46.

FREY, R. w. 1973. Concepts in the study of biogenic sedimentary structures. J. Sedim. Petrol. 43, 6-19.

FRITZ, w. H. 1978. Fifteen stratigraphic sections from the Lower Cambrian of the Mackenzie Mountains, North-
western Canada. Geol. Surv. Pap. Can. 77-33, 19 pp.

and crimes, t. p. 1985. Lithology, trace fossils, and correlation of Precambrian-Cambrian boundary beds,

Cassiar Mountains, north-central British Columbia. Ibid. 83-13, 24 pp.

NARBONNE, G. M. and GORDEY, s. p. 1983. Strata and trace fossils near the Precambrian-Cambrian

boundary, Mackenzie, Selwyn and Wernecke Mountains, Yukon and Northwest Territories. Ibid. 83-lB,
365-375.

1984. Strata and trace fossils near the Precambrian-Cambrian boundary, Mackenzie, Selwyn

and Wernecke Mountains, Yukon and Northwest Territories: Reply. Ibid. 84-lB, 409-412.

674

PALAEONTOLOGY, VOLUME 30

GIBSON, G. G., TEETKR, s. A. and EEDONKiN, M. A. 1984. Ediacarian fossils from the Carolina Slate Belt, Stanly
County, North Carolina. Geology, 12, 387-390.

GLAESSNER, M. F. 1969. Trace fossils from the Precambrian and basal Cambrian. Lethaia, 2, 369-393.

1979. Precambrian. In robison, r. a. and teichert, c. (eds.). Treatise on invertebrate paleontology. Pt. A, 79-

1 18. Geological Society of America and University of Kansas Press, Boulder, Colorado and Lawrence, Kansas.

1984. The dawn of animal life. A biohistorical study, 244 pp. Cambridge University Press, Cambridge,

London, New York, New Rochelle, Melbourne, Sydney.

and DAILY, B. 1959. The geology and Late Precambrian fauna of the Ediacara Lossil Reserve. Rec. S. Aust.

Mas. 13, 369-401.

and WADE, M. 1966. The Late Precambrian fossils from Ediacara, South Australia. Palaeontology, 9,

599-628.

GOLDRING, R. and CURNOW, c. N. 1967. The stratigraphy and facies of the Late Precambrian of Ediacara, South
Australia. J. geol. Soc. Aiistr. 14, 195-214.

GNiLOVSKAYA, M. B. 1971. The oldest aquatic plants of the Vendian of the Russian Platform. Paleont. Zh. 1,
101-107. [In Russian.]

GORDEY, s. i>. 1980. Stratigraphic cross-section, Selwyn Basin to Mackenzie Platform, Nahanni Map Area,
Yukon Territory and District of Mackenzie. Geol. Surv. Pap. Can. 80-lA, 353-355.

GRABAU, A. w. 1922. The Sinian system. Bull. geol. Soc. China, 1, 44-88.

GUREEV, YU. A. 1985. Vendiata- primitive Precambrian Radialia. In sorolov, b. s. and zhuravleva, i. t. (eds.).
Prohlematiki pozdnego dokewmriya: paleozoya. Akad. nauk SSSR, Sib. Otd. Inst. Geol.; Geofiz, Tr. v.632,
pp. 92-103. Nauka, Moscow. [In Russian.]

— VELIKANOV, V. A. and iVANCHANEKO, V. YA. 1985. Nonskeletal fauna in the regions of the Baltic and
Berezhkov series of Podolia. Depovidi Akad. Nauk Ukr. RSR, ser. B, geol. khini. biol. Nauki (for 1985), 6,
pp. 10 13. [In Ukrainian.]

HAKES, w. G. 1976. Trace fossils and depositional environments of four clastic units. Upper Pennsylvanian
megacyclothems, northeast Kansas. Paleont. Contr. Univ. Kans. Art. 63, 46 pp.

HANTZSCHEL, w. 1975. Trace fossils and problematica. In teichert, c. (ed.). Treatise on invertebrate paleontology,
Pt. W. suppl. 1, 269 pp. Geological Society of America and University of Kansas Press, Boulder, Colorado and
Lawrence, Kansas.

HARRINGTON, H. J. and MOORE, R. c. 1956. Scyphomedusae; Trachylindae; Medusae incertae sedis and
unrecognizable forms. In moore, r. c. (ed.). Treatise on invertebrate paleontology, Pt. F, 38-53, 68-76, 153-161.
Geological Society of America and University of Kansas Press, Lawrence, Kansas.

heer. o. 1865. Die Urwelt der Schweiz, 622 pp. Zurich.

HOFMANN, H. J. 1981. First record of a Late Proterozoic faunal assemblage in the North American Cordillera.
Lethaia, 14, 303-310.

1984. Organic-walled microfossils from the latest Proterozoic and earliest Cambrian of the Wernecke

Mountains, Yukon. Geol. Surv. Pap. Can. 84-lB, 285-297.

^ FRITZ, w. H. and narbonne, g. m. 1983. Ediacaran (Precambrian) fossils from the Wernecke Mountains,
northwestern Canada. Science, N.Y. 221, 455-457.

mountjoy, e. w. and teitz, m. w. 1985. Ediacaran fossils from the Miette Group, Rocky Mountains,

British Columbia, Canada. Geology, 13, 819-821.

international code of ZOOLOGICAL NOMENCLATURE. 1964. 176 pp. London.

JENKINS, R. J. F. 1981. The concept of an Ediacaran Period and its stratigraphic significance in Australia. Trans.
R. Soc. S. Aust. 105, 179-194.

1984. Interpreting the oldest fossil Cnidarians. Paleontogr. Am. 54, 95-104.

FORD, c. H. and GEHLiNG, J. G. 1983. The Ediacara Member of the Rawnsley Quartzite: the concept of the

Ediacara assemblage (Late Precambrian, Flinders Ranges). J. geol Soc. Aust. 29, 101-119.

and GEHLING, J. G. 1978. A review of the frond-like fossils of the Ediacara assemblage. Rec. S. Aust. Mus. 17,

347-359.

KELLER, B. M., MENNER, V. V., STEPANOV, V. A. and CHUMAKOV, N. M. 1 974. New hnds of fossils in the Precambrian
Valday Series along the Syuzma River. Izv. Akad. Nauk SSSR, 12, 130-134. [In Russian.]

KSiAZiEwicz, M. 1977. Trace fossils in the flysch of the Polish Carpathians. Palaeont. pol. 36, 208 pp.

LENZ, A. c. 1972. Ordovician to Devonian history of northern Yukon and adjacent District of Mackenzie. Bull.
Can. Ass. Petrol. Geol. 20, 321-361.

MCMENAMiN, M. A. s. 1986. The garden of Ediacara. Palaios, 1, 178-182.

NARBONNE, G. M., HOFMANN, H. J. and AiTKEN, J. D. 1985. Precambrian-Cambrian boundary sequence, Wernecke
Mountains, Yukon Territory. Geol. Surv. Pap. Can. 85-lA, 603-608.

NARBONNE AND HOFMANN: EDIACARAN BIOTA FROM CANADA

675

MYROW, p., LANDING, E. and ANDERSON, M. M. 1987. A Candidate stratotype for the Precambrian-Cambrian

boundary. Fortune Flead, Burin Peninsula, southeastern Newfoundland. Can. J. Earth. Sci. 24, 1277-1293.

NICHOLSON, H. A. 1873. Contributions to the study of the errant annelides of the older Palaeozoic rocks.
Proc. R. Soc. 21, 288-290.

and HiNDE, G. J. 1875. Notes on the fossils of the Clinton, Niagara and Guelph Formations of Ontario,

with descriptions of new species. Can J. Sci. Lit. Hist. ser. 2, 14, 137-160.

NOWLAN, G. s., NARBONNE, G. M. and FRITZ, w. H. 1985. Small shelly fossils and trace fossils near the
Precambrian-Cambrian boundary in the Yukon Territory, Canada. Lethaia. 18, 233-256.

PALij, V. M. 1976. Remains of soft-bodied animals and trace fossils from the Upper Precambrian and Lower
Cambrian of Podolia. In Paleontologiya i stratiyrafiya verkhneyo clokenihriya i nizhneyo kembriya ynyo-zapada
Vostochno-Evropeiskoi platfonny, 63-76. Naukova dumka, Kiev. [In Russian.]

POSTi, E. and eedonkin, m. a. 1979. Soft-bodied Metazoa and animal trace fossils in the Vendian and Early

Cambrian. In Upper Precambrian and Cambrian paleontoloyv of the East European Platform, 49-82. ‘Nauka’,
M oscow. [In Russian; English translation published in 1983 under joint sponsorship of Inst. Geol. Sci.,
Polish Acad. Sci., and Geol. Inst., Acad. Sci. USSR by Publishing House Wydawnictwa Geologiczne, Warsaw,
56-93.]

PARK, J. K. and AiTKEN, J. D. 1986. Paleomagnetism of the Katherine Group in the Mackenzie Mountains:
implications for post-Grenville (Hadrynian) apparent polar wander. Can. J. Earth Sci. 23, 308-323.

PEMBERTON, s. G. and FREY, R. w. 1982. Trace fossil nomenclature and the Planolites-Palaeophycus dilemma.
J. Paleont. 56, 843-881.

PiCKERii.L, R. K. 1981. Trace fossils in a Lower Paleozoic submarine canyon sequence— the Siegas Formation of
northwestern New Brunswick, Canada. Maritime Seds Atlantic Geol. 17, 36-58.

PRANTL, F. 1945. Dvezazhadne, zkameneliny (stoopy) Z vrester chrustenickych— d kz, Rozpravy II. T fidv Ceske
Akad. 55, (3), 3-8.

RICHTER, R. 1937. Marken und Spuren aus alien Zeiten. I-II. Senckenberyiana, 19, 150-169.

1955. Die altesten Fossilien Siid-Afrikas. Senckenberyiana Lethaea, 36, 243-289.

SEiLACHER, A. 1960. Lebensspuren als Leitfossilien. Geol. Rdsch. 49, 41-50.

1984. Late Precambrian and early Cambrian metazoa: Preservational or real extinctions. In Holland, h. d.

and TRENDALL, A. F. (eds.). Patterns of Chanye in Earth Evolution, 159-168. Dahlem Konferenzen, Springer-
Verlag, Berlin, Heidelberg, New York.

SOKOLOV, B. s. 1952. On the age of the oldest sedimentary cover of the Russian Platform. Izv. Akad. Nauk
SSSR, 5, 21-31. [In Russian.]

1972a. The Vendian Stage in Earth History. Int. Geol. Congr. 24, 114-123. (Dokl. Sov. Geol., Probl. 7.

Nauka, Moscow. [In Russian.]

19726. The Vendian Stage in Earth History. Int. Geol. Conyr. 24, Sect. 1, 78-84. Montreal.

1973. Vendian of northern Eurasia. Mem. Amer. Ass. Petrol. Geol. 19, 204-218.

1976. The organic world of the Earth on the way to Phanerozoic dillerentiation. Vestnik Akad. Nauk

SSSR, I, 126-143. [In Russian.]

and FEDONKiN, M. A. 1984. The Vendian as the terminal system of the Precambrian. Episodes, 7, 12-19.

SOUTHCOTT, R. V. 1958. South Australian jellyfish. S. Aust. Nat. 32, 53-61.

SPRiGG, R. c. 1947. Early Cambrian (?) jellyfishes from the Flinders Ranges, South Australia. Trans. R. Soc.
S. Aust. 71, 212-224.

1949. Early Cambrian ‘jellyfishes’ of Ediacara, South Australia, and Mount John, Kimberley District,

Western Australia. Ibid. 73, 72-99.

VON STERNBERG, K. M. 1833. Versuch einer geognosisch-botanischen Darstellung der Flora der Vorwelt, 5-6,
1-80.

SUN, w. G. 1986. Precambrian medusoids: The C yclomedusa plexus and C yclomedusa-Uke pseudofossils. Pre-
cambrian Res. 31, 325-360.

TIMOFEEV, B. V. 1969. Proterozoic Spheromorphida, 146 pp. Inst. Precambrian Geol. and Geochronology, Acad.
Sci. USSR. Nauka, Leningrad. [In Russian, with English abstract.]

TORELL, o. M. 1870. Petrifacta Synecana Formationis Cambricae. Acta Univ. land, 6, pt. 2, no. 8, 1-14.

VIALOV, o. s. 1962. Problematica of the Beacon Sandstone at Beacon Height West, Antarctica. N.Z. J. Geol.
Geophys. 5, 718-732.

WADE, M. 1968. Preservation of soft-bodied animals in Precambrian sandstones of Ediacara, South Australia.
Lethaia, 1, 238-267.

1969. Medusae from uppermost Precambrian or Cambrian sandstones, central Australia. Palaeontology,

12, 351-365.

676

PALAEONTOLOGY, VOLUME 30

WADE, M. 1972. Hydrozoa and Scyphozoa and other medusoids from the Precambrian Ediacara fauna. South
Australia. Ibid. 15, 197-225.

WEBBY, B. D. 1970. Late Precambrian trace fossils from New South Wales. Lethaia, 3, 79-109.

XING, Y. 1984. The Sinian and its position in the Geological time scale. Int. Geol. Congr. 27, 1, p. 212.
ZAiKA-NOVATSKiY, V. s. and PALiJ, V. M. 1968. New data on problematic imprints from the Vendian of Podolia.
Paleont. Sb. Lvov Uiiiv. 5, 130-135. [In Russian.]

GUY M. NARBONNE
Department of Geological Sciences
Queen’s University
Kingston, Ontario
Canada K7L 3N6

HANS J. HOFMANN

Typescript received 2 June 1986
Revised typescript 22 February 1987

Department of Geology
University of Montreal
Montreal, Quebec
Canada H3C 3J7

CARADOCIAN BIVALVE MOLLUSCS
FROM WALES

by S. P. TUNNICLIFF

Abstract. The occurrence of a widespread, distinctive, and sometimes well-preserved bivalve fauna from the
Lower Longvillian (Ordovician, Caradoc) rocks of North Wales is recorded, especially from the Allt-Tair-
ffynon Beds at Allt y Gadair, south of Llanfyllin, Powys. Six species are described, of which Cynuitonota
verisimilis sp. nov., Myodcikryotus deigryn gen. et sp. nov., and Psendarca celtica sp. nov. are new. The new
family Myodakryotidae is established to accommodate the new genus Myodakryotus.

The Allt-Tair-ffynon Beds (Caradoc, Soudleyan-Longvillian: Williams et al. 1972) of Allt y Gadair
(SJ 145 175; text-fig. I), south of Llanfyllin, Powys, Wales, yielded a number of well-preserved and
unusual bivalves to early Geological Survey collectors in the late 1840s. These were studied by
J. W. Salter, and were referred to by him in his palaeontological contribution to Ramsey’s (1866)
Memoir, ‘The Geology of North Wales’.

Some of the specimens listed in BGS records can no longer be accounted for, possibly because,
in the past, such collections were often distributed for teaching purposes. However, the significant
bivalves have survived and other specimens have been collected since.

The locality is recorded variously as Galt-y-Gader, Gallt y gader, Alt-y-gader, and Allt-y-Gader
and is referred to as near Meifod or near Llanfyllin, both of which were used as field stations by
Geological Survey collectors. There is no doubt, however, that these all refer to the same hill,
though not necessarily exactly the same localities. On the old series Ordnance Survey maps it was
spelt ‘Gallt y Gader’ but on new 1 :50000 series maps it is ‘Allt y Gadair’, the form adopted herein.
A copy of the old series geological map in the Biostratigraphy Research Group, BGS, has MS
locality numbers and some notes on the reverse side; for Allt y Gadair, they are as follows, with
grid references added:

Llanfyllin (station) 19 Gallt y Gader, Bala Lime? (SJ 1460 1715); 20 N. End of Gallt y Gader,
Bala Lime? (SJ 1500 1815); 21 E. Elank of Gallt y Gader, Bala Lime? (SJ 1520 1715).

M‘Coy knew of the locality and described Area EdmomUaformis [.y/c] from there in 1851 (p. 52).
Subsequently, in Sedgwick and M‘Coy (1852, p. 352), he listed Orthis expaiisa (J. de C. Sowerby),
O. porcata (M‘Coy) and A. Edtnomiiiformis [i/c] (M‘Coy) as coming from Allt-y-Gader, near
Llanfyllin, and he redescribed A. Edmondiiformis (p. 283, pi. \k, figs. 2 and 3).

Wedd et al. (1929, pp. 42-45) recorded a single Caradoc bivalve from the area of Geological
Survey sheet 137; an 'lOrtIumota (see Orthodesma sp. herein) from Bryngwyn Camp, 3 km east of
Allt y Gadair. Whittington (1938) recorded no bivalves from the nearby Lluest Quarry but
established a lower Longvillian age for the highest Allt-Tair-ffynon Beds. Although Drs Pickerill
and Brenchley have collected widely over the area, their collections include no bivalves from Allt
y Gadair (Brenchley, pers. comm., Feb. 1981). Dr N. J. Morris (pers. comm.) has also searched
in vain for the bivalve-rich horizon. With the help of Dr D. E. White, I have examined exposures
on the northern end of Allt y Gadair without finding any weathered rock precisely comparable
with the known specimens. I conclude, therefore, that the original specimens came from lenses or
beds now obscured.

I Palaeontology, Vol. 30, Part 4, 1987, pp. 677-690, pis. 76-77.|

© The Palaeontological Association

678

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 1. Distribution of bivalve elements of the Caradocian Allt y Gadair fauna in North Wales.

DISTRIBUTION OF THE BIVALVE EAUNA
The species described here from the Allt-Tair-ffynon Beds at Allt y Gadair are:

Cymatonota verisimilis sp. nov.

Myodakryotus deigryn gen. et sp. nov.

Goniophora sp.

Orthodesma sp.

Pseudarca cel tic a sp. nov.

Vanuxemia bulla (Salter)

At least two of these species, Goniophora sp. and V. bulla also occur in the Bryn Siltstone
Formation (Longvillian: Brenchley 1978, pp. 146-149) at Nant lorwerth (SJ 216 364) near Glyn
Ceiriog, south of Llangollen, some 20 km NNE of Allt y Gadair. The Nant lorwerth specimens
are associated with single specimens of Lyrodesma and ITancrediopsis and with more numerous
dalmanellid and sowerbyellid brachiopods. No Lyrodesma has been seen from the Allt y Gadair
collections, nor any palaeotaxodont, although the old lists do record Ctenodonta. Specimens of M.
deigryn are recorded from as far to the north-west as Betws-y-Coed and Roman Bridge, and as
far south-east as the Welshpool area.

Also present in the collections from Allt y Gadair are the monoplacophoran Cyrtolites nodosus
(Salter), the gastropod Cyclonema crebristria (M‘Coy), and a number of other fragmentary
unidentified gastropods, as well as the brachiopod Macrocoelia and dalmanellid brachiopod and
trilobite fragments.

TUNNICLIFF: LATE ORDOVICIAN BIVALVES FROM WALES

679

All the bivalve specimens from the Allt-Tair-ffynon Beds at Allt y Gadair are preserved in a
highly fossiliferous, fine-grained rotten mudstone which is chocolate brown in colour when
weathered. The specimens from Bala, Nant lorwerth, and the other localities mentioned in this
paper are also from fine-grained mudstones which are grey in colour.

All of the occurrences of these distinctive bivalve faunas are from approximately contemporaneous
horizons of early Longvillian age. Indeed, the widespread and sudden appearance of these bivalve
faunas with such unusual genera as Myodakryotus and Pseudarca, at lower Longvillian horizons
is notable because over much of North Wales the appearance of a number of genera and species
is taken as indicative of the onset of Longvillian times (e.g. Whittington 1962). In particular, the
trilobites Kloucekia and Estoniops and brachiopods such as Plowellites autiquior (M'Coy) are taken
as indicators of Longvillian age. Although scant attention has been paid previously to the molluscan
faunas, the bivalves have a contribution to make to Caradocian biostratigraphy as well as
palaeoecology.

Pseudarca is known principally from France, but the rest of the Allt y Gadair bivalves are of
eastern North American aspect. There is, however, a notable lack of taxodont or other smaller
infaunal forms such as the actinodontoid Lyrodesma. All the specimens available are of similar
size and this feature and the fragmentary nature of many of the associated trilobite and brachiopod
fragments suggests that they are part of a well-sorted thanatocoenosis in which burrowing forms
such as Cymatonota and Orthodesma are mixed with semi-infaunal Vaniixemia and presumably
epifaunal Goniophora and Myodakryotus.

Specimen numbers bearing the prefix BGS are housed in the collections of the British Geological Survey,
Keyworth; those with the prefix SM are in the Sedgwick Museum, Cambridge; those with NMW are in the
National Museum of Wales, Cardiff; and BM indicates the British Museum (Natural History).

SYSTEMATIC PALAEOTOLOGY

In general, the classification used by Pojeta (1978) is adopted. Synonymies follow the recommen-
dations of Matthews (1973).

Subclass ISOFILIBRANCHIA Ircdalc, 1939
Family modiomorphidae Miller, 1877
Genus Goniophora Phillips, 1848

Type species. Goniophora cymbaefonnis (J. de C. Sowerby), by original designation of Phillips (1848, p. 264).
The genus has been discussed recently by Liljedahl (1984).

Goniophora sp.

Plate 76, figs. 6 and 9

Material. BGS Zv 2018a and b (internal moulds of two valves, an almost complete left valve, and a
fragmentary right valve), from Allt y Gadair, Llanfyllin.

Description. The material available shows the shell to be equivalve, prosogyrate, and apparently
rhomboidal with a pronouneed sigmoidal carina extending from the umbo towards the posteroven-
tral margin. Dorsal margin straight, anterior margin rounded, posterior margin possibly truncated
obliquely. The area dorsal from the carina has one or two eoarse but faint radial ribs, and the rest
of the shell shows a fairly coarse concentric ornament. Musculature, dentition, and ligament
unknown.

Remarks. Of the two specimens of Goniophora sp. the right valve, BGS Zv 20186, is fragmentary
but the left, Zv 2018a, is nearly complete. From Britain, Reed (1905, p. 500, pi. 24, fig. 15) and
Hind (1910, p. 539, pi. 4, figs. 24-27) have described and illustrated Ordovician Goniophora spp.,
but both appear more elongate than the present speeimens. G. carinata (Hall) from the Ottawa

680

PALAEONTOLOGY, VOLUME 30

Formation (Black Riveran-Trentonian) from Canada, redescribed by Wilson (1956, p. 76, pi. 9,
fig. 20), is closer in size and shape to the Welsh specimens. A single specimen of a Goniopliora
(BGS Zs 2751) from the Killey Bridge Formation (Ashgill, Cautleyan) of the Pomeroy Inlier, Co.
Tyrone, was described by Tunnicliff (1982, p. 80, pi. 12, fig. 15), and has proportions similar to
the Welsh specimens.

A slab (BGS Zv 2010) from the Bryn Siltstone Formation at Nant lorwerth (SJ 216 364) bears
a fragment of the posterior part of a left valve closely corresponding to the Allt y Gadair fragments
and probably represents the same form. The Nant lorwerth specimen is associated with sowerbyellid
and dalmanellid brachiopods and with single specimens of Lyrodesma and ITancrediopsis.

Subclass PTERiOMORPHiA Beurlan, 1944
Family cyrtodontidae Ulrich, 1894
Genus Vanuxemia Billings, 1858

Tvpe species. Cyrtodonta rugosa Billings, 1858, by subsequent designation of Williams and Breger (1916,
p.' 149).

Vanuxemia hidla (Salter, 1866)

Plate 76, figs. 1115; Plate 77, fig. 11; text-fig. 2a-c

V* 1866 Palaearcal bulla Salter, p. 344, woodcut 13, fig. 3 (two views), p. 270 (in list).

Type material. BGS GSM 12397 is here selected as lectotype. This is the specimen figured by Salter [m
Ramsey 1 866), an internal mould of a left valve from the Allt-Tair-ffynon Beds at Allt y Gadair. Paralectotypes
are BGS GSM 12398, Zv 2018, internal moulds of right valves from the same horizon and locality. Other
material consists of BGS GSM 22190, an internal mould of a left valve, and NMW.27.1 10.G609, an internal
mould of a right valve, from the Bryn Siltstone Formation (Longvillian) at Nant lorwerth (SJ 216 364), and
two specimens, SM 53554-53555, from the same horizon at Bryn Quarry, Glyn Ceiriog, which are internal
moulds of left and right valves respectively. There are also three specimens in BM(NH), BM 42791, a right
valve, internal mould, recorded as from Meifod; BM LI 3 156, a large right valve, internal mould, from Glyn
Ceiriog; BM PL4436 labelled ‘W.pool 21 m2’, left valve, internal mould, probably from Allt y Gadair.

Description. Strongly inflated Vanuxemia (maximum inflation of 17 mm in a valve 26 mm long), in which
the height is one-tenth to one-fifth greater than the length and the umbo is at about the anterior one-tenth
to one-fifth. The hinge line is straight with sets of anterior and posterior teeth separated by an edentulous
area. There are three anterior teeth on a hinge plate, the most posterior being hoseshoe shaped in the left
valve. There are apparently two ridge-like posterior teeth in the left valve and one in the right. The anterior

EXPLANATION OF PLATE 76

Figs. 1 -5. Orthodesma sp. 1 and 3, dorsal and left lateral views of internal mould, conjoined valves, BGS
GSM 22055. 2 and 4, dorsal and right lateral views, internal mould, conjoined valves, BGS GSM 24287.
Both from Allt-Tair-ffynon Beds, Allt y Gadair, Llanfyllin. 5, left lateral external view of gaping conjoined
valves, BM 44492, from Meifod.

Figs. 6 and 9. Goniophora sp., dorsal and lateral views of internal mould, left valve, BGS Zv 2018, from Allt-
Tair-ffynon Beds, Allt y Gadair, Llanfyllin.

Figs. 7, 8, 10. Pseudarca celtica sp. nov. Holotype, BGS GSM 24179, internal mould, left valve; 7 and 10,
latex cast, 8, lateral view.

Figs. 11-15. Vanuxemia bulla (Salter 1866). 11 and 15, lectotype, BGS GSM 12397, anterior and lateral
views of internal mould, left valve. Allt-Tair-ffynon Beds, Allt y Gadair, Llanfyllin. 12 and 13, lateral and
anterior views, internal mould, right valve, NMW.27. 1 10.G609, from the Bryn Siltstone Formation, Nant
lorwerth, Glyn Ceiriog (NGR SJ 216364). 14, latex of hinge area of internal mould, left valve, BGS GSM
22190, horizon and locality as for 12 and 13.

All x2 except Fig. 10, which is x 4.

PLATE 76

• . ,#/>4 ,V *.

TUNNICLIFF, Orthodesma, Goniophora, Pseudarca, Vanuxemia

682

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 2. Vanuxemia bulla (Salter), x 2. a, NMW 27.I10.G609, latex, right valve, h, BGS GSM 22190,
latex, left valve, c, BM (NH) PL4436, dorsal view, internal mould, left valve, u = umbo, aam = anterior
adductor muscle scar, ad = anterior dentition, pd = posterior dentition, dvl = duplivincular ligament.

adductor muscle scar is strongly impressed and large, about one-quarter of the height of the shell in height.
Posterior adductor muscle scar not discernible. Ornament of coarse concentric undulations only faintly seen
on the internal moulds.

Remarks. The lectotype no longer shows the dentition illustrated by Salter {in Ramsey 1866,
woodcut 13, fig. 3) but this is seen in one specimen from Nant lorwerth (BGS GSM 22190,
PI. 76, fig. 14; text-fig. 2b and NMW.27.1 10.G609, PI. 76, figs. 12 and 13; text-fig. 2a). A single
posterior tooth is seen in a specimen of the right valve (BM 42791).

The strongly inflated valve and general appearance of V. hiiUa corresponds well with that
generally accepted for Vanuxemia. However, its dentition does not readily compare with the better
preserved North American material of Vanuxemia and Cyrtodonta such as that illustrated by
Pojeta (1971, pis. 6-9; 1978, pis. 8 and 9).

EXPLANATION OF PLATE 77

Figs. 1-10. Myodakryotiis deigryn gen. et sp. nov. 1-4, holotype, BGS GSM 22040. 3, internal mould, right
valve; 1, 2, 4, oblique ventral, oblique anterodorsal and lateral views of latex cast. 5, internal mould, left
valve, BGS GSM 22041. All from the Allt-Tair-ffynon Beds, Allt y Gadair, Llanfyllin. 6, external lateral
view of right valve, BGS GSM 22053, from Gelli-Grin Formation, Pont Rhiwedog, Bala, Gwynedd. 7 and
8, external lateral views of two right valves, BGS DJ 3338, 3337 from beds of Longvillian age near Roman
Bridge, Gwynedd (NGR SH 7163 5121). 9, external lateral view of right valve, BGS Zv 2032, from Allt-
Tair-ffynon Beds, near Cefn Lleyfnog, Llanfyllin. 10, external lateral view of right valve, BGS Zv 1468,
from Gelli-Grin Formation, Cwm Chwilfod, Bala (NGR SH 955 400).

Fig. 1 1. Vanuxemia bulla (Salter 1866). Internal mould, right valve, BGS Zv 2018, Allt-Tair-ffynon Beds, Allt
y Gadair, Llanfyllin.

All x2.

PLATE 77

■'-AS.

TLFNNICLIFF, Myodaktryotus, Vanuxemia

684

PALAEONTOLOGY, VOLUME 30

Family myodakryotidae nov.

Type genus. Myodakryotus gen. nov. here designated.

Diagnosis. Fleteromyarian, pectiniform pteriomorphs with variable ornament.

Description. Equivalve, pectiniform, heteromyarian, prosogyrate pteriomorphs with small anterior and
posterior ears and variable ornament.

Known stratigraphic range. Restricted to Ordovician rocks of Caradoc age in Britain and their approximate
equivalents (Black Riveran, Trentonian: see Ross et al. 1982) in North America.

Remarks. A new family is established here to accommodate Myodakryotus gen. nov., which is
considered to be synonymous with New Genus 10 of Pojeta (1978), subsequently referred by Pojeta
and Runnegar (1985) to Prolohellcik and generically similar to those specimens described by Wilson
(1956, p. 56) as ActinoptereHal tessellata Wilson.

The affinitites of Myocktkryotiis are far from clear, but the combination of dimyarian musculature
with the pectiniform shape, distinctive, almost cardinal dentition, and apparent lack of a
duplivincular ligament precludes its being placed in any of the three families of Pteriomorphia
recognized from the Ordovician: Cyrtodontidae, Ambonychiidae, and Pterineidae (Pojeta 1978,
p. 235). The shell form is closest to that of a pterineid (if the pectinacean resemblance is discounted)
but the dentition is strongly reminiscent of the anterior dentition in some cyrtodontids (e.g. C.
saffordi (Hall), Pojeta 1978, pi. 8, fig. 11).

Pojeta and Runnegar (1985, p. 327) placed the Ordovician limiform shells which they assigned
to Prolohella in the Limidae Tor now’, and (Pojeta 1985, fig. 2) suggested that the Limacea might
be derived from the Cyrtodontidae approximately during Caradoc times. This seems likely as
outlined below, but the presence of the distinct anterior adductor muscle in Myodakryotus precludes
its placement in the strictly monomyarian Limidae and I am obliged to propose the family
Myodakryotidae to bridge the gap between the Cyrtodontidae and the Limidae.

I suggest that the Myodakryotidae could be derived from a cyrtodontid form by the acquisition
of pectiniform shape, involving a development of the anterior portion of the shell, leaving the
anterior teeth and adductor muscle close to the umbo, and the loss of the posterior dentition and
reduction of the ligament. This would presumably reflect a change in life-style from the infaunal
cyrtodontid type to a byssate epifaunal existance as suggested for the Ambonychiidae and some
pterineids (Pojeta 1971, pp. 32-34).

Whether the Myodakryotidae have any direct bearing on the possible origins of the Pectinacea
has yet to be fully determined.

Genus Myodakryotus nov.

1978 New genus 10 Pojeta, p. 239, pi. 12, figs. 1-8.

Derivation of name. Compounded from Greek myo-dakrytos. muscle, tearful, and rendered euphonically in
Latin form with masculine gender. The name refers to the teardrop-shaped anterior adductor muscle scar
and the shape of some internal moulds (e.g. PI. 77, fig. 5).

Type species. M. deigryn here designated.

Diagnosis. As for family.

Description. As for family.

Other species. M. hennione (Billings), from the upper Middle Ordovician (Pojeta 1978, p. 239, pi. 12, figs. I -
8; Pojeta and Runnegar 1985, p. 327, fig. 17) and M. tesselatiis (Wilson) from the Ottawa Formation (Black
Riveran Trentonian age) of eastern Canada. Judging from the published information the three species may
be distinguished by their ornaments. The figures by Pojeta and Runnegar (1985, fig. 17a, b) of M. arctica
(Schuchert) from the Upper Ordovician of Baffin Island, Canada, reveal no ornament.

TUNNICLIFF: LATE ORDOVICIAN BIVALVES FROM WALES

685

TEXT-FIG. 3. Myodakryotus deigryn gen. and
sp. nov., hoiotype, BGS GSM 22040,
internal mould right valve, x 2. cd = cardi-
nal dentition, pi = pallial line, aam = an-
terior adductor muscle scar, pam =
posterior adductor muscle scar.

cd

pam

pi

1cm

Myodakryotus deigryn gen. et sp. nov.

Plate 77, figs. 1 10; text-fig. 3

Derivation of name. Welsh deigryn (m), a tear; from the teardrop-shaped anterior adductor muscle scar and
the shape of some internal moulds (e.g. PI. 77, fig. 5).

Material. Hoiotype, BGS GSM 22040, internal mould, right valve (a second, less well-preserved internal
mould, right valve, on the same slab is a paratype) and paratype, BGS GSM 22041, internal mould,
left valve, all from Allt-Tair-ffynon Beds, Allt y Gadair, Llanfyllin. Other paratypes are BGS GSM 22052,
22053, right valves from the Gelli Grin Formation (Longvillian), Pont Rhiwedog, near Bala, Gwynedd (exact
locality uncertain but in the area of SH 947 348) and BGS Zv 2032, 2033, slabs with left and right valves
from the Allt-Tair-ffynon Beds near Cefn Lleyfnog, Llanfyllin; BGS DJ 3337, 3338, two right valves from
rocks in the Cwm Eigiau Formation of Longvillian age near Roman Bridge, Gwynedd (SH 7163 5121); BGS
Zv 1468, external of right valve from the Gelli Grin Formation at Cwm Chwilfod, north-east of Bala (SH
955 400); BGS Zv 913, external of left valve (incomplete) from the Gelli Grin Formation at Bryn Cut, Bala
(c. SH 951 344); BGS Zv 646 from Waterloo Bridge, Betws-y-Coed; BGS Zv 2557, 2558 and Zv 2541 from
the area of Moel y Garth and the Quakers’ Burial Ground, north-west of Welshpool; NMW.77. 1 1G.307
(Brenchley Collection) from the Cwm Rhiwarth Siltstone Formation at Cwm Rhiwarth (SJ 020 296). All
specimens are from horizons within the Longvillian stage of the Caradoc.

Description. Orthocline, prosogyrate, subcircular, equivalve, equilateral shell of pectiniform appearance, a
little higher than long (h; 1 1 0 to 1 T ), with small anterior and posterior ears. Inflation of a single valve about
5 mm in a shell 30 mm high. Teeth a little anterior from beak and with the form of a single chevron or
horseshoe-shaped tooth with a simple tooth posterior to and parallel with the posterior limb of this. Ligament
uncertain, but possibly opisthodetic, from the nature of the dorsal margin (but a duplivincular ligament is
possible: see remarks below). A shallow groove extends from below the beak towards the posterior ear.
Musculature heteromyarian: a subcircular posterior adductor muscle scar, about one-fifth of the height of
the shell in diameter and situated well within the body of the shell; the anterior adductor muscle scar is tear-drop

686

PALAEONTOLOGY, VOLUME 30

shaped, close to the umbo and anterodorsal margin, to which its axis is subparallel. Accessory muscles
unknown. Pallial line complete, non-sinuate and co-marginal (generally at a distance of about one-sixth of
the width of the shell from the margin) extending from the anterior end of the anterior muscle scar, touching
the outer edge of the posterior adductor muscle scar, where it shows a slight embayment, and returning
towards the umbo. Ornament of generally fine concentric lines, occasionally coarser especially on or near
the ears; some specimens show faint radial ornament. The area dorsal of the anterior muscle scar is separated
from it by a strong ridge, a weaker extension of which bisects the area and meets the anterodorsal angle of
the shell.

Remarks. It is perhaps surprising that such a distinctive species, collected so widely in North Wales
at a particular horizon has not been described before and it is especially remarkable that Salter
(1866) made no mention of it. The only forms known to resemble M. deigryn are those figured by
Pojeta ( 1978, pi. 12, figs. 1-8) as New genus 10 hermione Billings from the upper Middle Ordovician,
and specimens described by Wilson (1956, p. 56, pi. 7, figs. 8-11) from the Ottawa Formation
(Black Riveran-Trentonian age) of eastern Canada, as A.l tessellata Wilson. Pojeta subsequently
(Pojeta and Runnegar, 1985, p. 327, fig. 17) reassigned New genus 10 hermione and another species,
arctica (Schuchert), from the upper Ordovician of Baffin Island, Canada, to Prolohella Ulrich with
a qualifying question mark. He stated (1978, p. 239) that the musculature of his New genus 10
was poorly known and that the ligament and dentition were unknown, but that the ornament is
of co-marginal growth lines and fine radial ribs. By 1985 (Pojeta and Runnegar 1985, p. 327) he
had more internal detail including (pers. comm.) evidence of a duplivincular ligament, but the
dentition remained unknown and he referred to a single, posterior adductor muscle suggesting that
the anterior adductor muscle was not recognizable in the specimens before him. In some of the
Welsh specimens radial ornament is apparent but faint. With the larger, strongly impressed anterior
muscle scar, they differ from the American species which in turn seem, from Pojeta’s illustrations
(1978, pi. 12, fig. 6) to have a more strongly impressed posterior muscle scar. However, the
similarities are otherwise so great that one must conclude that the two species are closely allied
and place them in the same genus. Wilson’s figures show most of the features of the Welsh
specimens, but have a tessellated ornament and poorly known musculature and dentition.

Ulrich’s original figure of his P. striatula (1897, pi. 35, fig. 27) suggests a more prosocline form
lacking any posterior auricle. His description (1897, p. 532) as ’very inequilateral’ seems contrary
to the evidence presented by Myodakryotm. In the absence of a revised description of the type
material of Prolohella I feel the establishment of the new genus is justifiable.

Subclass ORTHONOTiA Pojeta, 1978

This subclass was established by Pojeta (1978, p. 240) to accommodate five genera: Cymatonota,
Orthodesma, Palaeosolen, Psiloconcha, and an unnamed genus. Taxonomy at the level of order,
superfamily and family remains unestablished.

Genus Cymatonota Ulrich, 1893
Type species. By original designation C. typicalis Ulrich (1893, p. 661).

Cymatonota verisimilis sp. nov.

Text-fig. Aa, b

v. 1866 Orthonota verisimilis MSS Salter, p. 270 (in list only), ?iom. mtd.

Type specimens. Holotype designated here, BM 42792, an internal mould of conjoined valves, complete except
for the extreme posterior portion, recorded as from Meifod, but possibly from Allt y Gadair (obtained from
W. Prosser); paratypes BGS GSM 24292, 24293 from Allt y Gadair, Llanfyllin.

Description. Valves very elongate with parallel dorsal and ventral margins. Height less than a quarter of the
length. The inflation of a single valve is about a quarter of the height. The umbo is almost terminal anteriorly.

TUNNICLIFF: LATE ORDOVICIAN BIVALVES FROM WALES

687

TEXT-FIG. 4. Cymatonota verisimilis sp. nov., x 2. holotype, BM 42792, left lateral
view, internal mould, conjoined valves, recorded as from ‘Meifod'. b, BGS GSM 24292,
lateral view, posterior portion, right valve, Allt-Tair-ffynon Beds, Allt y Gadair,
Llanfyllin. The richly fossiliferous nature of the bivalve-bearing rocks at this locality
can be seen in this hgure; present are trilobite, brachiopod, and gastropod fragments
including Cyrtolites noclostis (Sailer).

and the positions of the muscle scars are uncertain, although an area in the preumbonal region of the shell
may represent an anterior adductor muscle scar about half as high as the shell. Details of the ligament and
dentition are unknown. The ornament consists of two zones of ornament separated by a line running from
the umbo to the posteroventral angle (no carina or ridge is developed) as seen in recent Eiisis spp. Ventral
to this line, the ornament is of parallel, fairly coarse (c. 1 per mm) co-marginal lines which undulate gently
in section. The posteroventral ornament is of similar co-marginal lines but with a stronger and more irregular
undulation. There was a permanent posterior gape.

Remarks. The dimensions of the holotype are: height 12-0 mm, length 54-0 mm, and inflation of
a single valve 3-5 mm. Ornament is best seen in the paratypes, both of which are fragmentary.
BGS GSM 24293 is an internal mould of a section of the right postumbonal shell about 25 mm
long and showing the dorsal margin with a thin fragment of the left valve in place. BGS GSM
24292 also shows the internal mould, postumbonal section (c. 50 mm) of a right valve but with
the umbo itself missing. Both specimens have a height of about 13 mm and both show the nature
of the posterior ornament. The posterior end of the shell in BGS GSM 24292 suggests the presence
of a permanent gape.

Although Salter (1866, p. 270) clearly intended to describe this species as Ortiwuota verisimilis,
he never did so and his use of that taxon becomes invalid as a nomen nudum. It is reintroduced
here to retain the association with Salter’s work and to avoid future confusion. It was recorded

688

PALAEONTOLOGY, VOLUME 30

in one of the old Geological Survey lists as "Orthonota vessimilis (sic) MSS exactly like Solen .
Compared with other species of Cymalonota (e.g. Ulrich 1893, pi. 55, figs. 1-14; Pojeta 1978, pi.
15, figs. 9, 11-13), C. verisimilis is distinctive in its remarkably parallel dorsal and ventral margins
and its almost terminal umbones.

Genus Orthodesma Hall and Whitfield, 1875

Type species. By original designation, Orthodesma rectum Hall and Whitfield (1875, p. 93).

Orthodesma sp.

Plate 76, figs. 1-5

V . 1866 Orthouota sp. Salter, p. 270 (in list).

Material. Three internal moulds of conjoined valves, BGS GSM 22055, 24286, 24287 all from Allt y Gadair,
Llanfyllin; BGS WK 339, conjoined valves from Bryngwyn Camp, near Llanfyllin; BM 42793, 44492-44494,
BM LI 3526, 13547, left, right, and conjoined valves all recorded as from Meifod and purchased from
W. Prosser.

Description. Elongate, modioliform or soleniform shell with height about half the length, tapering rapidly
towards the anterior end, with the highest part of the shell towards the posterior end. The umbones are at
about the anterior quarter. Maximum inflation of the two valves coincides with the highest part of the shell
and is 14-5 mm in a shell 42-7 mm long (BGS GSM 24287). Posterior dorsal margin straight, anterior dorsal
margin drops away from umbo. Posterodorsal margin truncate, posteroventral margin rounded, anterior
margin rounded, ventral margin nearly straight or with a sinus a little posterior to the umbo and, in some,
anterior to the umbo. Adductor muscle scars of subequal size and round to ovate; anterior adductor scar
almost terminal, posterior adductor scar faint, below posterodorsal angle and in the upper half of the shell.
Pallial line not distinguishable. Details of ligament and dentition not known. Ornament of concentric
undulations about 2 mm apart, faint on the internal moulds, and finer comarginal lines (c. 5 per mm).

Remarks. Two of the specimens (BGS GSM 22055, 24287) appear to be inequivalve, but this is
probably the result of distortion, for they are inequivalve in opposite senses, and most of the other
specimens suggest an equivalve condition. None of the specimens shows a clear gape at either end
as is suggested for the genus by Ulrich (1894, p. 516). Pojeta (1971, p. 24) also expressed doubt
over this feature.

Subclass ACTiNODONTiA Pojeta, 1978
Family lyrodesmatidae Ulrich, 1894
Genus Pseudarca Tromelin and Lebesconte, 1875

Type species. Pseudarca typa Tromelin and Lebesconte. This genus was treated as a nuculoid by McAlester
(1968, p. 47), who redescribed the type species and discussed synonyms.

Pseudarca celtica sp. nov.

Plate 76, figs. 7, 8, 10

? 1866 Orthonota sp. Salter, p. 270 (in list).

Material. Holotype, BGS GSM 24179, an internal mould, left valve, from the Allt-Tair-ffynon Beds at Allt
y Gadair, Llanfyllin. A slab in the Sedgwick Museum, SM A53554, from the Bryn Lormation, Bryn Quarry,
Glyn Ceiriog, bears what appears to be the anterior portion of an internal mould of conjoined valves of
P. celtica. On the same slab is a specimen of Vanu.xemia bulla (q.v.).

Description. Elongate, with a rounded anterior end and probably a rounded posterior end (the posterodorsal
angle is missing). The dorsal margin is straight behind the umbo and curves continuously with the anterior
margin before the umbo. The ventral margin is gently curved such that the highest part of the shell is behind
the umbo, a little anterior to the midpoint. Dimensions of the holotype: length 22-5 mm, height 7-4 mm.

TUNNICLIFF: LATE ORDOVICIAN BIVALVES FROM WALES

689

h : 1 0-33, maximum inflation of the single valve c. I -5 mm. The umbo is at about the anterior one-fifth, it is
small and opisthogyrate. A hinge plate behind the umbo is just over one-tenth of the height of the shell high
and 013 of the length of the shell in length, and it tapers towards the posterior. It bears eight radiating,
possibly striated teeth which are longer towards the posterior. A single tooth lies anterior to the umbo.
Immediately behind the umbo is a faint, poorly-developed sulcus extending perpendicularly downwards and
fading rapidly. Ligament unknown but probably opisthodetic. An elongate anterior muscle scar lies
immediately anterior to the umbo with a strongly impressed dorsal edge and bounded on the anterior edge
by a rounded ridge running close to the anterodorsal margin. The posterior adductor muscle scar is unknown.
Ornament unknown.

Remarks. I know of no previous record of Pseudarca from Britain. Babin (1966, pp. 243-244,
pi. 9, figs. 4-6, 8) has described Siliquarca [= Pseudarca] typa Tromelin and Lebesconte, the type
species, and has placed several other French species in synonymy with it. However, P. typa has
almost twice as many posterior teeth as P. celtica and appears to taper more towards the posterior.
Despite the conventional view of the genus as a nuculoid (McAlester, 1968, p. 47; 1969, p. N234),
Babin (1966, p. 237) placed Siliquarca in the family Lyrodesmatidae and the features shown in
this species tend to corroborate his opinion.

Acknowledgements. Access to collections in their care and loans of material were kindly provided by Mr
R. J. Cleevly and Dr N. J. Morris (British Museum, Natural History), Mr M. Dorling (Sedgwick Museum,
Cambridge), and Dr R. M. Owens (National Museum of Wales). Dr Morris, Dr D. E. Butler, Dr A. W. A.
Rushton, and Dr J. Kriz discussed aspects of the paper and offered constructive advice. The paper benefited
from critical examination and constructive comment by Dr J. Pojcta and an unknown referee. Mr M. P.
Brett drew the map. This paper is submitted as a contribution to the British Geological Survey, North Wales
Regional Survey Programme, and is published with the permission of the Director, BGS (NERC).

REFERENCES

BABIN, c. 1966. Mollusques bivalves et cephalopodes dti Paleozoique armoricain, 412 pp., 18 pis. Brest.

BILLINGS, E. 1858. New genera and species of fossils from the Silurian and Devonian formations of Canada.
Canadian Naturalist and Geologist, 3, 419 444.

BRENCHLEY, p. J. 1978. The Caradocian rocks of the north and west Berwyn Hills, North Wales. Geological
Journal, 13, 137-164.

HALL, J. and WHITFIELD, R. p. 1875. Descriptions of Silurian Fossils. Geological Survey of Ohio, 2 (2), 65 161,
pis. 1 9.

HIND, w. 1910. The lamellibranchs of the Silurian rocks of Girvan. Transactions of the Roved Society of
Edinburgh, 4,1 O), 419 54%, ph. \ 5.

LILJEDAHL, L. 1984. Silurian silicified bivalves from Gotland. Sveriges Geologiska Undersokning, 78, I 82.

MCALESTER, A. L. 1968. Type species of Paleozoic Nuculoid bivalve genera. Memoir of the Geological Society
of America, 105, 1 143, pis. 1 36.

MATTHEWS, s. c. 1973. Notes on open nomenclature and on synonymy lists. Palaeontology, 16, 713-719.

1969. In MOORE, R. c. 1969. Treatise on invertebrate paleontology, pt. N. Mollusca 6, Bivalvia, N1 N952.

Geological Society of America and University of Kansas Press, Boulder, Colorado and Lawrence,
Kansas.

M'COY, E. 1851. On some new Silurian mollusca. Annals and Magazine of Natural History, Ser. 2, 7, 45-63.

1852. In SEDGWICK, a. and m'coy, f. 1851 1855. .4 synopsis of the classification of the British Pcdaeozoic

rocks, with a systematic description o f the British Palaeozoic fossils in the geological Museum of the University
of Cambridge, xcviii -1-661 pp., 25 pis. London and Cambridge.

PHILLIPS, J. 1848. Memoir on the Malvern Hills compared with the Palaeozoic Districts of Abberly etc.
Memoir of the Geological Survey G.B. 2, 1 386, pis. 1-30.

POJETA, J. 1971. Review of Ordovician Pelecypods. Professional Papers of the U.S. Geological Survey, 695, 1
46, pis. 1 20.

1978. The origin and early taxonomic diversification of pelecypods. Philosophical Transactions Royal

Society London, B 284, 225-246, pis. 1-15.

1985. Early evolutionary history of diasome Mollusks. In broadhead, t. w. (ed.). Mollusks: notes for

a short course, 102 130. Paleontological Society, Orlando, Florida.

690

PALAEONTOLOGY, VOLUME 30

POJETA, j. and RUNNEGAR, B. 1985. The early evolution of diasome molluscs. The Mollusca, 10, 295-336.

RAMSEY, A. c. 1866. The Geology of North Wales. Memoir of the Geological Survey G.B. 3, 1-381, pis. 1-28.

REED, F. R. c. 1905. New fossils from the Haverfordwest district. V. Lamellibranchiata. Geological Magazine,
5 (2), 492-501, pi. 24.

ROSS, R. J. and 27 others, 1982. The Ordovician System in the United States: correlation chart and explanatory
notes. I.U.G.S. Publication no. 12, 73 pp., 3 loose tables.

TROMELiN, G. DE, and LEBESCONTE, p. 1875. Note sur quelques fossiles des gres siluriens de Saint-Germaine-
sur-Ille etc. Compte rendu du Congres de T Association Francaise pour f Avancernente des Sciences en 1875,
1876, 683-687. [Reprinted from a privately published pamphlet: ^V/e McAlester 1968, p. 65.]

TUNNiCLiFF, s. p. 1982. A revision of late Ordovician bivalves from Pomeroy, Co. Tyrone, Ireland.
Palaeontology, 25, 43-88, pis. 7 13.

ULRICH, E. o. 1893 [1895]. New and little known Lamellibranchiata from the Lower Silurian rocks of Ohio
and adjacent states. Report of the Ohio Geological Survey, 7 (2), 627-693, pis. 45-56.

1894. The Lower Silurian Lamellibranchiata of Minnesota: from vol. 3 of the hnal report, Minnesota

Geological and Natural History Survey, 475-628, pis. 35-42. (Published under separate cover prior to the
entire vol. T. fide Pojeta 1971, p. 41.)

1897. The geology of Minnesota, vol. 3, pt. 2, of the final report: Paleontology, i-cliv+ 1081 pp., 48 pis.,

133 figs.

WEDD, c. B., KING, w. B. R. and WRAY, D. A. 1929. The geology of the country around Oswestry. Memoir of
the Geological Survey U.K., Sheet 137, i xix+ 1-234.

WHITTINGTON, H. B. 1938. Fauna of the Lluest Quarry, Llanfyllin, (Wattsella horderleyensis superzone) and
its correlation. Proceedings of the Geological Association, 49, 49-54.

1962- 1968. The Ordovician trilobites of the Bala area, Merioneth, Parts 1 4. Palaeontographical Society

[Monograph], 1 138, pis. 1-32.

WILLIAMS, A., STRACHAN, L, BASSETT, D. A., DEAN, W. T., INGHAM, J. K., WRIGHT, A. D. and WHITTINGTON, H. B.
1972. A correlation of Ordovician rocks in the British Isles. Geological Society, London, Special Report 3,
1-74.

WILLIAMS, H. s. and BRHGER, c. L. 1916. The fauna of the Chapman Sandstone of Maine, including descriptions
of some related species from the Moose River sandstone. Professional Papers of the U.S. Geological Survey,
89, 1-347, pis. 1-27.

WILSON, A. E. 1956. Pelecypoda of the Ottawa Formation of the Ottawa-St. Lawrence Lowland. Bulletin
Geological Survey Canada, 28, 1 102, pis. 1 9.

S. P. TUNNICLIFF

Biostratigraphy Research Group
British Geological Survey

Typescript received 12 August 1985 Keyworth

Revised typescript received 8 March 1987 Nottingham NG12 5GG

SKELETAL STRUCTURE, DEVELOPMENT AND
ELEMENTAL COMPOSITION OF THE
ORDOVICIAN TREPOSTOME BRYOZOAN

PERONOPORA

by DAVID R. HICKEY

Abstract. ED spectroscopy of Ca and other elemental densities indicate differences in the rates of secretion
and growth among crystallite ultrastructures, acanthostyles, the median lamina, zooecial wall, and stages of
ontogeny and astogeny in Peronopora. Growth rate (inversely proportional to Ca density) was highest within
the granular skeleton of the median lamina and ‘A-type’ acanthostyles (paurostyles). Growth rate decreased
exponentially from the median lamina through the recumbent zone, endozone, and exozone. Paurostyle cores
were deposited more rapidly than endozonal wall but less rapidly than recumbent zone wall and the median
lamina. Zones of disordered irregular crystallites and laminar growth alternate in zooecial wall axes, and
were respectively linked with increased secretion rates and increased episodicity during the cystiphragm/
mesozooecial tabulae emplacement cycle. Cu and Mn densities support the sequence of relative growth rates
inferred from the Ca data. Paurostyles and the median lamina contain more Cu than other structures. Mn
is also most abundant in the median lamina and declined monotonically with reduced growth rate. The
median lamina is structurally continuous with the basal lamina but was not secreted against cuticle. Formation
of the median lamina and paurostyle cores may be explained by differentiation of inner epithelium for high
rates of secretion.

Bryozoan skeletal ultrastructures, if not diagenetically altered, are most informatively viewed as
products of the physiology of growth (Sakagami et al. 1984; Sandberg 1977, 1983; Tavener-Smith
and Williams 1972). This study seeks to link differences in the skeletal ultrastructure of Peronopora
vera Nicholson and P. decipiens Rominger to the rates and modes of growth of astogenetic
sequences and skeletal structures. The ultrastructures of zooecial wall ‘flanks’ (that region of
proximally directed laminae lateral to the wall axis) and axes, acanthostyles (acanthopores), and
the median lamina (‘mesotheca’ of other authors; ‘median wall’ of Boardman 1983) are described
by means of SE microscopy and analyzed for compositional differences by means of ED
spectroscopy. Results of this study provide information about the relative growth rates of skeletal
structures and astogenetic zones which can be extended to other taxa and used as bases for
hypotheses concerning the evolution of skeletal structures.

The term ‘granular’ has been applied to two similar crystallite morphologies in many studies of
skeletal ultrastructure (e.g. Tavener-Smith and Williams 1972). Both types lack regular crystallite
surfaces. The term ‘granular’ is herein applied to large, approximately equidimensional, non-
tabular crystallites that appear optically hyaline and homogeneous. The term ‘irregular’ is applied
to smaller, non-tabular crystallites which are optically differentiable, irregular, and variable in
dimensions and shape.

Much of the ultrastructure of bifoliate Peronopora resembles that of other trepostomes described
by Tavener-Smith and Williams (1972), Tavener-Smith (1969o), Armstrong (1970), Bigey (1979,
1982), and Bigey and Lafuste (1982). However, the median lamina of bifoliate Peronopora is an
unusual and distinctive structure among trepostome genera; it is otherwise found only in Polyteicus
Pocta, Diplostenopora Ulrich and Bassler, Stenocladia Girty, Nipponostenopora Sakagami,
Petalotrypa Ulrich, rare variants of Amplexopora thomesi Ross, and Araxopora Morozova. A similar

(Palaeontology, Vol. 30, Part 4, 1987, pp. 691-716, pis. 78-79.|

© The Palaeontological Association

692

PALAEONTOLOGY, VOLUME 30

Structure is an important synapomorphy of bifoliate cryptostomes and many fistuliporines. Trepo-
stome median lamina ultrastructure has been studied only in Peronopora compressa Ulrich
(Tavener-Smith and Williams 1972) and Petalotrypa sp. (Bigey 1979). The astogenetic origin and
ultrastructure of the median lamina has an important bearing on the growth mode of trepostomes
(Tavener-Smith and Williams 1972) and the evolutionary origin of Peronopora (Hickey, in press).
Additional observations demonstrate that the median lamina of the primary frond is structurally
continuous with the basal lamina (‘epitheca’ of other authors) but is none the less interior-walled
skeleton. The median lamina contains style-like structures (Boardman and Utgaard 1966) which
optically resemble acanthostyle cores but differ in ultrastructure. The styles and granular layer
of the median lamina exhibit ultrastructural details as yet unreported in the mesothecal skeleton
of other taxa.

Few studies of the elemental composition of bryozoan skeletons have been undertaken. Phillips
(1922) reported Cu, Zn, Fe, and Mn in one Recent bryozoan. Clark and Wheller (1917) found Si,
Al, Fe, Mg, P, and S in a survey of 9 Recent bryozoans. Schopf and Manheim (1967) reviewed
the chemical composition literature and reported Ca, Mg, Sr, Ba, Fe, P, Zn, and C in 29 Recent
bryozoan species. Soule and Soule (1981) analyzed heavy metal uptake in 7 Recent bryozoa and
discovered traces of Fe, Cr, Hg, Ni, and Zn in exoskeletons, Fe and Hg in tissues and As, Cd, Cr,
Cu, Fe, Hg, Mn, Ni, Pb, and Zn in bulk samples. Morrison and Anstey (1979) provide one of the
few studies of the elemental composition of fossil bryozoans. They detected C, Fe, S, Si, Al, and
K in the brown bodies of Heterotrypa Nicholson and Peronopora.

Minor and trace elements, aside from Mg and Sr, have received little attention in biogeochemical
studies of bryozoans or other organisms. Most studies have sought correlations between composition
and temperature or salinity variation in bivalves (e.g. Dodd 1967; Eisma et al. 1976; and others),
or links between mineralogy and phylogenetic affinities among bryozoans (Schopf and Manheim
1967). A few studies have focused on ontogenetic variation of elemental abundances in bivalves
(Crisp 1975, 1983; Goreau 1977; Rosenberg 1980; Rosenberg and Jones 1975). The latter aspect
of skeletal chemistry could provide data pertinent to studies of the physiology of skeletal secretion
and growth in bryozoans. No compositional studies of this kind have been undertaken with fossil
bryozoans. However, Sakagami et al. (1984) have defined categories of skeletal composition
indicative of equilibrium and nonequilibrium skeletal formation in some Recent and fossil
bryozoans. They concluded that the presence of Mn, Fe, Zn, Co, Cu, and Ni within cheilostome
skeletons was a product of physiological control over skeletal composition.

Mg, Cu, Fe, Mn, Si, Yb, and Nb were found in P. vera Ulrich. This investigation is concerned
with the systematic variations of Ca and trace element concentrations during astogeny and
relative element abundances among major skeletal components and different microstructures. The
abundance distribution of Ca, usually ignored in compositional analyses (Rosenberg 1980), is of
particular interest in this study. Ca abundance could provide a measure of relative rates of skeletal
and crystallite growth (Rosenberg, pers. comm. 1984). EDS analyses of Ca indicate that relative
skeletal growth rates decreased with increasing age from the median lamina through the exozone.
Granular crystallite morphologies typical of the median lamina and acanthostyles are indicative
of higher growth rates than crystallites comprising laminar wall. The median lamina was deposited
most rapidly. The development of the median lamina may be explained in terms of the differentiation
of inner epithelium to produce regions of increased, and relatively continuous crystallite secretion.
Acanthostyles were deposited more rapidly than endozonal and exozonal skeleton. Differences in
individual crystallite shape, thickness, orientation, and laminae number suggest differential growth
rates and episodicity between zooecial wall flanks and axes and cyclical zonation within axes.
Exterior-walled skeleton (Boardman et al. 1983; ‘single-walled’ skeleton of many authors) of the
basal lamina is inferred to have been deposited more rapidly than most interior-walled (Boardman
et al. 1983; ‘double-walled’ skeleton of many authors) skeleton.

Astogenetic trends and differential abundances of trace elements could reflect the effects of
growth rates (Rosenberg 1980), or other aspects of physiological change during astogeny. The
abundance variations of Mn, Cu, and Fe among skeletal structures and astogenetic zones support

HICKEY: ORDOVICIAN TREPOSTOME BRYOZOAN

693

the inferences based on Ca abundance. Results support the inferences concerning relationships
between skeletal ultrastructures and growth rates proposed by Tavener-Smith (1969^/, b; 1975).

If similar crystallite morphologies in the interior-walled skeleton of other trepostome taxa are
not products of alteration, the evolution of differences in zooecial wall structure among astogenetic
zones and acanthostyle microstructures among closely related taxa could be interpreted in terms
of heterochronic processes. Studies of compositional differences among astogenetic growth zones
provide a basis for the construction of growth curves for taxa like Peroiwpora that possessed
periodically deposited intrazooecial structures such as cystiphragms. Results also suggest that the
evolution of mesothecal structures in other taxa involved a differentiation of inner epithelium
crystallite secretion rates along the colony margin.

MATERIALS AND METHODS

Specimens used in this study belong to P. vera and P. decipiens from the Kope (Eden Shale) and
Dillsboro Formations of the Cincinnatian Series (Late Ordovician) from the Ohio Valley region.
Several preparation methods were used in the SEM investigation. Specimens were either fractured,
or etched in 01% formic acid for 60-90 seconds or 1-25% EDTA for 30 minutes. Electron
micrographs of acetate peels were also made. Specimens were coated with gold-palladium or
carbon. Etched specimens were carbon coated for energy dispersive X-ray (EDS) analysis of
elemental composition.

Electron micrographs were taken with an ISI (International Scientific Instruments) S-III SEM
and the EDS analyses were made with a JSM-35C SEM, manufactured by Japan Electron Optics
Limited. The latter machine is equipped with a dual annular photolithographic disc back-scattered
electron detector and a Tracor Northern Energy Dispersive X-ray analysis system. EDS analyses
were done at 20 kv, over time intervals of 60 seconds. Beam strength was held constant to prevent
systematic bias in the relative abundance measures. Possible biases due to the variable surface
topography of etched specimens are believed to have been minimal (Flegler, Director, MSU
Electron Optics Center, pers. comm.). Analyses were limited to elements with atomic numbers
above that of carbon because specimens were carbon coated.

Two elemental analyses were made. Changes in element abundances were measured with a
narrow beam probe in a proximo-distal series of thirty-seven points taken alternately from laminar
regions of the zooecial wall axis, and from cystiphragms periodically emplaced along the wall of
a single zooecium. This analysis was designed to determine if there exist any elemental abundance
periodicities associated with the cycle of cystiphragm emplacement. A second analysis was
performed to determine if elemental abundances differed in systematic manners among astogenetic
growth zones and among different skeletal structures. Abundances were measured with a line scan
at magnifications of 1,000 to 9,000 for the exozone, endozone, acanthostyles, recumbent zone, and
median lamina for each of fifteen zooecia of a single specimen. Magnifications that differ by less
than an order of magnitude should not have a significant effect on measured abundances (Flegler,
pers. comm.). Only the latter analysis provides replicate data amenable to statistical treatment and
reliable inference. Mean elemental abundances were statistically compared among skeletal structures
and among astogenetic zones. The Student’s t-test was used to evaluate differences in the mean
abundances; it was preferable to other parametric tests because sample sizes were small (n^ 15).

Skeletal structures and different crystallite ultrastructures, as well as matrix, could respond
differentially to diagenesis. Trace element distributions may be greatly influenced by the distribution
and abundance of organic material in the earliest stages of diagenesis following burial. Mean
abundances within skeletal structures were statistically compared with those of the zooecial void
cement in order to determine whether diagenetic alteration significantly influenced skeletal
composition. Some idea of the effects and extent of diagenesis on the results of this study is
indicated by comparisons between element concentrations within the diagenetic zooecial void
matrix and individual skeletal elements. Several elements are significantly enriched within and
among various skeletal structures in comparison to mean levels found in the void cement.

694

PALAEONTOLOGY, VOLUME 30

Distributions of Ca and some of the trace elements are consistent with independent predictions of
compositional behaviour whieh would be expected in the relative absence, or limited influence of
diagenesis on skeletal material. Of these, Mn, Fe, and Cu occur in cheilostomes with nonequilibrium
growth (Sakagami et al. 1984) and are thus inferred to have been little influenced by diagenesis.
Yet, a differential response of skeletal materials to diagenesis and/or the effects of early-stage
diagenesis could have affected the observed variations and trends. Thus, any conclusions regarding
astogenetic trends and structural differences are tentative and subject to confirmation by additional
studies and alternative methods.

SKELETAL STRUCTURE

The autozooecia of bifoliate Peronopora are subcircular to circular in cross-section, short, and
arranged in fairly well defined longitudinal ranges. Recumbent zones are well developed where
autozooecia diverge from either side of a median lamina (if present) in bifoliate species (Boardman
and Utgaard 1966). The median lamina is a distinctive intracolonial structure, unusual among
trepostomes. Cystiphragms form continuous overlapping series throughout the autozooecial tubes.
The area subtended by cystiphragms decreases monotonically during autozooecial ontogeny.
Cystiphragm overlap increases distally as size and spacing decrease. Acanthostyles are generally
abundant, large, and may inflect autozooecial walls. The polygonal mesozooecia (mesopores) may
be abundant, large, and closely tabulated.

Zooecial walls

The majority of the interior-walled skeleton is laminar, like that typical of most trepostomes (e.g.
Armstrong 1970; Bigey 1979, 1982; Bigey and Lafuste 1982; Tavener-Smith 1969r/; Tavener-Smith
and Williams 1972). Autozooecial and mesozooecial wall laminae are of variable thickness, length,
and continuity (PI. 78, figs. 1 and 2). Exozonal laminae may occasionally extend the length of a
wall flank and intercalate with laminae from the adjacent wall in the compound wall axis. Crystallite
shape and discontinuity of laminae render it difficult to define discrete, continuous laminations.
The location of actual growth surfaces is therefore obscured.

Edgewise crystallite growth can also produce the appearance of a laminated wall without
synchroneity of laminae and growth surfaces (Boardman and Cheetham 1969; Boardman and

EXPLANATION OF PLATE 78

Figs. 1-8. Peronopora vera, Eden Shale, Ohio. 1, exozonal autozooecial wall; wall flank laminae (F) continuous
with laminae comprising cystiphragm (distal to outlined laminae) and larger core crystallites (C) proximal
and distal (arrows) to cystiphragm (bottom centre). Unusually long laminae and intercalation of laminae
are outlined in region of wall axis between zones of larger crystallites. Long., MSU 220335-00245, Eden
Shale, Ohio, x 1,000. 2, exozonal autozooecial wall. Large, irregular, disordered wall core crystallites (C)
(right) proximal to wall flank laminae (F) continuous with cystiphragm (bottom left, out of field). Long.,
MSU 220335-00251, Eden Shale, Ohio, x 5,000. 3, pluricupular zooecial wall flank crystallites. Fractured
section. Long., MSU 220335-00241, Eden Shale, Ohio, x 2,600. 4, fine ultrastructure of median granular
layer. Compare granular-rhombic crystallites of median lamina with pluricupular crystallites of zooecial
walls in fig. 3. Fractured section, Transv., MSU 220335-00241, x 1,000. 5, granular acanthostyle core,
pluricupular crystallites of thin laminae proximal to core and larger crystallites of laminar zooecial wall
to far left, right and top centre. Tang., MSU 220335-00245, Eden Shale, Ohio, x 1,000. 6, granular
crystallites of acanthostyle axial core (C) and pluricupular crystallites of sheath laminae (S). Fractured
section. Transv., MSU 220335-00241, x 1,500. 7, acanthostyle core and sheath laminae within autozooecial
wall. Long., MSU 220335-00245, x 400. 8, acanthostyle axial core (centre) and flanking laminae. Fractured
section. Transv., MSU 220335-00241, x 2,400.

Fig. 9. P. decipiens, median lamina; outer laminar layer (OLE), median granular layer (MGL), zooecial wall
(ZW), and median styles (MS; arrows). Transv., MSU 220314-00415. Dillsboro Fm., Indiana, x 240.

PLATE 78

HICKEY, Peronopora

696

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. I. Peronopora vera, mesozooecial wall (axis parallel to long
axis of photo) and adjoining tabula (bottom centre). Note disordered
wall core and laminar flanks. Acetate peel (negative) replica. Long.,
MSU 220335-00239, Eden Shale, Ohio, x 2,000.

Towe 1966). Armstrong (1970) suggested the possibility of edgewise growth in Stenopora crinita
Lonsdale. Undamaged distal tips of zooecial walls in Peronopora reveal continuity of laminae
across wall cores with no evidence of concerted edgewise growth. Therefore structural laminae and
growth surfaces are considered equivalent, and no evidence suggests non-laminar growth.

Wall laminae of cystiphragms and mesozooecial tabulae intercalate with zooecial walls and often
become attenuated and terminate before reaching the wall core (PI. 78, fig. 1). Zooecial laminae
are generally thickest, and crystallites largest and most irregularly shaped within wall axes proximal
to zones of laminae continuous with cystiphragms or tabulae (PI. 78, figs. 1 and 2). Laminae which
extend into zooecial walls from the distal surfaces of cystiphragms, and those typical of outer
zooecial wall flanks, tend to be thinner and composed of smaller, and thinner ‘pluricupular’ (PI.
78, fig. 3; see definition below) crystallites. A more pronounced alternation of axial crystallite size
and shape occurs within walls shared by mesozooecia (text-fig. 1).

Wall laminae within the endozone are few in number, steeply to vertically inclined, and arch
sharply over an undifferentiated, laminar wall axis (PI. 78, fig. 9). The laminae and component
crystallites throughout most of the endozone are pluricupular and somewhat larger than those of
the wall flanks within the exozone. The most proximal portions of the recumbent zone are composed
of granular crystallites which are short distal extensions of the median lamina (PI. 79, fig. 4).

Crystallites

Individual crystallites of zooecial walls do not display the planar surfaces or uniform thickness
and size of the tabular crystallites found in many Tubuliporata (formerly cyclostomes; Boardman
1983) and Trepostomata described by Brood (1976), Sandberg (1977), Tavener-Smith and Williams
(1972), and others. For example, the laminar wall crystallites of Stenopora (Armstrong 1970) and
Leioclema asperum Hall (Tavener-Smith 1969a) are more uniformly tabular than those of
Peronopora. Fractured sections (PI. 78, fig. 3) and light formic acid etches reveal that most
crystallites of Peronopora are pluricupular-shaped like those of Leptotrypella Vinassa (Bigey and
Lafuste 1982). Pluricupular crystallites are wafer-like tablets with undulating proximal and distal
surfaces and attenuated margins (PI. 78, figs. 1-3). Laterally adjacent crystals often exhibit
intercalated, interlocking boundaries. Exozonal axial crystallites are generally larger and more

HICKEY: ORDOVICIAN TREPOSTOME BRYOZOAN

697

irregularly shaped than those of wall flanks (PI. 78, figs. 1-3). Wall axes contain large, irregular
crystallites which may envelop smaller crystallites. The smallest crystallites could be relicts of
diflFerential etching in response to minor compositional differences. Large irregular crystallites
disrupt lamina continuity and thickness and give a generally disordered appearance to regions of
the wall axis deposited between cystiphragms and tabulae (PI. 78, figs. 1 and 2). Wall axes shared
by mesozooecia appear more disordered than those of autozooecia (text-fig. 1).

Acanthostyles

The acanthostyles of Peronopora have a paurostyle-like (terminology of Blake (1983) extended to
trepostomes) morphology (PI. 78, figs. 5-8). Sheath laminae (Blake 1983) adjacent to the
acanthostyle axial core are composed of nearly tabular crystallites. Crystallites become thicker,
longer, and more pluricupular away from the core (PI. 78, figs. 5 and 7). Crystallites in regions of
distally concave laminae between axial cores and zooecial walls are generally larger, longer, and
more tabular in shape than those of zooecial walls. Fractured sections reveal large granular
crystallites within the axial core (PI. 78, figs. 6 and 8). The boundary between cores and flanking
laminae is irregular yet distinct (PI. 78, fig. 7). Laminae do not extend across the cores and no
growth discontinuities were found within the core.

Median lamina

The median lamina varies in structure and continuity within and, possibly, between species. It
consists of one to three components: 1, an outer laminar layer; 2, a median granular layer; and,
3, granular ‘tubules’ or median styles (median ‘tubules’ or ‘rods’ of other authors) (PI. 78, fig. 9).
The outer layer contains a variable number of horizontally disposed laminae composed of small
pluricupular crystallites (PI. 79, figs. 2-6). The distalmost laminae are continuous with the zooecial
walls of the recumbent zone (PI. 79, figs. 4 and 5). These laminae arch over a short distal extension
of granular median lamina at the bases of zooecial junctions between adjacent walls. The median
lamina may be composed entirely of two outer laminar layers, particularly in very thin walled
regions and near the bases of secondary fronds. In the former case the median lamina is optically
visible as a parting plane between layers of laminar skeleton separating oppositely oriented zooecia.
For this reason the outer laminar layer is considered a component of the median lamina.

The median granular layer is optically hyaline, but granular in etched section (PI. 78, fig. 7; PI.
79, figs. 2-7). The granular layer may vary irregularly in thickness or become thin between regular
lensoidal swellings (PI. 78, fig. 9). The median layer is comprised of large, tightly interlocking
granular crystallites (PI. 79, figs. 2-6). The granules are composed of small, rhombic crystallites
(PI. 78, fig. 4; PI. 79, fig. 1). These crystallites are broadly similar to those of the exterior walls
(fixed-wall) of some cheilostomes (Sandberg 1983, figs. 121, 123-125) and the ‘secondary layer’ of
some tubuliporates (Brood 1976; Tavener-Smith and Williams 1972). The major dilferences between
these crystallites in Peronopora and those of the ‘secondary layer’ in tubuliporates are their lack
of organization into discrete laminae, poorly defined boundaries between individual crystallites,
and variable crystallite size in the former. Comparable ultrastructural detail could not be observed
in the ‘primary granular layer’ of tubuliporates or ptilodictyines figured by Tavener-Smith and
Williams (1972) or in the median lamina of the trepostome Petalotrypa sp. (Bigey 1979) or the
cystoporate Cystodictya (Healy and Utgaard 1979).

The granular layer may be either internally continuous (PI. 78, fig. 9), discontinuous, lensoidal
(PI. 79, fig. 6), or separated into segments by folding of the laminar layer around the distal end of
each segment (PI. 79, fig. 2). Segmentation of the granular layer indicates periods of slowed or
interrupted growth at the frond terminus. This segmentation demonstrates the direction of growth
of the medial layer and provides evidence about the location of the secretory epithelium. Regions
of discontinuity and/or thinning are often associated with unusually wide zones of thin-walled
(‘endozonal’) growth (text-fig. 2) and recumbent zooecia which lack cystiphragms (Boardman and
Utgaard 1966). These observations suggest a model to explain local discontinuities in median
lamina formation (see discussion). The median laminae of primary and secondary fronds are

698

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 2. Peroiiopora vera, discontinuity of median
lamina (ml) in region of prolonged endozonal growth
(e). Note endozonal (e)-exozonal (x) growth cycles,
growth check at former frond margin (large arrow),
absence of median lamina within endozonal region
and lack of cystiphragms within recumbent zooecia.
Long.-Transv., MSU 220314-00253, Eden Shale,
Indiana, x 13.

always discontinuous; they are separated by exozonal growth in the parent frond and thin-walled
growth at the bases of secondary fronds.

The rounded to elliptical, optically hyaline tubes described by Boardman and Utgaard (1966),
and herein termed median styles, display an ultrastructure identical to that of the granular layer
(text-fig. 3). The median styles are composed of tightly interlocking granules with an ultrastructure
identical to that of the median granular layer. Like acanthostyles and the median rods of
ptilodictyines (Karklins 1983), they were not hollow tubes, but continuously deposited, solid
structures. Similar structures occur in the mesotheca of the trepostome P. perforata Nekhoroshev
(Volkova 1974). In Peronopora they are generally centered on the junctions between adjacent
zooecial ranges and extend the length of one to several zooecial bases in longitudinal section.
Oblique sections through the mesostyles could produce the ‘mesothecal lenses’ (PI. 79, figs. 3 and
4) described by Tavener-Smith and Williams (1972). Where mesostyles occur within the granular
layer, they may be undifferentiated from it or separated by a thin parting. Median styles may
occur in the absence of the medial granular layer.

EXPLANATION OF PLATE 79

Fig. 1. Peronopora vera. Fine ultrastructure of median granular layer. Fractured section, Transv., MSU
220335-00241, x 1,000.

Figs. 2-6. P. decipiens, MSU 220314-00415, Dillsboro Fm., Indiana. 2, Growth check in median lamina
showing outer laminar layer (OFF) folded around distal tip of median granular layer (MGL). Long.,
X 720. 3, lensoidal region (median style (MS)) of median granular layer (MGL) between compound

zooecial walls (ZW). Outermost laminae of outer laminar layer (OFF) continuous with recumbent zooecial
wall (centre, top, and bottom). Transv., x 940. 4, elliptical median style. Zooecial wall at top right.
Labelled as above. Long., x 1,300. 5, median lamina; pluricupular crystallites of outer laminar layer

(OFF) and median granular layer (MGL). Transv., x 3,200. 6, median lamina. No medial parting present.
Note thinning of median granular layer at right. Labelled as above. Transv., x 660.

PLATE 79

HICKEY, Peronopora

700

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 3. Peronopora vera, fine ultrastructure of median style.
Fractured section, Transv., MSU 220335-00241, x 200.

TEXT-FIG. 4. Reconstruction of the median lamina showing distribution of median styles, lensoidal
swellings, and local thinning of medial granular layer.

HICKEY: ORDOVICIAN TREPOSTOME BRYOZOAN

701

The distribution of mesostyles and thickness variation within the median lamina suggest that
the mesostyles and lensoidal regions served as space-filling structural supports between adjacent
zooecia (PI. 78, fig. 9). Tavener-Smith (1969b) suggested that median rods in fenestellids may have
served to anchor the median lamina skeleton to the outer epithelium or cuticle. The median styles
of Peronopora could have assumed a similar function, although evidence indicates (see below) that
they were secreted by inner, not outer epithelium, and thus not directly associated with outer
epithelium or cuticle. Short discontinuities of the median layer may represent local thinning or
non-deposition of the median granular layer beneath recumbent zooecial bases rather than an
actual discontinuity of the skeletal layer. A three-dimensional reconstruction of the median lamina
is illustrated in text-figure 4.

TEXT-FIG. 5. A, Peronopora decipiens, colony base showing median lamina (arrows), basal
lamina not preserved, MSU 220314-00415, x 5. b, P. decipiens, colony base showing
junction of median lamina (arrows) and basal lamina. Latter not preserved, but presence
indicated by epizoans underlying (on opposite side of) preserved ornamentation of bivalve,
upper surface of which was encrusted by young Peronopora, MSU 220314-00486, Dillsboro

Fm., Indiana, x 5.

An understanding of the skeletal growth of Peronopora depends on the relationship between the
exterior-walled basal lamina, median lamina, and interior-walled skeleton (Tavener-Smith and
Williams 1972). Complete basal portions of early colonies demonstrate that the median lamina
was continuous with the basal lamina of the primary frond (text-fig. 5). The median lamina arose
after the formation of an elliptical ancestrular disc (Hickey, in press). Completion of the disc was
followed by reorientation of the zooecia at the proximal and distal margins of the disc so that the
normally basal portions of adjacent zooecial tubes became lateral walls separating oppositely
diverging zooecia. The median lamina was formed by vertical extension of the basal lamina
following the completion of the recumbent zones of these zooecia. Thus the median lamina was
generated at the terminal margins of the primary frond and arched over the ancestrular disc with
continued vertical frond growth. Text-figure 6 illustrates the primary features of early astogeny
and relationships among the major skeletal structures in bifoliate Peronopora.

702

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 6. Growth model for bilaminate species of bifoliate Peronopora. Left, reconstruction of
relations between median and basal laminae and ancestrular disc; Centre, relationships between
epithelial layers and skeletal layers, median lamina and basal lamina; Right, detail of growth model
showing region of epithelial differentiation for high and reduced growth rates, granular and laminar

skeleton.

Continuity of the median and basal laminae implies, but does not require, deposition in contact
with the cuticle. A tendency for fractured specimens to split along separate linear series of zooecia
or lamina could be taken as evidence of deposition against a cuticle (Sandberg 1977). Although
zoaria of Peronopora readily split along the median lamina, the parting tends to be irregular; it
usually passes along alternate sides of the central layer rather than through its centre. The inner
surface of the median lamina does not exhibit an ultrastructure comparable to the planar spherulitic
crystallites of fixed-walled skeleton in cheilostomes (Sandberg 1983).

Tavener-Smith and Williams (1972) concluded that the median lamina of P. milleri was not
deposited against a cuticle because the median lamina consisted of discontinuous ‘mesothecal
lenses’ with no evidence of a medial parting. This study found no ultrastructural evidence of a
medial parting within the continuous medial layer (PI. 78, fig. 9; PI. 79, figs. 5 and 6) despite the
common occurrence of a dark medial line in thin section. Folding of laminar internal-walled
skeleton around the distal ends of median granular layer ‘segments’ (PI. 79, fig. 2) indicates that
a secretory epithelium of the median lamina could not have been located between the median
granular layer and the outer laminar layer as suggested by Tavener-Smith and Williams (1972).
The discontinuity of the median laminae of primary and secondary fronds of Peronopora also
shows that median lamina skeleton could not have been deposited against cuticle or within a fold
of (basal) epithelium. Thus, the median lamina was continuous with the upper portion of the basal
lamina but must have been deposited by inner epithelium at the growing margin of the colony
(text-fig. 6). The inner epithelium appears to have been capable of secreting skeleton which was
ultrastructurally very similar or identical to exterior-walled skeleton (see discussion).

HICKEY: ORDOVICIAN TREPOSTOME BRYOZOAN

703

ELEMENTAL COMPOSITION
Elemental rhythms and ontogenetic trends

Elemental abundances in some organisms other than bryozoans are known to vary with tidal
(Rosenberg and Jones 1975), diurnal (Schmalz and Swanson 1969), or seasonal cycles (Bryan 1973;
Goreau 1977; Crisp 1975). If the physiological process of skeletal growth do interact with ambient
levels of trace elements in the environment, or if non-equilibrium modifications in elemental
composition occurred in conjunction with the periodic formation of structures, such processes could
be evinced in abundance cyclicities. Cystiphragms are periodically formed structures which could
have been deposited in association with tidal cycles (Bartley and Anstey 1983). A single zooecial
wall of P. vera was analysed for periodic abundance fluctuations of Ca, Cu, Si, Fe, Mg, Mn, and
Nb between the laminae of cystiphragms and intervening regions of laminar zooecial wall deposited
between cystiphragms. This analysis was designed to determine whether elemental abundances
were distributed in a systematic manner in association with the periodic emplacement of
cystiphragms. With the possible exception of Ca, none of the elements displayed periodic structural
variations.

There is some indication of a periodic fluctuation of Ca within the endozones and exozones of
the autozooecial wall. Ca abundance changed in a systematic manner with ontogeny (text-fig. 7).
Minimal values in the thin-walled recumbent zone increased abruptly during transition from the
recumbent zone to the lower endozone. Abundance gradually decreased from the endozone through
the latest exozone. Within the exozone, Ca levels of cystiphragms and contiguous zooecial wall
laminae are generally lower than those of intervening laminae of axial positions in the wall, but
the difference is not statistically significant. This structural fluctuation is also present, though
weakly developed, within the endozone.

Astogenetic Stage

TEXT-FIG. 7. Distribution of Ca in zooecial wall of Peronopora vera. MSU 220335-00252. RZ
= recumbent zone. ‘Astogenetic Stage’ approximated by datum points located at equal
intervals, alternately within wall axis and on each successive cystiphragm wall.

704

PALAEONTOLOGY, VOLUME 30

Astogenetic Stage

TEXT-FIG. 8. Distribution of Ca among median lamina
(ML), recumbent zone (RZ), endozone (ENDZ),
exozone (EX), and acanthostyles (A, circled) for
fifteen zooecia of Perouopora verci. MSU 220335-
00252. Circled datum indicates Ca density value
differing significantly from void cement. C.L= con-
fidence interval for Ca density levels of zooecial void
cement. Solid line: Ca density distribution. Dashed
line: relative growth rates of astogenetic zones and
skeletal structures.

Rosenberg (1980) has stressed the need for replicated analyses in studies of this kind. His review
of similar studies of bivalves noted that element abundances may vary among planes of cross-
section and among individuals, suggesting that allometric analyses of ontogenetic trends may
explain much of the variation in compositional trends. Thus, although the above results are
interesting, no general inferences should be drawn from them. In contrast, the second analysis is
based on the mean values of Ca and trace element abundances from three astogenetic zones,
acanthostyles, and the median lamina measured in 15 individual autozooecia of a single colony.
Statistical treatment of these data permits inferences about the relationships between compositional
change and growth. The difference in exozonal Ca abundance between the former and following
analyses were probably due to the absence of replicate data from other zooecia and the collection
of data from points of very limited surface area (small beam size) in the former.

Astogenetic trends

Analysis of fifteen zooecial walls revealed a Ca trend that is broadly similar to that of text-figure
7, but differs critically in the relative values of the endozone and exozone. Ca levels are minimal
in the median lamina and become progressively greater in the recumbent zone, endozone, and
exozone (text-fig. 8). All pair-wise comparisons between astogenetic zones and other structures
show significant differences in Ca abundance. Ca abundance within the exozone is significantly
greater than that of the endozone at p = 010. Acanthostyle axial cores contain significantly less
Ca than the recumbent zone, endozone and exozone (p < 0 05), but significantly more Ca than
the median lamina (p < 0-05). The median lamina contains significantly less Ca than acanthostyles
and astogenetic zones of zooecial wall skeleton (p < 0 05).

Ca densities of the endozone and exozone do not differ significantly from that of the zooecial
void cement. Thus it is possible that Ca abundance within void matrix could have varied in a
proximo-distal manner similar to that of the skeleton. However, some aspects of the Ca distribution
suggest that the observed trend is not a product of diagenesis. Significant differences (p < 0 05)
do occur among the void cement and acanthostyles, recumbent zone and median lamina. Ca
concentration within void cement was measured at approximately the mid-point of each zooecial
tube; a position equivalent to that of the inner exozone. Acanthostyle Ca abundance was measured
within the exozone, yet is significantly lower than that of the matrix, exozonal, and endozonal
wall. In addition, petrologic criteria indicate that the recumbent zone of Perouopora and other

HICKEY: ORDOVICIAN TREPOSTOME BRYOZOAN

705

ML RZ ENDZ EX ML RZ ENDZ EX

Astogenetic Stage Astogenetic Stage

TEXT-FIG. 9. Trace element distributions in Peronopom vera. MSU 220335-00252. Labelled as for text-
fig. 8. Circled points indicate densities significantly dilferenl from void cement.

taxa is less-well calcified than endozone or exozone (Boardman and Utgaard 1966). Thus, it
appears that the trend in skeletal Ca abundance is not a product of diagenesis. Additional studies
of this kind are needed to confirm or reject the results of this study.

Mg abundance declined slightly over the course of astogeny in P. vera (text-fig. 9). Because Mg
readily substitutes for Ca, it is expected to vary inversely with Ca. The overall trend is toward
decreasing Mg concentration, yet Mg is most abundant within the median lamina and endozone,
and least abundant in the recumbent zone and exozone. Mg concentrations of the median lamina
and endozone are significantly greater than that of the void cement (p < 0-05).

Mn concentration decreased monotonically with autozooecial age in P. vera (text-fig. 9). Mn
abundance is inversely proportional to Ca density across astogenetic zones, a relationship which
could reflect elemental substitution. Mn concentrations within the median lamina and recumbent
zone are significantly greater than that of the void matrix (p < 0 05).

Cu and Fe exhibit parallel fluctuations over the course of astogeny; both generally decline with
increasing age (text-fig. 9). Fe and Cu reach maximal abundance in the recumbent zone and
acanthostyles respectively. The minimum values for Fe and Cu occurred in acanthostyles and the
endozone respectively. Levels of Fe in the endozone and acanthostyles are significantly less than
that of the void cement. Cu abundances in the endozone and exozone are significantly lower than

706

PALAEONTOLOGY, VOLUME 30

that of the void cement. Cu concentration in acanthostyles is significantly greater than that of the
void cement. Si gradually increased with age through the endozone but declined somewhat within
the exozone. Si concentration is significantly less abundant in the median lamina than in the void
cement. Yb concentration remained essentially constant with increasing age but differed significantly
from the void cement in the endozone.

Structural variation

Acanthostyle cores contain significantly more Cu than the median lamina, endozone, or exozone.
The concentrations of Fe, Cu, Si, and Yb in acanthostyles differ significantly from that of void
cement. The median lamina contains significantly more Mn than the zooecial wall skeleton and
void cement. Si and Mg concentration are also significantly greater in the median lamina than in
the void cement.

DISCUSSION

Astogenetic growth zones

Tavener-Smith (1969c/, 6, 1975) and others have proposed that differences in crystallite morphology
reflect differences in the continuity and rates of crystallite and skeletal secretion and growth among
astogenetic zones. These inferences were based on consideration of the effects of the thermodynamics
of crystallite growth on ultrastructure, the differential response of skeletal structure to episodic
versus continuous growth, and the astogenetic succession of crystallite/skeletal wall morphologies.
These hypotheses of growth rate are important for two reasons. Firstly, they can increase our
understanding of growth, development, and skeletal formation. Secondly, knowledge of the relative
growth rates of skeletal structures can be used to explain the evolution of zooecial wall and other
skeletal structures. The results of elemental analyses (particularly Ca) in Peronopora support the
inferred associations between crystallite/skeletal morphology and secretion rates. The data also
support inferences about the relative rates of growth among astogenetic zones.

If it can be assumed that EDS measurements of Ca abundance reflect the density distribution
of calcium within the skeleton (and that this distribution is not a result of diagenesis), relative
levels of Ca can be used as a gauge of relative growth rates. It is assumed that the relative
abundance values obtained from a given surface area can be extrapolated to proportional values
of density when comparing surfaces of approximately equal area. Rosenberg (1983) argued that
this assumption was a valid interpretation of the distribution of Ca in fossil brachiopods. This
interpretation is based on the thermodynamics of crystal growth which predict an inverse
relationship between growth rate and Ca density. Thus zones of continuous and rapid growth
should display lower Ca densities than zones of slower, laminar (episodic) growth. An equivalent
relationship should exist among Ca densities and astogenetic zones of predominately laminar
growth. Less ‘densely’ laminated recumbent zone and endozone should display lower Ca densities
than thicker-walled exozone. In addition, granular and irregular crystallites should contain less Ca
than laminar skeleton and tabular crystallites. A corollary of this relationship predicts that trace
elements which commonly substitute for Ca should be greater within skeleton with lower Ca
densities and by inference, higher rates of crystallite secretion and skeletal growth. Those which
easily complex with organic materials could be expected to be more common in more densely
laminated wall and/or zones of disordered growth in wall axes.

The thin-walled recumbent zone and endozone are commonly believed to have been faster
growing than the thick-walled exozone. The relative densities of Ca in these astogenetic zones are
in accord with theoretical predictions of Ca density behaviour with growth rate variation (text-fig.
4). These data also support the petrologic evidence that autozooecial recumbent zones in Peronopora
were lightly calcified (Boardman and Utgaard 1966). Therefore Ca density data support the
hypothesis that growth rate decreased exponentially with increasing age from the recumbent zone
through the exozone in Peronopora. The recumbent zone formed more rapidly than the endozone

HICKEY: ORDOVICIAN TREPOSTOME BRYOZOAN

707

TEXT-FIG. 10. Growth curves based on cy-
stiphragm distribution per unit wall length for
Peronopora vera (MSU 220314-00508, Eden
Shale, Indiana) and P. paiica (paratype,
IU8252G 160-1170, Indiana University speci-
men, Whitewater Fm., Indiana).

and the endozone formed more rapidly than the exozone. The results also allow inferences about
the relative growth rates of other structures. The median lamina was secreted most rapidly, followed
by acanthostyle cores and the recumbent zone. Granular crystallites of optically hyaline skeleton
generally reflect higher rates of secretion and growth than tabular crystallites within internal-walled
skeleton. Presumably, these results could be extrapolated to other taxa which show various modes
of skeletal differentiation within and among recumbent/endo- or exozones, mesothecal, and exterior-
walled skeleton. Further EDS analyses on other taxa should provide interesting data on bryozoan
growth rates.

If rates of wall growth gradually decrease from the recumbent zone through the exozone, and
if cystiphragms or other intrazooecial structures were formed with a regular periodicity (Bartley
and Anstey 1983) then it is possible to construct growth curves for bryozoans. The area subtended
by cystiphragms and the spacing between cystiphragms decreases (degree of cystiphragm overlap
increases) from the recumbent zone through the endozone in bifoliate Peronopora (Hickey, in
press). Spacing is minimal and appears to remain relatively constant throughout the exozone. The
size/spacing distributions are consistent with the assumption that cystiphragms were formed with
a regular periodicity while growth rates of zooecial wall decreased with age. Thus, a plot of the
cumulative number of cystiphragms per unit length of zooecial wall should reveal changing rates
of wall growth. Text-figure 10 illustrates preliminary results of such growth curves for single
colonies of P. vera and P. pauca. That these growth curves are of the same shape as that of the
Ca density data (and the inverse of the Ca growth rate curves) lends support to the hypotheses
that zooecial wall growth rates generally decreased with increasing age. Knowledge of the actual
time lapse (e.g. fortnightly cycles, etc.) between cystiphragm formation is not necessary for such
a reconstruction of relative growth rates. The growth curves for P. pauca and P. vera suggest that
the growth rate of the former species declined more rapidly than that of the latter; a conclusion
consistent with the thicker exozonal walls of the former species. The ability to construct relative
growth curves could add to the understanding of the palaeobiology and systematics of bryozoans.
Such growth curves could be used as taxonomic descriptors.

Trace element data may also have a bearing on interpretations of skeletal growth rates. Several
studies have found that Mg and Mn concentrations can vary with growth rate. Goreau (1977)

708

PALAEONTOLOGY, VOLUME 30

found an increase in Mg with decreasing growth rate in the coral Montastrea annularis. However,
Zolotarev (1974) documented a decline in Mg with reduced growth rate in the bivalve Mytilus
yessoensis. A gradual increase with age and reduced growth rate is evident in the data of Moberly
(1968) for the bivalve Aequipecten irradians. Crisp (1983) reported a decreasing Mn concentration
with increasing age in freshwater bivalves. However, he also found that Mn coneentrations were
greatest in the slowest growing portions of the oldest growth increments. Crisp (1975) found an
apparent habitat dependence in Mn ontogenetic trends; Mn decreased with increasing age among
individuals of the bivalve Lanipsilis sp. from one habitat, but increased with age among individuals
from a different habitat. The bivalve Anodonta corpiilenta displayed an increase in Mn with
inereasing age.

The monotonic decrease of Mn density with increasing age in Peronopora is consistent with the
expected distribution of trace elements with respect to Ca and inferred growth rates (text-fig. 9).
The Mn trend could reflect a direct physiological function of skeletal growth rates and/or a function
of elemental substitution for Ca. Substitution is expected to have been greater in faster growing
zones of lesser Ca density such as the median lamina and recumbent zone. However, simple
substitution is probably an inadequate explanation for the Mn trend given that the incorporation
of Mn in cheilostome skeletons is under physiological control (Sakagami et al. 1984). Mg also
shows a slight decrease in density with increasing age and decreasing growth rate; however, these
data are much less definitive. Mg concentration is expected to have been a simple equilibrium
function of crystallite growth rate (Sakagami et al. 1984).

The meaning of the astogenetic trends of other trace elements is not entirely clear. Each of the
elements analyzed show significant density differences between one or more astogenetic zones and
the void cement. Some of the variation could be attributed to substitution for Ca and thus, could
be interpreted as rate related phenomena. The data suggest that some elements eould have been
preferentially concentrated/excluded at various stages of growth or among skeletal structures. By
analogy with the composition of cheilostomes (Sakagami et al. 1984), the abundances of Fe, Cu,
and Mn eould represent products of differential concentration or exclusion through physiological
control over ambient levels. Alternatively, the data could indicate that astogenetie stages and
skeletal structures were differentially influeneed by diagenesis as a funetion of crystallite type
and/or ultrastructural variation within laminated zooecial wall. Additional studies could conceivably
determine whether such results reflect ontogenetic variation, differential response to diagenesis, or
the influence of environmental factors.

The differences in Ca density among growth zones of laminar skeleton, crystallite types and
skeletal structures in P. vera are consistent with independent inferenees about relative rates of
skeletal growth based on observations of microstructural variation in this and other studies. Such
compositional variations appear to reflect real differences in the physiology of skeletal growth. The
distributions of most trace elements are not as easily interpreted as is that of Ca. However, the
distributions of Ca, Cu, and Mn among crystallite types are generally consistent with predictions
of thermodynamics and a variety of qualitative observations and may be extended to skeletal
growth rates during astogeny.

Zooecial wall growth

Tavener-Smith and Williams (1972) suggested that differences between thicknesses of the laminae
of wall axes and flanks within some ‘amalgamate’ trepostomes are due primarily to oblique
orientations of laminae within the axial zones. The ‘bimodal’ nature of axial crystallite dimensions
and lamina frequencies in Peronopora produce indistinct zones of laminae which resemble the
independent wall units discussed by Boardman (1960). These zonal microstructural variations
indicate that axial crystallite shape and thickness differences are not entirely artefacts of lamina
orientation. Zonal differences in erystallite morphology and lamina number could be characteristic
of taxa with cyclical growth rates.

Some aspects of the central wall strueture of the Recent tubuliporate Heteropora pelliculata
Waters (Ross 1976, 1977) provide a developmental analogue for interpretation of differential wall

HICKEY: ORDOVICIAN TREPOSTOME BRYOZOAN

709

TEXT-FIG. 11. A, zooecial wall structure of Heteropora pelliculata, decalcified long. sec. Redrawn from
photograph in Ross (1976). b, Peronopora vercu exozonal autozooecial wall. Disordered central wall structure
(C), laminar wall flanks (F). Acetate peel. Long., MSU 220335-00239, Eden Shale, Ohio, x 3,000.

flank/axis ultrastructure in Peronopora and similar trepostomes. Distal portions of zooecial walls
of H. pelliculata display irregularly shaped, disordered crystallites in a relatively dense organic
matrix within central longitudinal ‘canals’ (Ross 1976, text-fig. 1). Ross (1977) suggested that the
‘canals’ may serve as loci for locally accelerated growth and conduits for the resorption and
distribution of skeletal and other materials. Acetate peels (negative replicas) of the axial zone of
zooecial walls in Peronopora bear a striking textural resemblance to the distribution of the organic
matrix in the interior of the longitudinal canals of decalcified specimens of H. pelliculata (compare
text-fig. 11a, h). It is not suggested that the axial zones in Peronopora contained longitudinal
‘canals’ as Ross proposed for PI. pelliculata and it should also be noted that the zooecial wall
flanks of H. pelliculata are proximally oriented, unlike those of Peronopora. However, the textural
similarity between the axial zones of the two taxa strongly suggests that the axial zones of
Peronopora could have been characterized by relatively rapid, disordered crystallite growth and
higher organic content similar to H. pelliculata. The dynamics of crystallite growth in general
suggest that irregularly shaped, disordered crystallites were more rapidly formed than those with
uniform shapes (Tavener-Smith 1969/)). This interpretation is supported by Ca data which indicates
that the similar granular crystallite morphologies of the median lamina and acanthostyle cores
were a product of increased rates of crystallite secretion.

Large, granular crystallites also occur between the laminar annular thickenings (monilae) of
Stenopora crinita (Armstrong 1970) and the laminar monilae of Tahulipora Young. Evidence of
fortnightly tidal cycles has been found within the monilae of Tahulipora (Rabbio and Regalbuto
1985). Constricted zones of granular zooecial wall occur between the fortnightly cycles recorded
within the monilae of Tahulipora. The limited width and granular morphology of the constricted

710

PALAEONTOLOGY, VOLUME 30

regions suggests that these portions of the wall were formed more rapidly than the laminar monilae.
Thus wall growth in Tahulipora appear to have varied between slow, laminar (episodic) and rapid,
continuous modes in conjunction with tidal cycles.

Microstructural zonation of the crystallite morphology of Peronopora is inferred to reflect
periodic variations in growth rates and episodicity of laminae formation. Although most individual
laminae are difficult to trace for long distances within a wall, the number of laminae intersected
by a line drawn parallel to the zooecial wall growth axis can provide rough estimates of laminae
frequency. Counts of laminae within a given unit wall length made by independent observers differ
significantly from the null hypothesis of no difference in means. However, differences in laminae
frequency between axial and flank regions within a given unit length differ significantly (p < 0-05)
despite the variation attributable to observational error. Lamina frequency consistently differ by
a factor of two or more between zooecial wall flanks (mean = 19-3) and axes (mean = 27-4) per
unit (distal) length. In addition, laminae appear to be at least twice as numerous— per unit wall
length — within the laminar zones of wall axes than within zones of irregular axial crystallites. Axial
laminae within intervening zones of larger, irregular crystallites are often less distinct than those
of wall flanks and laminar zones of the axis. These frequency distributions of laminae reflect a
greater degree of intercalation between the laterally discontinuous laminae of axial zones than
among laminae in the adjacent wall flank, and an alternation between laminar and irregular
crystallite growth within the wall axis. It is inferred that the episodicity of laminae formation was
greater (and rate of skeletal formation thus lower) within laminar portions of axial zones than on
wall flanks. Irregular crystallite morphology can be attributed to increased rates of crystallite
secretion and decreased episodicity by inference from the similarity between irregular and granular
crystallite ultrastructures.

The irregular and disordered crystallite morphology and approximate doubling of laminae
number within different zones of the zooecial wall axis of Peronopora strongly suggests periodic
variation of growth rate and episodicity within the axis. Structural differences between wall axes
and flanks suggest periodic alternations between increased growth rates and increased episodicity,
respectively. Structural zonation is most pronounced within mesozooecial walls. Zones of irregular
crystallites within the axial wall occurring proximal to cystiphragm/tabula emplacements could
indicate increased rates of crystallite secretion preceding cystiphragm/tabula formation. The
increased frequencies of laminae composed of more tabular-shaped crystallites occur in zones
which are contiguous with cystiphragm and tabula walls. This structural variation suggests slower
rates of crystallite secretion and increased episodicity of lamina formation during periods of
cystiphragm formation.

Periodic variation in ultrastructure could be linked to periodicities of element abundances.
Rosenberg and Jones (1975) found regular fluctuations in Ca and S abundance in living specimens
of the bivalve Cardiiim edule which were correlated with tidal cycles. Individual lamina of bryozoan
skeletal walls are presumed to be daily deposits (Tavener-Smith 1969/?; Rabbio and Regalbuto
1985). Laminae are nested within periodic cystiphragm emplacements; events of a higher temporal
order that could be linked with tidal cycles (Bartley and Anstey 1983). Variation in Ca density
could be expected in association with the structural variation of zooecial wall which occurred
during the cystiphragm formation cycle in P. vera. The observed variation of Ca density between
periods of cystiphragm emplacements and normal wall growth in P. vera could reflect cyclic growth
phenomena associated with periodic changes in environmental variables (text-fig. 7). However, the
data are equivocal because the difference between mean Ca levels in cystiphragms and zooecial
walls is not statistically significant. The results do indicate that further studies of relationships
between struetural periodicities, Ca variations and growth rhythms are warranted.

Evolution of wall structure

If erystallite morphologies and astogenetic zones of growth can be linked to differential growth
rates, the evolution of wall structure among closely related taxa may be described in terms of
heterochronic change. Tavener-Smith (1969r/) and Brood (1976) have noted the presence of short

HICKEY: ORDOVICIAN TREPOSTOME BRYOZOAN

711

distal extensions of granular exterior-walled skeleton at the base of zooecial walls in free-walled
taxa. Larwood and Taylor (1979) proposed that the evolution of free-walled taxa from primitive
fixed-walled taxa in the Palaeozoic could be interpreted in terms of paedomorphic heterochronic
changes which prevented the coalescence of free-walled skeleton with outer epithelium. The early
termination and/or reduced growth rate of the granular external-walls could have been the principal
heterochronic change leading to the origin of paedomorphic free-walled taxa. This interpretation
of skeletal growth rates supports the hypothesis of Larwood and Taylor (1979). Given either early
termination (progenesis) or reduced growth rate (neoteny), the phylogenetic phenomenon would
have been reverse recapitulation (terminology of Alberch et al. 1979).

Tavener-Smith and Williams (1972) noted that the optical differences between the zooecial wall
microstructures of the former trepostome suborders Amalgamata and Integrata reflect the increased
‘granularity’ (irregularity) and greater dimensions of axial wall crystallites in the latter. These
suborders may have grouped together taxa with homogeneous rates of laminar growth, and taxa
with locally accelerated rates of axial growth, respectively. Given a hypothetical ancestral taxon
with homogeneously laminar growth, accelerated axial growth in a derived taxon could be
interpreted as a peramorphic product representing acceleration. The evolution of the ‘amalgamate’
Atactotoechidae from the ‘integrate’ Amplexoporidae (Astrova 1965, p. 114) provides a concrete
example. In this case, the derivation of uniformly laminar walls from primitive granular walls
could be viewed as an example of paedomorphosis representing the process of neoteny in a
phylogenetic pattern of reverse recapitulation.

The evolution of wall structure among Peronopora, Prasopora, and Atacloporella has been
interpreted in terms of heterochronic processes (Hickey, in press). Prasopora can be considered
primitive and probably ancestral to Atacloporella (Astrova 1978) and Atactoporella appears to be
the ancestral sister group of Peronopora (Hickey, in press). Prasopora is characterized by a
‘granular’ zooecial wall structure, while zooecia of Atactoporella and Peronopora have predominately
laminar walls. The derived laminar wall structures of the latter two genera appear to have evolved
through a paedomorphic decrease in the rate of wall growth during astogeny.

Acantliostyle development and evolution

The epithelial origins of acanthostyles are uncertain. The optically hyaline appearance of the axial
cores of trepostome acanthostyles and portions of ‘heterostyles’ (Blake 1983) resembles that of
exterior-walled skeleton. This suggests that the axial crystallites of trepostome acanthostyles could
have been secreted by outer epithelium as Tavener-Smith (19696) proposed for fenestellid skeletal
rods and ‘primary skeleton’. However, secretion of core crystallites by outer epithelium in
trepostomes appears to be precluded by the lack of any examples in which acanthostyle cores are
proximally continuous with exterior-walled skeleton. Overgrowth of the axial core terminus by
laminar skeleton in the trepostome Leptotrypella? praecox Boardman (Boardman 1983, text-fig.
51-lc, p. 104), the complex growth mechanics needed to explain the alternation of hyaline
and laminar skeleton in trepostome ‘heterostyles’ (Boardman 1983, p. 105), and the different
ultrastructures of acanthostyles and median lamina in Peronopora also argue against secretion of
axial core skeleton by outer epithelium. Yet, the microstructural contrasts between axial cores and
sheath laminae indicate there were distinct growth differences of some kind.

Armstrong (1970) postulated that acanthostyle axial cores in Stenopora could have been deposited
by patches of specialized epithelium. This hypothesis is supported by four lines of evidence in
Peronopora as well: 1, the distinct boundary between cores and sheath laminae; 2, the contrasting
morphologies of core and sheath crystallites; 3, laminar versus continuous growth modes; and, 4,
differential Ca and trace element concentrations between acanthostyle core crystallites and laminar
exozonal skeleton.

Several observations provide clues for an explanation of these microstructural differences. The
solid nature of acanthostyle axial cores has been well established (Armstrong 1970; Blake 1973,
1983; Boardman 1983; Tavener-Smith 1969a). Infrequent lamination of the axial core skeleton in
most taxa indicates that the deposition of granular crystallites was relatively continuous. The

712

PALAEONTOLOGY, VOLUME 30

topographic prominence of acanthostyles and the relationship between axial cores and sheath
laminae indicate that granular crystallite deposition within axial cores was locally accelerated
relative to laminae formation (Tavener-Smith 1969a, h). Occasional preservation of ‘brown bodies’
and pyrite grains within acanthostyle axial cores (Boardman 1983) suggests a high organic content;
organic material could presumably have been more readily incorporated within core skeleton if
growth were rapid. Thus axial core ultrastructure appears to be a product of accelerated crystallite
secretion.

Armstrong (1970, p. 584) found iron carbonate minerals in the acanthostyles of Stenopora. This
was interpreted as an additional indication of a specialized epithelial origin for acanthostyles. The
results of this investigation can be interpreted in a similar manner. Ca densities are lower, and by
inference, growth rates were higher in acanthostyles than within endozone and exozone (text-fig.
8). Substitutional elements such as Cu, Fe, and Mn are generally most abundant in acanthostyle
cores (text-fig. 9). Acanthostyles are enriched in Cu, Fe, Si, and Yb relative to the zooecial void
cement. High Cu levels in acanthostyle cores satisfy the predicted increase in substitutional element
abundances within zones and structures with low Ca densities. The high levels of Cu are consistent
with non-equilibrium skeletal formation (Sakagami et al. 1984). These data support the hypothesis
that paurostyle cores were secreted more rapidly than endozonal and exozonal laminar skeleton.
Furthermore, they suggest a more specific explanation for the skeletal differences between
acanthostyle cores and laminar skeleton.

Paurostyle morphology could be explained by development of an extreme differentiation of
laminar and granular crystallite growth modes. This differentiation could have been accomplished
by patches of inner epithelium with particularly high rates of crystallite secretion. The disordered
axial growth and irregular crystallites typical of zooecial walls in ‘integrate’ taxa could be considered
a developmental analogue with the early stages of acanthostyle evolution in trepostomes. Paurostyles
could have evolved from laminar zooecial wall by a highly localized acceleration of crystallite
secretion rates. Irregular crystallites of zooecial wall axes could be invoked as a transitional step
in this hypothetical transformation series.

Given the above model of acanthostyle origins, it is possible to describe the evolution of different
acanthostyle types among closely related taxa in terms of heterochronic changes in rates of skeletal
growth. Paurostyle morphology is assumed to be the primitive state (Tavener-Smith 19696).
Heterostyles, and heterostyle-like structures of some trepostomes, are composed of alternating
layers of hyaline and laminar skeleton (Blake 1983). Most rhabdomesid aktinotostyles are composed
of distally (and laterally) deflected laminar skeleton (Blake 1983). Aktinotostyles could have been
derived from paurostyles by a decreased rate of crystallite secretion coupled with an increase in
lamination episodicity. Similarly, heterostyle structure could be explained in terms of periodic
alternation between the primitive paurostyle growth mode and decreased secretion rates coupled
with increased episodicity. Thus both aktinotostyles and heterostyles (in part) could be considered
paedomorphic products of neoteny. In this model, heterostyle evolution must have involved
temporal signals which induced switching from one growth mode to another. The presumed ease
with which such heterochronic modifications could occur suggests that homoplasic evolution of
all acanthostyle types may have been common.

The median lamina

The median lamina arose as a distal extension of basal lamina skeleton. However, there is no
evidence to indicate that the median granular layer was deposited against cuticle or within a distally
directed fold of basal epithelium. The absence of a medial parting argues against the first
interpretation. The second interpretation is precluded by occasional growth checks in the median
granular layer which were followed by deposition of the outer laminar layer around the distal
margin of the median layer, and the fact that the median granular layers of primary and secondary
fronds are discontinuous. These factors, in concert with a very low Ca density indicate that the
skeletal structure of the median granular layer was a product of very high, relatively continuous,
rates of crystallite secretion by inner epithelium at the distal margins of the colony (text-figs. 6-

HICKEY: ORDOVICIAN TREPOSTOME BRYOZOAN

713

8). The generally high levels of Fe, Mn, and Cu within the median lamina (text-fig. 9) satisfy the
predicted increase in substitutional element abundances and/or the presence of non-equilibrium
growth products within zones and structures with low Ca densities. The similarity between granular
median and basal lamina skeleton suggest that granular exterior-walled skeleton also grew more
rapidly than laminar interior-walled skeleton.

The ultrastructural character of basal lamina skeleton does not appear to have been a consequence
of deposition against cuticle per se, but rather diflferentiation of the crystallite secretion rates of
inner epithelium (text-fig. 6). Because both granular median lamina and laminar zooecial wall
skeleton are products of secretion by inner epithelium, some mechanism for differentiation of
cellular processes and skeletal products is required. It is evident from the astogenetic succession
of skeletal types and their common epithelial origin that new epithelial cells must have been added
at the colony margin in a manner similar to the ‘conveyor-belt’ model of skeletal formation
proposed by Tavener-Smith (\9b9b). Thus, it could be hypothesized that the stimulus for the
developmental differentiation of epithelial processes/skeletal products lay in positional and/or
temporal regulation of cellular activity and products. As cells produced at the colony margin
migrated proximally, or as the colony margin advanced a predetermined distance, the mode and
rate of inner epithelial secretion switched from high, continuous rates of growth and production
of granular crystallites to slower, episodic growth rates, tabular crystallites and laminar skeleton.

Local discontinuities in the median lamina of Peronopora may be explained in terms of a growth
rate model. Local discontinuities of the median lamina in Peronopora often occur in association
with regions in which thin-walled (‘endozonal’) zooecial growth was prolonged (text-fig. 2). The
recumbent zones of zooecia in these regions are also unusually long and lack cystiphragms
(Boardman and Utgaard 1966, p. 1096). The periodic interruption of median granular layer
formation was followed by relatively wide regions of endozonal growth. These widened endozonal
regions occur not only at the base of young secondary fronds, but also periodically throughout
frond growth along the margin of young fronds. Periodic prolongation of the relatively high
growth rates typical of the recumbent zone and endozone is indicative of a colony-wide increase
in growth rate. That a high growth rate was prolonged beyond its normal duration is also supported
by the atypical lack of cystiphragms within the recumbent zones of the autozooecia in these regions.
In other words, growth rate appears to have been high enough in these instances to have produced
a relatively long recumbent region and some degree of endozonal growth before the first period
of cystiphragm emplacement.

Rapid growth is energetically more expensive than slow or normal growth rates. Thus, if
endozonal growth is indicative of high growth rates, the association of ‘endozonal’ growth with
discontinuities in the median lamina suggests that there could have been a limited availability of
energy for skeletal formation. High rates of growth along the colony margin may have precluded
formation of the median lamina during limited periods of accelerated growth. The median lamina
was formed during periods of normal growth. When growth rates increased, presumably energy
that would otherwise have been allocated to median lamina formation was diverted to permit
somewhat longer periods of ‘endozonal’ growth. Such an explanation could be considered a special
case of generally poor developmental regulation of median lamina formation.

Diagenesis

The compositional differences between skeletal structures and the diagenetic matrix, coupled with
the results of Sakagami et al. (1984), and readily interpretable astogenetic trends consistent with
inferences based on alternative arguments suggest that elemental data could reflect real differences
in the physiology of growth. Significant differences between the matrix and skeletal structures and
among skeletal structures do not unequivocally demonstrate the absence of diagenetic alteration.
However, the primary results of this study are consistent with independent predictions of
compositional behaviour which would be expected in the relative absence, limited or differential
influence of diagenesis on skeletal material and are consistent with inferences based on alternative

714

PALAEONTOLOGY, VOLUME 30

evidence. Further studies of this kind are warranted because they could contribute to knowledge
about skeletal diagenesis, the physiology of skeletal growth and growth rates in extinct organisms.

CONCLUSIONS

The skeleton of P. vera contains Cu, Si, Fe, Mg, Mn, Yb, and Nb, many of which occur in
concentrations significantly different from that of the zooecial void matrix. Ca density is inversely
related to rates of skeletal growth and crystallite secretion. The density distributions of Ca and
other trace elements among crystallite types, skeletal structures, and astogenetic growth zones
support inferences concerning relative rates of crystallite secretion and skeletal growth based on
alternative lines of evidence in bifoliate Peronopora. EDS data indicate differences in the rates of
secretion and growth among crystallite types, acanthostyle cores, the median lamina, zooecial wall,
and stages of zooecial ontogeny and colonial astogeny. Growth was most rapid within the granular
skeleton of the median lamina and ‘A-type’ acanthostyles (paurostyles). Growth rate decreased
from the median lamina through the recumbent zone, endozone, and exozone. Paurostyle cores
were deposited more rapidly than endozonal wall but less rapidly than recumbent zone wall and
the median lamina. Cu and Mn abundance distributions also support the succession of relative
growth rates based on Ca data. Paurostyles contain more Cu than other structures. Mn is also
most abundant in the median lamina and declined monotonically with reduced growth rate. Zones
of disordered, irregular crystallites and laminar growth alternate in zooecial wall axes and
are linked with increased secretion rates and increased episodicity, respectively during the
cystiphragm/tabula emplacement cycle.

Results of this study are preliminary but suggest that compositional studies can provide
information about relative rates of skeletal growth. The astogenetic growth rate curve derived
from EDS data supports the results of growth curves based on an alternative method presented
herein; intrazooecial cystiphragm frequency distribution per unit autozooecial wall length. A
combination of these methods permits the reconstruction of astogenetic growth rates and their use
as taxonomic descriptors. If differential growth rates can be identified on the basis of ultrastructural
morphology, the evolution of skeletal wall microstructure and acanthostyle types may be interpreted
in terms of heterochronic change.

The median lamina of bifoliate Peronopora is composed of one to three structures: 1, an outer
laminar layer; 2, a medial granular layer; and 3, granular mesostyles. The mesostyles and medial
layer are composed of smaller rhombic crystallites which have not been described in mesothecae
of other stenolaemates. The median lamina is structurally continuous with the basal lamina but
was not secreted against cuticle and is therefore interior-walled skeleton. Eormation of the median
lamina and paurostyle cores may be explained by differentiation of inner epithelium for high rates
of secretion and growth. Local discontinuities in the median lamina can be explained by an energy
budget-growth rate hypothesis.

Acknowledgements. This research was conducted as part of a Ph.D. dissertation at Michigan State University.
Acknowledgement is made to the Donors of the Petroleum Research Fund, administered by the American
Chemical Society, for support of this research. Discussions with Robert L. Anstey contributed greatly to this
research. I thank Robert L. Anstey, Daniel B. Blake, Frank K. McKinney, Joseph F. Pachut, Gary D.
Rosenberg, and two anonymous referees for their critical reviews of the manuscript. I thank Stanley L.
Flegler of the Electron Optics Center, Michigan State University for performing the X-ray analyses, some
of the photography, and for the use of their facilities. Kenneth Harrison assisted in specimen preparation.
Special thanks to Erin O’Brien for preparation of the illustrations.

REFERENCES

ALBERCH, p., GOULD, s. J., OSTER, G. E. and WAKE, D. B. 1979. Size and shape in ontogeny and phylogeny.
Paleobiology, 5, 296-317.

HICKEY: ORDOVICIAN TREPOSTOME BRYOZOAN

715

ANSTEY, R. L. 1981. Zooid orientation structures and water flow patterns in Paleozoic bryozoan colonies.
Lethaia, 14, 287-302.

1987. Astogeny and phylogeny: heterochrony and morphological evolution in Paleozoic bryozoans.

Paleobiology, 13, 20-43.

ARMSTRONG, J. 1970. Zoarial microstructures of two Permian species of the bryozoan genus Stenopora.
Palaeontology, 13, 581-587.

ASTROVA, G. G. 1965. Morphology, evolutionary history, and systematics of Ordovician and Silurian bryozoans.
Akad. Nauk SSR, Paleontol. Inst. 106. [In Russian.]

1978. The History of Development, System and Phylogeny of the Bryozoa. Order Trepostomata. Ibid.

169.

BARTLEY, J. M. and ANSTEY, R. L. 1983. CycHc growth in Lower Paleozoic Stenolaemate bryozoans. Geol. Soc.
Am. Abstr. w. Prog. 15, 523.

BiGEY, F. 1979. Microstructure of the main types of Devonian Bryozoa from the Ferques area, Boulonnais,
Northern France. In larwood, g. p. and abbott, m. b. (eds.). Advances in Bryozoology, 153-178. Syst.
Assoc. Spec. Vol. 13. Academic Press, London, England.

1982. Bryozoaires tubulaires Paleozoiques et actuels: Morphologie et microstructure squelettiques

(Trepostomida, Cyclostomida). Bull. Soc. Zool. France, 107, 297 306.

and LAEUSTE, j. 1982. Morphologie et microstructure chez Leptoptrypella Bryozoaire trepostome du

Devonien saharien. Bull. Geol. Soc. France, 24, 711 716.

BLAKE, D. B. 1973. Acanthoporc ultrastructure in the Paleozoic bryozoan family Rhabdomesidae. In larwood,
G. p. (ed.). Living and Fossil Bryozoa, 221-230. Academic Press, London.

1983. Introduction to the suborder Rhabdomesina. In boardman, r. s. et al. (eds.). Treatise on

Invertebrate Paleontology. Part G, Bryozoa, Revised 1, G530-G549. Geological Society of America and
University of Kansas Press, New York and Lawrence, Kansas.

BOARDMAN, R. s. 1960. Trepostoiiiatous Bryozoa of the Hamilton Group of New York State. Prof. Pap. U.S.
geol. Surv. 340, 1 87.

1983. General features of the class Stenolaemata. In boardman, r. s. et al. (eds.). Treatise on Invertebrate

Paleontology. Part G, Bryozoa, Revised 1, G49-GI37. Geological Society of America and University of
Kansas Press, New York and Lawrence, Kansas.

and CHEETHAM, A. H. 1969. Skeletal growth, intracolony variation, and evolution in Bryozoa: a review.

J. Paleont. 43, 205-233.

and TOWE, k. m. 1966. Crystal growth and lamellar development in some Recent Cyclostome Bryozoa.

Spec. Pap. geol. Soc. Am. 101, 20.

and UTGAARD, J. 1966. A revision of the Ordovician bryozoan genera Monticulipora, Peronopora,

Heterotrypa, and Dekayia. J. Paleont. 40, 1082-1 108.

BROOD, K. 1976. Wall structure and evolution in cyclostome Bryozoa. Lethaia, 9, 377-389.

BRYAN, G. w. 1973. The occurrence of seasonal variation of trace metals in the scallops Pecten ma.ximus and
Chlamys opercularis. J. Mar. Biol. Assoc. U.K. 53, 145-166.

CLARK, F. and WHELLER, w. c. 1917. The inorganic constituents of marine invertebrates. Prof. Pap. U.S. geol.
Surv. 102, 1-62.

CRISP, E. L. 1975. The skeletal trace element chemistry of freshwater bivalves. Ph.D. thesis (unpubl.). University
of Indiana, Bloomington, Indiana.

1983. Shell structural, phylogenetic, and ontogenetic related variations in skeletal trace element

concentrations within fresh-water aragonitic bivalve shells. Geol. Soc. Am. Abstr. vr. Prog. 15, 550.
CUMINGS, E. R. 1912. Development and systematic position of the monticuliporids. Bull. geol. Soc. Am. 23:
357-370.

DODD, J. R. 1967. Magnesium and strontium in calcareous skeletons: A review. J. Paleont. 41, 1313-1329.
eisma, d., mook, w. g. and das, h. a. 1976. Shell characteristics, isotopic composition, and trace element
contents of some euryhaline molluscs as indicators of salinity. Paleogeog., Paleoclim., Paleoecol. 19, 39-
62.

GOREAU, T. J. 1977. Coral skeletal chemistry: physiological and environmental regulation of stable isotopes
and trace metals in Montastrea annularis. Proc. R. Soc. B 196, 291-315.

HEALY, N. D. and UTGAARD, J. 1979. Ultrastructure of the skeleton of the cystoporate bryozoans Ceramophylla,
Crassaluna and Cystodictya. In larwood, g. p. and abbott, m. b. (eds.). Advances in Bryozoology, 179-
194. Syst. Assoc. Spec. Vol. 13. Academic Press, London.

HICKEY, D. R. In press. Bryozoan astogeny and evolutionary novelties: their role in the origin and systematics
of the Ordovician monticuliporid trepostome genus Peronopora. J. Paleont.

716

PALAEONTOLOGY, VOLUME 30

KARKLINS, o. L. 1983. Introduction to the suborder Ptilodictyina. In boardman, r. s., et al. (eds.). Treatise
on Invertebrate Paleontology. Part G, Bryozoa, Revised I, G453-G488. Geological Society of America and
University of Kansas Press, New York and Lawrence, Kansas.

LARWOOD, G. p. and TAYLOR, p. D. 1979. Early structural and ecological diversification in the Bryozoa. In
HOUSE, M. R. (ed.). The Origin of Major Invertebrate Groups, 209-234. Syst. Assoc. Spec. Vol. 12. Academic
Press, London.

MOBERLY, R., JR. 1968. Composition of magnesium calcites of algae and pelecypods by electron microprobe
analysis. Sedimentology, 11, 61 82.

MORRISON, s. J. and anstey, r. l. 1979. Ultrastructure and composition of brown bodies in some Ordovician
trepostome bryozoans. J. Paleont. 53, 943-949.

PHILLIPS, A. H. 1922. Search for metals in Tortugas marine organisms. Pap. Dept. Mar. Biol. Carnegie Instit.
312, 95-99.

PODELL, M. E. and ANSTEY, R. L. 1979. The interrelationship of early colony development, monticules and
branches in Paleozoic bryozoa. Palaeontology, 22, 965 982.
rabbio, s. e. and regalblito, d. p. 1985. Tidal growth cycles in skeletal laminations of the upper Paleozoic
bryozoan Tabulipora. Geol. Soc. Ainer. Abstr. vr. Prog. 17, 322.

ROSENBERG, G. D. 1980. An ontogcnctic approach to the environmental significance of bivalve shell chemistry.
In RHOADS, D. c. and lutz, r. a. (eds.). Skeletal Growth of Aquatic Organisms, 133 167. Plenum, New
York.

1982. Growth rhythms in the brachiopod Rafinesquina alternata from the Late Ordovician of Southeastern

Indiana. Paleobiology, 8, 389 401.

1983. Facies and physiologic control of growth patterns in living and fossil brachiopods. Geol. Soc. Am.

Abstr. H'. Prog. 15, 674.

and JONES, c. b. 1975. Approaches to chemical periodicities in molluscs and stromatolites. In rosenberg,

G. D. and RUNCORN, s. K. (eds.). Growth Rhythms and the History of the Earth's Rotation, 223-240. John
Wiley, London.

ROSS, J. R. p. 1976. Body wall ultrastructure of living Cyclostome Ectoprocts. J. Paleont. 50, 350-353.

1977. Microarchitecture of body wall of extant Cyclostome Ectoprocts. Am. Zool. 17, 93 105.

sakagami, s., masuda, e., kawabe, t. and ntshizawa, y. 1984. Recent and fossil bryozoan skeletons: their
chemical composition. Chigaku Zasshi, 93, 61 76. [In Japanese, English summary].

SANDBERG, p. A. 1977. Ultrastructurc, mineralogy, and development of bryozoan skeletons. In wollacott,
R. M. and ZIMMER, R. L. (eds.). Biology of Bryozoans, 143-183. Academic Press, London.

1983. Ultrastructure and skeletal development in Cheilostome Bryozoa. In boardman, r. s., et al. (eds.).

Treatise on Invertebrate Paleontology. Part G, Bryozoa, Revised L G238 G286. Geological Society of
America and University of Kansas Press, New York and Lawrence, Kansas.

SCHMALZ, R. F. and SWANSON, F. J. 1969. Diurnal variations in the carbonate saturation of seawater. J. Sed.
Pet. 39, 255 267.

SCHOPF, T. J. M. and manheim, f. t. 1967. Chemical composition of Ectoprocta. J. Paleont. 41, 1197-1225.
SOULE, J. D. and soule, d. f. 1981. Heavy metals uptake in bryozoans. In larwood, g. p. and nielsen, c.

(eds.). Recent and Fossil Bryozoa, 227-235. Olsen and Olsen, Fredensborg.

TAVENER-SMiTH, R. 1969r/. Wall structure and acanthopores in the bryozoan Leioclema asperum. Lethaia, 2,
89-98.

19697). Skeletal structure and growth in the Fenestellidae (Bryozoa). Palaeontology, 12, 281-309.

1975. The phylogenetic affinities of fenestelloid bryozoans. Palaeontology, 18, 1-18.

and WILLIAMS, a., F.R.S. 1972. The secretion and structure of the skeleton of living and fossil Bryozoa.

Phil. Trans. R. Soc. B 264, 97 -159.

VOLKOVA, K. N. 1974. Devonian Bryozoa of the South-Eastern Atali. Acad. Nauk S.S.S.R. Sibirisk. Otdel.
Trudy Inst. Geol. and Geof. 199, 1-182 [In Russian].

ZOLOTAREV, V. N. 1974. Magnesium and strontium in the shell calcite of some modern pelecypods. Geochem.
Interntl. 11, 347-353.

Typescript received 11 November 1985
Revised typescript received 21 February 1987

DAVID R. HICKEY

Department of Geological Sciences
Michigan State University
East Lansing, MI 48824-1 1 14, USA

A NEW ENCHODONTID FISH GENUS FROM THE
UPPER CENOMANIAN OF JERUSALEM, ISRAEL

by MENAHEM YAEL CHALIFA

Abstract. Parenchodits longipterygius n. g. et sp. belonging to the Enchodontidae is described on the basis of
eight specimens. These were found in platy limestones of the late Cenomanian Kefar Shaul Formation, in the
vicinity of Jerusalem, Israel. The species is characterized by a short, high body, a fenestra-less premaxilla, and by
the long and narrow postcleithrum reaching the ventral margin of the belly. Other characteristics are the axial
skeleton with thirty vertebrae, of which only seven are abdominal; dorsal and anal fins with long bases; a ventral,
well-developed pectoral fin; a naked body, devoid of scales except for two dorsal postoccipital and three lateral
scutes at the base of the tail; and with a reduction and fusion of the endoskeletal elements of the caudal fin. The
fish was a small, fast swimming predator. The structure of the head, the absence of scales, and the fusion of the
caudal fin elements suggest that this genus is related to Enchoilus.

F I N D s of fossil fish in the vicinity of Jerusalem have been recorded since the beginning of the century
(Blanckenhorn 1905; Shalem 1925, 1927), however, they have never been studied in detail. The fossil
fish come from the Laminated Limestone Member, which constitutes the upper part of the Kefar
Shaul Formation and overlies its Lower Argillaceous Member.

The Laminated Limestone Member consists of bufif to yellowish-white, red stained, thin-bedded,
laminated to platy (up to 1 5 cm), fine-grained limestone with argillaceous material between the plates
(Arkin et al. 1965, p. 20; Braun 1970, p. 33). The formation is overlain by the dolomitic, partly
calcareous Weradim Formation, and overlies the dolomitic-calcareous Amminadav Formation. The
complex of strata belongs to the Judea Group which is widely extended in Israel.

The distribution of the fish-bearing Laminated Limestone Member around the type locality is very
restricted. It extends over less than 10 sq. km, in varying thicknesses up to 15 m, and is confined to the
eastern flanks of the Judea anticline, in the western outskirts of Jerusalem.

Another restricted occurrence of fish-bearing strata, of a similar lithology and more or less in a
similar stratigraphic position, is exposed near the town Ramallah, some 20 km north of Jerusalem.

The underlying Argillaceous Member is thicker, up to 60 m thick, and though of greater extent than
the Laminated Limestone Member, it wedges out in all directions, within a distance of up to a few tens
of kilometres.

The overlying Weradim Formation is much more extended than the Kefar Shaul Formation. Its
thickness ranges from 25 to 115 m, and where the Kefar Shaul Formation is missing, it directly
overlies the Amminadav Formation (Braun 1970). These field relationships may be an indication that
the Weradim Formation gradually replaced the Kefar Shaul Formation.

The fish-bearing Laminated Limestone Member is quite poor in fossils of other groups. The fish
assemblage comprises several holosteans, and many teleosts, which have not yet been studied. Some
pelecypods, echinoids, and especially ammonites, though badly preserved, enable dating the member
as of late Cenomanian age (Shalem 1925, 1927; Picard 1938). The fishes are also accompanied by
crustacean and terrestrial plant remains (Lewy and Raab 1976).

Thin sections examined microscopically revealed finely laminated micrites, rich in faecal pellets,
filamentous algae, and fish teeth and scales. Rare planktonic and benthonic foraminiferids also occur
(Hamaoui and Raab, 1965 pp. 34-35).

A low energy, marine environment of restricted, shallow, partly closed basins was suggested (Lewy
and Raab 1976, p. 32.5).

I Palaeontology, Vol. 30, Part 4, 1987, pp. 717-731.]

© The Palaeontologieal Association

718

PALAEONTOLOGY, VOLUME 30

The underlying Argillaceous Limestone Member of the Kefar Shaul Formation is very rich in
invertebrate megafossils. Assemblages comprising pelecypods, echinoids, and especially ammonites
(e.g. Turrilites costatus Passy and Neolobites vibrayeanus (d’Orbigny)) assign a late Cenomanian age
to the member (Avnimelech and Shoresh 1962; Lewy and Raab 1976).

The overlying Weradim Formation, which partly replaces the Kefar Shaul Formation (see above),
is poor in fossils, most probably due to extensive dolomitization. Flowever, in places, especially in
quartzolitic lenses, various pelecypods, particularly rudists (known as reef builders) were recorded
(Braun 1970, p. 33; Shalem 1927: ‘calcari superior! a Radioliti’).

METHODS AND TECHNIQUES

The exposed side of the specimens, which were embedded in fine-grained limestone slabs, was mechanically
cleaned under a stereoscopic binocular, with the aid of needles and fine brushes. The slabs were then embedded in
polyester resin and later soaked in 10% acetic acid to remove the enclosing calcareous matrix of the underside
(Rixon 1976).

SYSTEMATIC PALAEONTOLOGY

Suborder enchodontoidei Berg, 1940
Family enchodontidae Lydekker, 1889

Diagnosis. (Emended, after Goody 1969, p. 71.) Head deepened, especially posteriorly; body may be
deepened in thoracic region. Posttemporal fossa unroofed. Lower jaw long and deep behind the
constricted symphysis; articular facet visible in lateral view. Opercular convex posteriorly, and deeper
than it is broad; preopercular without prominent ventral spine. Pectoral fins larger than pelvic fins
and extremely low on body. Pelvic fins abdominal. No posterior extension of the cleithrum. Lateral
line scales present or absent, if present they do not overlap; mid-dorsal scutes reduced and not
overlapping.

Genus parenchodus n.g.

Diagnosis. A fish with a short, high, and laterally compressed body. The length of the head exceeds one
third of the standard length, slightly exceeding also the depth of the head at the occipital region. The
maximal depth of the body is half the standard length. The obit is large, its length reaching one third of
the length of the head. The premaxilla lacks a fenestra for reception of the end of the large dentary
tooth. Postcleithrum well developed, long and narrow, reaching the ventral margin of the belly. Axial
skeleton with thirty vertebrae, twenty-three of them caudal. Dorsal fin with a long base, occupying
most of the back, reaching the caudal peduncle.

Body naked, except for two unequal, not overlapping scutes on the mid-dorsal, postoccipital
region, and three overlapping lateral scutes on each side of the caudal peduncle. The bases of the
caudal fin rays deeply embrace the hypural plates.

Etymology. Para (Greek), near— resemblance to Enchodiis.

Type species. Parenchodus longipterygius n. sp.

Parenchodus longipterygius n. sp.

Diagnosis. Parenchodus of up to 75 mm standard length, with fin rays counting as follows: dorsal fin,
30; pectoral fin, 14; pelvic fin, 5; and anal fin, 22.

Elolotype. P.163, The Hebrew University of Jerusalem. An almost complete fish embedded in resin (text-figs. 1
and 2).

Referred material. Seven specimens: P.143, P.150, P.1502, P.1503 (text-fig. 3), P. 1504, P.1508 (text-fig. 7),
HUJ. DY-15, deposited in The Hebrew University of Jerusalem.

THXT-FiG. 1. Parencliodiis longipteryghis n. g. et sp. Holotype, P.163, Upper Cenomanian, Kefar Sliaul
Formation, Jerusalem, Israel. An almost complete specimen. Left side.

TKXT-FiG. 2. Parenchothis longipterygius n. g. et sp. Restoration of the skeleton, lateral view.

720

PALAEONTOLOGY, VOLUME 30

Locality ami horizon. Givat-Shaul, a western suburb of Jerusalem. The Upper Cenomanian, laminated lime-
stone Member of the Kefar-Shaul Formation.

Etymology. Longipterygius — having a long fin.

DESCRIPTION

General features. The body is short, deep, and laterally compressed. Its maximal standard length is
75 mm, slightly more than twice its maximum depth. The head is short and deep; its length slightly

TEXT-HG. 3. Parenchodus longipterygius n. g. et sp. P.1503, Upper Cenomanian, Kefar Shaul Formation,

Jerusalem, Israel. Lateral view of the skull.

me

le

pmx

mt

rt

ss

RAAB AND CHALIFA: ENCHODONTID FISH FROM ISRAEL

721

Table 1. Dimensions (in mm) and ratios of six specimens of Parenchodits longipterygius n. g. et sp.

Specimen

P.1502

P.163

P.143

HUJ. DY-15

P.1503

P.1504

Mean

Standard length

61

67

69

74-5

Maximum depth of trunk
Depth of trunk as percentage of

23

35

34

31

34

standard length

37

52

49

41

44

Length of head

Length of head as percentage of

23-5

29

31

29

31

28

standard length

38

43

44

38

40-75

Depth of head

Depth of head as percentage of head

23

25

28-5

25

27-5

24-5

length

97

86

91

86

88

87

89

Length of orbit

Length of orbit as percentage of the

8

8

1L5

9-5

9

head length

28

26

40

31

32

31-4

exceeds its maximal depth at the occipital region, and is more than one third of the standard length.
The oral gape is wide, the length of the lower jaw is about three times its maximum depth. The orbit is
large, occupying half the length of the neurocranium. The dimensions of 6 specimens are given in
Table 1. The post- and preorbital regions are narrow. Three infraorbitals exist. The postcleithrum is
narrow and long, reaching the ventral margin of the belly. There are 30 vertebrae of which 23 are
caudal. The dorsal fin is long and consists of 30 fm rays. The pectoral fin is in a ventral position and
consists of 14 fin rays. The pelvic fin is smaller than the pectoral fin, in abdominal position, and
consists of 5 fin rays. The anal fin has a long base and consists of 22 fin rays. The caudal fin is deep,
deeply forked with a very narrow peduncle, carrying flat neural and haemal spines, strongly inclined
backwards. The anterior haemal spines have a wide and shallow proximal part. Three supraneurals
exist. Dentition well developed. The palatine carries one large, striated conical tooth. The
ectopterygoid bears large teeth. The dentary has 2 rows of conical teeth. The typical anterior
mandibular tooth is exceptionally large— about a quarter of the total length of the entire lower jaw
and longer than the palatine tooth. The maxilla is toothless. One row of narrower and high conical
gill-rakers exists on the floor of the anterior part of the gill chamber.

The body is naked, except for two large, not overlapping, unequal scutes, at the mid-predorsal.
Three lateral scutes exist at the base of the caudal fin; they are large, overlapping and have crenulated
margins.

Neiirocraniuin and the skull ?-oq/ (text-figs. 3 and 4). The skull roof is flat, anteriorly narrow becoming
wider posteriorly, widest above the posterior part of the orbit. Its posterior margin is notched by a
wide, unroofed posttemporal fossa. The neurocranium is deep and wide posteriorly, becoming
narrower anteriorly. The orbit is large, occupying more than half the length of the neurocranium. The
posterior margin of the skull roof extends backwards laterally.

The frontals form most of the skull roof, meeting along the median line in a sinuous suture. At the
posterodorsal corner of the orbit, the frontal widens abruptly laterally, forming a wing-like process,
inclined ventrally in front of the sphenotic. This process builds the front wall of the posttemporal fossa
(text-figs. 3 and 4). The frontals reach the premaxillae anteriorly. Their surface is ornamented by
prominent ridges, radiating from the centre of ossification, which is situated above the posterior part
of the orbit. The ridges in front of and behind the centre of ossification are tuberculated, whereas
laterally they are replaced by closely spaced tubercles. The lateral margins of the frontals are
crenulated. A branching point of the supraorbital sensory canal occurs in front and close to the centre
of ossification of the frontals. One branch is directed anteriorly along the bone; a second runs

722

PALAEONTOLOGY, VOLUME 30

pop

TEXT-FIG. 4. Parenchodus longipterygius n. g. et sp. Restoration of the skull. Lateral view.

posterolaterally along the lateral expansion to the pterotic, and a third turns backward, mesial to the
posttemporal fossa, passing to the parietal.

The parietals are small and narrow, in close contact with the posterior margin of the frontals. They
are located mesial to the posttemporal fossae. The dorsal surface of the bone is tuberculated.

The supraoccipital is small, median, separating the parietals. A low crest extends along the median
line of the supraoccipital from the anterior to the posterior end, without protruding posteriorly. The
surface of the bone is smooth.

The sphenotic is situated at the posterodorsal corner of the orbit, posteroventral to the wing-like
lateral process of the frontal. The lateral surface of the sphenotic is triangular and slightly
tuberculated. The exposed part of the pterotic is very thick, forming the posterior continuation of the
lateral process of the frontal. A ridge runs along the middle of the bone, bordering the posterolateral
side of the posttemporal fossa. The dilatator fossa lies lateral to the ridge.

The parasphenoid is beam-like in lateral view, with a low ventral carina along it. Beyond the orbit
the carina expands ventrally and the beam dorsally to meet the prootic. The parasphenoid terminates
anteriorly in front of the anterior end of the maxilla.

RAAB AND CHALIFA: ENCHODONTID FISH FROM ISRAEL

723

The features of the prootic, including the arrangement of its foramina, generally fits the
enchodontid prootic, as described by Goody (1969, p. 85, fig. 38b).

The mesethmoid is well developed and ossified. Its dorsal side is covered posteriorly by the frontal
and anteriorly by the premaxilla.

The lateral ethmoid is a beam-like bone. Its ventral face is grooved to accommodate the dorsal
process of the palatine.

The circumorbital bones are represented by only three thin infraorbital plates, which housed the
infraorbital sensory canal. They extend from the centre of the posterior margin of the orbit to the
centre of its ventral margin, without touching each other. The anterior one is the longest and the
posterior one is the shortest.

The intercalar is the only bone that could be identified in the posterior wall of the neurocranium, as
a well-developed thickening below the pterotic.

An incomplete interorbital septum forms two wings, penetrating into the anterior and posterior
regions of the orbit, but not meeting.

Hyopalatine bones (text-figs. 3 and 4). The hyomandibular is narrow and long, forwardly inclined. The
upper half of the bone is flat, carrying a prominent arched ridge, running obliquely from anterodorsal
to posteroventral. It is wide, plate-like, with a thick, wedge-like lower part.

The symplectic is small and wedge-like with about half its length inserted in the groove of the
quadrate.

The quadrate is a narrow, high triangle. The ventral apex of the triangle is rounded, slightly arched
posteriorly. The bone is relatively thick and its surface is striated, except for its anterior and dorsal
margins, which are thin and smooth. A deep groove exists in the upper part of the inner side of the
bone, for the symplectic.

When the mouth was closed the lower jaw covered the anterior margin of the quadrate and only its
thickened part remained exposed in lateral view.

The endopterygoid is thin and large, forming most of the infraorbital region. It carries tubercle-like
teeth. Its posterodorsal part is covered by the metapterygoid.

The dorsal margins of the ectopterygoid are exposed dorsal to the maxilla. The ectopterygoid
continues forward, covered by the maxilla and the premaxilla. It is long and flat, and bears along its
entire margin a row of high conical teeth (similar to the premaxillary teeth).

The metapterygoid is thin with rounded margins, except for the posterior ones which extend into
two lobes, a dorsal one, prominently ridged and a ventral one, thin and flat. Both lobes overlap the
hyomandibular, below its mid-length.

The palatine is thick and short and bears posteriorly a dorsal process which fits a groove in the
ventral side of the lateral ethmoid. In lateral view the palatine is recognized as a thickening behind the
premaxilla. It bears at its anterior end a very large, somewhat compressed, striated tooth.

Upper jaw. The premaxilla is triangular, with a posteriorly extended base. The frontal area of the
premaxilla is tuberculated and its lateral area is smooth. The premaxilla forms more than
three-quarters of the margin of the upper jaw and bears thin, high conical teeth, similar in size to the
smaller teeth of the lower jaw (see below p. 724). The foremost tooth is the largest, the others becoming
gradually smaller backwards. The fenestra in the premaxilla, so typical of enchodontids (Goody 1969,
p. 71) was not observed.

The maxilla is long, rod-like, thick anteriorly, wide and flat posteriorly, toothless. Its anterior part is
situated along the dorsal margin of the premaxilla, crossing the gape at an angle of about 45°. The
anterior end of the maxilla is slightly bent ventrally, and is situated below the premaxilla, probably in
a groove in its inner side. The surface of the maxilla is smooth.

Lower jaw. The lower jaw is relatively high, especially in its posterior part; its maximum height is
about one third of its total length (text-fig. 3).

The angulo-articular is well developed, triangular, with ventral margin extending forward almost
to the mid-length of the jaw. It is ornamented by strong tuberculated ridges, radiating sparsely from a

724

PALAEONTOLOGY, VOLUME 30

centre near the articulation of the jaw. It bears a postarticular process, which fits in a notch at the
ventroanterior end of the preopercular. The retroarticular is reduced to a small bone situated in the
posteroventral corner of the lower jaw.

The dentary is smooth, except for a fine radial striation originating from a centre near the
symphysis. The dorsal margin of the dentary thickens into a wide flange, reaching the symphyseal
region, and bearing two rows of teeth. The external row contains numerous, closely spaced, high,
conical, smooth, sharp teeth, which are somewhat irregular in size, though the anterior ones are
larger. A second, internal row, consists of seven to eight high, thick, slightly laterally compressed and
striated teeth. The size of these teeth is almost twice that of those of the outer row. The foremost tooth
of the inner row is the largest. This large tooth (text-figs. 3 and 5a)— so typical for the genus
Tnc/mc/ns— protrudes just beyond the symphysis. It is very high, its height slightly exceeding a quarter
of the length of the jaw. Its base is wide and its surface smooth. It is situated on the margin of the jaw,
with a slight forward inclination. The tooth has an asymmetric cross-section, with a convex posterior
face, and a flat to concave anterior face. It has two cutting edges, one anterolateral and the other
posteromedial. A smaller tooth could be observed laterally and behind the foremost fang in specimen
P.163. A smaller tooth (relatively to the fang) and in a similar position and direction is illustrated by
Woodward (1903, pi. 14, fig. 7). Such teeth, so close to the fangs, do not seem functional and hence
might be regarded as replacement teeth. The symphyseal region is relatively high, and its dorsal
margin lipless. The symphysis is built of three interlocking finger-like processes on each dentary. A
conspicuous foramen of the mandibular sensory canal is situated just behind the symphysis.

Branchial arches and hranchiostegal rays. The ceratohyal and the hypohyal (seen on the right side of
specimen P. 143) constitute one beam-like element, extending backwards from the middle of the lower
jaw to the posterior end of the preopercular.

The beam bears eleven posteriorly arched hranchiostegal rays. These are long, thin, and have a
thickened base. A pair of high and thin beam-like first ceratobranchials (text-fig. 3) carry high, slender,
slightly anteriorly inclined gill-rakers.

Operculum. The opercular bones form together a narrow cover, twice as high as its maximum width.

The opercular is narrow and high. Its dorsal, posterior, and ventral margins form a continuous
arched line. The anterior margins are almost straight, perpendicular to the body axis. Its articular
condyle with the hyomandibular is located at about two-thirds of its height. An inner, supporting
beam extends from the condyle backwards, and another along the anterior margins, just beyond the
area covered by the preopercular. The surface of the opercular is covered by finely tuberculated
prominent ridges, radiating from the centre of ossification, which is situated near the condyle.

The triangular subopercular is slightly overlapping the ventral margins of the opercular. It is
ornamented by finely tuberculated ridges, radiating from the anterodorsal corner.

The preopercular is thick, high-triangular, with a posteriorly and anteriorly extended ventral base.
In its anteroventral corner there is a socket for the postarticular process of the lower jaw.
Tuberculated ridges radiate from a centre beyond the articulation socket, changing dorsally, and
especially along the anterior margin, into fine striae. The preopercular sensory canal ran along the
preopercular, opening through three large pores in the anteroventral area of the bone. From the
anteriormost opening, which is directed forward, the lateral line canal continued to the mandible.
Interopercular bone absent.

Axial skeleton. The vertebral column numbers 30 vertebrae, of which only 7 are abdominal and
23 caudal. Eadh centrum is almost quadrangular and strongly constricted in the middle. A finely
striated ridge runs along the lateral side of each centrum. The centra are approximately equal in length
along most of the body, however an abrupt shortening occurs at the caudal peduncle. A high neural
arch extends all along each centrum, not fused with it (text-fig. 5b). The neural arch comprises a
narrow prezygapophysis, and a high and wide postzygapophysis, both with a delicately striated
surface. The centre of the lateral side of each neural arch is pierced by the foramen for the spinal nerve.

In the abdominal region the neural spines are paired. They originate at the anterior part of the

RAAB AND CHALIFA: ENCHODONTID FISH FROM ISRAEL

725

TEXT-FIG. 5. Parenchodus longipterygius n. g. et sp. Restoration of: A, anterior part of right lower jaw. b. abdominal
vertebra no. 7 {left) and precaudal vertebra no. 7 {right), c, pterygiophores and dorsal fin rays, d, siipraneurals and

dorsal scutes, e, left pectoral girdle in lateral view.

726

PALAEONTOLOGY, VOLUME 30

neural arches. Each spine is high, acute, striated along its entire length and is slightly inclined
backwards. The highest neural spines occur at about the middle of the dorsal fin, from where they
start to shorten gradually, becoming more inclined backwards. Near the base of the caudal fin the
spines become short, stout, and much more inclined.

The pleural ribs are very long and longitudinally furrowed. They are inclined backwards and
arched forwards. They are carried by enlarged parapophyses, not fused to the centrum. Epineurals
occur up to the centre of the dorsal fin. Epipleurals occur along the anterior half of the caudal region
(text-figs. 1 and 3).

The haemal arches are relatively long, at least in the anterior part of the caudal region. The first
caudal centrum bears only a haemapophysis. The haemal spines increase gradually in length, reaching
a maximum in caudal vertebra 7, where the distal end of the haemal spine reaches the base of the anal
fin. Hereafter backwards, the spines become shorter. Each of the first six haemal spines is flat and
wide at its base, becoming sharp distally.

The caudal peduncle is short. The neural and haemal spines are flat, expanded, and sharp at their
distal ends. They are strongly inclined backwards and fit tightly to each other. Three large oval scutes,
with indented margins, are situated along the four terminal vertebrae, each overlapping the following
scute, and together forming a lateral keel-like structure (text-fig. 7).

Three supraneurals are situated in front of the dorsal fin (text-fig. 5d). The first one is oval, with a
wide supporting ridge along it. It supports the first dorsal scute from below. The second is
high-triangular. It bears a longitudinal supporting ridge, expanding backwards and forwards at
the distal end, and split at its proximal end, at about half its length. This element supports the
second dorsal scute. The third supraneural is similar in form to the second, but does not support
any scute.

Pectoral fin and girdle. The pectoral fin is ventral, carrying 14 unsegmented, distally branched fin rays.
Its length is about that of 6 5 successive vertebrae, and almost twice the length of the pelvic fin.
Complete pectoral girdles, including their dermal and endoskeletal elements, can be observed in
specimens P.143 and P.163.

The cleithrum is flat and smooth supported all along its midline by a heavy, striated ridge. It is
arched forward and terminates somewhat ventral to the posteroventral end of the preopercular. The
bone becomes narrower towards its extremities.

The supracleithrum is thick and flat, rod-shaped, ornamented by longitudinal, closely spaced
grooves. It touches the outer dorsal surface of the cleithrum, and is directed obliquely from
posteroventral to anterodorsal. Its dorsal end is expanded posteriorly. The expansion is flat, rounded,
and closely covered by tubercules. It is traversed by the lateral line canal on its way from the body to
the head (text-fig. 5e).

The postcleithrum is exceptionally developed. It is thick, long, rod-like slightly arching in the
posteroventral direction. The bone is longitudinally striated. Its proximal end is in contact with the
posterior inner region of the supracleithrum. Ventrally it terminates at the horizontal level of
the posteroventral corner of the mandible. The distal end is sharply bent posteriorly, almost reaching
the anterior end of the pelvic girdle. The pair of postcleithra look like very large ribs, and most
probably served as a support to the body wall of the very deep belly.

The dorsal arm of the posttemporal is well developed, wide, and triangular. The posterior corner of
the triangle articulates with the supracleithrum. The apex of the triangle reaches the dorsal mid-line.
The outer surface of the posttemporal is ornamented by prominent tuberculated ridges, scattered over
the bone in a fan-like fashion, centred at the posterior corner of the triangle. The ventral arm of the
posttemporal has the shape of a flattened rod with an enlarged base and is densely striated. It starts in
front of the articulation point of the posttemporal with the supracleithrum. The lateral line canal
traversed the posttemporal on its way from the supracleithrum to the head. The pores of the canal are
distinct near to the articulation area.

The coracoid is a large thin trapezoidal plate, with a horizontal base extending slightly backwards.
Its anterior margins are firmly jointed to the posteroventral margins of the cleithrum. A prominent

RAAB AND CHALIFA: ENCHODONTID FISH FROM ISRAEL

727

ridge extends from the centre of the coracoid to its anterior corner. The coracoid thickens near its
contact with the scapula.

■ The scapula is small and thick relative to the coracoid, and has a large and distinct scapular
foramen. Four radials can clearly be counted in the pectoral girdle (specimen P.163).

Pelvic fin and girdle. The pelvic fin is abdominal. It originates at the level of the sixth caudal vertebra. It
consists of five to six unsegmented fin rays. These are longitudinally grooved and distally branched.
The pelvic fin is half the length of the pectoral fin. The pelvic girdle is flat and thin, wedge-like, with a
few longitudinal thickenings. Four radials exist.

Dorsal fin. The dorsal fin has a long base, and occupies most of the back. It is as long as nineteen
successive vertebrae, and contains thirty fin rays. Its anterior part is high, shortening gradually
posteriorly. Each lepidotrich is longitudinally grooved whereas the last two are completely bifurcated.
The fin rays are unsegmented. The first pterygiophore is very expanded and forwardly inclined. The
proximal segments of the pterygiophores are flat, truncated ventrally, with a longitudinal ridge. Their
distal end is expanded in a nail-head form, with two articulation facets. The middle radial is wide and
thick. The distal radial is also preserved as a small rounded element, embraced between the bases of
the hemitrichia. The fin ray is attached anteriorly by its base to the distal radial of the preceding
pterygiophore, and by its ventral side to the proximal radial of the underlying pterygiophore (text-
fig. 5c). The head of each proximal radial is anteriorly in close contact with the middle element of the
preceding pterygiophore, dorsal to the base of the fin ray and posterior to its middle radial. The
middle and distal radials form together a continuous flexible support for the dorsal fin.

Anal fin. The anal fin extends along about eleven vertebrae on the posterior half of the venter. It starts
opposite the mid-length of the dorsal fin and contains twenty-two blunt, longitudinally grooved,
unsegmented fin rays. The lepidotrichia of the last several rays are split from their bases. The anterior

ns

TEXT-FIG. 6. Parenchodus lonyipterygius n. g. et sp. Restoration of the endoskeleton of the caudal tin.

728

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 7. Parenchodus longipterygius n. g. et sp. P.1508, Upper Cenomanian, Kefar Shaul Formation,

Jerusalem, Israel. Lateral view of the tail.

part of the anal fin is relatively high (as high as the length of four vertebrae), becoming lower
posteriorly. The proximal elements of the pterygiophores have one or two longitudinal grooves and
are truncated at their proximal end. The two anterior elements are thick and long, the following ones
become gradually shorter. The distal elements of the pterygiophores are small, thin, with an
undefined shape.

Caudal fill (text-figs. 6 and 7). The caudal fin is large, deeply forked. The dorsal lobe is somewhat
longer than the ventral one. The dorsal and the ventral lobes contain ten and nine principal fin rays,
respectively. These are segmented and branched distally. Nine segmented, undivided accessory fin
rays precede the principal fin rays of each lobe.

The proximal ends of the principal fin rays, except for the two middle ones, are narrow and sharp.
They embrace the hypural plate in such a way that their proximal ends almost meet. The two middle
rays have a truncated, thickened base, only touching the posterior edge of the hypural plate, not
embracing it.

The endoskeleton of the caudal fin consists of four vertebrae, namely, two separate preurals (pU2,
puj) and a ural vertebra fused to the first preural, to form a single triangular centrum. The hypurals
are fused into one wide plate, which is divided by a relatively wide longitudinal furrow— the
continuation of the vertebral axis— into a ventral and a dorsal part. An additional long and narrow

RAAB AND CHALIFA: ENCHODONTID FISH FROM ISRAEL

729

hypural occurs in close contact dorsally to the hypural plate. The terminal centrum (Ui +puj) bears a
high, narrow stegural. The neural and the haemal spines of pu2 and PU3 are thick, wide, their distal
ends are truncated, and they are not fused to their centra. The surface of these elements is finely
striated. The truncated distal margins of the haemal spines, the parhypurals, the hypurals, the
stegural, and the neural spines of pu2 and puj, form a continuous arched line constituting the
posterior margin of a wide plate supporting the fin rays. A distinct semilunar ridge, arching ventrally,
occurs on the ventrolateral area of the terminal centrum (text-fig. 7). Its ventral position as well as its
direction hinders its identification as the uroneural of ural 1. It is probably only a thickening of the
centrum of the vertebra, intended to enlarge the adhesion surface of the caudal fin muscles. A
functionally similar element was recorded by Nursall (1963, p. 459). However, the two elements are
not homologous, since NursalFs hypurapophysis is a lateral process of the anterior hypural whereas
the element described here is a part of the terminal centrum. Summarizing the main features of the
caudal skeleton, it is characterized by ( 1 ) fusion of the hypurals, (2) epaxial and hypaxial elements not
fused to the centra— a primitive feature, (3) reduction in elements— epurals and uroneurals missing
and, (4) the loss of ural 2.

Scales. The body is naked, without scales except for two dorsal scutes and a row of three lateral scutes
at the base of the tail. The dorsal scutes are large, oval, not overlapping. They are situated along the
dorsal mid-line, from the occiput to near the anterior end of the dorsal fin. The first scute is more than
twice as large as the second. The scutes are supported by the two first supraneurals (text-fig. 5d).

The posterior part of each dorsal scute is ornamented by scattered tubercles whereas the anterior
part bears tuberculated ridges, radiating from the centre of the scute. An inner supporting ridge runs
along the posterior half of the scute.

Three relatively large scutes occur laterally along the four terminal vertebrae. They are thin, with
indented margins, successively overlapping each other, to form a lateral keel-like structure (text-fig. 7).

DISCUSSION

The basic features of Parenchodiis agree well with those used by Goody (1969) in defining the family
Enchodontidae: Head and body high; posttemporal fossa unroofed; lower jaw long and posteriorly
high; articular facet exposed laterally; opercular high, with convex posterior margins; preopercular
without a ventral spine; pectoral fin ventral, larger than the pelvic fin; pelvic fin abdominal; existence
of a few, non-overlapping dorsal scutes. Only one feature, namely the total reduction of the lateral line
scales in Parenchodiis, disagrees with Goody’s diagnosis of the family. It seems that this feature is not
enough for the erection of a separate family.

The tendency towards a short and deep body, indicated as one of the features defining the genus
Enchodiis (Goody 1976), and characterizing especially the old world species of Enchodiis, is exhibited
in extreme form in Parenchodiis.

This tendency correlates well with other structural features, so typical of short and deep fish,
namely: (1) reduction in number of vertebrae; (2) reduction in number of abdominal vertebrae (in
Parenchodiis, the abdominal vertebrae constitute only 25% of the total number of vertebrae);
(3) shortening of the post- and preorbital regions; (4) enlargement of supraneurals, serving as a
support to the high epaxial, post-cranial region; (5) long, rib-like postcleithra, supporting the belly;
(6) long-based dorsal and anal fins, and a well-developed pectoral fin, providing manoeuvering ability.
All of these features can be observed in other deep-body fishes such as Pharmacichthys and Aipichthys
from the Cenomanian of Lebanon (Gayet 1980) or Exellia and Ceratoichthys from the Middle Eocene
of Monte Bolca, Italy (Blot 1969).

Such a body-shape could be the result of a different habitat and mode of life of Parenchodiis
compared to the other genera of the family. Enchodiis and Palaeolyciis were active, mid-water, or
pelagic predators (Goody 1969, 1976), whereas Parenchodiis was a predator which probably lived
among and in vertical reefal elements. Although the body of Parenchodiis is not streamlined, the
structure of the caudal fin indicates a good swimming ability. These features are as follows: (1) a
well-developed and deeply forked caudal fin; (2) a stout peduncle; (3) a reduction in the elements of the

730

PALAEONTOLOGY, VOLUME 30

caudal skeleton as manifested by the fusion of the hypurals, and the lack of uroneurals and epurals;
(4) a large contact area between the bases of the caudal fin rays and the hypurals; (5) a lateral
expansion of the centre of the fused terminal vertebrae (u^ + pui) enlarging the adhering area of the
caudal fin muscles; (6) the lateral keel-like structure formed by the overlapping lateral scutes at the
base of the tail (typical of fast-swimming fishes, Hildebrand 1976).

Parenchodus exhibits extremely the tendency of the family Enchodontidae towards a short body.
This feature existed, though to a lesser degree, in the Old World Enciwdiis-species.

The short and high body of Parenchodus seems to be an adaptation to the special ecological
conditions which prevailed in the sea in the area of Givat Shaul during the late Cenomanian.

Since no evidence for a reefal facies is exhibited in the Kefar Shaul Formation, the existence of such
an environment should be looked for in the partly contemporaneous, rudist-containing Weradim
Formation.

Acknowledgements. The authors wish to thank Messrs M. Dvorachek and Y. Levy for their help in photography,
as well as Mrs B. Katz for editing the English text. The authors are grateful to Dr C. Patterson who critically
reviewed the manuscript.

REFERENCES

ARKiN, Y., BRAUN, M. and STARiNSRY, A. 1965. Type sections of Cretaceous formations in the Jerusalem-Bet
Shemesh area. I. Lithostratigraphy. Geol. Surv. Israel, Stratigr. Sections, 1, 1-37.

AVNiMFXECH, M. A. and SHORESH, R. 1962. Les cephalopodes cenomaniens des environs de Jerusalem. Bull. Soc.
geol. Fr. 7e ser., 4, 528-535.

BLANCKENHORN, M. 1905. Geologie der naeheren Umgebung von Jerusalem. Dte. Geol. Ges. 57, 35-43.

BLOT, J. 1969. Les poissons fossiles du Monte Bolca, classes jusqui’ici dans les families des Carangidae,
Menidae, Ephippidae, Scatophagidae. Memorie Mus. civ. Stor. nat. Verona, 2, 1-525.

BRAUN, M. 1970. Facies changes of the Judea Group in Judea and Samaria. Israel Geol. Soc. Annual Meeting
1970, Jerusalem, 29-35.

GAYET, M. 1980. Contribution a I’etude anatomique et systematique des poissons cenomaniens du Liban,
anciennement places dans les acanthopterygiens. Mem. Mus. natn. Hist. nat. Ser. C, 44, 1-150.

GOODY, p. c. 1969. The relationships of certain Upper Cretaceous teleosts with special reference to the
myctophoids. Bull. Brit. Mus. nat. Hi.st., Geol. Suppl. 7, 1-255.

1976. Enchodus (Tdeostei: Enchodontidae) from the Upper Cretaceous Pierre Shale of Wyoming and South

Dakota, with an evaluation of the North American enchodontid species. Palaeontographica A, 152, 91-112.

HAMAOUi, M. and raab, m. 1965. Type sections of Cretaceous formations in the Jerusalem-Bet Shemesh area. II.
Biostratigraphy. Geol. Surv. Israel, Stratigr. Sections, 1, 1-37.

HILDEBRAND, M. 1976. Analysis of vertebrate structure, 1-654. John Wiley and Sons.

LEWY, z. and raab, m. 1976. Mid-Cretaceous stratigraphy of the Middle East. Annls Mus. Hist. nat. Nice, 4,
32.1-32.20.

NURSALL, J. r. 1963. The hypurapophysis. An important element of the caudal skeleton. Copeia, 2, 458-459.

PICARD, L. 1938. The Geology of new Jerusalem. Bull. Geol. Dept. The Hebrew University, Jerusalem, 2 (1), 1-12.

RIXON, A. E. 1976. Fossil animal remains. Their preparation and conservation, 1 -304. The Athlone Press, University
of London.

SHALEM, N. 1925. II Cenomaniano ad occidente di Gerusalemme. Boll. Soc. geol. ital. 44, 1-18.

1927. La Creta Superiore nei dintorni di Gerusalemme. Ibid. 46, 171-192.

WOODWARD, A. s. 1903. The fossil fishes of the English Chalk. Pt. 2. Palaeontogr. Soc. [Monoryr.], 57, 57-96.

M. RAAB

Geological Survey of Israel, 30 Malkhe Yisrael Street
Jerusalem, 95501, Israel

Y. CHALIFA

Typescript received 24 September 1985 Department of Zoology, The Hebrew University of Jerusalem

Revised typescript received 7 April 1987 Jerusalem, 91904, Israel

RAAB AND CHALIFA: ENCHODONTID FISH FROM ISRAEL

731

ABBREVIATIONS USED IN TEXT-FIGURES

an

angular

pci

postcleithrum

cb

ceratobranchial

Pf

pectoral fin

cl

cleithrum

pfr

pectoral fin radial

CO

coracoid

ph

parhypural

d

dentary

pmx

premaxilla

dpt

dorsal limb of the posttemporal

pop

preopercular

ec

ectopterygoid

ps

parasphenoid

en

endopterygoid

Pt

pterotic

Ids

first dorsal scute

ptf

posttemporal fossa

fpn

first postneural

pu

preural

f

frontal

q

quadrate

hi -6

hypurals 1-6

r

pleural rib

ha

haemal arch

ra

retroarticular

hm

hyomandibular

rd

distal radial

hs

haemal spine

ri

ridge on the centrum Uj + pui

ht

hemitrichium

rm

median radial

ic

intercalar

rp

proximal radial

io

infraorbital

rt

replacement tooth

le

lateral ethmoid

sc

scapula

11

lateral line canal

scf

scapular foramen

Is

lateral scute

scl

supracleithrum

me

mesethmoid

snf

spinal nerve foramen

mn

foramen of the mandibular branch of the

so

supraoccipital

lateral line

sop

SLibopercular

mpt

metapterygoid

sp

sphenotic

mt

largest mandibular tooth

ss

symphyseal suture of the mandibular

mx

maxilla

St

stegural

na

neural arch

sy

symplectic

ns

neural spine

te

external teeth row

op

opercular

ti

internal teeth row

P

parietal

u

ural

pa

palatine

vpt

ventral limb of the posttemporal

..

)

I

r.

Mi ^ ^ h

A

BIOGEOGRAPHY AND EVOLUTION OF AFRICAN
FRESHWATER MOLLUSCS: IMPLICATIONS
OF A MIOCENE ASSEMBLAGE FROM
RUSINGA ISLAND, KENYA

by P. w. KAT

Abstract. A freshwater mollusc fauna from the Early Miocene Gumba beds of Rusinga Island, Kenya, is
re-examined. The fauna consists largely of mutelid bivalves, and two new species of the genus Pleiodoiu P-
lanceolalum and P. rusingae, are described. The Gumba fauna is compared with those of the contemporary
Mohari Formation of the Edward-Albert Rift and the early Miocene Turkana Grits in northwestern Kenya.
None of the early Miocene freshwater faunas thus far described contains representatives of the presently
widespread and diverse bivalve family Unionidae and gastropod family Viviparidae. Ancestors of the recent
African viviparids and some of the unionids are proposed to have been introduced when Africa and Eurasia were
joined by closure of the Tethys Sea. The Miocene faunas are also uninformative as to the origins of the distinctive
endemic mollusc fauna of Lake Tanganyika. Collectively, however, these faunas provide valuable insight into
African freshwater molluscan species diversity during the Miocene, and relationships between Miocene and
extant taxa.

The pre-Pleistocene fossil record of African freshwater mollusc faunas remains little investigated
despite their potential to provide information on molluscan biogeography, evolution, and origins of
endemic assemblages of the older great lakes such as Tanganyika and Malawi (Fuchs 1936; Dartevelle
and Schwetz 1948; Gautier 1965, 1970; Lepersonne 1970). Faunas of the Miocene are perhaps best
known, and have been collected from three main areas: the Edward-Albert Rift (the Mohari
Formation; early Miocene by mammalian faunal inference; see Gautier and Van Damme 1973), near
Lake Turkana (the Turkana Grits, age estimated between 17 and 23 my; see King and Chapman 1972;
Van Couvering and Van Couvering 1976), and around the Winam Gulf area of Lake Victoria
(especially the Gumba Beds on Rusinga Island, age estimated at 19 my; see below) (Verdcourt 1963;
Gautier and Van Damme 1973; Van Damme 1984). These Miocene faunas largely predate the
establishment of the major African rift systems, and are thought to be contemporary with taxa
ancestral to the distinctive endemic mollusc fauna of Lake Tanganyika (Fuchs 1936; Gautier 1967;
Beadle 1974; Andrews and Van Couvering 1975). Also, pre-rift drainage patterns are thought to have
linked the Zaire (Congo) River with large areas of eastern Africa (e.g. Cooke 1958; Fryer and lies 1972;
Beadle 1 974), and taxa of the Mohari mollusc fauna have been proposed to be conspecific with taxa of
the Gumba Beds (Gautier 1967; Gautier and Van Damme 1973).

The Miocene and largely Pleistocene freshwater molluscs of the Edward-Albert Rift have been well
described (e.g. Adam 1957, 1959; Gautier 1965, 1966, 1970; Lepersonne 1970; Gautier and Van
Damme 1973). In contrast, the Rusinga freshwater fossils were only examined by Verdcourt (1963),
and proposed similarities between the two faunas were based largely on incomplete or fragmentary
specimens. Sinee Verdcourt’s (1963) initial description of this fauna, several other collections have
been made and deposited in the Division of Palaeontology of the National Museums of Kenya. These
subsequent collections constitute a marked improvement over the poor material available to
Verdcourt, and have both clarified relationships between the two faunas and their levels of
diversity.

IPalaeontology, Vol. 30, Part 4, 1987, pp. 733-742.|

© The Palaeontological Association

734

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 1. Location of the Gumba Beds on Rusinga Island, Lake Victoria, Kenya.

ENVIRONMENT AND AGE OF THE GUMBA BEDS

The freshwater molluscs examined here occur in the Gumba Beds on the western part of Rusinga
Island (text-fig. 1). The fossils occur in lenticles of coarse sand/gravel consisting of very poorly sorted,
angular particles indicative of a relatively high-energy depositional environment such as a stream
channel. Most of the shells are fragmentary. In situ ‘reefs’ of the freshwater bivalve Etheria elliptica
occur in the same and underlying deposits, all of which lie between the Rusinga Agglomerate (19-5 my;
Basal Agglomerate of Kent 1944) and the Hiwegi Formation (17 to 18-5 my; Tuffaceous Series of Kent
1944), and are contemporary with the ‘Unnamed Formation’ (19 my; Argillaceous Series of Kent
1944) (dates in Van Couvering and Miller 1969). In addition to the gravels, the beds also contain a
variety of other aquatic deposits such as sandstones and deep water greenish siltstones. Kent (1944),
Bishop (1963), and Temple (1969) generally agree on the existence of one or a series of lakes in the
region formed by volcanic damming of drainages. These Miocene lakes, however, were not ancestral
to Lake Victoria, which was formed much later as a result of the development of the African rift
systems, uplifting, and the formation of volcanic fields which reversed some of the westward flowing
drainages of the region (Bishop 1965; see Kendall 1969 for a general background description of the
formation of Lake Victoria).

KAT: MIOCENE FRESHWATER MOLLUSCS FROM KENYA

735

SYSTEMATIC PALAEONTOLOGY

Superfamily unionacea
Family mutelidae
Genus etheria

Etheria elliptica Lamarck, 1807
Text-fig. 2f

Discussion. This is the most common bivalve species present, occurring in in situ ‘reefs’, and as
articulated and fragmentary valves within the gravels. The species is presently widely distributed
within Africa, occurring in the Nile, Zaire (Congo), and Niger River basins as well as on Madagascar.
It is not known from Lakes Tanganyika and Malawi, but occurs in large numbers in Lake Victoria.

Material. Twenty-two single and three articulated valves.

Genus pleiodon

Pleioclon moliarensis (Gautier, 1965)

Text-fig. 2e

1963 Iridina sp., Verdcourt, p. 31, fig. 63a, h.

1963 Miitela sp., Verdcourt, p. 30, fig. 62.

1965 Iridina (Pliodon) moliarensis Gautier, pp. 144-146, pi. 6, figs. 5-7; pi. 7, figs. 1, 3, 5.

1965 Iridina (Pliodon) siihelongata Gautier, pp. 146-147, pi. 7, figs. 7 and 8; pi. 8, fig. 3.

Discussion. This is the largest Pleiodon species present, with an elongate ovate shell. The valves are
compressed to subcompressed, and solid in large specimens. The anterior end is regularly rounded,
and the posterior end often squared or bluntly rounded. The ventral margin is straight to slightly
concave in large specimens; the dorsal margin is generally straight. A broadly rounded posterior ridge
is present. The umbos are low and situated in the anterior third of the shell. Shell sculpture consists of
concentric striae. The dentition is taxodont and well developed, consisting of straight teeth set at a
slight angle to the hinge anterior to the umbo. Anterior adductor and retractor scars are deep, and
dorsal retractor scars are set below the umbo. The posterior adductor scar is distinct, and the
posterior retractor scar is well separated from that of the adductor. The pallial line is usually distinct.

Material. Sixteen partial valves, one complete valve.

Occurrence. This species is also known from the Mohari Formation in the Edward-Albert Rift (Gautier and Van
Damme 1973), and was postulated by Gautier (1966) to be closely related (ancestral?) to recent P. spekii from
Lake Tanganyika.

Pleiodon tavernieri (Gautier, 1965)

Text-fig. 2b, c

1963 Caelatiira sp. Verdcourt, p. 30, fig. 61.

1965 Iridina (Pliodon) tavernieri Gautier, pp. 138-140, pi. 4, figs. 5 and 7; pi. 5, figs. 1, 2, 4, 5, 7.

1970 Pliodon (Pliodon) tavernieri Gautier, p. 123, pi. 5, fig. 7.

Discussion. This is the smallest bivalve species represented, and is variable in shape between
subtriangular and ovate. Specimens are generally subinflated, with a straight or slightly curved
ventral margin and a rounded anterior margin. A posterior ridge is well defined, especially in the
subtriangular specimens, where the posterior end is pointed. The umbos are full, project above the
hinge line, and are located in the anterior one-third to one-half of the shell. Shell sculpture consists of
fine radial lines. The dentition is taxodont, and the hinge plate is relatively straight posteriorly but
curved anteriorly. The teeth are irregular anterior to the umbo, and become chevron-shaped
posteriorly. The beak cavities are deep, and dorsal muscle scars are visible under the hinge plate. The

736

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 2. The mutelid bivalve species of the Gumba Beds, a, Pleiodon lanceolatiiin (holotype, KNMI-RU-
16814). B, c, P. tavernieri (KNMI-RU-14840, 14835). D, P. rusingae (holotype, KNMI-RU-16815). E, P.
moharensis (KNMI-RU-1681 1). e, Etheria elliptica (KNMI-RU- 16809a). All scale bars = 1 cm.

KAT: MIOCENE FRESHWATER MOLLUSCS FROM KENYA

737

anterior adductor and retractor muscle scars are deep and the pallial line distinct. The shell can
become thick in large specimens.

Material. Fifteen partial and three articulated valves.

Occurrence. This species also occurs in the Mohari Formation (Gautier and Van Damme 1973), where the
specimens are apparently less triangular than those from Rusinga.

Pleiodon lanceolatimi sp. nov.

Text-hg. 2a

Description. A medium-sized, very elongate ovate shell more than three times as long as high, moderately inflated.
The anterior and posterior ends are pointed, and the dorsal and ventral margins are straight. A well-developed
posterior ridge is present. The umbos are low and situated in the anterior third of the shell. The hinge plate is very
narrow and set with fine taxodont dentition especially anterior to the umbo. Posterior to the umbo, the teeth
consist of delicate, rather rounded chevrons. The beak cavities are shallow, and the dorsal retractor muscle scars
are located posterior to the umbo. The anterior retractor and adductor muscle scars are distinct, the pallial line is
faint, and the posterior adductor and retractor scars not present in the type specimen. The species diflers from
other known Pleiodon by its elongate shape, pointed anterior end, and delicate dentition.

Material. Holotype KNMI-RU-16814 consisting of a single, partial left valve.

Dimensions (in mm). Length Width

KNMI-RU-16814 58 18

Pleiodott riisingae sp. nov.

Text -tig. 2d

Description. A medium-sized, subcompressed, elongate ovate shell more than twice as long as high. The anterior
end is regularly rounded, and the posterior end bluntly pointed. The ventral margin is straight or very slightly
curved. The dorsal margin is straight and the posterior ridge well defined. The umbos are low, situated in the
anterior quarter of the shell. Dentition is taxodont, and the hinge plate short and curved anterior to the umbo.
Posterior to the umbo, the hinge plate is straight and set with regular, strong teeth lying at an angle to the hinge
line. The beak cavities are shallow, and the dorsal retractor muscle scars are set posterior to the umbo. The
anterior adductor and retractor muscle scars are deep and the pallial line is well defined. The posterior adductor
scar is well defined, and separated from the posterior retractor scar. The species dififers from other Pleiodon
described mainly by its distinctive short and curved anterior hinge plate.

Material. Holotype KNMI-RU-16815; Paratypes KNMI-RU-16817, 16818. 14818, 14834.

Dimensions (in mm). Length Width

KNMI-RU-16815 70 25

Order mesogastropoda
Family ampullariidae
Genus pii.a

Pila ovata Olivier, 1804

Discussion. The single gastropod species from the Gumba Beds is only known from opercula.
Verdcourt (1963), however, mentions that shells of this species and those of the related ampullariid
Lanistes carinatus were collected from other localities on Rusinga Island. Both species are also known
from the Mohari Formation (Gautier and Van Damme 1973). The present geographic range of PUa
ovata extends from Egypt to northern Mozambique and westwards to Nigeria (Brown 1980).

Material. Eight opercula.

738

PALAEONTOLOGY, VOLUME 30

Table 1. Freshwater Molluscan faunas of the Miocene East African sites
mentioned in the text

Rusinga Island
(c. 18-23 my)

Turkana Grits
(c. 1 7-23 my)

Edward-Albert Rift
(early Miocene)

BIVALVES

Etheria elliptica
Pleiodon moharensis
P. tavernieri
P. rusingae
P. lanceolatum

Pleiodon sp.
Mutela sp.

Etheria elliptica
Pleiodon moharensis
P. tavernieri

Aspatharia triangulata
Corhicula consohrina

GASTROPODS

Pila ovata
Lauistes carinatus
Melanoides tuherculata

Pila ovata
Lcmistes carinatus
Melanoides tuherculata
Saulea sp.

Pila ovata
Lanistes carinatus
Melanoides tuherculata

Sources; Verdcourt (1963); Gautier and Van Damme (1973); Van Damme
(1984).

DISCUSSION

East African Miocene freshwater faunas appear to consist mainly of mutelid bivalve taxa such as
Aspatharia, Pleiodon, and Etlieria, and prosobranch gastropod taxa such as Lauistes, Pila, Gahbiella,
Cleopatra, and Melanoides (Table 1). All these taxa with the exception of Pleiodon remain widely
distributed in Africa. Among the bivalves, Etheria is an apparently monotypic genus with a wide
distribution, although soft part anatomy and molecular genetics of widely separated populations
have not been compared. Aspatharia has diversified into an as yet unknown number of species (48
according to Pilsbry and Bequaert 1927; 9 according to Haas 1969), some of which are ostensibly
widely distributed, while others are endemic to single lakes such as Victoria and Malawi. Among the
gastropods, Pila, Lauistes, and Melanoides are also encountered outside Africa (e.g. Brandt 1974;
Smith and Kersaw 1979; Brown 1980). Lauistes (19 African species) and Melanoides (27 African
species) are especially diverse (Brown 1980). The 20 species of Cleopatra and 22 species of Gabbiella
listed by Brown (1980) are essentially African; a few species in each genus are widely distributed, but
many are confined to particular localities or lakes.

Pleiodon is poorly represented in extant African freshwater faunas. The fossil record of the genus
dates back to the Upper Cretaceous, where two taxa are encountered in the Nubian Sandstone near
Aswan in Egypt (Cox 1955). With the addition of two new species from the Gumba Beds, the number
of Pleiodon taxa described from the Miocene has increased to four (an undescribed species from the
Turkana Grits could increase this total; see Van Damme 1984). P. inoharensis and P. tavernieri
apparently were widely distributed in pre-rift drainages. The more restricted occurrence of P.
lanceolatum and P. rusingae could either be indicative of habitat differences between the sites or the
occurrence of endemic taxa in the Miocene Winam Gulf area lakes. Pleiodon had further diversified by
the Pleistocene: early Pleistocene Edward-Albert rift deposits contain six species, including P.
tavernieri and P. moharensis (Gautier 1970). Elsewhere, Pleiodon sp. occurs in the Plio-Pleistocene
Koobi Fora formation near Lake Turkana (Williamson 1981). This diverse assemblage declined

KAT: MIOCENE FRESHWATER MOLLUSCS FROM KENYA

739

during the later Pleistocene, and Pleiodon is now represented by two species with relictual
distributions: P. spekii in Lake Tanganyika (previous distribution: Edward-Albert rift (Gautier 1970)
and Lake Kivu (Dartevelle 1948)) and Pila ovata in West Africa (previous distribution: Edward-
Albert rift (Gautier 1970)). Pain and Woodward (1964) describe an additional West African species, P.
water St oni, but I agree with Gautier (1966) that this is probably an ecotypic variant of P. ovata.

The Rusinga deposits lack representatives of Aspatharia, Gabhielkp Cleopatra, and Melanoides,
taxa reasonably common in the Edward-Albert rift (Gautier and Van Damme 1973). These differences
could be due to dissimilarities in the habitats represented: the Mohari Formation has been
hypothesized to represent a marshy environment close to a river (Gautier and Van Damme 1973), and
the fauna is dominated numerically by taxa such as Lanistes and Cleopatra, which presently occur in
swamps and slowly flowing rivers (Brown 1980). In contrast, the coarse, poorly sorted sediments and
the predominance of E. elliptica (which is very rare in the Mohari) indicates a fluviatile and higher
energy environment for the Gumba Beds. Melanoides occurs in contemporary sediments at Nyakach,
about 50 km east of Rusinga (pers. obs.).

Entirely absent from all early Miocene deposits examined to date are representatives of the
presently widespread and diverse bivalve family Unionidae and gastropod family Viviparidae. It is
possible that ancestors of some of the present African unionids (especially the genus Caelatura) and all
viviparids were introduced after these deposits were formed, when Africa and Eurasia were joined by
the closure of the Tethys Sea (e.g. Dewey and Bird 1970; Andrews and Van Couvering 1975). This
event is also thought to have mediated the introduction of several mammal taxa (Cooke 1968), all
African cyprinid fishes (Bowmaker et al. 1978), a large number of the present Pan-Ethiopian
freshwater invertebrates other than molluscs (Harrison 1978), and many flowering plant taxa (Smith
1973).

Newton (1920) and Prashad (1928), however, consider the Viviparidae to have been present on
ancient Gondwanaland, based on the occurrence of an upper Cretaceous (Tertiary according to
Dartevelle 1948) ‘viviparid’ in southern Africa, and a high degree of similarity between African and
Indian viviparid species. Eater, Rennie (1943) described ‘viviparid’ taxa from the Early Jurassic of
Mozambique. All these fossils, however, are poorly preserved, and seem more likely to be
representatives of Cleopatra (Thiariidae) which Prashad (1928) included in the Viviparidae. Cleopatra
was widespread in the Miocene, and now occurs throughout Africa and Madagascar. The earliest
unambiguous viviparid {a\a(Bellatnya and N eothaiima) occur in Pliocene deposits (e.g. Gautier 1970;
Williamson 1981; Van Damme 1984).

The present distribution (text-fig. 3) and fossil occurrences of the Viviparidae outside Africa also
argue against their presence on Gondwanaland. In South America, viviparids are only known from
the Palaeocene, and these representatives became extinct a short time later. Parodiz (1969) believes
these viviparids, as well as other elements of the South American Palaeocene molluscan fauna, to have
been derived from North American ancestors at the end of the Cretaceous. Similarly, viviparids are
unknown from Madagascar, and apparently have no fossil record in Australia (Prashad, 1928). Smith
and Kersaw (1979) believe the present viviparid fauna, as well as several other Australian fresh-
water molluscs, to have been derived from Asian ancestors in relatively recent times. The similarity
between Recent Indian and African viviparids is more likely to have resulted from shared ancestry
with a taxon that invaded Africa relatively recently than shared ancestry with a Gondwanaland
taxon. Molecular genetic and karyologic data support a relatively recent divergence among African
viviparid taxa.

The taxonomy of the numerous recent Caelatura species and their relationships to unionid taxa
outside Africa is poorly understood. Unambiguous fossil Caelatura are, like viviparids, first
encountered in Pliocene and Pleistocene deposits. Recently, Heard and Vail (1976) placed southern
African Cafferia caffra in the genus Unio (Unionidae) based on shell sculpture, some soft-part
anatomy, and larval and reproductive characteristics. Unio s.s. is otherwise only known from the
Palaearctic and Heard and Vail suggest the south African taxon to be a ‘biogeographic relict’.
Regardless of its generie placement (see Mandahl-Barth 1982), Cafferia appears to have a fossil record
dating to the Upper Cretaceous in the Transvaal (Modell 1964), which would imply a long-standing

740

PALAEONTOLOGY, VOLUME 30

presence of some unionid taxa in Africa. Similarly, three endemic and monotypic unionid genera
(Brazzaea, Pseiidospatlia, and Moncetia) occur in Lake Tanganyika, and one monotypie genus
(Prisodontopsis) is endemic to Lake Mweru (Pain and Woodward 1968; pers. obs.). The Tanganyika
unionids exhibit anatomical evidence of shared ancestry (pers. obs.), but their relationship to
Prisodontopsis and Cafferia is unknown. All five taxa, however, clearly represent a distinct and
separate group of unionids from Caelatnra, but much more fossil evidence is necessary to elucidate
their origin and past distributions.

TBXT-riG. 3. Distribution of extant members of the gastropod family Viviparidae (based on Starobogatov 1970).

Also missing from the East African Miocene faunas are gastropods that could elucidate the origin
of the thalassoid gastropod assemblage of Lake Tanganyika. This highly distinctive group consists of
about twenty-five eonchologically diverse species distributed among sixteen largely monotypic
endemic genera (no other African lake appears to contain even a single endemic snail genus), and is
thought by Brown (1980) to have originated from Miocene thiariid taxa living in eentral Africa before
formation of the lake. Brooks (1950) and subsequent authors such as Boss (1978) believe the thalassoid
snails to have evolved entirely within Lake Tanganyika, and thus to represent an ancient endemic
radiation in which morphologic differentiation has transcended the species level. While the fossil
record indicates that several mollusc taxa now endemic to the lake previously had more extensive
geographic ranges (e.g. the viviparid Neothauma tangcmyicense whieh is found in Pleistocene deposits
of the Edward- Albert rift and Lake Rukwa (Cox 1939; Gautier 1970); species of the thiariid genus
Lavigeria which are also found near Lake Rukwa (Cox 1939); and Pleiodon spekii mentioned above),
there is to date no fossil evidence of thalassoid species having occurred outside the Tanganyika basin.
The soft-part anatomy of these snails is presently too imperfectly known to determine if there exist
shared anatomical characters (e.g. see Davis’ ( 1 979) analysis of the eonchologically variable Mekong
River triculine gastropods) indicative of a truly endemic radiation, and biochemical data which could
indicate a possible shared genetie identity are in the early stages of being gathered. The origin and
evolution of the thalassoid snails must therefore remain speculative at this point.

KAT; MIOCENE FRESHWATER MOLLUSCS FROM KENYA

741

Acknowledgements. I am grateful to Blaire Van Valkenburgh, Mark Teaford. Kamoya Kimeu, Musombi
Kibberene, and especially Alan Walker for their help in collecting the specimens that made this study possible.
Comments by B. Verdcourt and an anonymous reviewer improved the manuscript. Funding was provided by a
grant from the American Philosophical Society and NSF 83-15195.

REFERENCES

ADAM, w. 1957. Mollusques Quaternaires de la region du Lac Edouard. Expl. Parc Nat. Albert. Mis.sion J. de
Heinzelin de Braucourt (1950), 3, 1-172.

1959. Mollusques Pleistocenes de la region du Lac Albert et de la Semliki. Ann. Mas. Ro\>. Congo Beige, Sci.

Geo/. 25, 1-149.

ANDREWS, p. and van couvering, j. a. h. 1975. Palaeoenvironments in the African Miocene. Contrih. Primatol. 5,
62-103.

BEADLE, L. c. 1974. The inland waters of tropical Africa. 274 pp. Longmans, London.

BISHOP, w. w. 1963. The later Tertiary and Pleistocene in eastern equatorial Africa. In howell, c. and
BOURLIERE, R. (eds.), African ecology and human evolution. 246-275. Wenner Gren Foundation, New York.

1 965. Quaternary geology and geomorphology in the Albertine rift valley, Uganda. Geol Soc. America Spec.

Pap. 84, 293-321.

BOSS, K.. J. 1978. On the evolution of gastropods in ancient lakes. In fretter, v. and peake, j. (eds.). Pulmonates.

Vol. 2 A. Systematics, evolution, ecology. 385-428. Academic Press, London.

BOWMAKER, A. p., JACKSON, p. B. N. and JUBB, R. A. 1978. Freshwater fishes. In werger, m. j. a. (ed.). Bio-
geography and ecology of southern Africa. 1 183 1219. Junk, The Hague.

BRANDT, R. A. 1974. The non-marine aquatic Mollusca of Thailand. Arch. Molluskenk. 105, 1 423.

BROOKS, J. L. 1950. Speciation in ancient lakes. Quart. Rev. Biol. 25, 131-176.

BROWN, D. s. 1980. Freshwater snails of Africa and their medical importance, 487 pp. Taylor and Francis,
London.

COOKE, H. B. s. 1958. Observations relating to Quaternary environments in East and Southern Africa. Trans.
Proc. Geol. Soc. S. Afr. 60, Annexure 5, 1-73.

1968. Evolution of mammals on southern continents. II. The fossil mammal fauna of Africa. Quart. Rev.
Biol. 43, 234-264.

cox, L. R. 1939. Mollusca from the Quaternary deposits of Lake Rukwa (Tanganyika Territory). Proc. Malacol.
Soc. London, 23, 242-252.

1955. Lamellibranchiata from the Nubian Sandstone series of Egypt. In attia, m. t. (ed.). Topography,

geology, and iron ore deposits of the district ea.st of Aswan. 237-244 (Appendix 1). Min. Res., Dept. Geol.
Surv., Cairo, Egypt.

DARTEVELLE, E. 1948. Contribution a la faune malacologique des terrasses de la region des lacs Edouard et Kivu.
Bull. Serv. Geol. Congo Beige et Ruanda-Urundi, 3, 97 145.

1952. Viviparidae vivants et fossiles d’Afrique. Ann. Soc. roy. Zook Beige. 83, 153-181.

and SCHWETZ, j. 1948. Le Lac Tanganyika. Mem. Inst. roy. Congo Beige, Sci. Nat. Med. 14, I I 18.

DAVIS, G. M. 1979. The origin and evolution of the Pomatiopsidae, with emphasis on the Mekong River
Triculinae. Monogr. Acad. Nat. Sci. Phila. 20, 1-120.

DEWEY, J. E. and BIRD, J. M. 1970. Mountain belts and the new global tectonics. J. Geophys. Res. 75, 2625
2647.

FRYER, G. and ILES, T. D. 1972. The cichlid fishes of the great lakes of Africa, 641 pp. Oliver and Boyd,
Edinburgh.

FUCHS, V. E. 1936. Extinct Pleistocene Molluscs from Lake Edward, Uganda, and their bearing on the
Tanganyika problem. Zool. J. Linn. Soc. 40, 93-106.

GAUTIER, A. 1965. Geological investigation in the Sinda-Mohari (Ituri, NE Congo). 161 pp. Publ. Ganda-Congo,
Univ. Ghent.

1966. Geschiedenis en evolutie van de zoetwatermolluskenfauna in de Albert en Edwardmeren slenk.

Natiiurwet. Tijdschr. 48, 3-24.

1967. New observations on the Later Tertiary and Early Quaternary in the Western Rift: The strati-
graphic and palaeontological evidence. In bishop, w. w. and clarke, j. d. (eds.). Background to evolution in
Africa, 73-87. Univ. Chicago Press, Chicago.

1970. Fossil freshwater molluscs of the Lake Albert-Edward rift. Doc. Zook, Mus. Row Afr. Centr.. ser 8.

67, 1-160.

742

PALAEONTOLOGY, VOLUME 30

GAUTIER, A. and VAN DAMME, D. 1973. A revision of the Miocene freshwater mollusks of the Mohari Formation.
Ann. Mils. Roy. Afr. Centr. Sci. Geol. 677, 1-160.

HAAS. F. 1969. Superfamilia Unionacea. In wermuth, a. (ed.). Das Tierreich, Vol. 88, 1-663. Walter de Gruyter
and Co., Berlin.

HARRISON, A. D. 1978. Freshwater invertebrates (except molluscs). In werger, m. j. a. (ed.). Biogeography and
ecology of southern A frica, 1 139-1152. Junk, The Hague.
heard, w. h. and vail, v. a. 1976. The systematic position of Unio ( Pelecypoda: Unionoida; Unionidae).
Zoo/. T/r. 11, 45-58.

KENDALL, R. L. 1969. All ecological history of the Lake Victoria basin. Ecol. Monogr. 39, 121-176.

KENT, p. E. 1944. The Miocene beds of Kavirondo, Kenya. Quart. J. Geol. Soc. London, 100, 85 1 18.

KING, B. D. and CHAPMAN, G. R. 1972. Volcanism of the Kenya Rift Valley. Phil. Trans. R. Soc., A 271, 185-208.
LEPERSONNE, J. 1970. Revision of the fauna and the stratigraphy of the fossiliferous localities of the Lake
Edward-Lake Albert rift (Congo). Ann. Mus. Afr. Centr., Sci. Geol. 67, 169-207.

MANDAHL-BARTH, G. 1982. Studics on African freshwater bivalves. I. The Unionidae. Unpubl. manuscr., Danish
Bilharziasis Laboratory, Charlottenlund.

MODELL, H. 1964. Das naturliche Sytem der Najaden. 3. Arch. Molluskenk. 93, 71-126.

NEWTON, R. B. 1920. On some freshwater fossils from central South Africa. Ann. Mag. Nat. Hist. 5, ser. 9,
241-249.

PAIN, T. and WOODWARD, F. R. 1964. A monograph of the African bivalves of the genus Pleiodon Conrad. Ann.
Mus. Roy. Afr. Centr., Sci. Zool. 130, 1-33.

1968. A monograph of the African bivalves of the genera Brazzaea Bourguignat, Mweruella Haas,

Prisodonlopsis Tomlin, and Pseudospatha Simpson. Rev. Zool. Bot. Afr. 77, 190-220.

PARODiz, J. J. 1969. The Tertiary non-marine Mollusca of South America. Ann. Carnegie Mus. 40, 5-239.
PILSBRY, H. A. and BEQUAERT, J. 1927. The aquatic molluscs of the Belgian Congo. Bull. Anier. Mus. nat. Hist. 53,
69-602.

PRASHAD, B. 1928. Recent and fossil Viviparidae. A study in distribution, evolution, and palaeogeography.
Mem Indian Mus. 8, 153-251.

RENNIE, J. v. L. 1943. Fossils from the Lebombo volcanic formation. Boll. Serv. Indust. Minas Geol. Mozambique,
5, 120-123.

SMITH, A. c. 1973. Angiosperm evolution and the relationship of the floras of Africa and America. In beggers,
B. J., AYENSU, E. s. and DUCKWORTH, w. D. (eds.). Tropical forest ecosystems in Africa and South America:
a comparative review, 49-61. Smithsonian Institution, Washington, DC.

SMITH, B. J. and KERSAW, R. c. 1979. Field guide to the non-marine molluscs of southeastern Australia, 139 pp.
Australian National University Press, Canberra.

STAROBOGATOV, YA. I. 1970. Mollusc fauna and zoogeographical partitioning of continental water reservoirs of
the world. Akad. Nauk. SSSR Zool. Inst. Leningr. 1970, 3-373 [In Russian].

TEMPLE, p. H. 1969. Some basic implications of a revised geological history for Lake Victoria. Biol. J. Linn.
Soc. 1, 363-371.

VAN couvERiNG, J. A. and MILLER, J. A. 1969. Miocene stratigraphy and age determinations, Rusinga Island,
Kenya. Nature, 221, 628-632.

VAN COUVERING, J. A. H. and VAN COUVERING, J. A. 1976. Early Miocene mammal fossils from East Africa.
In FAGAN. B. (ed.). Human origins: Louis Leakey and the East African evidence, 155-208. W. A. Benjamin
Series of Anthropology, Berkeley.

VAN DAMME, D. 1984. The freshwater Mollusca of Northern Africa. Distribution, biogeography, and paleoecology,
164 pp. Dr W. Junk Publishers, Dordrecht, Boston, and Lancaster.

VERDCOLiRT, B. 1963. The Miocene non-marine Mollusca of Rusinga Island, Lake Victoria and other localities
in Kenya. Palaeontographica, 121, 1-37.

WILLIAMSON, p. G. 1981. Palaeontological documentation of speciation in Cenozoic mollusks from Turkana
Basin. Nature, 293, 437-443.

PIETER W. KAT

Malacology Section
National Museums of Kenya

Typescript received 14 February 1986 Pq 40658

Revised typescript received 10 October 1986 Nairobi, Kenya

MENISCATE TRACE FOSSILS AND THE
MUENSTERIA-TAENIDIUM PROBLEM

Zjj^assuntad’alessandro and richard g. bromley

Abstract. The systematics of meniscate trace fossils are in need of revision. Most authors follow Seilacher and
call unbranched, unlined meniscate burrow fills Miiensteria. However, Sternberg’s original description of
Muensteria is highly confused, involving true algae, a coprolite, and forms of Chondrites, and the name cannot be
considered available for trace fossils. Re-examination of the original description of Taenidium Heer has revealed
that the earliest ichnospecies erected, T. serpentinum, corresponds closely to the forms called Muensteria in the
more recent literature. Subsequent ichnospecies placed in Taenidium include septate branched forms that require
a new name. It is these branched meniscate burrows that commonly have been referred to Taenidium. We
designate the branched forms Cladichnus ichnogen. nov. on the basis of T.fischeri Heer. T. satanassi ichnosp. nov.
is described from the lower Tertiary of Italy. Other ichnogenera of meniscate trace fossils are briefly discussed,
including Nereites, Keckia, Scolicia, Psammichnites, Scoyenia, Beaconites, Phoehicimus, and Ancoricimus.

Trace fossils having a cylindrical form and meniscate structure are common in many settings.
Generally they represent the active packing of sediment behind animals moving through the loose
substrate for food, or vertical escape traces. Although meniscate structures are particularly
characteristic of shallower marine deposits, they also occur in continental and deep marine basinal
settings. Authors today tend to follow Seilacher (1958, 1962, 1964) in considering that Muensteria
embraces these ‘stuffed linear burrows’, but this conflicts with the original diagnosis of Sternberg
(1833, p. 31): "Frons coriacea, fistulosa, cylindracea aut simplex caespitosa aggregata, aut dichotoma,
transverse elevato-striata, striis interruptis creberrimis. Sporangia punctiformis, sparsa, creberrima,
inter strias laminae frondis immersa' (Approximate translation: Leathery frond, pustulose, cylindrical,
or simply aggregated as a bush, or dichotomous; transversely striated in relief, with dense,
interrupted striae. Sporangia punctiform, scattered, very dense, immersed between the laminae of
the frond.)

In contrast, "Taenidium' is generally used for branched meniscate structures, although this conflicts
with Heer’s (1877) original diagnosis: "Frons cylindrica, fistulosa, plerumque simplex, rarius ramosa,
annulata, dissepimentis instructa' (Cylindrical fronds, pustulose, usually simple, rarely branched,
annulated, provided with partitions.)

A close inspection of the status of these names is clearly overdue (see e.g. Frey and Howard 1985,
p. 378).

STRUCTURE OF MENISCATE TRACE FOSSILS

There are several reasons for the confusion that at present characterizes the systematics of meniscate
trace fossils. Among the most important is the heterogeneous nature of the taxobases used. These
comprise: ( 1 ) a meniscate backfill, together with the presence or absence of (2) discrete wall structure,
(3) wall ornament, and (4) true branching. Not all of these features are suitable or easy to use as
diagnostic characters. Another cause of taxonomic difficulty is that many of the taxa were introduced
as algae and were thus described in non-ichnological terms. These problems, of course, are far from
restricted to the meniscate trace fossils, but are of general occurrence in older ichnological systematics
(Pemberton and Frey 1982).

IPalaconlology, Vol. 30, Part 4, 1987, pp. 743-763|

© The Palaeontological Association

744

PALAEONTOLOGY, VOLUME 30

Backfill. Meniscate backfill occurs in many morphologically distinctive groups of trace fossils. It is
eonsidered an essential feature of such ichnogenera as Ancoriclmus and Scolicia, but also occurs as an
accessory feature, albeit rarely, in ichnospecies of Ophiomorpha and Teichichnus (Frey and Bromley
1985), Chondrites, and many others. As a basic result of any tunnelling that involved active stuffing
behind an animal, meniscate backfill alone is not particularly suitable as an ichnological taxobase.

Wall structure. The boundary between burrow fill and undisturbed surrounding sediment may
display a host of different features. The boundary may be defined by a clean discontinuity surface
caused by a different grain orientation on either side of thejunction (e.g. Heinberg 1973, 1974), or by a
difference in consistency of the two sediments. Diagenesis may enhance this difference to produce a
coloured halo or local concretionary effect. A thin film of clay minerals or carbonaceous material may
occur at the boundary, but this is commonly difficult to detect and is lost in weathered material; we do
not regard this film as a true ‘lining’. Alternatively, the burrowing organism may have actively
deposited a true lining on the open wall, as in Ophiomorpha ichnospp. (Frey et al. 1978), or the animal’s
burrowing technique may have produced a superficial zone around the periphery of the fill with a
special construction, as in A. ancoriclmus (Heinberg 1974). In these cases, therefore, the boundary is
marked by a zone of sediment having a distinctive structure. Finally, local compaction of sediment
caused by the construction of the burrow may have produced a preservable boundary feature. In a
variable lithology, the characteristics of the burrow boundary may change locally, reflecting the
original sediment consistency and consequent changes in burrowing technique. This is also the case
for the scratch ornament, which may well be preserved in a firm, fine-grained lithology but is lost in a
soft, coarse sand (Fiirsich 1974, p. 22).

A corollary of this problem of substrate variation is that of preservation in general. The ‘same
burrow’ may exhibit widely different morphologies in different preservational situations, e.g. in full
relief within a sandstone, or epirelief at the sandstone boundaries. In some cases these widely different
morphologies have been lumped together as a single ichnotaxon on the basis of their belonging to the
‘same burrow’ (see the Scalarituha/ Phyllodocites/ N ereites/ N eonereites problem: Chamberlain 1971,
1978).

Branching. The main problem here is that of ramification. Burrows may branch in several distinct
fashions (Bromley and Frey 1974, fig. 6):

A. On irregular rock surfaces, simple intersection of burrows can be mistaken easily for true
branches (text-fig. 1a); this is particularly so where a high degree of compaction has brought burrows
that originally lay at different levels in the sediment into close juxtaposition (text-figs. 2 and 3).

B. Commonly a deposit feeder reworks the backfill of a pre-existing burrow and then leaves the
earlier course (text-fig. 1b) at an angle to produce a side branch. We suggest calling this type

TEXT-FIG. 1. Diagrammatic representation of four distinct modes of producing branched biogenic structures, a,
false branching, where accidental intersections on an irregular surface can give the spurious impression of
branching, b, secondary successive branching, where an animal followed the fill of an earlier burrow for some
distance, giving the false impression of branched burrows, c, primary successive branching, where the same
animal produced a systematically branched structure, actively filled, d, simultaneous branching, where a network
of open burrows is produced; the simultaneous stuffing of both branches would produce this pattern, but it is a

process that is presumably not possible.

D'ALESSANDRO AND BROMLEY: MENISCATE TRACE FOSSILS

745

TEXT-FIG. 2. Taenidiiim satanassi ichnosp.
nov. Field photograph of material displaying
the opposite colour-play to the usual: the
pelleted sediment here is the paler. Scale,
1 cm.

‘secondary successive branching’, as it is produced by two individuals in successive burrowing
operations. In many such cases the direction of movement (meniscus orientation) is opposite in the
two burrows, which reveals the type of structure immediately (e.g. Brady 1947, pi. 69, fig. 1; herein
text-fig. 2). In other cases, however, where the menisci are concordant (e.g. Bromley and Asgaard 1979,
fig. 10) it is not easy to distinguish this branching category from the next.

C. In burrow systems where successive probings are made after backfilling previous probes (e.g.
Chondrites'. Seilacher 1957, fig. 2; Ferguson 1965), a single individual produces a branched structure
having a characteristic backfill pattern (text-fig. Ic). We call this type of bifurcation ‘primary
successive branching’ (text-fig. 4).

D. Finally, where an open gallery is produced, side branches and dichotomous bifurcations exist
simultaneously as empty habitations. This type we call ‘simultaneous branching’. The fill might be
emplaced either in successive operations, i.e. one branch filled at a time, as in Ophiomorpha (Bromley
and Frey 1974, fig. 1 1), or alternatively, the two branches might be filled at more or less the same time.
In the latter case, a backfill pattern is produced (text-fig. Id), but although this structure has been
illustrated in sketches (e.g. Seilacher 1955, fig. 5.83) it is not easy to understand the mechanics
of the process. Probably ‘simultaneous backfill’ does not exist. It is easy to envisage this structure
arising in a growing alga, however, the menisci representing successive terminations of the
thallus (e.g. Fischer-Ooster 1858, pi. 7, fig. 3). We emphasize these difficulties because branching
patterns play an important role in understanding the history of the taxonomy of meniscate
trace fossils.

Algal origins. Most ichnogenera erected in the nineteenth century were initially described as body
fossils of fucoid algae (Hantzschel 1975, p. 14). This fact renders interpretation of the initial
descriptions very difficult, owing to the terminology used and the unintentional bias of original
authors. ‘Sporangia’ usually may be interpreted as pellets, ‘epithelia’ as walling material, etc. A greater

746

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 3. Taenidium satanassi ichnosp. nov. showing
false branching (cf. text-fig. 1a). Scale, 1 cm.

problem, however, is the tendency to assume that 'fucoid’ fossils were branched in the same way as the
extant seaweed Fiicus (text-fig. 3), a tendency that was encouraged by discoveries of truly branched
trace fossils (e.g. Chondrites). Thus even unbranched trace fossils such as Cruziana were described as
branched, owing to the misinterpretation of intersections (Osgood 1975, p. 6).

Meniscate ichnotaxa share all these problems. On account of the 'branching bias’, authors tended
to regard crowded, intersecting trace fossils as more ‘complete’ than isolated ones, and secondary
successive branching and false branching were readily interpreted as true ramifications. In this way,
crowded or bunched trace fossils were selected for illustration, and consequently as type material,
whereas in trace fossils such material preferably should be avoided. As explained above, branching in
trace fossils is a complicated and critical matter; it is not sufficient merely to repeat the original
author’s statement regarding branching in algal terms when emending the diagnosis of a fucoid taxon
in terms of trace fossils.

Taxohases for meniscate trace fossils

Ichnotaxa are based upon morphological features of trace fossils, primarily those that reflect
behavioural qualities of the excavating organism. Morphology is also influenced, however, by
external factors such as stratinomic conditions and diagenesis. Such morphological features are
unsuitable as taxobases and primary features reflecting behaviour should be chosen if possible
(Seilacher 1953; Fursich 1973; Bromley and Frey 1974). For example, as Smith and Crimes (1983)
pointed out, Scolicia and Subphyllochorda represent closely similar if not identical patterns of
behaviour. Their preservation, however, is different — Scolicia being a concave epirelief and
Subphyllochorda a full relief expression of the same structure. Hantzschel (1975) suggested, therefore,
that Subphyllochorda be considered a junior synonym of Scolicia. In contrast, the structures
Ophiomorpha, Thalassinoides, Spongeliomorpha, and Gyrolithes may all be found in mutual
connection as a single burrow system (Fursich 1973; Bromley and Frey 1974; Pemberton and Frey
1982); yet they represent different behavioural responses of the burrowing organism to environmental
and biological factors, and so should not be placed in synonymy. In this way, the ichnotaxon unites
behavioural patterns of the same type. This is not to question the importance of preservational
aspects: taphonomic processes are closely linked to sedimentary facies and the dominance of a
particular preservational style in an ichnotaxon in a given occurrence provides useful information on
depositional environments.

D’ALESSANDRO AND BROMLEY: MENISCATE TRACE FOSSILS

747

CONCLUSIONS

The erection of several ichnogenera for the septate or meniscate trace fossils is clearly justified on the
basis of their wall details and presence or absence of branching (see below): walled structures include
Phoebichmis, Ancorichnus, and an unnamed branching group; unwalled structures containing
longitudinal canals or ridges are distinguished as Scolicia and Psammichnites', Nereites has a unique
disturbance zone around it, whereas Scoyenia has no wall but is ornamented by longitudinal
scratches; Beaconites is dubious, having a weak wall or none at all, and it should probably be included
in Taenidium as defined herein; Keckia appears to be a fwmen dubium and is probably a heterogeneous
taxon, including elements resembling several other taxa.

The unwalled meniscate trace fossils pose the greatest problem. Muensteria, in spite of the general
use of this name today, must be regarded as unavailable for trace fossils. On the basis of its first
ichnospecies, Taenidium is available for the unbranched ichnospecies of Muensteria of authors, and T.
serpentinum remains its type ichnospecies. However, Heer’s T. fischeri and T. lusitanicum must be
transferred to the flabelliformly branched and annulated Cladichnus ichnogen. nov.

SYSTEMATIC PALAEONTOLOGY

In our attempt to clarify the meaning of Muensteria and Taenidium, we reinterpret their original diagnoses in
terms of ichnotaxobases and must therefore examine briefly the distinguishing features of related meniscate
ichnogenera. We suggest synonyms among these ichnogenera (although we have not examined type material)
which are treated below in chronological order.

Muensteria Sternberg, 1833

Discussion. Many authors have used the taxon Muensteria in recent years to define simple, unlined,
unbranched meniscate structures (e.g. Seilacher 1962, 1964; Fiirsich 1974; Chamberlain 1977;
Pemberton and Frey 1984; Frey and Howard 1985), despite the fact that Hantzschel (1962, 1975)
clearly demonstrated the weakness of the taxon. Thus, although Muensteria for these authors is a
fairly well-defined concept, Muensteria Sternberg remains largely uninterpreted: it is, as Hantzschel
(1975) stated, a heterogeneous genus.

Sternberg ( 1 833) based his algal genus Muensteria on six species. We have not been able to locate his
type material. The generic diagnosis has been given above. The diagnoses and illustrations of the six
species are uneven. The first three were based on Jurassic material, from Solnhofen lithographic
limestone.

The first species, M. clavata, was not illustrated, but reference was made to Brongniart’s (1828)
Fucoides encoelioides (Ekdale et al. 1984, fig. 1-1, top left). Sternberg disagreed with Brongniart that
the species was closely related to the recent alga Encoelium buUosum, and therefore altered its species
name. Sternberg’s diagnosis for M. clavata is a shortened version of Brongniart’s (1828, p. 55) for F.
encoelioides. Brongniart’s sample was collected in the Jurassic limestone of Solnhofen and probably
comprised poorly preserved plant remains, as did Sternberg’s fossils; in fact the trace fossil
"Muensteria' of authors is unknown in the Solnhofen sediments, which were deposited in largely
anoxic environments (F. T. Fiirsich, pers. comm. 1984).

Sternberg’s second species, M. vermicularis, differs from M. clavata only in being solitary. The
illustration supports the description, showing a particularly unimpressive specimen. The palaeo-
botanist Andrews (1955, p. 191) chose this to be the type species of the genus, identifying the plant as
an ‘alga?’.

The third species, M. lacunosa, appears to be a coprolite, and was considered as such by Schenk
(1864, p. 296). The remaining three species derive from Vienna and are branched. M. hoessii
(Sternberg, 1833, pi. 6, fig. 4; cf. text-fig. 4) appears to be a Chondrites, and its diagnosis describes it as
dichotomous. M. flagellaris Sternberg, 1833, pi. 8, fig. 3, also strongly resembles Chondrites; his
diagnosis distinguishes it on the basis of short side branches. Finally, M. geniculata has a
characteristic spreite and has been designated type ichnospecies of Hydrancylus.

748

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 4. Chondrites hoessii (Sternberg, 1833), MGUH
17555, from the same horizon as Taenidium satauassi
ichnosp. nov., showing the faintly meniscate fill and primary
successive branching. Scale, 1 cm.

We conclude that Sternberg’s genus has no single claim for validity, so Muensteria must be
considered unavailable as an ichnogenus. As a plant taxon it is probably a nomen diibium.

The next author to use Muensteria was Schafhiiutl (1851) whose M. annulata is a clearly branched
trace fossil (apparently in a primary successive manner) but extremely poorly defined.

Fischer-Ooster (1858) greatly complicated the issue by erecting three subgenera for Muensteria.
The only one that concerns us is M. (Keckia), into which he placed M. annulata Schafhautl and,
mysteriously as a synonym, K. annulata docker, 1841. Also in M. (Keckia) he placed Sternberg’s M.
hoessii.

Heer (1877) further confused matters by erecting a simple branched species, M. antiqua, and
by introducing a further genus, Taenidium, that mainly differs from Muensteria in possessing
menisci (Heer 1877, p. 117), although Heer’s illustration of M. hoessii (1877, pi. 69, fig. 3) clearly
shows menisci.

Nereites MacLeay, 1839

[= Phyllodocites Geinitz, 1867; Scalarituha Weller, 1899; Neonereites Seilacher, 1960;

Maldanidopsis Plicka, 1973.]

Discussion. The close relationship between Nereites, Neonereites, and the meniscate burrow
Scalarituha was demonstrated by Seilacher and Meischner (1964). Although these ichnogenera have
highly distinctive morphologies, they appear to be merely preservational variants of the same
behavioural expression (Ekdale et al. 1984, figs. 2-5). This equivalence was emphasized by
Chamberlain (1971) and Seilacher (1983).

Thus, although Scalarituha has the appearance of a simple meniscate structure corresponding more
or less to the concept of Muensteria of authors, in fact it is merely the cylindrical septate axis of a more
complex structure, as is demonstrated by the common presence of a more or less visible halo. In Frey
et fl/.’s (1984, fig. 1a) illustration of Weller’s holotype of S. missouriensis, this halo is visible and
corresponds to the disturbance zone surrounding the central meniscate fill that is emphasized in
Nereites preservation. Seilacher (1983) advocated placing Scalarituha and Neonereites in synonymy
with Nereites, thus completing the trend started by Chamberlain (1971, 1978). The more important
ichnogenera included by Seilacher in this group of synonyms are Scalarituha, Phyllodocites,
Neonereites, and Helminthoida; however, the last-named has a non-septate central fill and therefore is
distinct.

D’ALESSANDRO AND BROMLEY: MENISCATE TRACE FOSSILS

749

Keckia Glocker, 1841

Discussion. Glocker (1841) erected Keckia annulata on the basis of two specimens from Upper
Cretaceous sandstone (Germany). Despite his beautiful colour plate (1841, pi. 4), it is not clear what
the genus comprises. Described in terms of plants and compared to lycopodiaceans, Glocker (1841,
pp. 318-319) considered the crescent-shaped rings visible in parts of his specimens to be atypical of
fucoid structures. On this basis, he distinguished his genus from Muensteria Sternberg. Furthermore,
K. ammlata was considered to be more densely branched than M. hoessii. From his plate the
branching pattern appears to be primarily successive in most cases. However, the structure of the
fossil varies so extensively from part to part that possibly it is composed of intersections of several
distinct taxa. Although one branch of the larger specimen and the small second specimen appear to be
unwalled meniscate burrows, we regard this confusingly based ichnotaxon as a nomen diihiuni.

Scolicia de Quatrefages, 1849

[Bolonia Meunier, 1886; Taphrhelminthopsis Sacco, 1888; SuhphyUochorda Gotzinger and Becker, 1932.]

Discussion. Seilacher (1955, 1983) stated that the morphological diflferences between the ichnogenera
listed above (among many others) represent various preservational expressions of the same
fundamental type of burrowing activity. Bolonia and SuhphyUochorda were mentioned by Hiintzschel
(1962, 1965, 1975), together with other ichnotaxa, as belonging to the "Scolicia group’ even if not
classifiable as true synonyms, but Taphrhelminthopsis was described separately. Hiintzschel (1975),
Ksi§zkiewicz (1977), Smith and Crimes (1983), and others have shown by means of sketches the
possible relationship between biostratinomic processes and the morphology of the same structures
belonging to the Scolicia group, to which the effects of preservation must be added (D’Alessandro
1980, fig. 3). Thus the ichnogeneric names under discussion appear to reflect different taphonomic
features of the same type of trace, and should therefore be placed in synonymy. Nevertheless, some
authors (e.g. Ksi^zkiewicz 1977; Crimes 1977; Smith and Crimes 1983) preferred to retain distinct
ichnogeneric names even in cases where such names could with reasonable safety be ascribed to
taphonomic variation.

Thus we consider the ichnogenus Scolicia to include those trace fossils in which the sediment is
backfilled to form densely packed lamellae bearing one or two parallel ridges, strings, or grooves lying
below or within the lamellar fill. The concave epirelief expression takes the form of a bilaterally
symmetrical laminated trail divided into two or three lobes by longitudinal grooves; as a slightly
convex epirelief variant, Scolicia displays a plait-like pattern (Ksi^zkiewicz 1977, fig. 24/; Smith and
Crimes 1983, fig. 4d). Scolicia always lacks the median vertical discontinuity plane that is
characteristic of Curvoiithus and the distinctively different Gyrochorte (see Heinberg 1973).

Psammichnites Torell, 1870

[= Olivellites Fenton and Fenton, 1937a; Aiiliclmites Fenton and Fenton, 19376;

‘ILaminites Ghent and Henderson, 1966.]

Discussion. The original material of Olivellites was described (Fenton and Fenton 1937u) as a
meandering trace fossil preserved in convex epirelief It appears to be bilaterally symmetrical and
divided into two zones by a median solid string. The fill is composed of concave thin backfill lamellae;
in the string itself, the sediment was described as ‘not closely packed’ (Fenton and Fenton 1937u,
p. 453).

The original specimen of Aulichnites is a full relief structure that seems to differ from Olivellites
exclusively in lacking the string; thus it appears to be divided into two lobes by a single median furrow.
Hakes (1977) demonstrated that the lower surface of the original specimen of Aulichnites is regularly
convex. It is reasonable to suppose, therefore, that both ichnogenera represent preservational variants
of the same structure. Furthermore, the morphology of Olivellites closely corresponds to that of
Psammichnites (see Hantzschel 1975, fig. 62, 2c), of which we consider it a junior synonym.

750

PALAEONTOLOGY, VOLUME 30

In its original diagnosis, Laminites was described as a subcylindrical backfilled trace having
concave or biconcave laminae alternately dark and light; these laminae were interpreted as having
been produced by separate packets of faeces extruded into the burrow (Ghent and Henderson 1966).
The trace fossil is preserved as convex epirelief Further examples of such a structure were given by
Chamberlain (1971, 1978). In longitudinal section it is not always possible to distinguish Scolicia from
Laminites. At present, our knowledge of this form is still limited as regards its morphological range
and preservational variants; the construction of its under surface, so critical in the Scolicia group, is
unknown. Therefore, provisionally we consider Laminites to be a synonym of Psammichnites, until
its construction is better understood.

non 1823
non 1833
1858
1877
? 1877
1877
1947
1958
1958
1962
1962
1964
1971
1974
1974
? 1977
1977
1977

1983
non 1984

1984

1985

1985

1986

Taenidium Heer, 1877

Fucoides encoelioides Brongniart, p. 55, pi. 6, fig. 4.

Muensteria Sternberg, pp. 31-32, pi. 6, fig. 4; pi. 7, fig. 3.

Muensteria hoessii {non Sternberg) Fischer-Ooster (partini). p. 62, pi. 16, fig. 4.
Taenidium Heer (partim). p. 117, pi. 45, figs. 9 and 106; pi. 50, figs. 1 and 2.
Muensteria flageilaris {non Sternberg) Heer, pi, 66, figs. 4 and 5.

Muensteria hoessii {non Sternberg) Heer, p. 164, pi. 69, fig. 3.

Scolecocoprus Brady, p. 471, pi. 69, fig. 1.

Muensteria Sternberg; Seilacher, p. 1071, table 2, fig. 28.

Muensteria hoessii {non Sternberg) Seilacher, p. 1071, table 2, fig. 40.
Muensteria Sternberg; Hantzschel {partim). p. 205.

Muensteria hoessii {non Sternberg) Seilacher, p. 229, pi. 2, fig. 6.

Muensteria Sternberg; Seilacher, p. 309, fig. 7 (27), table 1 (27).

Taenidium serpentium Heer; Chamberlain, p. 241, pi. 32, fig. 10.

Muensteria Sternberg; Fiirsich {partim). p. 34, fig. 29a.

Muensteria Sternberg; Heinberg, p. 17, figs. 1b and 9c.

Keckia cf. hoessii (Sternberg); Ksi^zkiewicz, pi. 3, fig. 12.

Muensteria Sternberg; Chamberlain {partim), p. 14, figs. 2e and 5f.

Taenidium serpentium Heer; Chamberlain, p. 18, figs. 2a, h, 3f, i, 7a.
Muensteria Sternberg; Wetzel {partim). p. 290, fig. 2.

Muensteria Sternberg; Bracken and Picard, p. 485, fig. 12.

Muensteria Sternberg; Pemberton and Frey, p. 291.

Muensteria Sternberg; Frey and Howard, pp. 378-379, figs. 10.12, 16.3(b), 19.6.
Entradichnus Ekdale and Picard, p. 8, pi. 2a, b.

Muensteria ichnosp. D’Alessandro et at., p. 299, fig. 5b.

Preliminary remarks. Taenidium is generally considered today to be a branched meniscate structure
(Hantzschel 1975, fig. 70.1). However, the terms in which the genus was introduced originally (Heer
1877) are very similar to the original diagnosis of Muensteria (see introduction above). Its original
Latin diagnosis is unsatisfactory owing to the linguistic bias caused by an algal interpretation.

Heer’s first species, T. serpentinum (text-fig. 5) has been designated subsequently by Andrews (1955,
p. 243) as type species of ‘alga?’; and by Hantzschel (1975, p. 112) as type species of a branched,
root-like system of burrows, even though this species is unbranched. Heer’s diagnosis is as follows; ‘T.
fronde semplici, valde incurvata, serpentina. 3 mm lata, evidenter articulata, articulis I mm longis'.
(Simple Taenidium frond, strongly bent, serpentine, 3 mm wide, evidently articulated, the joints 1 mm
long.)

Heer’s second and third species, T. gillieroni and T. convolutum, are also unbranched and differ
only in size from the type species. Their illustrations show no true branching, but indicate superficial
wrinkling and evenly spaced menisci (Heer 1877, pi. 45, figs. 9 and 106; pi. 50, figs. 1 and 2). Taenidium
is thus meniscate and unbranched, and on this basis it would be logical to consider Taenidium to be
the first valid name for Muensteria of authors. These three species were founded on Swiss Jurassic
material.

Later in the same work, Heer ( 1 877, p. 162) illogically named a strongly branched, radiating, septate
structure T. fischeri and he listed M. annulata of Schafhautl (1851) among its synonyms. Afterwards

D’ALESSANDRO AND BROMLEY: MENISCATE TRACE FOSSILS

751

TEXT-FIG. 5. Taenidium serpentinum Heer, 1877. Redrawn from Heer’s pi. 45,
fig. 9. About natural size.

(Heer 1881) he founded a new species, T. lusitanicum, for a palmate form. These later species differ
radically from the original ones in possessing ramification. Nevertheless, Hantzschel (1975, p. 112)
pronounced T.fischeri the ‘most typical’, and confirmed the general tendency to consider Taenidium a
much-branched ichnogenus. As pointed out above, however, this is not the original connotation.

Emended diagnosis. Unlined or very thinly lined, unbranched, straight or sinuous cylindrical burrows
containing a segmented fill articulated by meniscus-shaped partings.

Type species. T. serpentinum Heer, 1877.

Discussion. Fragments of Cladichnus that do not show their characteristic branching may be
confused with short lengths of Taenidium. However, individual branches of Cladichnus commonly are
less winding, have stronger annulation or moniliform development, or they may possess cone-in-cone
articulations, and consequently have a less linear border than Taenidium. Many species in the
literature are based on fragments too small to display sufficient characteristics for identification.

Ner cites, when only its central axis is preserved, may be mistaken for Taenidium; the disturbance
zone around Nereites is revealed, however, in such cases where two burrows cross (Frey et al. 1984,
fig. 1a).

Taenidium differs from Ancorichnus in lacking a ‘distinct mantle’ (Heinberg 1974). However, on the
basis of illustration, the trace fossils A. coronus (Frey et al. 1984, figs. Id, e and 3) and A. capronus
(Howard and Frey 1984, figs. 2 and 3a, b) appear to be thinly walled and thus less distinct from
Taenidium than from A. ancorichnus. McCarthy’s (1979, p. 263, fig. d) Keckia, a trace fossil having
‘poorly defined walls’, was compared to Frey and Howard’s (1970, fig. 8a) ‘chevron trails’; but Frey
and Howard (1985) placed the latter in synonymy with A. capronus.

There is a close resemblance between Taenidium and Beaconites. When the latter is better known it
may be revealed as a junior synonym of Taenidium. However, differences such as the irregularity and
flattened shape of the menisci, and the large size of the structures, may be used to perpetuate the
ichnospecies of Beaconites.

The morphology of Imhrichnus protuherans in bedding plane view (Marintisch and Finks 1982,
pi. 5, fig. 1) resembles Taenidium in the same view. Moreover, the vertical structure through the

752

PALAEONTOLOGY, VOLUME 30

)))))) DT7

T serpentinum

TEXT-KiG. 6. Essential features of the three valid ichno-
species of Taenidium.

T cameronensis

T satanassi

mid-line of I. protuherans (their pi. 5, fig. 2) shows imbricate structure that may represent the lower half
of the simple meniscate fill of Taenidium. On the other hand, /. wattonensis Hallam has an imbricated
structure (Hallam 1970, p. 198) that does not take the form of hemispherical menisci.

Interpretation. Different authors have interpreted active fill of burrows in different ways. Some have
assumed that the material packed behind the animal on its way through the sediment is entirely faecal,
having passed through the gut (Fursich 1974, p. 35). Others have emphasized that the sediment is
transported around the body during the animal’s progress through the substrate (e.g. Heinberg 1974,
p. 18; Bromley and Asgaard 1975, p. 276). The alternation of two types of sediment that produces the
meniscus structures may arise from physical sorting of the sediment, producing alternate fine-grained
and coarse laminae (Stanley and Fagerstrom 1974, p. 74), or from a combination of ingestion and
external transport (Bromley and Asgaard 1979; Pemberton and Frey 1984). It seems reasonable in
most cases to regard the sediment packages that resemble lithologically the surrounding sediment as
non-ingested, and the alternating compressed fine-grained material as coprolitic in origin. The
proportions of these two components vary in different types of burrow fills.

Included species. We recognize three valid ichnospecies (text-fig. 6) while four other ichnospecies must be
regarded as dubious: type ichnospecies— T. serpentinum Heer, 1877; other ichnospecies— T. cameronensis (Brady,
1947) and T. satanassi ichnosp. nov.; dubious ichnospecies — Muensteria cretacea Heer, 1877, T. carhoniferum
Sacco, 1888, T.? maeandriformis Muller, 1966, and M. planicostata Ksi^zkiewicz, 1977.

Taenidium serpentinum Heer, 1877
Text-figs. 6a and 7

? 1858 Muensteria hoessii Fischer-Ooster (partim), p. 62, pi. 16, fig. 4 (non fig. 5).

1877 Taenidium serpentinum Heer, p. 117, pi. 45, figs. 9 and 10b.

1877 Taenidium gillieroni Heer, p. 117, pi. 50, fig. 1.

1877 Taenidium convolutum Heer, p. 117, pi. 50, fig. 2.

? 1877 Muensteria fiagellaris Heer (non Sternberg), pi. 66, figs. 4 and 5.

+ 1877 Muensteria hoessii Heer (partim), p. 164, pi. 69, fig. 3 (non pi. 66, fig. 6).

+ 1958 Muensteria hoessii Heer; Seilacher, p. 1071, table 2, fig. 40.

-I- 1962 Muensteria hoessii Heer; Seilacher, p. 229, pi. 2, fig. 6.

1971 Taenidium serpentium Heer; Chamberlain, p. 241, pi. 32, fig. 10.

+ 1977 Muensteria cf. hoessii Sternberg; Chamberlain (partim), p. 14, figs. 2e and 5f.

Diagnosis (emended). Serpentiform Taenidium having well-spaced, arcuate menisci; distance between
menisci about equal to or a little less than burrow width. External moulds may show slight annulation
corresponding to menisci, or fine transverse wrinkling. Secondary subsequent branching and
intersections occur. Boundary sharp, lining lacking or insignificant.

D'ALESSANDRO AND BROMLEY: MENISCATE TRACE FOSSILS

753

TKXT-FIG. 7. Taenidium serpentinum Heer, 1877, MGUH 17552, from the Lower Jurassic Gule Horn member,
Neill Klinter Formation of Jameson Land, East Greenland. Scale. 2 cm.

Type area and age. Switzerland, Jurassic.

Discussion. This name is useful for the common form of ’’Muenstericr, for example, in the Lower
Jurassic of East Greenland (Surlyk et al. 1971, p. 29; Heinberg and Birkelund 1984, purf/m) (text-fig. 7),
where the form is generally preserved in full relief in micaceous sandstone. The wall of this trace fossil
is sharp and possibly lined by an extremely thin film of clay or mica grains. Similarity with the
illustrations of Heer (1877, pi. 50, figs. 1 and 2) is striking.

The citations in the synonymy list prefixed with a cross ( + ) may represent a distinct ichnospecies.
As far as can be judged from illustrations, these burrows are larger and the fill has more densely
alternating packets of two types of sediment than the typical T. serpentinum. Nevertheless, as we are
dealing with incomplete burrows, shown only in longitudinal section, we prefer temporarily to include
them in T. serpentinum.

If they are considered to comprise a separate ichnospecies, the once commonly used specific name
'hoessif would not be suitable because it was employed by Sternberg for two diflferent structures: a
Chondrites (1833, pi. 6, fig. 4) comparable with our text-fig. 4, and an incomplete burrow (1833, pi. 7,
fig. 3) of doubtful interpretation. The name, therefore, could be employed in the ichnogenus
Chondrites as C. hoessii (Sternberg). Fischer-Ooster (1858, pi. 16, figs. 4 and 5) applied the name to
meniscate traces, one of which cannot be included in Taenidium, being bordered by a wide band, and
the other being scarcely significant. Heer (1877) determined two different structures as M. hoessii: one
that could be an incomplete Spirophycus(p\. 66, fig. 6) and another that suggests a real Taenidium. M.
hoessii is thus a nomen catastrophicum, as is the genus itself

Hantzschel’s (1972) T. carhonicum (a lapsus by Hantzschel for T. carboniferum Sacco) has an
indistinct dark zone surrounding a well-preserved meniscate fill. The zone appears to cut other
burrow fills, indicating that this structure should be referred instead to Nereites.

754

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 8. Taenicliwn satanassi ichnosp. nov. in cleavage relief preservation, a, b, MGUH 17553, specimen
containing the holotype (most conspicuous, longest preserved burrow, detailed in b); the other meniscate
structures are paratypes. c, D, MGUH 17554, paratypes. Scale, 2 cm.

The non-marine Entradichnus meniscus may belong here, although the menisci appear to be closely
spaced and rather flat, as in B. antarcticus Vialov.

Taenidium cameronensis (Brady, 1947)

Text-fig. 6b

1947 Scolecocoprus cameronensis Brady, p. 471, pi. 69, fig. 1.

? 1979 Keckia sp. McCarthy, p. 363, fig. 3c.

Diagnosis (emended). Taenidium having intermeniscate segments generally longer than wide, and
deeply concave menisci; secondary successive branching and intersection occur.

Type area and age. Arizona, probably Permian.

Discussion. This ichnospecies is separated from T. serpentinum on the basis of its distinctly longer
packets of sediment between successive menisci in the backfill. The arc of the meniscus is also much

D’ALESSANDRO AND BROMLEY: MENISCATE TRACE FOSSILS

755

deeper, tending towards parabolic, so that each packet nests around the next. Hitherto, all
references to this ichnospecies have been from shallow water settings (e.g. Decourten 1978,
p. 491, fig. 1).

Taenidium satanassi ichnosp. nov.

Text-figs. 2, 3, 6c, 8a-d, 9

1958 Muensteria Seilacher, p. 1071, table 2, fig. 28.

1986 Muensteria ichnosp. D’ Alessandro et ai, p. 299, fig. 5b.

Derivation of name. Satanassus. Satan, after whom the type locality is named, a river that had disastrously
unpredictable floods.

Type material. Holotype, MGUH 17553 (text-fig. 8a, b) and paratypes, MGUH 17553, 17554 (text-fig. 8a, c, d)
deposited in the Geological Museum, University of Copenhagen, Denmark; from Fiumara Satanasso,
Villapiana, near Sibari, southern Italy; Saraceno Formation, middle-upper Eocene.

Other material. Several specimens from the lower part of the Saraceno Formation, deposited in the
Department of Geology and Geophysics, University of Bari, and Department of Earth Science, University of
Calabria.

756

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 9. Sketch of an example of Taenidiiim satanassi
ichnosp. nov., convex hyporelief, showing the tightest
observed curvature.

Diagnosis. Weakly sinuous to nearly straight Taenidium, the fill consisting of evenly alternating
meniscus-shaped packets of two types of sediment, of more or less equal thickness; sediment packets
considerably shorter than wide.

Description. The burrows lie parallel or weakly inclined to bedding. The fill in the type material is very
characteristic, the packets of sediment being alternately dark and pale. One of these two types of sediment,
usually the dark material, is clearly pelleted and grades on the concave side into the other type of sediment which,
in turn, is closely similar to the surrounding sediment. The packets of unpelleted sediment are usually a little
thinner than the pelleted segments, or may reach equal size. The curvature of the meniscus is relatively wide, and
the sediment packets consequently do not partially enclose or wrap around each other. In thin section the
unpelleted sediment packets seem to run continuously into the surrounding sediment without interruption, but
on weathered surfaces a very faint colour difference and boundary interface can be detected. There is thus no
detectable wall lining or special grain orientation at the junction.

The burrows have an average width of 8-82 + 0-73 mm (n = 33), showing an observed range of 4- 1 4 mm and a
unimodal size-frequency distribution. The type material is highly compacted; it is mainly preserved in (now
flattened) cleavage relief on upper surfaces or, more rarely, in semirelief on soles.

Discussion. T. satanassi differs from the other ichnospecies in its weakly arcuate menisci, more or less
equally sized packets of alternating sediment type, obviously pelleted form of the fill (text-fig. 8),
straight-sided boundary, and little-winding course. Greater affinity exists between T. satanassi and
the forms prefixed by a cross in the T. serpentinurn synonymy. The main differences here are in the
more or less equal quantities of alternating sediment types in the former, and the more arcuate menisci
of the latter so that each packet partially wraps around the next.

T. satanassi resembles B. antarcticus. However, the latter’s backfill appears to have alternating
sediment units of unequal thickness; furthermore, the septa appear flattened at the centre but curve
more or less abruptly as they approach the boundary (Vialov 1962; Gevers et al. 1971, p. 84; Bradshaw
1981, p. 631), whereas those of T. satanassi curve in a regular arc. Moreover, B. antarcticus seems to
have a wall.

T. satanassi also resembles the meniscate trace fossil of Stanley and Fagerstrom (1974, figs. 6-8),
recently included in Ancoriclmus by Frey et al. (1984). However, a clear difference can be observed in
the arrangement of the menisci when the direction of the burrow changes. In Taenidium the diameter
of the structure remains constant and the menisci tend to keep parallel to each other, fanning out
slightly where the burrow bends (text-fig. 9); when the traces described by Stanley and Fagerstrom
(1974) bend, the menisci are oriented at variable angles and truncate one another. Thus, this character
may be the morphological expression of a distinct burrowing pattern. A different form of irregular
meniscate development is shown by the form referred to as Muensteria ichnosp. by Bromley and
Ekdale (1984, figs. 3b and 4a-c) in chalk.

Dubious ichnospecies of Taenidium

M. cretacea Heer, 1877, p. 144, is characterized by extremely deeply concave menisci that divide the
cylindrical structure into pale and dark densely alternating units; each unit wraps around several of
the following. It is not easy to evaluate this fragment; it may represent part of a body fossil.

D’ALESSANDRO AND BROMLEY: MENISCATE TRACE FOSSILS

757

The original description of T. carboniferum Sacco, 1888, p. 162, includes "Frons . . . moniliformis,
suhcylindrica, . . . annulosa . . . articulata; articuli crassi, siihelliptici, inter se profimde disjimcti . . Sacco
added that the rounded segments are separated by 0-5 mm. This form was included in the algal genus
"Taenidium\ having segments distinctly separated; on this basis, we hesitate to consider it a Taenidium
and suggest that it could belong rather to the ichnogenus Hormosiroidea.

The single specimen T.? maeandriformis Muller, 1966, resembles T. serpentiniim in its winding
course and size. The flattened fill consists of segments about as wide as long, separated by fairly flat
menisci, distinguishing it from T. camaronensis. The meandering course and mode of preservation
are reminiscent of Nereites, but no lateral zones are visible. Owing to this uncertainty, therefore, we
reserve judgement on this ichnospecies. This name was used again by Plicka (1973) who described a
large example of unquestionable Nereites, but considered it to be the body fossil of a worm and
assigned it to a new genus, Maldanidopsis.

Muensteria pJanicosta Ksi^zkiewicz, 1977, was founded on a winding meniscate trace fossil,
preserved as hyporelief; the menisci are regularly arcuate and very dense (Ksi^zkiewicz 1977, p. 122,
pi. 13, fig. 1). The trace bends abruptly through a right angle, showing changes in width, characters
that are atypical of Taenidium but reminiscent of 'Laminites'.

Scoyenia White, 1929

Discussion. The morphology and ethology of this trace fossil and its environmental significance
have been analysed recently (Bromley and Asgaard 1979; Frey et al. 1984). Straight or gently curved
endogenic structures, limited by distinct boundaries and ornamented with external longitudinal
striae, are included here. The burrow contains a conspicuously meniscate backfill. Scoyenia is an
unbranched burrow that commonly exhibits cross-overs, intersections, or secondary successive
branching (Bromley and Asgaard 1979, figs. 9a and 10).

Beaconites Vialov, 1962

Discussion. The original description of B. antarcticus Vialov, 1962, was extremely weak. Re-
examination of topotypic material by Bradshaw (1981) has improved our understanding of this form
but, owing to the high degree of weathering, details of the burrow boundary are unclear. In the type
ichnospecies, Bradshaw (1981, p. 630) wrote of ‘a poorly developed sand lining’ and this is supported
by her fig. 15. The illustration by Gevers et al. (1971, pi. 18, fig. 3) similarly shows a boundary ridge in
cross-section, interpretable as a wall structure. The sediments were apparently deposited in a
non-marine environment. Giant meniscate structures have also been assigned to the ichnogenus
Beaconites (e.g. Gevers et al. 1971; Ridgeway 1974; Pollard 1976) and these have been placed in B.
harretti by Bradshaw (1981). The description of this ichnospecies also implies a weak wall structure
(Gevers et al. 1971, p. 84; Bradshaw 1981, p. 631) although in the illustrations the wall appears to be
composed only of the superposed nested menisci rather than being a discrete structure. This also
appears to be the case in the giant burrows illustrated by Allen and Williams (1981). The vertical
orientation of many of these continental occurrences (Allen and Williams 1981; Bracken and Picard
1984) may indicate an escape element that would further separate them from normal 'Muensteria'' of
authors.

Examination of unweathered material will be necessary to clarify the relationship between
Beaconites, Taenidium, and Ancorichnus (as emended by Frey et al. 1984).

Phoehichnus Bromley and Asgaard, 1972

Discus.sion. The strongly walled Phoehichnus trochoides is distinguished by its special wall
construction and radiating morphology (Bromley and Asgaard 1972; Bromley and Mork
1984).

758

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 10. MGUH 17556, a branch-
ing, annulated, meniscate structure in
hyporelief, from the Upper Permian,
Domkirken, Scoresby Land, central
East Greenland (L. Stemmerik Collec-
tion); arrow indicates the meniscate fill
where the thick wall is broken away.
Scale, 2 cm.

Ancorichmis Heinberg, 1974

Discussion. Among the unbranched meniscate trace fossils, Ancorichmis is distinctive in having a
well-developed, unornamented wall structure. In A. coronas Frey et ai, 1984, and A. capronus Howard
and Frey, 1984, the wall material isfar less well developed than in the type ichnospecies A. ancoriclmus,
and their ichnogeneric placing is therefore questionable.

Branched septate burrow having wall structures

Discussion. There is a distinctive group of branched trace fossils that has a resemblance to M.
annulata Schafhautl, possessing a complicated and thick wall structure. These walled trace fossils
have received heterogeneous taxonomic treatment, and it is not clear to which ichnogenus or
ichnogenera they should be referred (but see D’Alessandro, Bromley, and Stemmerik, in prep.).

As in the case of the Ner cites /Scalarituba/Neoner cites group, the walled, branched, septate burrows
have a complicated structure that is capable of producing widely varying morphologies according to
the stratinomic position and taphonomic processes. Thus, in semirelief, the annulated or striated
external sculpture of the walling material is emphasized (e.g. D’Alessandro 1982, pi. 40), whereas only
in axially sectioned material is the septate central backfill visible (text-fig. 10).

Fucusopsis was used by Ksi^zkiewicz (1970, 1977) for structures of this kind, and Radionereites by
D’Alessandro (1982). Neither ichnogenus has been demonstrated to have septate interiors, however.
Pemberton and Frey (1982) placed Fucusopsis in synonymy with the non-septate Palaeophycus.
Topotypic material of Radionereites (kindly supplied by M. Gregory) confirms that D’Alessandro’s
form does not belong to this ichnogenus.

Cladichnus ichnogen. nov.

1858 Muensteria (Keckia) annulata Schafhautl; Fischer-Ooster (partim). p. 37, pi. 7, fig. 4 (non fig.

3 = Chondrites)', pi. 12, fig. 8.

1877 Taenidium fischeri Heer, p. 162, pi. 67, figs. 1-5, 7.

1881 Taenidium lusitanicum Heer, p. 12, pi. 20.

1887 Muensteria annulata Squinabol, p. 554, pi. 17, fig. 3.

D’ALESSANDRO AND BROMLEY: MENISCATE TRACE FOSSILS

759

1947 Saportia striata Wilckens (non Squinabol), p. 41, pi. 7, fig. 3.

1947 Taenidium Insitanicum Heer; Wilckens, pp. 41-45, pi. 6, figs. 1 and 2; pi. 7, fig. 1.
1955 Taenidium fischeri Heer; Seilacher, fig. 5.83.

1958 Taenidium fischeri Heer; Seilacher, p. 1073, table 3, fig. 48.

1962 Taenidium Heer; Hantzschel, fig. 136.2.
non 1971 Taenidium annulatum Schafhautl; Chamberlain, pi. 32, fig. 12.

1975 Taenidium Heer; Hantzschel (partim), p. 1 12, fig. 70.1.

1977 Taenidium fischeri Heer; Ksi^zkiewicz, pi. 5, fig. 3.

1978 Taenidium Heer; Kern, p. 249, fig. 8f.

1983 Taenidium ichnosp. Pedersen and Surlyk, p. 53, fig. 12.

1986 Taenidium annulatum Schafhautl; D’ Alessandro et a/., p. 300, fig. 5c.

Derivation of name. Greek klados, branch.

Type species. Cladichnus fischeri (Heer, 1877).

Diagnosis. Annulated or monilliform burrow fills composed of meniscus-shaped segments, compris-
ing primary successively branched and radiating systems; wall lining lacking or very thin.

Discussion. These meniscate burrow fills have a smooth outer sculpture and are annulated or
monilliform. They show primary successive branching, generally at acute angles, and the menisci are
concave towards the proximal direction. Branching may produce a radiating plan (C. fischerf text-fig.
1 1), or be palmate (C. lusitanicum) or dendroid (Taenidium ichnosp. of Pedersen and Surlyk 1983). It
is uncertain whether or not the type species has a wall structure. If it has, it is no more than a film-like
or insignificant skin. In the figure of Heer (1877, pi. 67, fig. 6) in which a wall is indicated, the
illustration is said to represent the condition ‘in life’ (as an alga!) and is therefore a spurious
reconstruction. Heer (1881, pi. 20) and Pedersen and Surlyk (1983, fig. 12c) indicated a very thin wall
material.

Cladichnus fischeri (Heer, 1877)

Text-fig. 11

? 1858 Muensteria (Keckia) annulata Schafhautl; Fischer-Ooster (partim), p. 37, pi. 7, fig. 4 (non fig. 3);
pi. 12, fig. 8.

1877 Taenidium fischeri Heer, p. 162, pi. 67, figs. 1-5, 7.

1887 Muensteria annulata Squinabol, p. 554, pi. 17, fig. 3.

1955 Taenidium fischeri Heer; Seilacher, fig. 5.83.

1958 Taenidium fischeri Heer; Seilacher, p. 1073, table 3, fig. 48.

1962 Taenidium Heer; Hantzschel, fig. 136.2.

TEXT-FIG. 11. Field photograph of

fischeri ( Heer, 1 877) from the same horizon as

the Taenidium satanassi material. Scale, 1 cm.

760

PALAEONTOLOGY, VOLUME 30

1975 Taenidium Heer; Hantzschel (partim), p. 112, fig. 70.1.

1977 Taenidium fischeri Heer; Ksi^zkiewicz, pi. 5, fig. 3.

1978 Taenidium Heer; Kern, p. 249, fig. 8f.

1986 Taenidium annulatum Schafhautl; D’Alessandro et a/., p. 300, fig. 5c.

Diagnosis. Radiating and primary successively branched Cladichnus.

Discussion. Muensteria annulata Schafhautl, 1851, is not listed as a synonym because it is a dubious
structure. In the illustration it closely resembles the branching, septate, walled burrow, but it is not
clear whether or not a wall structure is present. Heer considered M. annulata a synonym of his T.
fischeri.

Most Cladichnus show a tendency to radial development. In some examples (text-fig. 11) the
radiating burrow fills are generally straight, originate at several levels from a central shaft, and show
little or no subsequent branching. In others, branching of the radial elements may occur but to what
extent is hard to discern. In early illustrations suggesting algal origin (e.g. Heer 1877, pi. 67, fig. 1),
simultaneous filling is suggested, but it is probable that most such branching is in fact intersection,
compacted close application of elements, and primary subsequent branching. In lower Tertiary
material from Italy (text-fig. 1 1), extreme compaction has given the impression of true branching in
cases where intersection can be demonstrated by careful inspection.

Other ichnospecies of Cladichnus

We place T. lusitanicum Heer, 1881, temporarily in Cladichnus owing to its primary successive
branching pattern and meniscate fill. However, its palmate form is highly distinctive and represents a
fundamentally different burrowing behaviour from the radiating forms. Pedersen and Surlyk’s ( 1983)
Taenidium ichnosp. has a root-like branching pattern and appears to be a distinct ichnospecies of
Cladichnus.

Acknowledgements. We thank M. Sonnino (University of Calabria) for his assistance in the field, and F. T.
Fiirsich (University of Brehmen) for helpful review of the manuscript. A partial grant from the Italian C.N.R. to
A. D’A. is acknowledged.

REFERENCES

ALLEN, j. R. L. and WILLIAMS, B. p. j. 1981. Beaconites antarcticus: a giant channel-associated trace fossil from the
Lower Old Red Sandstone of south Wales and the Welsh Borders. Geol. J. 16, 255-269.

ANDREWS. H. N. 1955. Index of generic names of fossil plants, 1820-1950. Bull. U.S. geol. Surv. 1013, 262 pp.
BRACKEN, B. and PICKARD, M. D. 1984. Trace fossils from Cretaceous/Tertiary North Horn Formation in central
Utah. J. Paleont. 58, 477-487.

BRADY, L. F. 1947. Invertebrate tracks from the Coconino Sandstone of northern Arizona. Ibid. 21, 466-472.
BRADSHAW, M. A. 1981. Palaeoenvironmental interpretations and systematics of Devonian trace fossils from the
Taylor Group (Lower Beacon Supergroup), Antarctica. N.Z. Jl Geol. Geophys. 24, 615-652.

BROMLEY, R. G. and ASGAARD, u. 1972. Notes on Greenland trace fossils. III. A large radiating burrow-system in
Jurassic micaceous sandstones of Jameson Land, East Greenland. Rap. Gronlands geol. Unders. 49, 23-30.

1975. Sediment structures produced by a spatangoid echinoid: a problem of preservation. Bidl. geol.

Soc. Denmark, 24, 261-281.

1979. Triassic freshwater ichnocoenoses from Carlsberg Fjord, East Greenland. Palaeogeogr.

Palaeoclimat. Palaeoecol. 28, 39-80.

and EKDALE, A. A. 1984. Trace fossil preservation in flint in the European chalk. J. Paleont. 58, 298-311.

and FREY, R. w. 1974. Redescription of the trace fossil Gyrolitlies and taxonomic evaluation of

Thalassinoides, Ophiomorpha and Spongeliomorpha. Bull. geol. Soc. Denmark, 23, 311-335.

and M0RK, A. 1984. Phoehichnus ichnofacies in the condensed Lower Jurassic succession of Svalbard. Rep.

Inst. Kontinentsokkelunders. Norway, no. 04.6341.03/84, 13 pp.

BRONGNIART, M. A. 1828. Histoire des vegetaux fossiles ou recherches hot uniques et geologiques sur les vegetaux
renfermes dans les diverses couches du Globe. 1, 136 pp. Fortin, Masson & Cie, Paris.

D’ALESSANDRO AND BROMLEY: MENISCATE TRACE FOSSILS

761

CHAMBERLAIN, c. K. 1971. Morphology and ethology of trace fossils from the Ouachita Mountains, southeast
Oklahoma. J. Paleont. 45, 212-245.

1977. Ordovician and Devonian trace fossils from Nevada. Bull. Nev. Bur. Mines Geol. 90, 24 pp.

-- 1978. Recognition of trace fossils in cores. In basan, p. b. (ed.). Trace fossil concepts. Short Course Soc. econ.
Paleont. Miner. Tulsa, 5, 119-166.

CRIMES, T. p. 1977. Trace fossils of an Eocene deep-sea sand fan, northern Spain. In crimes, t. p. and harper, j. c.
(eds.). Trace fossils 2. Geol. J. spec. Issue, 9, 71-90.

d’alessandro, a. 1980. Prime osservazioni sulla ichnofauna miocenica della ‘Formazione di Gorgoglione’
(Castelmezzano, Potenza). Riv. ital. Paleont. Stratigr. 86, 357-398.

1982. Process! tafonomici e distribuzione delle tracce fossil! nel Flysch di Gorgoglione (Appennino

meridionale). Ibid. 87, 511-559.

EKDALE, A. A. and SONNINO, M. 1986. Sedimentologic signihcance of turbidite ichnofacies in the Saraceno

Formation (Eocene), southern Italy. J. sedim. Petrol. 56, 294-306.

DECOURTEN, E. L. 1978. Scolecocoprus cameronensis Brady from the Kaibab Limestone of northern Arizona: a
re-interpretation. J. Paleont. 52, 491-493.

EKDALE, A. A., BROMLEY, R. G. and PEMBERTON, s. G. 1984. Ichnology: the use of trace fossils in sedimentology and
stratigraphy. Short Course Soc. econ. Paleont. Miner. Tulsa, 15, 317 pp.

and PICARD, M. D. 1985. Trace fossils in a Jurassic eolianite, Entrada Sandstone, Utah, U.S.A. Spec. Pubis

Soc. econ. Paleont. Miner. Tulsa, 35, 3 12.

EENTON, c. L. and FENTON, M. A. 1937u. OUvelUtes, a Pennsylvanian snail burrow. Am. Midi. Nat. 18,
452-453.

19376. Burrows and trails from Pennsylvanian rocks of Texas. Ibid. 1079-1084.

FERGUSON, L. 1965. A note on the emplacement of sediment in the trace fossil Chondrites. Proc. geol. Soc. Lond.
1622, 79-82.

FISCHER-OOSTER, c. VON. 1858. Die fossilen Fucoiden der Schweizer-Alpen, nebst Erorterungen iiher deren
geologisches Alter, 72 pp. Huber, Bern.

FREY, R. w. and BROMLEY, R. G. 1985. Ichnology of Selma-Group chalks (Upper Cretaceous), western Alabama. I.
Stratigraphy and biogenic sedimentary structures. Can. J. Earth Sci. 22, 801-828.

and HOWARD, j. d. 1970. Comparison of Upper Cretaceous ichnofaunas from siliceous sandstones and

chalk. Western Interior Region, U.S.A. In crimes, t. p. and harper, j. c. (eds.). Trace fossils. Geol. J. spec. Issue,
3, 141 166.

1985. Trace fossils from the Panther Member, Star Point Formation (Upper Cretaceous), Coal Creek

Canyon, Utah. J. Paleont. 59, 370 404.

and PRYOR, w. a. 1978. Ophiomorpha: its morphological, taxonomic and environmental signihcance.

Palaeogeogr. Palaeoclimat. Palaeoecol. 23, 199 229.

PEMBERTON, s. G. and FAGERSTROM, J. A. 1984. Morphological, ethoIogical,and environmental signihcance of

the ichnogenera Scoyenia and Ancorichnus. J. Paleont. 58, 511-528.

fOrsich, f. t. 1973. A revision of the trace fossils Spongeliomorpha, Ophiomorpha and Thalassinoides. Neues Jh.
Geol. Paldont. Mh. 1973, 719-735.

1974. Corallian (Upper Jurassic) trace fossils from England and Normandy. Stuttg. Beitr. Naiurk. (B) 13,

52 pp.

GEiNiTZ, H. B. 1867. Die organischen Uberreste im Dachschiefer von Wurzbach bei Lobenstein. In geinitz, h. b.
and LiEBE, K. T. Uber ein Aquivalent der takonischen Schiefer Nordamerikas in Deutschland und dessen
geologische Stellung. Nova Acta Acad. Caesar. Leap. Carol. 33, 1-24.

gevers, t. w., frakes, l. a., EDWARDS, L. N. and MARZOLF, J. E. 1971. Trace fossils in the Lower Beacon sediments
(Devonian), Darwin Mountains, southern Victoria Land, Antarctica. J. Paleont. 45, 81-94.

GHENT, E. D. and HENDERSON, R. A. 1966. Petrology, sedimentation and palaeontology of Middle Miocene graded
sandstones and mudstones, Kaiti Beach, Gisborne. Trans. R. Soc. N.Z. 4 (7), 147 169.

GLOCKER, E. F. 1841. Uber die kalkfiihrende Sandsteinformation auf beiden Seiten der mittleren March, in der
Gegend zwischen Kwassitz und Kremsier. Nova Acta Acad. Caesar. Leap. Carol. 19 (Suppl. 2), 309-334.

GOTZiNGER, G. and BECKER, H. 1932. Zur geologischen Gliederung des Wienerwaldhysches (neue Fossilfunde). Jh.
geol. Bundesanst. Wien, 82, 343-396.

HANTZSCHEL, w. 1962. Trace fossils and problematica. In moore, r. c. (ed.). Treatise on Invertebrate Paleontology-
Part W, 177-245. Geological Society of America and University of Kansas Press, Boulder, Colorado and
Lawrence, Kansas.

1965. Vestigia invertebratorum et problematica. In westphal, t. s. (ed.). Fossilium catalogus, 108, 1-140.

Junk, s’Gravenhage.

762

PALAEONTOLOGY, VOLUME 30

HANTZSCHEL, w. 1972. Lcbensspuren in den Kulm-Tonschiefern von Neustadt a. d. Weinstrasse. Oberrhein. geol.
Abb. 21, 107-115.

1975. Trace fossils and problematica. In teichert, c. (ed.). Treatise on Invertebrate Paleontology- Part W,

Siippl. 1, 269 pp. Geological Society of America and University of Kansas Press, Boulder, Colorado and
Lawrence, Kansas.

HAKES, w. G. 1977. Trace fossils in Late Pennsylvanian cyclothems, Kansas. In crimes, t. p. and harper, j. c.
(eds.). Trace fossils 2. Geol. J. spec. Issue, 9, 209-226.

HALLAM, A. 1970. Gyroclwrte and other trace fossils in the Forest Marble (Bathonian) of Dorset, England. In
CRIMES, T. p. and harper, j. c. (eds.). Trace fossils. Ibid. 3, 189-200.

HEER, o. 1876 1877. Flora fossilis Helvetiae. Die vorweltliche Flora der Schweiz, 182 pp. J. Wiirster & Co.

1881. Contributions a la flore fossil du Portugal. Section des Travaiix geologiques du Portugal, 5 1 pp. Lisbon.

HEiNBERG, c. 1973. The internal structure of the trace fossils Gvrochorte and Curvolithus. Lethaia, 6,
227-238.

1974. A dynamic model for a meniscus tilled tunnel {Ancorichnus n. ichnogen.) from the Jurassic Pecten

Sandstone of Milne Land, East Greenland. Rap. Gronlands geol. Unders. 62, 20 pp.

and BiRKELUND, T. 1984. The trace-fossil assemblages and basin evolution of the Vardekloft Formation

(Middle Jurassic, central East Greenland). J. Paleont. 58, 362-397.

HOWARD, J. D. and frey, r. w. 1984. Characteristic trace fossils in nearshore to oft'shore sequences. Upper
Cretaceous of east-central Utah. Can. J. Earth Sci. 21, 200-219.

KERN, J. p. 1978. Trails from the Vienna woods: paleoenvironments and trace fossils of Cretaceous to Eocene
Flysch, Vienna, Austria. Palaeogeogr. Palaeoclimat. Palaeoecol. 23, 231-262.

KSi,yzKiEWicz, M. 1970. Observations on the ichnofauna of the Polish Carpathians. In crimes, t. p. and harper,
J. c. (eds.). Trace fossils. Geol. J. spec. Issue, 3, 283-322.

1977. Trace fossils in the Flysch of the Polish Carpathians. Palaeont. pol. 36, 208 pp.

MCCARTHY, B. 1979. Trace fossils from a Permian shoreface-foreshore environment, eastern Australia. J. Paleont.
53, 345-366.

MacLEAY, w. s. 1839. Note on the Annelida. In Murchison, r. i. The Silurian System, 699-701. Murray,
London.

marintisch, e. j. and finks, r. m. 1982. Lower Devonian ichnofacies at Highland Mills, New York and their
gradual replacement across environmental gradients. J. Paleont. 56. 1050-1078.

meunier, s. 1886. Sur quelques empreintes problematiques des couches boloniennes du Pas-de-Calais. Bull. Soc.
geol. Fr. (3) 14, 564-568.

MULLER, A. H. 1966. Neue Lebensspuren (vestigia invertebratorum) aus dem Karbon und der Trias
Mitteldeutschlands. Geologie, 15, 712-725.

OSGOOD, R. G. 1970. Trace fossils of the Cincinnati area. Palaeontogr. am. 6 (41), 281-444.

1975. The history of invertebrate ichnology. In frey, r. w. (ed.). The study of trace fossils, 3-12.

Springer-Verlag, New York.

PEDERSEN, G. K. and SURLYK, F. 1983. The Fur Formation, a late Paleocene ash-bearing diatomite from northern
Denmark. Bull. geol. Soc. Denmark, 32, 43-65.

PEMBERTON, s. G. and FREY, R. w. 1982. Trace fossil nomenclature and the Planolite-Palaeophycus dilemma. J.
Paleont. 56, 843-881.

1984. Ichnology of storm-influenced shallow marine sequence: Cardium Formation (Upper

Cretaceous) at Seeby, Alberta. Mem. Can. Soc. Petrol. Geol. 9, 281-304.

PLiCKA, M. 1973. Maldanidopsis maeandriformis (Muller) (Annelida sedentaria from the lower Carboniferous of
Moravia— Czechoslovachia). Cds. slezsk. Mus. Opave, (A) 28, 37-47.

POLLARD, J. E. 1976. A problematic trace fossil from Tor Bay Breccias of south Devon. Proc. Geol. Ass. 87,
105-108.

RIDGEWAY, J. M. 1 974. A problematic trace fossil from the New Red Sandstone of south Devon. Ibid. 84, 5 1 1 -5 1 7.

SACCO, F. 1888. Note di paleoicnologia italiana. Atti Soc. ital. Sci. nat. 31, 151-192.

SCHAFHAUTL, K. E. 1851. Geognostische U ntersuchungen der Sudbayrischen Alpengebirges, 208 pp. Literarischen-
artistische Anstalt, Miinchen.

SCHENK, A. 1864. Beitrage zur Flora der Vorwelt. Palaeontographica, 11, 296-308.

SEiLACHER, A. 1953. Studien zur Palichnologie. 1 . Uber die Methoden der Palichnologie. Neues. Jb. Geol. Paldont.
Abh. 96, 421-452.

1955. Spuren und Facies im Unterkambrium. Abh. math.-naturw. Kl. Akad. kFiss. Mainz, 10, 373-399.

1957. An-aktualistisches Wattenmeer? Paldont. Z. 31, 198-206.

1958. Zur okologischen Charakteristik von Flysch und Molasse. Eclog. geol. Helv. 51, 1062-1078.

D’ALESSANDRO AND BROMLEY: MENISCATE TRACE FOSSILS

763

1960. Lebensspuren als Leitfossilien. Geol. Rdsch. 49, 41-50.

1962. Paleontological studies on turbidite sedimentation and erosion. J. Geol. 70, 227-234.

1964. Biogenic sedimentary structures. In imbrie, j. and Newell, n. (eds.). Approaches to paleoecology,

296-316. Wiley, New York.

1983. Upper Paleozoic trace fossils from the Gilf Kebir-Abu Ras area in southwestern Egypt. J. Afr. Earth

Sd. 1,21-34.

and MEISCHNER, D. 1964. Fazies-Analyse im Palaozoikum des Oslo-Gebietes. Geol. Rdsch. 54, 596-619.

SMITH, A. B. and CRIMES, T. p. 1983. Trace fossils formed by heart urchins — a study of Scolicia and related traces.
Lethaia, 16, 79-92.

SQUiNABOL, s. 1887. Contribuzioni alia flora fossile dei Terreni Terziari della Liguria. Boll. Soc. geol. ital. 6,
545-561.

STANLEY, K. o., CHAMBERLAIN, c. K. and STEWART, J. H. 1977. Depositional setting of some eugeosynclinal
Ordovician rocks and structurally interleaved Devonian rocks in the Cordilleran Mobile Belt, Nevada. In
STEWART, J. H., STEVENS, c. H. and FRiTSCHE, A. E. (eds.). Paleozolc paleogeography of the western United States.
Pacific Coast Paleogeography Symposium f 259-274. Pacific Section Soc. econ. Paleontol. Mineral., Los
Angeles.

and FAGERSTROM, J. A. 1974. Miocene invertebrate trace fossils from a braided river environment, western

Nebraska, U.S.A. Palaeogeogr. Palaeoclimat. Palaeoecol. 15, 63-82.

STERNBERG, K. M. VON. 1833. Vcrsuch eiiier geognostisch-hotanischen Darstellung der Flora der Vorwelt. Parts 5-6,
80 pp. Fleischer Fr., Leipzig, Prague.

SURLYK, F., BROMLEY, R. G., ASGAARD, u. and PEDERSEN, K. R. 1971. Preliminary account of the mapping of the
Mesozoic formations of south-east Jameson Land. Rap. Gronlands geol. Unders. 37, 24-32.

VIALOV, o. s. 1962. Problematica of the Beacon Sandstone at Beacon Height West, Antarctica. N.Z. Jl Geol.
Geophys. 5, 718-732.

WELLER, s. 1899. Kinderhook faunal studies. 1. The fauna of the vermicular sandstone at Northview, Webster
County, Missouri. Trans. Acad. Sci. St. Louis, 9 (2), 9-51.

WETZEL, A. 1983. Biogenic structures in modern slope to deep-sea sediments in the Sulu Sea Basin (Philippines).
Palaeogeogr. Palaeoclimat. Palaeoecol. 42, 285-304.

WHITE, D. 1929. Flora of the Hermit Shale, Grand Canyon, Arizona. Pubis. Carnegie Instn, 405, 221 pp.
wiLCKENS, o. 1947. Palaontologische und geologische Ergebnisse der Reise von Kohl-Larsen (1928-29) nach
Siid-Georgien. Abh. senckenb. naturforsch. Ges. 474, 75 pp.

ASSUNTA D’ALESSANDRO

Departimento di Geologia e Geofisica
Campus Universitario
Via Re David 200
70125 Ban, Italy

RICHARD G. BROMLEY

Geologisk Centralinstitut
Kobenhavns Universitet
0ster Voldgade 10

Typescript received 30 June 1986 1350 Copenhagen K

Revised typescript received 16 January 1987 Denmark

AMMONITES FROM THE TYPE SANTONIAN AND
ADJACENT PARTS OF NORTHERN AQUITAINE,

WESTERN FRANCE

by W. J. KENNEDY

Abstract. The ammonite fauna of the Santonian stage in its type area and adjacent parts of northern
Aquitaine is revised, and five species recognized: Placenliceras polyopsis (Dujardin, 1837), P. paraplarmm
Wiedmann, 1978, Texaniles (Texanites) gallicus Collignon, 1948, Eulophoceras auslriacuni (Summesberger,
1979) and Boehmoceras loescheri Riedel, 1931. The Santonian corresponds to the range zone of P. polyopsis,
the P. syrtale of European authors, a view already expressed by A. de Grossouvre in 1901; no finer zonal
division based on ammonites is recognizable.

The Santonian is the middle division of the now redundant Senonian Stage. The latter term was
introduced by Alcide d'Orbigny in 1843; the former by Henri Coquand in 1857. Saintes was
mentioned by d’Orbigny as a typical area for the Senonian in the Cows Eleinentaire (1852, pp.
669-691); it and its environs— Saintoinge— give the Santonian its name. The original definition of
the Santonian was as follows (Coquand 18576, p. 749):

Deuxieme etage. — Santonien.

Craie tendre avec silex (Pleurotomaria scmtonesa, d’Orb., Rhyncbonella vespertilio, d’Orb., R. intermedia,
Coquand, Terebratnla Nanclasi, Coq., Micraster laxoporus, d’Orb., Hemiaster Stella, Desor, Salenia geometrica,
Agass.).

The stratotypes of the Santonian are localities cited by Coquand in 1856 for his ‘Troisieme sous-
etage’ of the ‘Premiere etage’ (which he subsequently named Santonian). They are ‘les bords de la
Charente Jusqu’ aux coteaux qui, a partir de Gimeux, Gente, Segonzac, dessinent une bourrelet
saillant parallele aux dernieres rides de la craie inferieur. Ce plain, qui prend le nom de Petit-
Champagne . . . Une excellent etude de ce systeme peut se fair dans les environs du Chateau de
Malberchie’ (Coquand 1856, pp. 86-87). ‘Road-side sections at Javrezac’ are not the only locality
mentioned specifically by Coquand, as van Hinte (1979) claims.

In 1859, Coquand listed the fauna of his Santonian, recording the following ammonites (p. 978),
none of which has been traced: (1) Ammonites Bowgeoisi d’Orbigny, 1850; this is a Coniacian
Protexanites (Kennedy 1984a p. 105, pi. 23, figs. 1-4, 7-9; pi. 24, figs. 1-8; pi. 26, figs. 4 and 5;
text-figs. 32, 33, 34, 35a-e, 36b, c, e, f). (2) A. Orbignyi d’Archiac, 1843; this is unrecognizable.
(3) A. polyopsis Dujardin, 1837; this is an exclusively Santonian Placenliceras (Kennedy and Wright
1983 p. 856, pis. 86-88; text-figs. 1-4; see below). (4) A. santonensis D’Orbigny, 1850; this is a
Neoptychites cephalotus (Courtiller, 1860) of the Turonian (Kennedy and Wright 1979 p. 676, pi.
86, figs. 4 and 5). (5) A. coniaciensis Coquand, 1859; this was described as follows:

‘Hauteur, 65 millim.; diametre transversal, 55 millim. Coquille comprimee, assez largement ombiliquee,
spire formee de tours convexes, ornee de cotes alternalivement simples et bifides, chaque tour possede cinq
rangees de tubercules; la premiere disposee autour de I’ombilic, deux medianes et deux pres du pourtour
externe; celle qui est la plus rapprochee du dos et qui termine les cotes est saillante. Dos carene, carene
tranchante logee entre deux sillons.

Cette espece a ete toujours confondue avec T A. varians.— Cognac, Malberchie, Epagnac.’

IPalaeontology, Vol. 30, Part 4, 1987, pp. 765-782, pis. 80-82|

© The Palaeontological Association

766

PALAEONTOLOGY, VOLUME 30

COQUAND 1856

*1^® nomenclature'
1®'' ETAGE

5e horizon de Rudistes

3® sous-6tage
Craie tendre, a silex

1857,

COQUAND 1858
*2® nomenclature'

2® ETAGE:

SANTONIEN

Craie tendre avec silex
Pleurotomaria Santonesa.
Janira Truellei. Soondvlus
hibpuritorum. Rhynchonella
vespertilio. Rhynchonella
intermedia. Terebratula
Nanclasi, Micraster
laxoporus, Hemiaster Stella,
Salenia geometrica

COQUAND 1857

*3® nomenclature'

2® ETAGE:

SANTONIEN
7e horizon de Rudistes

Micrasier brevis, Spondylus
truncates, Rhynchonella
vespertilio, Radiolites
Arnaudi

ARNAUD 1877

Calcaire noduleux,
glauconieux
Conoclypeus ovum
Marnes et gres supfirieurs
Ostrea acutirostris,
Sphaerulites Hoeninghausi
Marnes a Ostrea
vesicularis et Ostrea
proboscidea
Marnes et gres inf6rieurs
Rhynchonella detormis,
Botrioovous

TABLE I. Correlation of previous subdivisions of the Santonian in northern Aquitaine.

This is a very clear description of a Texanites ( Texanites) but is insufficiently detailed to determine
which species is described. In the absence of type specimens I regard it as a nomen diihium. (6)
BacuHtes incurvatus Dujardin, 1837; this is a Coniacian-Santonian species (Kennedy 1984r7 p. 143,
pi. 32, figs. 12, 15-19; pi. 33, figs. 1-22; text-figs. 41 ‘And Alv-u). (1) Scaphites constrictus diOrhigny,
1842; the correct author is J. Sowerby, 1817; the species is a Maastrichtian Hoploscaphites (Kennedy
1986r? p. 64, pi. 13, figs. 1-13, 16-24; pi. 14, figs. 1-38; pi. 15; text-figs. 9 and 11a-h).

Arnaud (1862-1897) refined the stratigraphy of the Cretaceous of the Aquitaine Basin, and his
vast collections form the basis of the present study.

His initial division of the stage (18776, p. 3) was:

2*“ serie: Santonien

18. Marnes et gres inferieurs: Rhynchonella cieformis', Botriopygus

19. Marnes a Ostrea vesicularis et O. proboscidea

20. Marnes et gres superieurs: Sphaerulites Hoeninghausi, Ostrea acutirostris;

21. Calcaire noduleux glauconieux; Conoclypeus ovum.

A letter system replaced this in the summary diagram at the end of this work, and was the basis
of all later modifications, summarized in Table 1 herein.

De Grossouvre (1894) used Arnaud’s Collections for his revision of the ammonites of the ‘Craie
Superieur’, referring specimens to the Arnaud letter divisions and recognizing a Santonien Inferieur
and Superieur in his tabulation of ranges. In 1901 de Grossouvre repeated Arnaud’s letter
system in his account of the ‘Craie de F Aquitaine’ (pp. 351 et seq.). He records Mortoniceras
serralomarginatwn Redtenbacher, M. texanum Roemer, Placenticeras syrtcde, and BacuHtes
incurvatus from M\ M. texanum and P. syrtale from M^, the latter extending to N^. In his
summary chapter 22 (Classification des couches supraeretacees, p. 751 et seep), the Santonian is
defined as follows (p. 795); ‘Le Santonien comprend les couches situees au-dessus du Coniaeien et
inferieurs a la zone a Placenticeras hidorsatum, par laquelle debate le Campanien; PI. syrtale existe
sur toute la hauteur de cet etage, on pourrait encore le definir comme eonstituant I’assize du PI.
syrtale.' From his work in the Corbieres, de Grossouvre reeognised what would now be called

KENNEDY : AMMONITES OF THE TYPE SANTONIAN

767

ARNAUD 1877

ARNAUD 1883

DE GROSSOUVRE
1901

ZONES

AQUITAINE

Placenticeras

syrtale

Calcaire sableux a
Rudistes

Marnes a
Ostracees

Mortoniceras

texanum

Banc a
Botriopygus

Couches a
Ammonites texanus

Marnes et grfts

santoniens

superieurs

Calcaire glauconieux a

Conoclypeus ovum

Banc a Ostrea

acutirostris

Gres a Sphaerulites

Hoeninqhausi

Marnes a Ostrea vesicularis et Ostrea
proboscidea

Marnes et gres santoniens intdrieurs
Botriopyqus Toucasanus

Grds ou calcaire
2 rnarneux, avec ou sans
N silex. et g6odes de
quartz Ammonites
polyopsls, etd.

Calcaire rnarneux avec
silex, passant a des
marnes grises ou
verd^tres, p6tries
d’Ostrac6es
O.vesicularis. etc.

Calcaire rnarneux ou

M2 noduleux gres

calcaritere Ammonites
polyopsis. etc.

Calcaire rnarneux
arenac^ Ammonites
coniacensis, ^c.

assemblage zones within the Santonian, a lower one characterized by the occurrence of Mortonicenis
texammi, and an upper one with P. syrtale and Pachydiscus iscidensis as indices (table 35,
following p. 830) or with Placenticeras syrtale alone as index (tables 36-39, following p. 830). A
texaimm -F syrtale Zone were recognized by Haug ( 1 9 1 1 ), while the Colloque sur le Cretace Superieur
held in 1959 (Dalbiez 1960) recognized a texamis Zone below and a syrtale-iscideiisis Zone above.

As I have noted elsewhere (Kennedy 19846, 1985) neither P. syrtale nor M. texaniim occur in
France, while in Aquitaine it is only possible to define the Santonian in ammonite terms by the
appearance of Texanites (Texanites) and as the range zone of P. polyopsis Dujardin, 1837 ( =
syrtale of European authors); I have traced only five surviving texanitids from the Santonian in
Aquitaine: they are recorded from M', M“, and N^. I have seen no specimens referable to
Paratexanites serratomarginatus (Redtenbacher, 1873), recorded by de Grossouvre and by Gillard
(1944) from M'; elsewhere I have suggested it to be exclusively Upper Coniacian in France
(Kennedy 1984r<f, p. 117, pi. 23, figs. 5 and 6; pi. 26, figs. 1-3; pi. 27, figs. 1-7; text-fig. 35f, g),
and de Grossouvre (1901) ultimately concluded that Santonian P. serratomarginatus were actually
P. emsclieris (Schliiter, 1876), I have equally seen none of these from Aquitaine, nor of the
Coniacian Protexanites hourgeoisi (d'Orbigny, 1850) noted from the Lower Santonian by de
Grossouvre in 1894 (table, p. 86).

The ammonite distributions from the present work can be summarized as follows:

Placenticeras polyopsis (Dujardin, 1837)

M‘ M2 N‘

N2

P. paraplanum Wiedmann, 1978

— — —

N2

T. (Texanites) gallicus Collignon, 1948

Ml M2 -

—

Eidophoceras austriacum (Summesberger, 1979)

— — —

N2

Boehmoceras loescheri Riedel, 1931

- M2 N>

—

I see no possibility of subdividing the Santonian in ammonite terms in Aquitaine, a disappointing
result when compared with the results from my revision of the Coniacian Kennedy (1984«) and
Campanian of the region (Kennedy 19866).

768

PALAEONTOLOGY, VOLUME 30

Low diversity and small numbers of individuals are an obvious explanation for this. There are
striking differences in stratigraphic ranges of taxa common to Aquitaine and the richly ammoniti-
ferous sequences of the Corbieres (Aude), as noted by de Grossouvre (1901) and Bilotte and
Collignon (1983), so that some environmental limitation on ammonite occurrence might be
involved. To this must be added the marked coarsening in resolution of ammonite zones in
successive stages of the Upper Cretaceous. Taking the time scales of Odin (1985) and Snelling
(1985) and the standard zonations for the Upper Cretaceous stages in north-west Europe proposed
by Kennedy (1984(?), average zonal durations are as follows: Cenomanian (4 Ma) 0-57 Ma;
Turonian (4 Ma) 0-75 Ma; Coniacian (2 Ma) 0-5 Ma; Santonian (3 Ma) 3 Ma; Campanian (11
Ma) 2-7 Ma. No zonal division based on ammonites can be recognized in the Maastrichtian (7
Ma) of north-west Europe at this time. A 3 Ma polyopsis Zone for all of the Santonian is thus of
the same order of duration as the succeeding Campanian zones recognized in Aquitaine.

Location of specimens. The following abbreviations are used to indicate the location of specimens mentioned
in the text:

EMP: Ecole des Mines, Collections, now at the Universite Claud Bernard, Lyon.

FSL: Universite Claude Bernard, Lyon.

GPIB: Geological and Palaeontological Institute, Bonn.

SP: Collections of the Sorbonne, now housed in the Universite Pierre et Marie Curie, Paris.

Suture terminology. The system of Wedekind (1916) as revised by Kullmann and Wiedmann (1970) is used
here. E = external lobe, L = lateral lobe, U = umbilical lobe, I = internal lobe.

Dimensions. All dimensions are given in millimetres; D = diameter, Wb = whorl breadth, Wh = whorl
height, and U = umbilicus; c = costal and ic = intercostal. Figures in parentheses refer to dimensions as a
percentage of diameter. The term rib index as applied to heteromorphs is the number of ribs in a distance
equal to the whorl height at the mid point of the interval counted.

Synonymies. Only citations which include illustrations of material or important systematic, stratigraphic, or
geographic information are included.

SYSTEMATIC PALAEONTOLOGY

Order ammonoidea Zittel, 1884, pp. 355, 392
Suborder ammonitina Hyatt, 1889, p. 7
Superfamily hoplitaceae H. Douville, 1890, p. 290

[Nom. correct. Wright and Wright, 1951, p. 21 {pro Hoplitida Spath, 1922a, p. 95, nom. transl. ex Hoplitidae

Douville, 1890, p. 290)]

Family placenticeratidae Hyatt, 1900, p. 585

[= Hypengonoceratinae Chiplonkar and Ghare, 1976, p. 2; Baghiceratinae Chiplonkar and Ghare, 1976, p. 3]

Genus placenticeras Meek, 1876, p. 462
[see Kennedy and Wright 1983, p. 869 for synonymy]

Type species. Ammonites placenta DeKay, 1828, p. 278, by original designation by Meek, 1876, p. 462.

Placenticeras polyopsis (Dujardin, 1837)

1837 Ammonites polyopsis Dujardin, p. 232, pi. 17, fig. 12.

1983 Placenticeras polyopsis (Dujardin, 1837); Kennedy and Wright, p. 856, pis. 86-88; text-figs.
1 4 (with full synonymy).

1983 Placenticeras syrtale (Morton); Collignon, p. 200, pi. 6, fig. 1.

KENNEDY: AMMONITES OF THE TYPE S ANTON I AN

769

1983 Stcmtonocems guaclaloupae (Roemer); Collignon, p. 201, pi. 6, fig. 3.

1983 Stantonoceras guadaloupae (Roemer) var. quadrata de Grossouvre; Collignon, p. 201, pi. 6,
fig. 2.

1983 Sumtonoceras depression (Hyatt); Collignon, pi. 202, pi. 7, fig. 2.

1985 Pkicenticeras polyopsis Dujardin; Kennedy, pi. 2, figs. 7 10.

Lectotype. The original of Dujardin 1837, pi. 17, fig. \2a, a juvenile macroconch from the ‘Craie Tufau’ of
Touraine, France; designated by Kennedy and Wright 1983, p. 856; present whereabouts unknown.

Material. Numerous specimens. SP unregistered e.\ Arnaud Collection, from Assise M’, Larat; M^, L’Ombre,
Miremont, Colombier du Miremont, Riberac, Perigueux, route d’Agonac (Perigueux), Beaulieu (Siolac,
Souterrain de Beaulieu), St Leon-sur-Vezere, Champagnac-de-Belair, Versannes, Epagnac, Puygaty, Rognac;
M' ^ at Angoumac, Cognac; N' St Caprais, at Arcivaux, Saintes, Foquebrune, and Sarlat, route
d’Eyzies. EMP e.\ Boucheron Collection, from Ea Valette. M. Seronie-Vivien Collection: SX62, from Mater
( = Stantonoceras guadaloupae in Seronie-Vivien 1972, text-fig. 26); 5579c, 80c, Castelfadeze ( = 5. guadaloupae
in Seronie-Vivien 1972, p. 88); SU2b, 2c, from I’Amblardie (= S. guadaloupae in Seronie-Vivien 1972, p. 78);
J. P. Plate! Collection: three specimens from Agonac.

Discussion. The Aquitaine material noted above adds nothing to the detailed description of the
species given by Wright and Kennedy (1983).

Occurrence. Widespread in Aquitaine, ranging from low in M' to N^. Also known from Touraine, Corbieres,
the Beausset Basin (Var) in Prance, Austria, the Germanics, and the Tombigbee Sand Member of the Eutaw
Formation in Alabama. See detailed discussion by Wright and Kennedy (1983).

Pkicenticeras paraplanuni Wiedmann, 1978
Plate 80, figs. 1-3, 8 10

1978 Placenticeras paraplanuni Wiedmann, p. 666, pi. 1, figs. 3 and 4; text-fig. 2a.

1979 Placenticeras paraplanuni Wiedmann; Summesberger. p. 152, pi. 13, figs. 53 57; text-figs. 38
and 39.

1985 Placenticeras aff. paraplanuni Wiedmann; Amedro and Hancock, p. 24 et seq., text-fig. lb/-c,
.kg-

Holotype. By original designation, the original of Wiedmann 1978 pi. 1, figs. 3 and 4; text-fig. 2a, from the
Upper Santonian Gosau Beds of the Gosau Basin, Austria.

Material. SP unregistered c.y Arnaud Collection, from Assise N^, Arcivaux, Saintes (Charente-Maritime);
from Assise N^, Beillant (Charente-Maritime); FSE, six specimens from the autoroute sections west of Saintes
(details in Amedro and Hancock 1985).

Dimensions.

SP, Arcivaux

D Wb Wh Wb : Wh U

84-2(100) 25-5 (30-3) 39-7 (47-1) 0-64 17-0 (20-2)

Description. All specimens are ill-preserved composite moulds. Coiling moderately evolute, with 55 % of
previous whorl covered. Whorl section compressed (Wb : Wh ratio 0-64), with greatest breadth at umbilical
shoulder. Umbilicus small (20 % of diameter), shallow, conical, with flat, outwards-inclined wall. Umbilical
shoulder narrowly rounded; inner flanks broadly rounded, outer flanks flattened, convergent. Venter narrow,
concave with sharp shoulders at smallest diameter visible, broad, with shoulders blunter and broadly arched
venter on outermost phragmocone whorl. Flank ornament distinctive, with markedly convex distant very
broad ribs with steep apical face but apertural face that merges imperceptibly with flank; a strong furrow
precedes each rib, and may develop into a marked constriction (PI. 80, fig. 2), so that the shell surface has
the step-like topography of a tiled roof. Flank ornament strengthens progressively through ontogeny, and,
on the outer phragmocone whorl, is accompanied by blunt outer ventrolateral clavi, 9 10 per half whorl.

Sutures ill-exposed, typical for genus.

Discimion. The striking convex course of the flank ornament, furrows/constrictions plus late
appearance of clavi separate this species from all other European forms referred to the genus. The

770

PALAEONTOLOGY, VOLUME 30

early whorls show a striking resemblance to those of the much larger species P. hidorsalum
(Roemer, 1841) (see revision in Kennedy 1986/?). Indeed, Amedro and Hancock (1985) regarded
some of the present specimens as transitional to P. caiialiculatum (Hyatt, 1903), a synonym of P.
bidorsatum. The adult ornament of the two is highly distinctive, but the similarity of early growth
stages suggests that bidorsatum is a hypermorphic giant derivative of paraplammi.

Occurrence. Assise N- in Aquitaine; high Santonian of the Corbieres (Aude); Upper Santonian of the Gosau
Basin, Austria.

Superfamily acanthocerataceae de Grossouvre, 1 894, p. 22

[nom. correct. Wright and Wright 1951, p. 24 {j)ro Acanthoceratida Hyatt 1900, p. 585, nom. transl. ex
Acanthoceratidae Hyatt 1900, p. 585, nom. correct, ex Acanthoceratides de Grossouvre, 1894)]

Family collignoniceratidae Wright and Wright, 1951, p. 30

[nom. siibst. pro Prionotropidae Zittel, 1895, p. 530; Prionotropis Meek, 1876, p, 453, non Fieber, 1853, p.
127; = Collignoniceras Breistroffer, 1947; = Prionocyclidae Breistroffer, 1947, ex Prionocyclus Meek, 1876,

p. 298, ineligible as family type]

Subfamily texanitinae Collignon, 1948, p. 54 (9)

[nom. transl. Wright 1957, p. L429 ex Texanitinae Collignon, 1948, p. 54(9)]

Genus texanites Spath, 1932, p. 379

Type species. Ammonites texanus Roemer, 1852, p. 31, pi. 3, fig. I, by original designation.

Subgenus texanites (texanites) Spath, 1932, p. 379
Texanites (Texanites) gcdlicus Collignon, 1948
Plate 80, figs. 4 7; Plate 81, figs. 1-6

1894 Mortoniceras texanum F. Romer sp.; de Grossouvre, p. 80, pi. 16, figs. 2 and 4; pi. 17, fig. 1.
1948 Texanites texanus Roemer var. gallica Collignon, p. 75 (30), pi. 8 (2), fig. 1.

1963 Texanites texanus gallica Collignon, 1948; Young, p. 81, pi. 38, figs. 3 and 4.

1966 Texanites gcdlicus Collignon; Collignon, p. 78, pi. 487, fig. 1964; pi. 508, fig. 2020; pi. 509,
fig. 2020.

1970 Texanites gallicus (Collignon); Matsumoto, pp. 267, 270, 272.

1980 Texanites texanus s. 1. (Roemer, 1852); Klinger and Kennedy, p. 162 (pars), text-figs. 124
and 125.

1982 Texanites gcdlicus Collignon; Martinez, p. 121, pi. 20, figs. 1 and 2.

1983 Texanites gallicus Collignon, 1948; Collignon, p. 203.

1985 Texanites (Texanites) gallicus Collignon, 1948; Kennedy, pi. 2, figs. 3-6.

Types. Collignon cites French, Venezuelan, Italian, Bulgarian, and Madagascan specimens in his text and
synonymy, but failed to select a holotype. Young (1963, p. 81 ) presumed the original of de Grossouvre (1894,

EXPLANATION OF PLATE 80

Figs. 13, 8-10. Placenticeras paraplanum Wiedmann, 1978. 1-3, FSL A362, from the Upper Santonian of
Autoroute Cutting S2, west of Saintes (Charente-Maritime). 8-10, SP unregistered ex Arnaud Collection,
from Assise N^, Arcivaux, Saintes (Charente-Maritime).

Figs. 4-7. Texanites (Texanites) gallicus Collignon, 1948. SP unregistered ex Arnaud Collection, two
paralectotypes, figured by de Grossouvre 1894, pi. 16, figs. 2 and 4, from Assise M^, Nieul-le-Virouil
(Charente-Maritime).

All figures are natural size.

PLATE 80

KENNEDY, Placenticeras, Texanites {Texanites)

Ill

PALAEONTOLOGY, VOLUME 30

pi. 17, fig. 1) to be the holotype, as did Martinez (1982, p. 121); in part 2 of his texanitid revision, Collignon
(1948, p. 42 (99)) refers to this as the Type', which I take to be a valid lectotype designation; it is from the
Santonian of the Corbieres. Two of the three small paralectotypes figured by de Grossouvre (1894, pi. 16,
figs. 2 and 4) have been traced and are shown in PI. 80, figs. 4-7; SP unregistered ex Arnaud Collection,
from Assise M-, Nieul-le-Virouil (Charente-Maritime).

Material. SP unregistered ex Arnaud Collection, an unregistered specimen from Assise M', Les Rentes,
Cognac (Charente) plus a fragment from M', Sergeac. EMP unregistered: two specimens labelled Lavalette
(Charente), without horizon indicated, ex Boucheron Collection.

Description. All specimens are composite moulds. Juveniles to a diameter of 66 mm are very evolute, whorls
overlapping only to the extent that the marginal tubercle (4) is just concealed below the umbilical seam of
the preceding whorl. Whorl section as wide as high at smallest diameter visible, becoming compressed as size
increases, with greatest breadth at lateral (2) tubercle. Umbilicus broad (up to 44 % of diameter) and shallow.
There are 32-36 ribs per whorl. They arise, singly or in pairs, at sharp umbilical bullae, of which there are
approximately 22 per whorl in the best-preserved specimen, with occasional intercalatories arising low on
the flank. Ribs are straight and prorsiradiate across the flanks, and bear a small conical inner lateral (2),
rounded to feebly clavate submarginal (3), larger, rounded to feebly clavate marginal (4) and strong clavate
external (5) tubercle. There is a sharp siphonal keel, flanked by deep, broad grooves.

A larger specimen (PI. 8, figs. 5 and 6) is badly crushed, and 123 mm in diameter. At this size, there are
26 straight, prorsiradiate primary ribs per whorl; the umbilical (1), lateral (2), and submarginal tubercles are
all markedly rounded, and the specimen closely resembles the lectotype.

Sutures not seen.

Discussion. Collignon (1948) correctly identified the differences between European specimens
referred to Texanites texaniis and the American lectotype, which is shown here in Text-figure 1.
He erected two varieties, gallica and hispanica, later (1966) elevating them to specific rank. T.
gallicits is certainly very distinct from T. texanwn, having 50 % more ribs, markedly crowded and
arising in pairs on the inner whorl. T. texamim twiningi Young, 1963 (p. 82, pi. 38, fig. 5; pi. 39,
fig. 1; pi. 41, figs. 2 and 5; pi. 48, fig. 4) differs from the nominate subspecies only in its slightly
higher rib density, and is equally distinet from T. gallicits. T. hispaniciis is held to differ from T.
gallicits by having the ribbing arising on the umbilical wall and a lateral (2) tubercle that migrates
outwards to mid-flank during ontogeny. It and T. gallicus are probably conspecific, but the present
material is inadequate for any firm conclusion on the point. There is also a close similarity to T.
{Texanites) qitincptenodosus (Redtenbacher, 1873), recently revised by Kennedy, Summesberger and
Klinger (1981) (p. 126, figs. 8-16). They differ most obviously in the markedly clavate tubercles of
T. quinqitenodosits compared to the rounded tubercles in T. gallicits.

Occurrence. Assise M' and in Aquitaine, Lower Santonian of the Corbieres; also recorded from Spain,
Bulgaria, ?Italy, Venezuela and Zululand and from the Lower and Middle Santonian of Madagascar.

Family sphenodiscidae Hyatt, 1900, p. 585
Subfamily sphenodiscinae Hyatt, 1900, p. 585
(= Libycoceratinae Zaborski, 1982, p. 306)

Genus Eitloplioceras Hyatt, 1903, p. 85

(= Praelibycoceras H. Douville, 1912, p. 315; Pelecodiscus Van Hoepen, 1921, p. 30; Sphenodiscoceras Spath,
1921, p. 242; Skoumalia Summesberger 1979, p. 141)

Type species. Eulophoceras natalense Hyatt, 1903, p. 86, pi. 11, figs. 2-6.

EXPLANATION OF PLATE 81

Figs. 16. Texanites (Texanites) gallicits Collignon, 1948. 1-4, EMP unregistered ex Boucheron Collection,
from Lavalette (Charente), without indication of precise horizon; 5 and 6, SP unregistered ex Arnaud
Collection, from Assise M^, Les Rentes. Cognac (Charente). All figures are natural size.

PLATE 81

KENNEDY, Texanites (Texanites)

774

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 1. Te.xaiiites (Texanites) texanus (Roemer, 1851). The lectotype, GPIB 45a, the original of Roemer’s
pi. 3, fig. 1, from Austin, Texas. The original is 144 mm in diameter.

EXPLANATION OF PLATE 82

Figs. 1 3. Eulophoceras austriacum (Summesberger, 1979). FSL A363, from the Upper Santonian of Autoroute
Cutting S2, west of Saintes (Charente-Maritime).

Figs. 4 16. Boehmoceras loescheri Riedel, 1931. All SP unregistered, ex Arnaud Collection; 4-6, 11 13, from
Assise N‘, Charmant (Charente); 7-10, 14-16, from M^, Nieul-le-Virouil (Charente-Maritime).

All figures are natural size.

PLATE 82

KENNEDY, Eulophoceras, Boehmoceras

776

PALAEONTOLOGY, VOLUME 30
Eulophoceras austriacum (Summesberger, 1979)

Plate 82, figs. 1 - 3

1979 Skoimialia austriaca Summesberger, p. 141, pi. 9, figs. 37-41; text-figs. 26-30.

1980 Skoumalia austriaca Summesberger; Summesberger, p. 280, pi. 2, figs. 5 and 6; pi. 3, figs. 7 and
8; text-figs. 5 and 6.

1985 Eulophoceras austriaca (Summesberger); Amedro and Hancock, p. 23, fig. \ \d, e.

Holotype. The original of Summesberger ( 1979, pi. 9, figs. 37 and 38) from the Upper Santonian of the Gosau
Basin (Sandkalk Member, Bibereck, Hofergraben).

Material. A specimen from the autoroute cutting.S2, west of Saintes (Charente-Maritime) (see Amedro and
Hancock 1985 for details); FSL collections, no. A363.

Description. The specimen is a composite mould. Oxycone. Umbilicus comprises less than 9 % of diameter.
A marked facet separates flank from a more obtuse, sharper venter. Low, faint, broad, prorsiradiate, feebly
concave ribs ornament the inner flank but decline by mid-flank, giving rise to low, falcoid, concave riblets
and striae which strengthen periodically into small inner ventrolateral tubercles.

Sutures not seen.

Discussion. The single specimen corresponds to the feebly ornamented form B of Skoumalia
austriaca Summesberger, 1979 (p. 141, text-figs. 29 and 30; pi. 9, figs. 39-41; 1982, p. 280, pi. 2,
figs. 5 and 6; pi. 3, figs. 7 and 8; text-figs. 5 and 6). Form A, to which the holotype belongs, is
much more strongly ornamented, with umbilical bullae.

Summesberger regarded Skoumalia as Diaziceratinae although uncertain as to the position of
the subfamily. His illustrations of the suture (1979, text-figs. 27, 28, 30; 1982, text-fig. 6) show
what may be the beginnings of an adventive lobe, suggesting Sphenodiscinae, to which I refer it,
as a synonym of Eulophoceras Hyatt, 1903, because some specimens of the highly variable type
species (of which E. amapondense (Van Hoepen, 1921), E. umzamhiense (Van Hoepen, 1921), E.
tenue (Spath, 1922/?) and E. minor (Spath, 1922/?) are synonyms) have umbilical bullae and feeble
ventral nodes (e.g. Spath 1922/? pi. 6, fig. 2).

The only other European occurrence of Eulophoceras is in the Corbieres (Aude), where Collignon
(1983, p. 205, pi. 7, figs. 4 and 5) has described two completely smooth specimens from the Upper
Santonian as E. cf. miloni Hourcq, 1949 and E. grossouvrei Collignon, 1983, differentiating them
on details of whorl section and the sutures, which were not illustrated. It is not possible to compare
them usefully with the present material without detailed restudy, although larger specimens from
the region (SP unregistered e.x Toucas Collection) clearly represent a different species.

Occurrence. Upper Santonian of Austria and northern Aquitaine.

Suborder ancyloceratina Wiedmann, 1966, p. 54
Superfamily turrilitaceae Gill, 1871, p. 3
Family baculitidae Gill, 1871, p. 3
[= Eubaculitinae brunnschweiler, 1966, p. 24]

Genus boehmoceras Riedel, 1931, p. 690

Type species. By the subsequent designation of Wright 1957, p. L220: Ancvloceras krekeleri Wegner, 1905
p. 210, pi. 8, fig. 2.

Diagnosis. Medium-sized rapidly expanding criocones with compressed oval whorl section, venter
flanked by shallow depressions, producing a blunt keel. Ornamented by straight to concave
prorsiradiate primary ribs, strengthened into crescentic umbilical bullae or not, shorter intercalated
ribs well-developed or not; ribs may split into secondaries on ventrolateral area giving a distinctive
corded keel. Aperture with short dorsal and long ventral rostrum. Suture with rectangular bifid
lobes and saddles.

KENNEDY: AMMONITES OF THE TYPE SANTONIAN

111

Discussion. Recent authors place Boehmoceras in the Family Phlycticrioceratidae Spath, 1926 (e.g.
Wright 1957, p. L220; Summesberger 1979, p. 117). It appears, rather, to be a recoiled baculitid,
as briefly noted elsewhere (Kennedy and Wright 1985, p. 142). The ornament of crescentic ribs,
strengthened into dorsolateral bullae in Boehmoceras loescheri Riedel, 1931, is a typical Baculites
feature, as is the rostrate aperture and suture with rectangular bifid elements. Specimens illustrated
here (PI. 82, figs. 7-10, 14-16) show remarkable resemblance to the Coniacian B. incurvatus
Dujardin, 1837 (p. 232, pi. 17, fig. 13; see Kennedy 1984r/, p. 143, pi. 32, figs. 12, 15-19; pi. 33,
figs. 1-22; text-figs. 41 and 42f-m); some body chambers of this species are markedly curved (e.g.
Kennedy 1984c/, pi. 33, figs. 4 and 7) and the genus Boehmoceras is yet another example (albeit
short-lived) of recoiling in heteromorphs. Resemblance to Phlycticrioceras Spath, 1926 is superficial
only. That genus has clearly differentiated ventrolateral and siphonal tubercles and a far more
intricately subdivided suture (De Grossouvre 1894, text-fig. 88; Kennedy 1984c/, text-fig. 42c). Two
species are currently referred to Boehmoceras, B. krekeleri (Wegner, 1905) known from the Upper
Santonian of the Munster Basin, North Germany and the Gosau Basin, Austria, and B. loescheri
Riedel, 1931 which has the same geographic distribution plus the present records from Arnaud’s
Assizes and N* at Nieuil-le-Virouil (Charente-Maritime) and Charmant (Charente) respectively.
Previous illustrations and the material before me provide too small a sample to determine whether
these species actually merit separation.

Occurrence. Mid-Upper Santonian. North Germany, Austria, and northern Aquitaine, France. Santonian
of Alabama and Texas in the USA (W. A. Cobban, pers. comm.).

1931 Boehmoceras loscheri Riedel, p. 692, pi. 78, figs. 3-6.

1979 Boehmoceras loescheri Riedel; Summesberger, p. 119, pi. 2, figs. 15, 16, 18; text-figs. 9-12.
1983 Boehmoceras Kennedy and Wright, p. 866.

1985 Boehmoceras sp. Kennedy, pi. 2, fig. 1.

Types. Riedel’s syntypes, if they survive, are in the collections of the Zentrales Geologisches Institut, Berlin.
They are from the Upper Santonian of the Munster Basin, German Federal Republic.

Material. Five unregistered specimens in the SP Collections, ex Arnaud Collection. Two (PI. 82, figs. 7 10,
14 16) are from Nieul-le-Virouil (Charente-Maritime), one labelled M^, the other M ?; three specimens
(PI. 82, figs. 4-6, 11-13 plus an unfigured fragment) are from Charmant (Charente), labelled N*.

Description. The smallest specimen available is from Charmant, and has a maximum preserved whorl height
of 12 mm. It is highly distorted and the original whorl section is compressed at the apical end but depressed
at the apertural. Coarse primary ribs are separated by interspaces approximately equivalent to the whorl
height. They are feeble and convex across the dorsum, strengthening into coarse concave crescentic dorsolateral
bullae that give rise to strong primary ribs that sweep across the flank before declining somewhat over the
venter.

Boehmoceras loescheri Riedel, 1931
Plate 82, figs. 4 16; text-fig. 2

TEXT-FIG. 2. External suture of Boehmoceras loesc.
Riedel, 1931.

778

PALAEONTOLOGY, VOLUME 30

One or two coarse, distant secondary ribs arise around mid-flank and cross the venter, without attaining
the strength of the primaries. The unfigured fragment from Charmant is crushed but shows the same style
of ribbing but with much weaker secondary ribs. The third fragment appears to be from close to the adult
aperture and shows ribs that coarsen markedly over ventrolateral shoulders and venter (PI. 82, figs. 4 6).
The specimens from Nieul-le-Virouil are much better preserved undeformed internal moulds, one wholly the
other part body chamber. The whorl section is compressed, with an intercostal whorl breadth to height ratio
of 0-67 0-72, the dorsum broadly, the venter more narrowly rounded.

The primaries are very distant, feeble, and convex on the dorsum, sweeping back over the dorsolateral
region and strengthening into crescentic dorsolateral bullae. These decline into concave ribs that project
strongly forwards and decline over the venter.

Between are much finer secondaries, mere striae on the inner flank but strengthening over the venter. They
appear distinctly scale-like in the body chamber fragment.

Suture line (text-fig. 2) with rectangular bifid lobes and saddles; moderately incised.

TEXT-FIG. 3a, b, Boelmioceras krekeleri (Wegner) Riedel, 1931, x I. Specimens are from the Santonian of the
Munster Basin, German Federal Republic, and in the Probenarchiv, Bernau (photographs courtesy of
Zentrales Geologisches Institut, per H. Summesberger). b is the original of Riedel 1931, pi. 77, fig. 4.

Discussion. Marked differentiation of ribbing in the Aquitaine material indicates it should be
referred to B. loescheri. The only other species, B. krekeleri (Wegner, 1905) (p. 210, pi. 8, fig. 2;
Riedel 1931, p. 691, pi. 77, figs. 3-5; pi. 78, figs. 1 and 2; Summesberger 1979, p. 118, pi. 2, fig.
14; text-figs. 7 and 8) presents certain problems of interpretation, as noted by Summesberger
(1979), for the type appears to be lost. Specimens subsequently referred to B. krekeleri by Riedel
(see text-fig. 3) show uniform, rather closely spaced ribs, without the same degree of differentiation
into primaries and secondaries. As noted above these differences may be within the limits of
intraspecific variation. Without more material, it is impossible to say.

Occurrence. Upper Santonian of the Miinster Basin (Recklinghauser Sandmergel), Gosau Basin, Austria
(Sandkalk Member, Bibereck, Ffofergraben 1); Assise M^ and N‘ in northern Aquitaine.

KENNEDY: AMMONITES OE THE TYPE SANTONIAN

779

Acknowledgements. I thank J. M. Hancock and R. Parish for assistance in the field, C. W. Wright, W. A.
Cobban, H. Summesberger, and U. Kaplan for valuable advice and criticism. Many museum curators allowed
me to study specimens in their care: I especially thank D. Pajaud, J. Sornay, J. Louail, G. Marie, A. Prieur,
R. Busnardo, H. Jaeger, H. Remy, D. Phillips, H. G. Owen, and M. K. Howarth. The technical assistance
of the stalf of the Geological Collections, University Museum and Department of Earth Sciences, Oxford is
gratefully acknowledged, as is the financial support of the Royal Society and Natural Environment Research
Council.

REFERENCES

AMEDRO, F. and HANCOCK, J. M. 1985. Les Ammonites de TAutoroute 'EAquitaine’, France (Turonien et
Santonien). Cret. Res. 6, 15-32.

ARCHiAC, A. D. 1843. Etudes sur la formation cretacee du Sud-Ouest de la France. Annls. Sci. geol. Paris, 2,
121 143.

ARNAUD, H. 1862. Note sur la craie de la Dordogne. Bull. Soc. geol. Fr. (2), 19, 465-500, pi. 11.

1873. Profils geologiques des chemins de fer d’Orleans traversant la craie du Sud-Ouest. Ibid. (3), 1,

405-408, pi. 9.

1877fl. Profils geologiques des chemins de fer des Charentes, region cretacee. Etudes pratiques de la

craie du Sud-Ouest, 2^ partie. Act. Soc. linn. Bordeaux, 31, 168-201.

18776. Memoire sur le terrain cretace du Sud-Ouest de la France. Mem. Soc. geol. Fr. (2), 10, 110 pp.,

pis. 21-28 (1-8), tables.

\Zllc. Profils geologiques des chemins de fer d’Orleans, region cretacee. Etudes pratiques sur la craie

du Sud-Ouest. 3® partie. Act. Soc. linn. Bordeaux, 31, 251 286.

1878a. Parallelisme de la Craie superieur dans le Nord et le Sud-Ouest de la France. Bull. Soc. geol. Fr.

(3), 6, 205-211.

18786. Synchronisme de I’etage du Turonien dans le Sud-Ouest et le Midi de la France. Ibid. 6, 233-

242.

1883a. Profils geologiques des chemins de fer de Siorac a Sarlat et de Perigueux a Riberac. Etudes

pratiques sur le craie du Sud-Ouest, 4*“ partie. Act. Soc. Linn. Bordeaux, 37, 34-48, pis. 1-3.

18836. Parallelisme entre la Craie du Nord et celle du Sud-Ouest. Bull. Soc. geol. Fr. (3), 11, 92.

1887a. Aperqu generale sur la craie du Sud-Ouest. Ibid. 15, 809-813, pi. 32.

18876. Compte-rendu de I’excursion a Saint-Cirq et de Saint-Cirq a Beaumont. Ibid. 15, 841-845.

1887c. Resume generale des observations sur la craie du Sud-Ouest. Ibid. 15, 884-91 1.

1887c/. Excursion a Montignac-sur-Vezere. Ibid. 15, 914-918.

1891. De la limite tracee par Coquand entre le Santonien et Campanien. Ibid. 19, 665-668.

1892. Profil geologique du chemin de fer d’Angouleme a Marmande, region cretacee. Etudes pratiques

sur la craie de Sud-Ouest. 5*“ etude. Act. Soc. Linn. Bordeaux, 45, 11-43, 1 pi.

1897. Divisions naturelles du Cretace superieur au-dessus du Santonien dans le Sud-Ouest et la region

pyreneenne. Bull. Soc. geol. Fr. (3), 25, 676-681.

BiLOTTE, M. and COLLIGNON, M. 1983. Biostratigraphie et Paleontologie des ammonites du Senonien Inferieur
de Rennes-les-Bains-Sougraine (Aude) Zone sous-Pyreneenne Orientale. Dociim. Lab. geol. Lvon hs 6, 175-
223, 7 pis.

BREiSTROFFER, M. 1947. Notes de nomenclature paleozoologique. Proc.-verb. mens. Soc. Sci. Dauphine, 26,
(195), 5 pp., (unpaged).

BRUNNSCHWEiLER, R. o. 1966. Upper Cretaceous ammonites from the Carnarvon Basin of Western Australia.
1. The heteromorph Lytoceratina. Bull. Bur. Miner. Resour. Geol. Geophys. Aust. 58, 1-58, 8 pis.

CHiPLONKAR, G. w. and GHARE, M. A. 1976. Palaeontology of the Bagh Beds— Part VII: Ammonoidea. Bull.
Earth Sci. for 1976, 4 and 5, l-IO, 2 pis.

COLLIGNON, M. 1948. Ammonites neocretacees du Menabe (Madagascar). I. Les Texanitidae. Annls geol. Serv.
Min. Madagascar, 13, 49-107 (1-63), pis. 7-20 (1-14); 14, 7-101 (64-120), pis. 15-23.

1966. Atlas des fossiles caracteristkptes de Madagascar (Ammonites). XIV, Santonien. x-l-134 pp., pis.

455-513. Tananarive, Service Geologique.

1983. In BILOTTE, M. and collignon, m., 1983. q.v.

COQUAND, H. 1856. Notice sur la formation cretacee du department de la Charente. Bull. Soc. geol. Fr. (2),
14, 55-98.

I857fl. Sur la craie superieur de EAquitaine. Ibid. 743-745.

780

PALAEONTOLOGY, VOLUME 30

coquand, h 1857A. Position des Ostrea columba et hiciiiriciilata dans le groupe de la craie inferieure. Bull.
Soc. geol. hr. (2) 14, 745 -766.

1858. Description physique, geologique, paleontologique et minercilogiqiie du department de la Charente

(imprime sous la direction du Conseil Generate). 542 pp., Map. Bcsangon, De Dodivers et Cie.

1859. Synopsis des animaux et des vegetaux fossiles observes dans la formation cretacee du Sud-Ouest

de la France. Bull. Soc. geol. Fr. (2), 16, 945 1023.

COURTILLER. 1860. Description de trois nouvelles especes d’ammonites du terrain cretace des environs du
Saumur. Mem. Soc. imp. Agr. Sci. Arts, Angers, 3, 246-252, 3 pis.

DALBiEZ, M. F. 1960. Correlations et Resolutions. C.r. emigres Societes Savantes-Dijon 1959: Colloque sur le
Cretace superieur franqais, 857-867.

DEKAY, J. E. 1828. Report on several fossil multilocular shells from the state of Delaware; with observations
on a second specimen of the new fossil genus eurypterus. Ann. Lyceum nat. Hist. 2, 273-279, pi. 5, figs.
2-5 only.

DOUViLLE, H. 1890. Sur la classification des ceratites de la craie. Bull. Soc. geol. Fr. (3) 18, 275-292.

1912. Evolution et classification des Pulchelliides. Ibid. (4), 11, 281 320.

UUJARDIN, F. 1837. Memoire sur les couches du sol en Touraine et description des coquilles de la craie et des
Faluns. Mem. Soc. geol. Fr. 2, 211-311, pis. 15-20.

FiEBER, I. X. 1853. Synopsis der europaischen Orthopteren mit besonderes Rucksicht der Bohmischen Arten.
Lotos, 3, 90 104; 115-129; 138-154; 168 176; 184 188; 201-207; 232-238; 252-261.

GILL, T. 1871. Arrangement of the Families of Mollusks. Smithson, misc. Colins, 227, xvi-f 49 pp.

GiLLARD, p. A. 1944. Contribution a la stratigraphic de la Saintoinge meridionale (feuille de Saintes au
1/80 000). C.r. Coll. Serv. Carte geol. Fr. 45, 137 170.

GROSSOUVRE, A. DE. 1894. Rccherches su la craie superieure, 2, Paleontologie. Les ammonites de la craie
superieure. Mem. serv. Carte geol. det. Fr. 1 264, 39 pis. (misdated 1893).

1901. Recherches sur la craie superieure I: stratigraphic generate. Ibid, vii-l- 1013 pp.

HAUG, E. 1907 191 1. Trade de Geologie, 2024 pp., 135 pis. Masson, Paris.

HiNTE, j. E. VAN. 1979. The Coniacian, Santonian and Campanian stratotypes. Lethaia, 12, 183-187.

HOEPEN, E. c. N. VAN. 1921. Crctaceous Cephalopoda from Pondoland. Ann. Transv. Mus. 8, I 48, pis. 1-11.

HOURCQ, V. 1949. Paleontologie de Madagascar. XXVIII.— Sur quelques ammonites du Senonien. Annls
Paleont. 35, 10 (87) 31 (117), pis. 11 (1) 13 (3).

HYATT, A. 1889. Genesis of the Arietidae. Smithson. Contr. Knowl. 673, xi + 238 pp., 14 pis.

1900. Cephalopoda, pp. 502-604, in zittel, k. a. von. 1896-1900, Textbook of Palaeontology, transl.

Eastman, C. R., Macmillan, London and New York.

1903. Pseudoceratites of the Cretaceous. Monogr. LIS geol. Surv., 44, 1-351, 47 pis.

KENNEDY, w. J. 1984«. Systematic palaeontology and stratigraphic distribution of the ammonite faunas of
the French Coniacian. Spec. Pap. Palaeont. 31, 1 160, 33 pis.

1984fi. Ammonite faunas and the ‘standard zones’ of the Cenomanian to Maastrichtian Stages in their

type areas, with some proposals for the definition of stage boundaries by ammonites. Bull. geol. Soc.
Denmark, 33, 147 161.

1985. Ammonite faunas of the Coniacian, Santonian and Campanian stages in the Aquitaine Basin.

Geol. Mediterran. 10, 103-113.

1986fl. The ammonite fauna of the Calcaire a Bacidites (Upper Maastrichtian) of the Cotentin Peninsula

(Manche, France). Palaeontology, 29, 25-83, pis. 1-16.

19866. Campanian and Maastrichtian ammonites from northern Aquitaine, France. Spec. Pap. Paleont,

36, M45, 23 pis.

SUMMESBERGER, H. and KLINGER, H. c. 1981. On Ammonites serratomarginatiis Redtenbacher, 1873 and

Ammonites quinquenodosus Redtenbacher, 1873. Pp. 117-137 in Kennedy, w. j., klinger, h. c. and
SUMMESBERGER, H. 1973. Crctaceous faunas from Zululand and Natal, South Africa. Additional observations
on the ammonite subfamily Texanitinae Collignon, 1948. Ann. S. Afr. Mus. 86, 115-155.

and WRIGHT, c. w. 1979. Vascoceratid ammonites from the type Turonian. Palaeontology, 22, 665-683,

pis. 82-86.

1983. Ammonites polyopsis Dujardin, 1837 and the Cretaceous ammonite family Placenticeratidae

Hyatt, 1900. Ibid. 26, 855-873, pis. 85-87.

1985. Evolutionary patterns in late Cretaceous Ammonites. Spec. Pap. Palaeont. 33, 131-143.

KLINGER, H. c. and KENNEDY, w. L 1980. Cretaceous faunas from Zululand and Natal, South Africa. The
ammonite family Texanitinae Collignon, 1948. Ann. S. Afr. Mus. 80, 1-357.

KULLMANN, J. and wiEDMANN, J. 1970. Significance of sutures in phytogeny of Ammonoidea. Paleont. Contr.
Univ. Kansas, 44, 1-32.

KENNEDY: AMMONITES OF THE TYPE S ANTONI AN

781

MARTINEZ, R. 1982. Ammonoideos cretacicos del Prepirineo de la Provincia de Lleida. Piiblicaciones de
Geologia, Universidad Autonoma de Barcelona, 17, T 197, 30 pis.

MATSUMOTO, T. 1970. A Monograph of the Collignoniceratidae from Hokkaido. Part VI. Mem. Fac. Sci.
Kyushu Univ., ser. D., Geol. 20, 225-304, pis. 30 47.

MEEK, F. B. 1876fl. A report on the invertebrate Cretaceous and Tertiary fossils of the upper Missouri country.

hi HAYDEN, F. V. Report of the United States Geological Survey of the Territories, 9, lxiv + 629 pp., 45 pis.
ODIN, G. 1985. Concerning the numerical ages proposed for the Jurassic and Cretaceous geochronology.
Mem. geol. Soc. Land. 10, 196 198.

ORBIGNY, A. d’. 1840 1842. Paleontologie franqaise: Terrains cretaces. 1. Cephalopodes. 1 120 (1840); 121
430 (1841); 431 662 (1842), 148 + 3 pis. Masson, Paris.

1842-1843. Ibid. 2. Gastropodes, 1 224 (1842); 225-456 (1843), pis. 149-236. Masson, Paris.

— 1850. Prodrome de Paleontologie stratigraphicpie universelle des animaux molluscjues et rayonnes. 2, 428
pp., Masson, Paris.

1852. Cours Elementaire de Paleontologie et de Geologic Stratigraphupie 2, 383-847. Masson, Paris.

REDTENBACHER, A. 1873. Die Ceplialopodeiifauna der Gosauschichten im den nordostlichen Alpen. Ahh.
K.-K. geol. Reischsanst. 5, 91 140, pis. 22-30.

RIEDEL, L. 1931. Zur Stratigraphic und Faciesbildung im Oberemscher und Untersenon am Siidrande des
Beckens von Munster. Jh. preuss. geol. Landesanst Bergakad. 51, 605 713, pis. 72 79.

ROEMER, A. 1840 1841. Die Versteinerungen des norddeutschen Kreidegehirges. 145 pp., 16 pis. Hahn’schen
Hofbuchhandlung, Hannover. (1 48, pis. I 7, 1840; 49 145, pis. 8 16, 1841).

ROEMER, F. 1852. Die Kreidebildungen von Texas und Hire organischen Einschliisse. 100 pp., 10 pis. Adolph
Marcus, Bonn.

SCHLUTER, c. 1871 1876. Cephalopoden der oberen deutschen Kreide. Palaeontographica, 21, 1-24, pis. 1-8
(1871); 21, 25-120, pis. 9 35 (1872a); 24, 1 144 (121 264) + x, pis. 36 55 (1876).

SERONIE-VIVIEN, M. 1972. Contribution a Tetude du Senonien en Acjuitaine septentrionale. Ses stratotypes:
Coniacien, Santonien, Campanien., 195 pp., 16 pis. Editions du Centre National de la Recherche Scientifique,
Paris.

SNELLiNG, N. 1985. All interim time-scale. Mem. geol. Soc. Loud. 10, 261-265.

SOWERBY, J. 1812 1822. The Mineral Conchologv of Great Britain. 1, pis. 1-9 (1812), pis. 10 44 (1813), pis.
45 78 (1814), pis. 79 102 (1815); 2, pis. 103 114 (1815), pis. 115 150 (1816), pis. 151 186 (1817), pis. 187-
203 (1818); 3, pis. 204-221 (1818), pis. 222 253 (1819), pis. 254 271 (1820), pis. 272 306 (1821); 4, pis.
307 318 (1821), pis. 319 383 (1822). London.

SPATH, L. F. 1921. On Cretaceous Cephalopoda from Zululand. Ann. S. Afr. Mus. 12, 217 321, pis. 19 26.

1922fl. On Cretaceous Ammonoidea from Angola, collected by Professor J. W. Gregory, D.Sc., FRS.

Trans. R. Soc. Edinb. 53, 91 160, 4 pis.

19226. On the Senonian ammonite fauna of Pondoland. Trans. R. Soc. S. Afr. 10, 113 147, pis. 5-9.

1926. On new ammonites from the English Chalk. Geol. Mag. 63, 77-83, table.

1932. A monograph of the Ammonoidea of the Gault. Monogr. Palaeontogr. Soc. Part 9, 379-410, pis.

37-42.

SUMMESBERGER, H. 1979. Line obersantone Ammonitenfauna aus dem Becken von Gosau (Oberosterreich).
Ann. Naturhistor. Mus. IVien, 82, 109 176, 15 pis.

1980. Neue Ammoniten aus der Sandkalkbank der Hochmoosschichten (Obersanton; Gosau, Oster-

reichs). Ibid. 83, 275-283.

WEDEKIND, R. 1916. Uber Lobus, Suturallobus und Inzision. Zentbl. miner. Geol. Palaont. for 1916, 185-195.
WEGNER, T. 1905. Die Granulatenkreide des Westlichen Miinsterlandes. Z. dt. geol. Ges. 57, 112-232, pis.
7 10.

wiEDMANN, J. 1966. Stammesgeschichte und System den Posttriadischen Ammonoideen; ein Uberblick. Neues
Jb. Geol. Palaont. Abh. 125, 49-79, pis. 12; 127, 13-81, pis. 3 6.

1978. Line palaogeographisch interessante Ammonitenfaunula aus der alpinen Gosau. Eel. Geol. Helv.

71, 663-675, 2 pis.

WRIGHT, c. w. 1957. hi MOORE, R. c. (ed.). Treati.se on Invertebrate Paleontology. Part L, Mollusca 4,
Cephalopoda Ammonoidea. xxii + 490 pp. Geological Society of America and University of Kansas Press,
Boulder, Colorado and Lawrence, Kansas.

and WRIGHT, E. v. 1951. A survey of the fossil Cephalopoda of the Chalk of Great Britain. Monogr.

Palaeontogr. Soc. 41 pp.

YOUNG, K. 1963. Upper Cretaceous ammonites from the Gulf Coast of the United States. Univ. Tex. Bull.
6304, ix + 373 pp., 82 pis.

782

PALAEONTOLOGY, VOLUME 30

ZABORSKi, p. M. p. 1982. Campanian and Maastrichtian splienodiscid ammonites from southern Nigeria. Bull.
Br. Mus. nat. Hist. (Geol.) 36, 303-332.

ZITTEL, K. A. 1884. Hamlbiich der Palaeontologie . . . Abt. 1, 2 (Lief 3), Cephalopoda, 329-522. R. Oldenbourg,
Munich and Leipzig.

1895. Gnmdzuge der Palaeontologie (Palaeozoologie). vii + 972 pp. R. Oldenbourg, Munich and Leipzig.

Typescript received 20 May 1986

Revised typescript received 23 November 1986

W. J. KENNEDY

Geological Collections
University Museum
Parks Road, Oxford 0X1 3PW

SILURIAN MYODOCOPID OSTRACODES: THEIR
DEPOSITIONAL ENVIRONMENTS AND THE
ORIGIN OF THEIR SHELL MICROSTRUCTURES

by DAVID J. SIVETER, JEAN M. C. VANNIER and DOUGLAS PALMER

Abstract. European Silurian myodocopid ostracodes form part of an almost exclusively pelagic, recurring
faunal association. In general, they are facies related to typically laminated mud-silt or carbonate sequences
deposited in relatively quiet, low energy, probably outer shelf to shelf slope environments. Their palaeogeo-
graphical and facies distributions may in part be due to depth control.

Reticulation, corrugation, and tuberculation are recognized as the true, genetically controlled external shell
ornament in Silurian bolbozoid and 'cypridinid’ myodocopids. In contrast, enigmatic polygonal and associated
radiate shell microstructures, described from a wide variety of Silurian myodocopids, are interpreted as the
surface expression of calcium carbonate platelets and possibly the result of in vivo shell calcification and
post-mortem ‘diagenetic’ changes; they cannot, therefore, similarly be used in taxonomy.

A model based on analogues in Recent myodocopids and on possible variation in the pattern of nucleation
centres in the shell is proposed to explain the mode, variety of styles, intraspecific, and ontogenetic variation of
shell calcification in typical Silurian myodocopids. Pore canals may be the nucleation centres of calcium
carbonate. If centres are numerous and densely arranged, this may account for the formation of closely packed,
polygonal microstructural patterns on {and through) the shell. If nucleation centres are sparser and distant,
calcification possibly terminates before all platelets can coalesce, thus accounting for observed patterns of
isolated to clustered platelets of various larger sizes. Radiate (‘rosette’) type platelets are possibly produced
when post-mortem acicular recrystallization aflects various sized platelets. Smooth valves may represent newly
moulted individuals; patches of platelets might indicate poor or an intermediate stage of calcification; specimens
with densely packed platelets may have completed cuticle calcification.

The last thirty years have witnessed substantial research on several major faunas of Silurian
ostracodes (e.g. see Siveter 1978, 1980) but no monographic study has encompassed Silurian
representatives of the extant order Myodocopida. Though treated in part by Canavari (1900) and
Boucek ( 1 936), our primary taxonomic knowledge of the group in the Silurian still relies largely on the
work of Barrande ( 1 872) and Jones (1861, 1873, 1884). However, as Siveter (1984) has shown, Silurian
myodocopids are important and unique ecologically in possibly representing pioneer pelagic
ostracode stock and they have unfulfilled biostratigraphical and biogeographical potential. This
paper is our first of an intended series on British, Czechoslovakian, French, and Sardinian Silurian
myodocopids (see also Palmer and Gnoli 1985) which will test this potential. We address a problem
fundamental to classification within the group: what is the nature of their true external ornament and
what is the nature, origin (biological or ‘diagenetic’?), and significance of enigmatic microstructures
which we have recognized in the shell of many types of Silurian myodocopids? Essentially, which of
these characters can, and which cannot, be used taxonomically? In addition, we outline the
distribution, depositional setting, and palaeoecology of Silurian myodocopids.

Herein myodocopid refers to Silurian myodocopines of two broad taxonomic groups, bolbozoids
and ‘cypridinids’ (text-fig. 2). Entomozoid ostracodes (Silurian-Carboniferous) have traditionally
been placed alongside the bolbozoids (in the Entomozoacea; Sylvester-Bradley in Moore 1961), with
which they often occur in space and time in the Silurian. However, we would support a notion that the
‘entomozoid’ group (as currently, but erroneously, understood and used in the literature, e.g. forms
such as Entomis migrans) is not closely related to typical myodocopines.

Silurian myodoeopids are restricted mainly to finely laminated, fine-grained clastic sediments.

(Palaeontology, Vol. 30, Part 4, 1987, pp. 783-813, pis. 83-88.|

© The Palaeontological Association

784

PALAEONTOLOGY, VOLUME 30

where they often occur in substantial numbers, with a recurring association of graptolites, nautiloids,
and bivalves. They are typically preserved as mould faunas and partly for that reason have been so
neglected.

By examination of our abundant, newly collected Silurian myodocopid material from Wales,
France, and Czechoslovakia, and comparison with Recent myodocopids, it is concluded that the
microstructures described herein from the Silurian shells are probably manifestations of both the
calcification process and, to varying degrees, a diagenetic overprint. Such microstructures cannot,
therefore, be used as characters in taxonomy. In order to explain the variety of observed shell
microstructures in Silurian myodocopids, a model outlining possible modes of shell calcification in
such groups is proposed.

EUROPEAN DISTRIBUTION OF SILURIAN MYODOCOPIDS

Both their morphology and pattern of distribution suggest a pelagic mode of life for typical Silurian
myodocopids (Siveter 1984). European Silurian myodocopids are currently known from the
Wenlock-Pridoli time interval and occur in the Welsh Basin of Britain, the Armorican Massif and
Montagne-Noire of France, the Barrandian basin of Czechoslovakia, Poland, Sardinia, Portugal, and
Spain (text-fig. 1). Although the range of facies within which they occur in these areas is somewhat
restricted, this is an otherwise widespread geographic distribution, and contrasts with a generally
more restricted distribution known from other Silurian ostracode assemblages. For example, Silurian
ostracode faunas from Czechoslovakia (Boucek 1936) are dominated by non-palaeocopes and are
substantially dififerent in overall composition from contemporaneous, beyrichiacean rich faunas of
northern Europe (e.g. Britain, Baltoscandia) and eastern North America (Siveter MS). In this case,
though the barrier of the incipient ‘mid-European’ Rheic Ocean was perhaps hindering the dispersal
of some invertebrate groups such as benthic palaeocopes and non-palaeocope ostracodes, the same
was certainly not true of myodocopid ostracode species. Moreover, some Silurian myodocopids also
occur in North Africa (unpublished information) and, for example, Australia (Jones 1861, 1873, 1884;
De Koninck 1876; Strusz, pers. comm.). Because of this ‘great extension of this old creature’s habitat
. . .’ Jones (1884, p. 393) suggested a ‘truely pelagic nature’ for some "Ent onus’ myodocopids.

Localities and material. The material studied amounts to several thousand specimens from the following areas
and localities. Information on the localities of all our collections will be given in later papers.

Armorican Massif, France. Localities in Ille-et-Vilaine, Mayenne, Sarthe, and Finistere. Their ages are
determined typically using graptolites or chitinozoans, but are generally imprecisely known. Current studies on
new graptolite collections may provide more finely resolved age determinations.

1. ‘Les Buhardieres’; exposure along D104 road between Saint-Germain-le-Fouilloux and Andouille,
Mayenne; Laval Synclinorium.

2. ‘La Cultais’; outcrops near ‘La Cultais’ farm, Vieux-Vy-sur-Couesnon, Ille-et-Vilaine; Menez-Belair
Synclinorium. See Paris 1977, 1981.

Localities 1 and 2 belong to the upper part of La Lande Muree Formation (Paris 1981 and pers. comm.). The
sediments (‘ampelites’) are grey to black, finely laminated shales. Trace and major element geochemical studies
show that, as with most Silurian black shales, these sediments are highly organic- and Vanadium-rich (Dabard
1983; Dabard and Paris 1986). The ostracodes occur as abundant, slightly tectonically deformed internal and
external moulds. Associated fossils include graptolites, small, thin-shelled bivalves, fragments of orthoconic
nautiloids and disarticulated eurypterid, and phyllocarid crustacean ("Ceratiocaris') remains.

The cleavage prevents accurate determination of the graptolites at locality 1 but there are clearly numerous
pristiograptids of the P. duhius type, some Monoclimacis cf. Iiaupti, and long distal fragments of slender
monograptids with simple, straight thecae and unelaborated thecal apertures of the nilssoni! progenitor type.
Until the proximal ends are found, confident identification cannot be made and the possible stratigraphic range
of this locality extends through a large part of the Ludlow (D. Palmer). Graptolites of locality 2 indicate an upper
Ludlow to lower Pridoli age (Jaeger, in Paris 1981).

3. ‘Le Moulin de Regereau’; exposure along the road between Origne and I’Huisserie, Mayenne; Laval
Synclinorium. See Peneau 1936; Paris 1981. The sediments consist of siltstones and mudstones of assumed

SIVETER ET AL.: SILURIAN MYODOCOPID OSTRACODES

785

TEXT-FIG. 1. Distribution of the main faunas of Silurian myodocopid ostracodes in Europe. 1, Welsh Basin
(Siveter 1984; Siveter and Palmer, in prep.). 2, Poland (Giirich 1896). 3, Bohemia (Barrande 1872; Boucek 1936;
Siveter and Palmer, in prep.). 4, Armorican Massif (Siveter, Vannier and Palmer, in prep.). 5, Montagne-Noire
(Chaubet 1937). 6, Portugal (unpublished information). 7, Spain (unpublished information). 8, Sardinia (Canavari

1900).

Pridoli age (Paris 1981). Rare ostracodes are associated with bivalves, the crinoid ^Scyphocrinites, and fragments
of 'Ceratiocaris\ all decalcified and preserved as moulds.

4. Chemire-en-Charnie, Sarthe; Laval Syncliorium. Siliceous nodules collected by Guillier (Guillier and de
Tromelin 1874; Guillier 1886) and deposited at the Musee d’Histoire Naturelle de Nantes, Loire-Atlantique.

5. Saint-Denis-d’Orques; outcrop at the intersection of the Laval to Le Mans motorway and the
Saint-Denis-d’Orques to Brulon road, Sarthe; Laval Synclinorium.

Localities 4 and 5 consist of siltstones and mudstones containing numerous, partly pyritized nodules in which
the ostracodes (myodocopids and palaeocopes) occur as normally well preserved and undeformed moulds of
single valves or closed carapaces. The diverse fauna includes crinoids. fragments of the phyllocarids Ceratiocaris
and Schugurocaris(\. Rolfe, pers. comm.), the rhombiferan echinoderm Lopliotocystis(i. Chauvel, pers. comm.),
and graptolites of Ludlow or Pridoli age (D. Palmer).

786

PALAEONTOLOGY, VOLUME 30

Welsh Basin. We have collected Silurian myodocopids from many localities in this region, mainly in the Long
Mountain district but also from the ‘Radnor’ and ‘Denbigh’ areas. Material studied specifically for the present
paper comes from just two localities (1, 2).

1. Cause Castle Farm (SJ 341 077), Long Mountain, Shropshire.

This is one of about 100 localities in the Long Mountain yielding infrequent myodocopid ostracodes, which
generally occur scattered throughout the laminated muddy siltstone facies of the Ludlow Long Mountain
Siltstone Formation and succeeding Causemountain Formation (Palmer 1970). At this locality (=LMS
Formation) some silt laminae are crowded with the moulds of finely comminuted shell debris, plus a few small but
entire shells (e.g. lingulids), crinoid columnals, and eurypterid and phyllocarid fragments. Associated are
occasional large myodocopid valves and numerous graptolites, Saetograptus leintwardinensis, the index for the
eponymous, basal Ludfordian zone.

2. Friends Meeting House Quarry (SO 138 641), Llandegley, Powys.

Apart from the Long Mountain, the ‘Radnorshire’ area (now in Powys) is the only part of the Welsh Basin to
have yielded a large number of myodocopid bearing localities, some twenty-two being cited by Kirk in
her unpublished thesis (1947). Most occur in her ‘Striped Flag Series’, which consists of finely laminated and
banded or ‘striped’ admixtures of mud and silt, generally with planar lamination or small scale cross-
stratification. The characteristic, autochthonous fauna of the Striped Flags consists of graptolites, orthocones,
and bivalves. Minor associates include myodocopids, rare crinoids, phyllocarids, and small brachiopods such as
Aegeria gray! and Lingula lata. Intercalated south to south-east derived slump sheets often introduce an
allochthonous, shelf shelly fauna into the area. The Striped Flags encompass the Ludlow N. nilssoni, L. progenitor
(see Palmer 1970), L. scanicus s.l., and P. tumescens/S. chimaera graptolite zones. However, Kirk also records rare
myodocopids in strata as old as the "vulgaris'' ( = ludensis) Zone and as young as the S. leintwardinensis Zone, and
immediately above. Thus the total range for myodocopids in ‘Radnorshire’ is essentially the same as in the Long
Mountain. Friends Meeting House Quarry comprises typical ‘Striped Flags’ lithology and mould fauna, and
lower Ludlow, C. scanicus Zone, graptolites including C. scanicus s.l., S. varians, S. chimaera semispinosus,
P. dubius, P. tumescens, and B. bohemicus (D. Palmer).

The geological survey of the ‘Denbigh’ area, Clwyd, resulted in finds of a few scattered Silurian myodocopid
localities in North Wales (Warren et ai. 1984). Here, myodocopids appear to be relatively very rare, but their
stratigraphic distribution is broadly as in the Long Mountain, specimens being scarce in the upper Wenlock
(C. lundgreni Zone), somewhat commoner in the Ludlow (TV. nilssoni to C. scanicus zones) and again less common
higher in the Ludlow (up to and including the S. leintwardinensis incipiens Zone; younger Silurian strata are not
preserved). Both their facies and faunal associations are the same as elsewhere in the Welsh Basin. The
myodocopids occur mainly in the ‘ribbon banded mudstone’ (Warren et al. 1984, pp. 44-46) of the Nantglyn
Flags Group (straddling the Wenlock-Ludlow boundary) and the higher Elwy Group (Ludlow), a facies
virtually identical to the ‘Striped Flags’ of the Radnor area. Its faunas are dominated by graptolites accompanied
by orthoconic nautiloids and bivalves (particularly cardiolids) and also include crinoids (especially scypho-
crinitids; Warren et al. 1984, p. 44) which are sometimes preserved complete and occasionally with orthocone
‘substrates’.

Czechoslovakia. Silurian myodocopids in Czechoslovakia are predominantly of Ludlow age. Some of Barrande’s
original localities were either small and isolated from the main outcrop, and are now unidentifiable, or were
poorly localized within thick sections so that they are now also unavailable. However, modern investigations by
Czech palaeontologists have resulted in new exposures at the same horizons, yielding faunas similar to those of
the ‘lost’ localities. For example, Barrande’s type locality for "Bolbozoe" bohemica can no longer be located at
Vyskocilka, but this species can still be obtained (DJS) from a similar horizon (C. scanicus Zone) on Holy Vrch
Hill at Lounin near Zdice. Here, as in other Czechoslovakian Ludlow localities, the myodocopids are often
abundant as moulds or as degraded carbonate shells in thin, platy, planar laminated, muddy limestones and are
accompanied by orthoconic nautiloids, graptolites, and occasional bivalves.

Preservation. The vast majority of myodocopids we have studied are preserved as moulds. This may
be as much the result of their occurrence in particular sedimentary environments or lithologies
(fine-grained elastics), where virtually all the calcareous-shelled fossils have been dissolved, and/or as
a product of the thinness of their shells. Exceptions are found where myodocopids occur in limestones
(e.g. Sardinia) and then, although recrystallized, shells can be preserved as carbonate; however, we
have not yet studied in detail this important mode of preservation.

SIVETER ET AL.\ SILURIAN MYODOCOPID OSTRACODES

787

Sedimentary matrix. That the types of deposit in which European Silurian myodocopids have so far
been found are rather restricted could in part be preservational due to their relative fragility. The
curiously impersistent stratigraphic record of the group since Silurian times supports this contention.
Nevertheless, the limited facies pattern of occurrence of our Silurian species requires further
explanation, especially since the group in general is considered to be pelagic.

The sediments in which our Silurian myodocopids are found are characteristically laminated and
lack any indication of bioturbation or hardly any of the contemporaneous representatives of the
typical brachiopod-dominated, shelly (shelf) faunas. Sedimentologically, they occur in a three
component system of muds, muddy silts, and carbonate. A main feature is the fine (silt or smaller)
grain size of the terrigenous material compared with the larger size of much of the carbonate ‘clast’
material, which is essentially made up of whole fossils and thus derived from within the depositional
environment. Most of the myodocopid-bearing deposits consist of closely spaced planar laminations
which result from either alternations of ‘clean’ silt laminae with muddy silts or, muddy silts with dark
hemipelagic muds. Less commonly the latter may predominate, as in the dark grey (originally black?;
but now chloritized) shales of Brittany. Because of the fine grain size, the moulds of the Brittany
myodocopids preserve exquisite details of shell structure. Increase in the bioclastic component of the
sediment in the Bohemian and Sardinian myodocopid-bearing strata tends to diminish but not
destroy lamination, as the fossils themselves are commonly arranged as parallel orientated, bedding
plane assemblages. This results in a coarse ‘grain’ parting lineation (see Gnoli et al. 1979, figs. 2
and 9a).

The bioclastic component varies greatly in volume and grain size, sometimes forming only a small
proportion of the sediment ( 1 % volume in some Welsh Borderland localities), especially where muddy
silts predominate and there are only very rare macrofossils which are almost invariably preserved as
moulds. In the middle of the range are the highly fossiliferous, but now decalcified, hemipelagic muds
of Brittany and the similarly fossiliferous, fine grained, laminated, and shaley micrites (microsparites)
typical of the Bohemian localities. In contrast, myodocopid-bearing coquinoid ‘cephalopod’
limestones of Sardinia have a small terrigenous clastic component (5 %; Garuti in Gnoli et al. 1979)
and consist largely of complete shells of all sizes in a micrite matrix.

Certain sedimentary characteristics appear common to almost all the myodocopid-bearing
deposits of this study, thus suggesting similar underlying parameters in their mode of deposition. In
particular, the fineness of the non-organic elastics and the prevalence of close spaced, planar laminae
seem to indicate a relatively ‘quiet’ but not static depositional environment (except perhaps as
represented by the black shales of Brittany) debarred from both the massive turbiditic input of
terrigenous material often associated with deep basins and from the rigours of high energy and coarse
(tidal) elastics of shallow shelves. Most of the inorganic sediment (muddy silt) was seemingly deposited
as turbid flows of low velocity and of no erosive power, which were frequently and intermittently
introduced into areas experiencing a low level ‘rain’ of hemipelagic mud. Gentle bottom traction
currents reworked the surface layers of the ‘homogenised’, structureless muddy silt layers to form
winnowed, often micaceous, ‘clean’ silt laminae. The overall result was to produce the ‘Striped Flag’ or
‘ribbon banded mudstone’ type of lithology, with its alternations of muds, muddy silts, and silts.
Currents were of the appropriate strength to orientate fossils (e.g. orthocones and graptolites in the
Welsh Basin localities) without any size sorting (e.g. of bivalved organisms) or much abrasion or
fracture (ostracodes and bivalves may be articulated), and sometimes to produce a parting
lineation or small scale, wispy cross-lamination but no other depositional or erosional structures.
In Bohemia, for example, it is common for the orthocones and graptolites to show parallel
orientation consistently throughout several centimetres of strata, indicating bottom currents of
fairly persistent velocity and direction over a considerable timespan (Petranek and Komarkova
1953; Turek 1983).

Although the volume of terrigenous silt and mud is relatively low in Bohemia and virtually absent
in Sardinia, there is similar evidence of both having been depositional areas again debarred from both
deeper water turbiditic and shallow shelf clastic depositional processes. The carbonate that is present
shows no reefal or other shallow water structures. Indeed, the dominant bioclastic components can

788

PALAEONTOLOGY, VOLUME 30

be diagnosed quite differently (see below) and their presence is possibly the result of both a primary
lack of sediment to ‘dilute’ the accumulating shelly remains and secondary accumulation by
currents.

Numerical occurrence and faunal associates of myodocopids. The frequency of occurrence of European
Silurian myodocopids is generally very low and is partly why they have received so little attention.
Moreover, they tend to occur in comparatively low diversity faunas, which lack the more commonly
studied members of the Silurian shelf benthos. Some idea of their frequency can be gained from studies
in the Long Mountain, a sequence more thoroughly investigated (by D. P.) in terms of its myodocopid
and associated fauna than any other Silurian myodocopid-bearing area. Their distribution
throughout the laminated siltstones of some 1 500 fossiliferous localities of Ludlow age is diffuse, only
about 7% yielding myodocopids, which generally comprise less than 10% of the fauna of each
locality. That a previous investigator of the Long Mountain sequence (Das Gupta 1932) did not
record any myodocopids, also reflects their relative scarcity.

Whilst the basis for comparing the various localities and areas (text-fig. 1) arises from the presence
of myodocopids, we also recognize a number of features in common in the associated faunas, although
the taxa may differ even at the family level and different members may dominate in particular
localities. We are essentially dealing with a known but little studied and relatively low diversity
graptolite-orthocone-bivalve faunal association, whose minor members include various arthro-
pods (ostracodes, phyllocarids, eurypterids, and rare, small trilobites) echinoderms (mainly
crinoids, small pisocrinids, and scyphocrinitids), and a few brachiopod species (small lingulids
and chonetids).

In Welsh Borderland myodocopid-bearing localities, graptolites and occasional small bivalves
tend to dominate the faunas, with orthocones or myodocopids next most abundant. All other
associated organisms tend to be very rare. The fauna from Cause Castle Farm, Long Mountain, is
graptolite dominated, with the associated myodocopids, lingulids, phyllocarids, and crinoid
columnals being current assembled on clean silt laminae. Generally in this area of the Long
Mountain, graptolites and occasional small pterineid bivalves (up to 400 individuals/m^) pre-
dominate. Myodocopids are slightly more abundant at Friends Meeting House Quarry, where even
so, they comprise only about 10% of a generally sparse, graptolite dominated fauna and are, once
more, scattered at very low densities throughout 3-4 m of laminated muddy silts. Here, fossils tend to
be more abundant in the hemipelagic muds, especially on certain bedding planes. They are also
present, but much less common, scattered throughout the muddy silts, their concentration being
‘diluted’ by the frequent incursions of sediment. Higher frequencies may again be encountered on the
winnowed silt laminae, where they may have been concentrated by the removal of the mud or by
current transport.

Compared with these Welsh occurrences, myodocopids can be much more abundant in the few
localities in Bohemia and Brittany that have yielded them since the pioneer work of Barrande (1872),
Barrois (1886), and Kerforne (1901). In the platy limestones of Bohemia graptolites and orthocones
are common but faunas tend to be predominantly myodocopids, with some bedding plane
assemblages achieving densities of 400 myodocopid individuals/m^. Even greater abundances occur
in the black shales of Brittany, where densities of 2000/m^ or more are common and where, at some of
the cited localities, ostracodes may dominate in comprising over 90% of individual assemblages.
However, myodocopine s.s. valves are not always dominant, since the same bedding planes in
Brittany may also have greater numbers of a smaller, thin shelled (and consequently poorly preserved)
entomozoid ostracode. In the coquinoid (cephalopod) limestones of Sardinia orthoconic nautiloids
tend to be the dominant group. However, some limestone blocks contain graptolite dominated
assemblages and/or abundant entomozoid myodocopids (see Palmer and Gnoli 1985 for geological
background). The accompanying faunas are slightly more diverse than elsewhere, with taxonomically
richer assemblages of bivalves and phyllocarids (Gnoli and Serpagli 1984). Again, the orthocones and
graptolites show parallel orientation (Gnoli et al. 1979) and the vagaries of faunal composition may
reflect fluctuations in bottom traction current strength.

PALAEONTOLOGY, VOLUME 30

789

Mode of life. European Silurian myodocopids are clearly part (albeit often a minor one) of a recurring
faunal association generally dominated by organisms which are usually considered to be pelagic,
either as virtually passive drifters (the planktonic graptolites and scyphocrinitids), or as active
swimmers (the nektonic cephalopods, myodocopids, phyllocarids, eurypterids, and small spinose
trilobites) at various levels within the water column. Only the bivalves and rare, small brachiopods
appear to have been benthic, although doubts have been raised over such an interpretation for the
mode of life of representatives of both groups (Watkins 1978, pp. 51-52; Kfiz 1969, pp. 1 1 1 - 1 14; Nye ct
al. 1975). The main bivalve groups present are praecardiaceans sensu Newell (1969) and pterineids.
The praecardiaceans are largely endobyssate suspension feeders and constitute the only possible
(semi-) infaunal representatives. The pterineids are epibyssate suspension feeders (Kriz 1984).

The occasional presence of abundant small pterineids is significant since they may have been the
only genuine, flourishing, autochthonous shelled benthos in areas where myodocopids are found.
Their abundance in all growth sizes, often with both valves, indicates that their associated bottom
conditions were unlikely to have been foul and toxic. Since they also occur in finely laminated muddy
silts with no indication of bioturbation, the near surface sediments may have been deficient in oxygen,
sufficient to prevent the establishment of an infauna. This possibility is supported by the presence of
disseminated, fine-grained, framboidal pyrite within the muddy silts. The few other benthic fossils,
found on clean silt laminae, are representatives of some of the most common epibenthic brachiopod
species characteristic of the shelf They are clearly allochthonous, transported individuals.

Epizoans are very rare, consisting mainly of spirorbids, bryozoans, and small crinoids attached to
orthoconic nautiloids in the Welsh material, and spirorbids attached to phyllocarids in some of the
French material. Since all surfaces of the orthocones tend to be colonized, it seems that colonization
happened either in vivo or post-mortem, but whilst the shell of the host was still buoyant, rather than
after it lay on the substrate.

Environments of deposition. All the evidence suggests that the relevant depositional areas were not
within normal epicontinental shelf water zones, but were more likely to have been in either outer shelf
to shelf margin or even shelf slope environments. Of the areas considered, the Welsh basin is perhaps
best known and understood, with its ‘Striped Flag’ and ‘ribbon banded mudstone’ lithologies
representing sedimentation against a background deposition of dark laminated hemipelagic mud.
This ‘tranquil’ depositional environment was frequently interrupted by incursions of turbid mud-silt
suspension flows, depositing occasionally graded but otherwise structureless thin muddy silt layers
(1-5 cm thick). The substrate surface was also reworked by low velocity bottom currents, which
winnowed out the fines and left thin, ‘clean’ silt laminae and shell lags, with the elongate fossils
becoming parallel orientated. The thinness and high frequency of repetition of these sediment layers is
suflrcient for them to be compared with present-day marine, ‘varve’-like (non-glacial) deposits. The
sedimentation model (Thornton 1984, p. 388 et seq.) for the latter has hemipelagic muds settling out
during the summers, the organic content being heightened by algal blooms. During winters, increased
river run off leads to an increased supply of terrigenous material of which the fines (mud and silt) are
carried out across the shelf in turbid suspension by storms. The resulting deposits are the thin,
structureless or graded mud-silt layers.

In the Long Mountain area, the frequent colonization of the sea floor only by abundant small
pterineid bivalves suggests that here Ludlow water depths were too great for the brachiopod
dominated faunas of the shelf. The viability of the bivalve faunas also shows that bottom conditions
were at times neither anoxic nor too turbid for suspension feeders to flourish. Yet (see above), oxygen
depletion may have occurred within and virtually up to the sediment-water interface. In ‘Radnor’ and
‘Denbigh’ the intercalated slumps in the ‘Striped Flags’ suggest an intermittently unstable slope
environment.

Compared with the Welsh area, Brittany includes what almost certainly was a stagnant anoxic
basin with only hemipelagic mud deposition and a planktonektonic fauna. The ostracode rich
Ludlow limestones of Bohemia are in many ways similar to the Welsh sediments but with a lesser
proportion of terrigenously derived silt. This fraction is replaced by fine-grained bioclastic carbonate

790

PALAEONTOLOGY, VOLUME 30

derived from the organically rich carbonate shelf. Recent work on the implosion of cephalopod shells
from this area (Westermann 1985) indicates outer shelf depths ( > 300 m) for these deposits. Certainly
there is again neither infauna nor shallow shelf epibenthos.

The Sardinian limestones differ in that they are almost wholly bioclastic carbonate (often entire
shells, not size-sorted) with a minute proportion of hemipelagic mud and minimal terrigenous
material. For the depths of deposition of the limestones Jaeger (1976) suggests deep water within
Palaeotethys whilst Gnoli et al. (1979) prefer shallower shelf conditions but with anoxic bottom
waters. Westermann’s criteria have yet to be applied to the orthoconic nautiloids of this area.

Geographical distribution. The apparent contradiction posed by the supposed pelagic mode of life and
the somewhat restricted geographical distribution of Silurian myodocopids exists not only in terms of
their facies occurrence, but also in a wider palaeogeographical context. Contemporaneous shelf
sediments in, for example, the Welsh Borderland of the Welsh Basin have probably been as closely
examined as any sediments for their fossil content and yet while almost no myodocopids have been
recorded other representatives of the faunal association are known. Furthermore, myodocopids are
unknown from the Silurian rocks of Scandinavia, the USSR, or North America. Apart from France,
England, Czechoslovakia, and Sardinia, myodocopids are also known (but unstudied) from Poland,
Spain, and Portugal (text-fig. 1), Morocco, Australia, and possibly China. Clearly the overall
distribution of Silurian myodocopids is enigmatic and needs explanation based on further studies. It is
possible, at least in the case of their best-known (narrow) belt of occurrence encompassing western
Britain, Brittany, Bohemia, and Sardinia, that myodocopids had some depth control or preferred
vertical range (e.g. 200-500 m) that resulted in their remains being restricted to marginal shelf
sediments. A variation on this would arise if they were nekto-benthic, as envisaged for longicone
nautiloids with strong septa (Westermann 1985, p. 91) and again restricted to substrates at water
depths of about 200-500 m.

MORPHOLOGY OF SILURIAN MYODOCOPID OSTRACODES

All of the Silurian myodocopids treated in detail and figured herein, from France and Britain, are
based on mould material. Casts of external moulds were prepared with Silicone rubber using the
technique described by Siveter (1982).

Pending current studies on their systematics, most of the Silurian myodocopids of this paper are
cited under open nomenclature. However, in interpreting the microstructural and sculptural features
of their shells it is necessary to outline features of their carapace morphology. The Silurian
myodocopids discussed herein fall into two broad taxonomic groups, bolbozoids and ‘cypridinids’
(text-fig. 2).

EXPLANATION OF PLATE 83

Examples of true ornament in Silurian myodocopid ostracodes.

Figs. 1, 3, 5, 7, 8. Bolbozoid gen. et sp. nov. A. Right valve, IGR 33040, La Lande-Muree Formation, Les
Buhardieres (Andouille, Mayenne), Armorican Massif, France. 1, lateral view, x 18. 3, detail of tubercles
with pores, posteroventral part of valve, x 180. 5, single pores and (?) double pore-like structures in the
anteroventral part of the valve, x 150. 7, detail of (?) double pore-like structure (from 5), x 505. 8, detail of
single pore (from 5), x 505.

Figs. 2, 4, 6. 'Bolhozoe' sp. nov. A. Left valve, IGR 33048, La Lande-Muree Formation, La Cultais
(Vieux-Vy-sur-Couesnon, Ille-et-Vilaine), Armorican Massif, France. 2, lateral view, x 36. 4, detail of
corrugated surface on either side of posterior sulcus, x 96. 6, detail of ventral part of the valve showing
corrugations surmounted by four tubercles, x 165.

All specimens are silicone rubber casts from external moulds. All SEM except fig. 2. IGR = Institut de Geologie
de rUniversite de Rennes.

PLATE 83

SIVETER, VANNIER and PALMER, Silurian myodocopid ostracodes

792

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 2. Morphological features of
two groups of Silurian myodocopid
ostracodes; a-d, a bolbozoid; e-h, a
‘cypridinid’. a, e, right lateral views;
B, F, dorsal views of right valves; c, G,
ventral views of right valves; d, h,
anterior views of carapaces.

Many of the Silurian myodocopids show plastic deformation of the valves and clear adductor
muscle scar impressions externally on their valves, both of which suggest the occurrence of a rather
thin, flexible but flimsy shell (Siveter 1984). Taken together with their geographical distribution and
the fact that many have an anterior rostrum and rostral incisure (suggesting the presence of
protruding, possible swimming appendages, e.g. see Kornicker 1975; fig. 141u, h), many Silurian
myodocopids such as bolbozoids can be reasonably interpreted as pelagic ostracodes (Siveter 1984).

Bolbozoids (text-fig. 2a-d). These unusually large ostracodes are characterized by having a huge
anterodorsal bulb delimited posteriorly and ventrally by a deep adductorial sulcus (e.g. PI. 83, figs. 1
and 2; PI. 84, figs. 2, 6, 7; PI. 85, fig. 3; PI. 86, figs. 1, 2, 5, 6), and are assigned to the Family Bolbozoidae
(e.g. Siveter 1984). A sigmoidal, posterior sulcus (e.g. PI. 84, figs. 2 and 7) and a caudal process (e.g.
PI. 83, fig. 2) occur in some forms. Undeformed, non-flattened material from the Armorican Massif
and Bohemia confirms that the lobal surface of most bolbozoids is quite strongly convex. The function
of the anterodorsal bulb, which is often larger and more hemispherical in instars, is speculative.
Bearing in mind its position, that it often shows plastic deformation which possibly indicates a thin
shell, and the fact that except for its ventral-most region it is consistently the only part of the lobal
surface to lack ornament in several types of bolbozoids (see text-fig. 5), it is possible that the bulb
on each valve was the site of some form of visual organ. Silurian bolbozoid and ‘cypridinid’
myodocopids clearly exhibit a prominent anterior rostrum and rostral incisure (identified in
three-dimensional, closed carapaces; see text-fig. 2) like those of Recent myodocopids (cf. PI. 84, figs. 1
and 2 with PI. 88, figs. 1, 3, 5-8).

A radiate adductor muscle scar pattern, consisting of a series of narrow, closely spaced, alternating

SIVETER ET AL.: SILURIAN MYODOCOPID OSTRACODES

793

ridges and grooves near the base of the vertical part of the adductorial sulcus, occurs in many Silurian
bolbozoids. In some the pattern is simple, consisting of a few ridges and grooves of equal width,
converging toward the bulb (e.g. PI. 83, fig. 1). Others have a more complex, feather-like arrangement,
composed of two unequal series of radiate ridges and grooves converging on a narrow, median
(external) ridge curved parallel to the margins of the adductorial sulcus (e.g. PI. 84, figs. 2 and 8; PI. 87,
fig. 6; text-fig. 3; and Gurich 1 896, pi. 1 5, fig. 12, for Bolhozoe polonica). A secondary ‘crenulation’ of the
ridges in some specimens is of possible comparable functional significance (oflfering increased
attachment area/bonding) to that of the intricate pattern of micro-attachment points of the adductor
muscle scars in some Recent cylindroleberidid myodocopids (see Kornicker 1975, fig. 265c-e).

In its biserial arrangement this distinctive muscle scar recalls that of platycopes such as cytherellids
(see Van Morkhoven 1962, fig. 40) and, more pertinently, shares strong morphological similarities
with Carboniferous entomoconchacean myodocopids (see Van Morkoven 1962, fig. 38; Kornicker
and Sohn 1976, fig. 35) and particularly with Recent, oceanic thaumatocypridacean myodocopids
(e.g. see Thaumatoconcha caraionae and T. elongata in Kornicker and Sohn 1976, figs. 35 and 44). In
this context it is also taxonomically significant to note that Kornicker and Sohn (1976, p. 4, fig. 2)
maintain that, based on cladistic analysis of the soft parts (eye; copulatory organ; appendages) of
Recent myodocopids, entomoconchaceans and thaumatocypridaceans are not closely related to
cypridinacean myodocopids.

"Cypric/iiiids' (text-fig. 2e-h). One group of Silurian myodocopids have simple, oval, dome-like shells
and differ considerably from bolbozoids in muscle scar pattern, in being relatively shorter and higher,
and in lacking sulci or an anterodorsal bulb (see PI. 84, fig. 1; PI. 85, fig. 1; PI. 87, figs. 1 and 2). The
general designation ‘cypridinid’ is employed herein for these forms. Their shape and outline is
comparable with Devonian and Carboniferous cypridinids (Bless 1973; Sohn 1977), Mesozoic
myodocopids such as Triadocypris (Weitschat 1983), and particularly with the Recent cypridinacean
families Cypridinidae, Philomedidae, and Cylindroleberididae (see PI. 88; Kornicker 1975, 1981;
Kornicker and Caraion 1978).

dorsal

lobal

surface

bulb

adductorial

sulcus

TEXT-FIG. 3. Schematic representation of adductor muscle scar of a Silurian bolbozoid
ostracode (see PI. 84, figs. 6 and 8); dorsal oblique view. Arrow indicates anterior.

794

PALAEONTOLOGY, VOLUME 30

TRUE ORNAMENT

Three types of external surface ornament have been recognized in Silurian myodocopids treated
herein: reticulation, corrugation, and tuberculation (text-fig. 4). Similar ornament is also known in
many fossil and Recent ostracode groups (see Sylvester-Bradley and Benson 1971 for details and
terminology).

Reticulation (text-fig. 4a). Reticulation is frequently developed in bolbozoids, often occurring on all
surfaces of the valve except for the upper part of the anterodorsal bulb, the rostrum, and the
adductorial and posterior sulci (e.g. PI. 84, fig. 7; text-fig. 5e, f). The reticulation consists of coarsely
elliptical, ovoid, polygonal or, more rarely reniform fossae, each 200-600 /im across in adults. The
deepest and widest fossae are normally in the ventral part of the valve (PI. 84, fig. 7); medially, fossae
often have an elliptical, elongate, or even sigmoidal shape ( PI. 84, figs. 4 and 7). The reticulum of young
instars appears relatively less strongly developed and in some mature instars affects the whole shell
thickness and is hence reflected on the internal mould. Reticulation also occurs in some Silurian
‘cypridinids’; in ‘Cypridinid’ gen. et sp. nov. A a regular pattern of shallow, elliptical to polygonal
fossae (each 100-300 /zm across) covers the posterior half of the valve (PI. 84, figs. 1 and 3).

Similar external reticulation is known from Ordovician to Recent in a wide variety of ostracode
groups such as binodicopes (see Schallreuter 1980, 1983), palaeocopes (see Schallreuter 1982u, b),
podocopes (see Benson 1971, 1973), and Recent myodocopids. The cypridinid Scleroconcha flexilis
(Brady, 1898) and the philomenid Anarthron chilensis (Hartmann, 1965) are just two Recent
myodocopids (Kornicker 1975, figs. 196 and 224) which illustrate similar reticulation to those of
Silurian bolbozoids such as "Bolbozoe' cf. bohemica Barrande, 1872 (PI. 84, fig. 7; PI. 87, fig. 6).

As observed in podocopids, the pattern and form of external (polygonal) reticulation in ostracodes
is closely determined by the arrangement of the underlying layer of individual epidermal cells (Okada
1981, 1982). The ‘individual cell: polygon’ correlation and the process of cell division were clearly
demonstrated using growth series of Recent and fossil ostracodes. From the size and ontogenetie
changes of their fossae, it is possible that the development of the reticulum was similarly controlled in
Silurian myodocopids.

Corrugation (text-fig. 4b). A series of sigmoidal to sinuous and occasionally bifurcated grooves and
ridges occurs in Silurian bolbozoids (e.g. PI. 83, figs. 2, 4, 6) and ‘cypridinids’ (unpublished material
from Moroeeo). Corrugation typically occurs throughout the lateral surface except on the sulei,
rostrum, and upper part of the bulb (e.g. PI. 83, figs. 2, 4, 6). Sometimes it seems best developed
posteriorly and ventrally (e.g. PI. 84, figs. 2 and 4). Juvenile bolbozoids have a relatively incomplete
pattern of corrugation; ridges are shorter and less densely packed.

EXPLANATION OF PLATE 84

Examples of true ornament and adductor muscle scar patterns in Silurian myodocopid ostracodes.

Figs. 1, 3, 5. ‘Cypridinid’ gen. et sp. nov. A. Left valve IGR 33035, La Lande-Muree Formation, Les Buhardieres
(Andouille, Mayenne), Armorican Massif, France. 1, lateral view, x51. 3, detail of posterior surface
reticulation, x 105. 5, reniform adductor muscle scar (and pattern of shell microstructures), x 225.

Figs. 2 and 4. 'Bolbozoe' sp. nov. A. Right valve BM OS 13060. Long Mountain Siltstone Formation, Ludlow
Series, Cause Castle Farm, Long Mountain, Powys, Wales. 2, lateral view, x 18. 4, detail of ventral part of the
valve showing a combination of tubercles, corrugation, and reticulation, x 45.

Figs. 6-8. ‘R’ cf. bohemica Barrande, 1872. 6 and 8, right valve, IGR 33067, La Lande-Muree Formation, La
Cultais (Vieux-Vy-sur-Couesnon, Ille-et-Vilaine), Armorican Massif, France. 6, lateral view, x 27. 8, radiate
adductor muscle scar, dorsal oblique view, x 105. 7, lateral view of open carapace, IGR 33057, La
Lande-Muree Formation, La Cultais (Vieux-Vy-sur-Couesnon, Ille-et-Vilaine), Armorican Massif, France,
x27.

All specimens are silicone rubber casts from external moulds. All SEM except fig. 7. BM = British Museum
(Natural History), London.

PLATE 84

SIVETER, VANNIER and PALMER, Silurian myodocopid ostracodes

796

PALAEONTOLOGY, VOLUME 30

tubercle

groove

sulcus

tubercle with
pore canal

TEXT-FIG. 4. Schematic representation
of individual types of true ornament in
Silurian myodocopid ostracodes. a, reti-
culation; B, corrugation; c, tuberculation.

In ''Bolhozoe sp. nov. A most corrugations appear to run parallel to the sulci but ventrally they are
aligned parallel to the ventral margin (PI. 83, fig. 2). This pattern is strikingly comparable to that of the
fossae in the dominantly reticulate bolbozoid ‘B.’ cf. bohemica (cf. PI. 84, figs. 2 and 7) and also mirrors
the alignment of tubercles in other Silurian tuberculate bolbozoids (e.g. PI. 83, fig. 1). Seemingly the
distribution of fossae, corrugations, and possibly tuberculation represent homologous patterns
within groups of Silurian bolbozoids (text-fig. 5).

Forms of corrugation are also known, for example, in Palaeozoic platycopes (e.g. Schallreuter
1978), Mesozoic and Tertiary podocopes (e.g. Clements 1974; Doruk 1974), Recent polycope
myodocopids (e.g. Hasan 1983), and Carboniferous entomozoids s.s. (e.g. Gooday 1983). Though
entomozoids were traditionally classified as myodocopine myodocopids (e.g. Sylvester-Bradley in
Moore 1961), as restricted to the ‘fingerprint’ ostracodes of current usage this group of Silurian-
Carboniferous ostracodes lack characteristic features (e.g. rostrum and rostral incisure) of typical
myodocopines and therefore belong outside that group. The arrangement and spacing of longitudinal
and bifurcated primary ribs of several entomozoid taxa (e.g. see IKuzminaella sp. of Gooday 1983,
fig. 10) are like the corrugated surface of Silurian bolbozoids. Gooday (1983) noted that shell ornament
may not be reflected on the internal surface in entomozoids from the Carboniferous, a characteristic

SIVETER ET AL.: SILURIAN MYODOCOPID OSTRACODES

797

TEXT-FIG. 5. Valves (a, c, e) illustrating similar patterns of alignment (b, d, f) of tuberculation (a, b), corrugation
(c, D), and reticulation (e, f) in bolbozoids. a, b, Bolbozoid gen. et sp. nov. A, based on right valve, IGR 33040
(valve reversed), c, d, "Bolhozoe sp. nov. A, based on left valve, IGR 33048. e, f, ‘B.’ cf. bohemica, based on left

valve, IGR 33057.

also observed in many Silurian myodocopids. Ribbing may be an adaptation in entomozoids to a
pelagic lifestyle (Gooday 1983; Siveter 1984) and it is possible, but again speculative, that the same
may be the case for corrugations.

Tuberculation (text-fig. 4c). Tubercles are common in Silurian bolbozoids, though their number, size,
and arrangement vary between taxa (cf. text-fig. 5a, c). In Bolbozoid gen. et sp. nov. A tubercles of
c. 70-300 fim diameter cover lobes but not the sulci, rostrum, or dorsal part of the bulb (PI. 83, figs. 1,
3, 5; text-fig. 5a). Most are regularly spaced, have a pore-like central perforation (PI. 83, fig. 3) and
show linear alignment (text-fig. 5a, b). Other bolbozoids such as ‘6.' sp. nov. A (PI. 83, figs. 2 and 6;
text-fig. 5c, d; PI. 84, figs. 2 and 4) have only a ventral row of three to four tubercles in adults. In ‘6.’
bohemica tubercles are more numerous and prominent in early instars, but are virtually lacking in
adults. Tubercles can have expression on both internal and external shell surfaces in Silurian
bolbozoids; the central perforation probably penetrates the shell thickness and may represent a
normal pore canal communicating in life with the epidermal cells. Tuberculation is common
throughout the Ostracoda as shown, for example, in palaeocopes (Martinsson 1962; Siveter 1976;
Schallreuter 1982c), podocopes (Sylvester-Bradley and Benson 1971), and Recent myodocopids
(Kornicker 1981, pi. 70).

Composite ornament (text-fig. 6c, d). Some Silurian myodocopids exhibit an intricate combination of
reticulation, corrugation, and tuberculation (e.g. PI. 84, figs. 2 and 4; text-fig. 6c, d). In ‘B.’ sp. nov. A
the reticulum centroventrally is surmounted by tubercles with pores and laterally merges with
corrugations. This ‘composite’ external ornament has counterparts in Recent myodocopids such as
S.frons (text-fig. 6a, b).

Normal pores. Though not ornament as such, the presence of normal pores in Silurian myodocopids is
worth recording. Apart from having pores within tubercles, in one well-preserved specimen of

798

PALAEONTOLOGY, VOLUME 30

M

TEXT-FIG. 6. Comparison of true ornament in a Recent and a Silurian myodocopid ostracode. a, b, Scleroconcha
Irons Kornicker, 1975, philomedid, right valve (redrawn from Kornicker 1975, fig. 209a-h). Recent, SW Pacific off
Chile; 1 190-1263 m depth, a, lateral view, x 50. b, detail of the external surface ornament (reticulation), x 200. Bl,
fossa; B2, elongate fossa; B3, normal pore canal, c, d, "Bolbozoe sp. nov. A, bolbozoid, right valve. Based on
BM OS 13060 (see PI. 84, figs. 2 and 4), Ludlow Series, Silurian, Long Mountain, Powys, Wales, c, lateral view,
X 20. D, detail of external surface ornament, x 75. Dl, posterior sulcus; D2, adductorial sulcus and muscle scar;

D3, tubercles with pores; D4, reticulation; D5, corrugation.

Bolbozoid gen. et sp. nov. A (PI. 83, figs. 1, 3, 5, 7, 8), pore-like structures were also observed
anteroventrally. They consist of several single, circular and also double, semicircular openings about
30 jttm in diameter, often with a central 'boss’. Comparable ‘normal pores’ occur in many Recent
myodocopids; Polycope choane has sparsely distributed single and double pores on its lateral surface
(Hasan 1983) and there are living cylindroleberidids with various types of minute, sparse, rimmed
pores, pore groups, or tubed pores whose number and position can vary within a species (see
Kornicker 1975, pis. 310, 320, 321).

In summary, the external sculptures of Silurian myodocopids described above are considered to be
true, genetically controlled ornament, similar to that of many Recent myodocopids. Such features
develop gradually and to a consistent pattern throughout the ontogeny of both Silurian and Recent
myodocopid species. Lastly, patterns of reticulation, corrugation, and tuberculation in groups of
Silurian myodocopids seem to be homologous (text-fig. 5) and the result of evolutionary processes
rather than the chance factors.

SIVETER ET AL.: SILURIAN MYODOCOPID OSTRACODES

799

SHELL MICROSTRUCTURES

Silurian myodocopids from the Welsh Basin, Armorican Massif, and Bohemia exhibit enigmatic,
surface polygonal and radiate microstructures, each usually with a central perforation (Pis. 85-87).

Radiate microstnictures. Subcircular radiate microstructures, herein termed rosettes (e.g. PI. 85, fig. 3,
B. cf. anomala), occur on both internal and external moulds, either individually; as clusters (e.g. PI. 85,
figs. 1, 2, 4, 5, ‘Cypridinid’ sp.; PI. 87, figs. 6 and 7, ‘6.’ cf. bohetnica); more rarely as rosettes covering
virtually the entire surface of the valve (e.g. PI. 85, figs. 9-11, B. cf. anomala)-, and sometimes as tiny
rosettes within a basically polygonal pattern of shell microstructure (see ‘composite microstructure’
below). Each rosette consists of some ten to thirty acicular radii and (excluding rosettes in composite
patterns) is c. 100-500 /<m across; most have a central perforation.

Polygonal microstructures. On the surfaces of both internal and external moulds of many of our
Silurian myodocopids a pattern of relatively small, polygonal microstructures is discernible (e.g.
Pis. 86, 87). The polygons vary in size (10-150 gm), shape (degree of elongation), outline (angular
to more rounded), and packing (dense to loose) to give different types of patterns on the surfaces of
the valves of several taxa. In three dimensions these structures are interpreted as polygonal
platelets (text-fig. 9).

Typical polygonal patterns are shown in text-fig. 7. A specimen of B. cf anomala (text-fig. 7a; PI. 86,
figs. 2 and 4) shows closely packed, centrally perforated polygons over almost all the external surface
except the anterodorsal bulb; they are mostly irregular in shape, size (60-150 ;um), and in the diameter
of their central perforation. Similar polygons occur in a specimen of ‘Cypridinid’ gen. et sp. nov. B
(PI. 87, figs. 2-4). In another specimen of B. cf anomcda (PI. 86, fig. 5, text-fig. 7b) the packing appears
looser and some polygons ill-defined, though most have central perforations. Generally, the upper
(outer) part of the central perforations of the polygons is often enlarged, such that each perforation
can appear as a tiny concavity housing a minute hole (e.g. PI. 86, fig. 4; PI. 87, fig. 4; text-fig. 9).

Another common pattern consists of closely packed but smaller, rounded to polygonal,
granule-like microstructures (each c. 10-60 gm diameter), some of which have a central perforation
(PI. 84, figs. 1, 3, 5; PI. 87, figs. 1 and 5).

Composite microstructures. Radiate and polygonal microstructures of various sizes often occur in the
same valve (text-fig. 9). For example, tiny radiate microstructures are often found intimately
associated within a pattern of perforated polygons (e.g. PI. 85, fig. 7, Bolbozoid gen. et sp. nov. A;
PI. 87, figs. 6 and 7, ‘6.’ cf hohemica). In many cases both the smaller, granule-like elements and the
perforated polygonal platelets (combined range: 10-150 gm diameter) occur as a gradational.

TEXT-FIG. 7. Pattern of perforated polygons on lateral surface of Bolbozoe cf anomala Barrande, 1872. a, left valve,
lateral view, IGR 33023 (see PI. 86, fig. 2), x 25. b, left valve, lateral view, IGR 33029 (see PI. 86, fig. 5), x 25.

800 PALAEONTOLOGY, VOLUME 30

intermingled pattern on individual valves (e.g. PI. 87, fig. 4, ‘Cypridinid’ gen. et sp. nov. B; PI. 87, fig. 5,
‘Cypridinid’ gen. et sp. nov. A).

The nature and origin of polygonal and radiate microstructures. The patterns of perforated polygons
and radiate microstructures have many, shared characteristics which indicate that such micro-
structures should not be interpreted as true external ornament.

1 . Their general distribution on valves appears to be random, with apparently no consistent trends
in individuals or species.

2. In some specimens the microstructures are observed to be ‘superimposed’ on true ornament such
as tubercles (PI. 85, figs. 6 and 7), as well as on the adductorial sulcus (PI. 87, figs. 6 and 7) and the
rostrum (PI. 84, fig. 1; PI. 87, fig. 2), two regions where ornament is either otherwise unknown
(adductorial sulcus of ‘6.’ cf. boheinica) or not expected to occur (rostrum of Silurian myodocopids in
general).

3. Critical evidence is found in some well-preserved carapaces which show a different polygonal
pattern (PI. 87, fig. 1) or arrangement of rosettes on the two valves. In one specimen a single radiate
structure is developed only on the left valve (PI. 85, fig. 3).

4. In contrast with the typical ornament of Silurian myodocopids (reticulation, corrugation)
perforated polygons and radiate microstructures manifest complimentary patterns on both external
and internal moulds.

5. Similar polygonal and radiate microstructures are found in various bolbozoid and ‘cypridinid’
myodocopid taxa from various localities of probably different ages. Compare, for example, the radiate
structures in the specimen of ‘Cypridinid’ sp. from ‘La Cultais’ (PI. 85, figs. 1, 2, 4, 5) with those of the
specimen of B. cf. anomaJa from ‘Les Buhardieres’ (PI. 85, figs. 3 and 8); or compare the polygonal
structures in the specimens of B. cf. anomala from ‘Les Buhardieres’ (PI. 86, figs. 1-6) with those of
‘Cypridinid’ gen. et sp. nov. B from Saint-Denis-d’Orques (PI. 87, figs. 2-4).

Accepting this evidence that the shell microstructures are not ornament, they may have originated
in one of several possible ways:

1. Microborings associated with endolithic fungi and algae occur in fragments of molluscs,
echinoderms, coralline algae, barnacles, serpulids, and foraminiferans from shelf sediments (Perkins
and Halsey 1971). Such activity produces different sizes (1-20 pm diameter), shapes, and patterns of
perforations which are not dissimilar to those of some of the Silurian myodocopids. However,
microborings in Recent material are never associated with complex, three-dimensional radiate or

EXPLANATION OF PLATE 85

Examples of polygonal and radiate microstructures of the shell of Silurian myodocopid ostracodes.

Figs. 1, 2, 4, 5. ‘Cypridinid’ sp. Right valve, IGR 33054, La Lande-Muree Formation, La Cultais (Vieux-Vy-sur-
Couesnon, Ille-et-Vilaine), Armorican Massif, France. 1, lateral view, x75. 2, anterior oblique view, x75.
4 and 5, details of acicular radiate structures (‘rosettes’) in anterior part of the valve (from 1 and 2
respectively; x 205, x2I5).

Figs. 3 and 8. Bolbozoe cf. anomala Barrande, 1872. Open carapace, IGR 33025, La Lande-Muree Formation, Les
Buhardieres (Andouille, Mayenne), Armorican Massif, France. 3, lateral view, x 56. 8, detail of perforated,
acicular radiate structure (‘rosette’) on ventral part of the left valve, stereo-pair, x 380.

Figs. 6 and 7. Bolbozoid gen. et sp. nov. A. Right valve, IGR 33040, La Lande-Muree Formation, Les Buhardieres
(Andouille, Mayenne), Armorican Massif, France. 6, lateral view, x 14. 7, detail of posteroventral part of
external surface showing an acicular radiate structure (‘rosette’) on the side of a tubercle {top right) and closely
packed perforated polygons (lower left), x 245.

Figs. 9-11. B. cf. anomala Barrande, 1872. Left valve, IGR 3300, La Lande-Muree Formation, Les Buhardieres
(Andouille, Mayenne), Armorican Massif, France. 9, oblique ventral view, x 17. 10 and 1 1, details of closely
packed radiate structures (‘rosettes’) on the lateral surface of the valve, x 200, x 150.

All specimens are silicone rubber casts from external moulds. All SEM except fig. 3.

PLATE 85

SIVETER, VANNIER and PALMER, Silurian myodocopid ostracodes

802

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 8. Reconstruction of acicular radiate structure (‘rosette’) based in part on observations on internal and
external moulds of IGR 33025 (Bolhozoe cf anomala Barrande, 1872; see PI. 85, figs. 3 and 8). Central perforation
may or may not completely penetrate shell thickness. True shell thickness unknown.

polygonal structures like those described from the Silurian myodocopids and such an origin is
considered untenable.

2. It is possible that some kind of acidic action within the sediment might have produced these
randomly distributed perforations and irregular polygonal patterns on the calcareous shells. Such
acid erosion is unlikely for several reasons. First, the radiate structures and perforated polygons are
lacking from associated faunas of palaeocope ostracodes, bivalves, eurypterids, and phyllocarids.
Secondly, delicate external features such as muscle scars and minute pores are remarkably well
preserved. Thirdly, the geochemistry of the Silurian black siltstones and mudstones indicates reducing
conditions for the substrate (Dabard and Paris 1986).

EXPLANATION OF PLATE 86

Examples of polygonal and radiate microstructures of the shell of Silurian myodocopid ostracodes.

Figs. 1-6. Bolbozoe cf. anomala Barrande, 1872. 1 and 3, right valve, IGR 33027. 1, lateral view, x 22. 3, detail of
packed, perforated polygons in ventral part of the anterodorsal bulb, stereo-pair, x 260. 2 and 4, left valve,
IGR 33023. 2, lateral view, x 22. 4, detail of packed, perforated polygons in posterodorsal part of the valve,
stereo-pair, x 260. 5, left valve, showing perforated polygons, IGR 33029, lateral view, x 22. 6, left valve, IGR
33026, lateral view, x 12, together with two other valves showing different types of external shell surfaces on the
same slab (smooth surface; small, closely packed polygons; larger, packed, perforated polygons). All from La
Lande-Muree Formation, Les Buhardieres (Andouille, Mayenne), Armorican Massif, France.

All specimens are Silicone rubber casts from external moulds. Figs. 1, 2, 5, 6 are light photographs; 3 and 4 are
SEM.

PLATE 86

SIVETER, VANNIER and PALMER, Silurian myodocopid ostracodes

804

PALAEONTOLOGY, VOLUME 30

3. The origin of the radiate and polygonal mierostruetures may be directly related to processes
concerned with shell calcification and its subsequent alteration. Radiate structures and perforated
polygons are generally expressed on both internal and external moulds of individual valves, thus
suggesting that the whole shell thickness is usually involved in the formation of these mierostruetures.
Indeed, in one specimen (PI. 86, fig. 5; text-fig. 7b) fracturing across the valve has clearly occurred in a
zig-zag path, indicating boundaries between polygons. Three-dimensional reconstructions (text-figs.
8 and 9) depict individual radiate structures and perforated polygons each as internal and external
surface expressions of isolated and coalesced platelets respectively. The central perforations may
represent former normal pore canals although there is no firm evidence that they completely
penetrate the thickness of the shell.

Bate and Sheppard (1982) have observed isolated and coalesced discs and spherulites of calcium
carbonate within the cuticle matrix of Recent myodocopids, structures that they interpret as
fundamental products of the in vivo calcification process responsible for producing the calcified
(‘shell’) euticle. Aeeording to these authors, growth of these struetures appears to take place either by
coalescence or by accretion around the disc (‘plate’) perimeter. The discs lie within and cut across the
layers of chitin and observations under plane polarized light clearly show their radiate, crystalline
nature (Bate and Sheppard 1982, pi. 2, figs. 1 and 2; herein PI. 88, fig. 4). Such discs and spherulites,
which are clearly present in the shell in many Recent myodocopids (see PI. 88, figs. 1 -8), share striking
similarities with the platelets described herein from Silurian myodocopids: e.g. in size, shape (cf. PI. 86,
figs. 1-6 with PI. 88, figs. 1-3), random distribution, coalescenee (cf. PI. 86, fig. 3 and PI. 87, fig. 4 with
PI. 88, fig. 2), and the occurrence of eentral perforations (see Bate and Sheppard 1982). In addition,
within their eutiele Recent myodocopids also show loosely packed (PI. 88, fig. 1) and isolated platelets
(PI. 88, fig. 5). It is, therefore, possible that the radiate structures and perforated polygonal platelets in
our Silurian myodocopids were produced by a similar process of shell calcification which may have
taken place in vivo. Further, presumably post-mortem changes of the calcium carbonate platelets
would then consist mainly of in situ recrystallization of the caleium carbonate, thus preserving most of
their original form and arrangment. Under certain chemical conditions, the calcium carbonate of
some platelets may have been rearranged to form acicular radiate structures (rosettes).

There is also, however, evidenee based on several observations of Recent material, for an apparently
entirely post-mortem crystallization of calcium carbonate matter in myodocopids. Discoidal,
spheroidal, and irregular-shaped ‘nodules’ (up to 300 /.im diameter) have been ‘produced’ within the
cuticle of Recent cypridinid valves simply by drying the valves and then soaking them in water (Sohn

EXPLANATION OF PLATE 87

Examples of polygonal and radiate mierostruetures of the shell of Silurian myodoeopid ostraeodes.

Figs. 1 and 5. ‘Cypridinid’ gen. et sp. nov. A. Open earapaee, IGR 33016, La Lande-Muree Formation, Les
Buhardieres (Andouille, Mayenne), Armoriean Massif, Franee. 1, lateral view, x 15. 5, detail of eentral area of
the left valve showing reniform musele sear, perforated platelets and elosely paeked mieroplatelets, x 150.

Figs. 2-4. ‘Cypridinid’ gen. et sp. nov. B. Right valve, IGR 33200, siltstones of Ludlow or Pridoli age,
Saint-Denis-d’Orques, Sarthe, Armoriean Massif, Franee. 2, lateral view, x 18. 3, detail of adduetor musele
sear (from 2), x 52. 4, paeked, perforated polygons in anteroventral part of the valve, stereo-pair, x 52.

Figs. 6 and 7. 'Bolbozoe' cf. bohemica Barrande, 1872. Right valve, IGR 33059, La Lande-Muree Formation, La
Cultais (Vieux-Vy-sur-Couesnon, Ille-et-Vilaine), Armoriean Massif, France. 6, lateral view, x 9; note
mierostruetures on dorsal part of the adductorial sulcus and on ornamented part of adjacent central lobal
area. 7, detail of dorsal end of the adductorial sulcus showing radiate mierostruetures (‘rosettes’) and
perforated polygons, stereo-pair, x 54.

Fig. 8. B. cf anomala Barrande, 1872. Left valve, lacking any discernible mierostruetures, IGR 33018, La
Lande-Muree Formation, Les Buhardieres (Andouille, Mayenne), Armoriean Massif France, x 19.

All specimens are Silicone rubber casts from external moulds. All SEM except fig. 6.

PLATE 87

'-fcT -s^'

'■^Wi

SIVETER, VANNIER and PALMER, Silurian myodocopid ostracodes

806

PALAEONTOLOGY, VOLUME 30

TEXT-HG. 9. Reconstruction of packed polygonal platelets, including some showing acicular radiate structures;
based on observations on internal and external moulds (e.g. see PI. 85, figs. 6 and 7; PI. 86, figs. 1-6). Central
perforation may or may not completely penetrate shell thickness. True shell thickness unknown.

and Kornicker 1969). These nodules, of monohydrocalcite (CaC03H20) and ealcite, are morpho-
logically comparable to the platelets in our Silurian specimens and, according to the authors, can
develop post-mortem in many Recent cypridinacean myodocopids. Sohn and Kornicker (1969, pi. 2,
figs. 3 and 4) demonstrated that the initial form of the ‘shell’ carbonate in such valves is that of an
amorphous gel evenly distributed throughout the cuticle; only subsequently is the carbonate
redistributed during a change in form to monohydrocalcite, with nucleation leading to ‘nodule’
production. In addition, Kornicker (1981, herein PI. 88, fig. 6) illustrated radiate, rosette-like
structures produced by post-mortem mineral rearrangement within the cuticle matrix of a Recent
myodocopid. Supposed post-mortem recrystallization (of possible platelets) occurs in Carboniferous
cypridinacean myodocopids (see Sohn 1977; Bless 1973), where the obvious radiate markings consist
of single or coalesced rosettes resembling those in Silurian myodocopids (cf. PI. 85, figs. 9-11 with
Sohn 1977, fig. l/i). Sohn (1977, p. 128) considered that the nodule ‘artefacts’ that Sohn and Kornicker
(1969) induced in Recent Vargula specimens are not similar to the radiating structures seen in the
Carboniferous "Cypridina radiata and other species, but we are inclined to take the opposite view.

In summary, the evidence for the role and timing of the calcification process in Recent myodocopids
is equivocal. Sohn and Kornicker (1969) emphasize the possibility of a very rapid posthumous
production of crystalline nodules. They state, however, that the same process can also occur (though
rarely) in vivo. Unfortunately they do not elaborate upon, or illustrate evidence for, this crucial
statement. By comparison. Bate and Sheppard, in describing the process of cuticular spherulite
( = nodule) development, appear to assume uncritically that it occurred in vivo, despite the fact that
‘some non or poorly calcified individuals were present’ (1982, p. 26). Furthermore, they seem to have
been unaware of Sohn and Kornicker’s earlier work and its fundamental implications for their own
study. From this, and also because they do not record the preservational medium or give the

SIVETER ET AL.: SILURIAN MYODOCOPID OSTRACODES

807

post-mortem age of their Recent specimens, it is difficult to assess the possible role that post-mortem
processes may have played in the formation of the spherulites in Bate and Sheppard’s material.

Ultimately this crucial question of the relative timing and role of the calcification process in the
myodocopid cuticle can only be resolved by the study of living or freshly killed specimens. With
regard to our fossil material we can say that it contains microstructures like those ‘produced’ in
Recent myodocopid taxa and we therefore conclude that the same calcite depositional processes may
well have been operative. Furthermore, the ubiquity of such structures in taxa separated in time by
some 400 m.y. emphasizes that they cannot be of use in low level systematics.

There is, however, a further point that may lend weight to the argument for in vivo calcification in
our fossil material. One of the main conclusions of Sohn and Kornicker’s study was that their
post-mortem interpretation for the calcification process explains ‘the scarcity of fossil Myodocopida’,
since after ‘the nodules formed, the protein and chitin framework of the shell usually disintegrated;
consequently, recognisable myodocopid fossils are rare’ (1969, pp. 104-105). However, our studies
show that at least in certain Silurian environments, myodocopids can occur comparatively often and
it could be that their relatively high preservational potential results from an early {in vivo) onset of the
calcification process.

We suggest that a distinction can be drawn between various possible sequences of the implied
calcification process and a model based on these diflferences is outlined below.

POSSIBLE MODE AND STYLES OF CALCIFICATION IN
SILURIAN MYODOCOPIDS

The inferred presence of calcium carbonate platelets, comparable with those of Recent myodocopids
(e.g. Bate and Sheppard 1982), within the shell of Silurian bolbozoids and ‘cypridinids’ implies a
particular mode of calcification. However, the occurrence (within and between individuals of the same
and different species) of a variety of sizes and styles of radiate and/or polygonal microstructures
(Pis. 85-87) attributable to such processes clearly requires explanation. Taking into account the
findings and in vivo interpretations of Bate and Sheppard ( 1 982), a model to explain the possible mode
of calcification and the formation of the various types of radiate and polygonal microstructures
of the Silurian myodocopid shells is proposed (text-fig. 10).

Nucleation centres have obvious importance at the beginning of any calcification process. If such
centres are numerous and densely packed, the resulting pattern of calcification will be very different
than if they were sparse and distant. Thus, based on possible variation in number, density, and
distribution of centres, four cases of shell calcification are considered in order to account for the
formation of the variety of microstructures observed (text-fig. 10a-d). The central perforations
present in most of the platelets of our Silurian myodocopids (text-figs. 7-9)— possibly normal
pore-canals (see above) — provide in some cases possible nucleation centres. This has parallels in
Recent myodocopids; Bate and Sheppard (1982) concluded that nucleation of calcium carbonate in
Halocypris inflata might preferentially take place around pores.

Case A (text-fig. 10a, 1-4). Here the nucleation centres for initial calcification would be distributed
regularly and densely across the cuticle matrix (Stage Al). Spherulites or discs of calcium carbonate
begin to form heterochronously and grow outwards in all directions (Stage A2). Further calcification
at new centres and peripheral accretion around existing discs produces a pattern of packed polygonal
platelets which then effectively prevents further ‘lateral’ calcification taking place (Stage A3). At this
stage the calcareous platelets are likely to be intimately associated with layers of chitin, as
demonstrated by Bate and Sheppard (1982, pis. 5 and 6) in Recent myodocopids. As a result of
possible microbiological action with the sediments the exuviae and post-mortem shells (Stage A4)
would then probably lose any outer and internal organic (chitin) layers they might have had. At the
same time and later some amounts of mineral recrystallization/rearrangement of the calcareous
platelets would occur. Such a process would account for the formation of the granule-like, finely
packed polygonal patterns (e.g. PI. 84, fig. 5; PI. 87, fig. 5).

808

PALAEONTOLOGY, VOLUME 30

Case B (text-fig. 10b, 1-4). Here, nucleation centres are less densely distributed and possibly
correspond to normal pore canals (Stage Bl). The calcification process is essentially similar to that in
Case A, but with the possible growth of larger platelets. In Stage B4, a pattern of perforated polygons
results, as seen after the degradation of organic matter (including the possible resultant enlargement of
pore-canals) and a mineral rearrangement which included the development of a few acicular radiate
microstructures (see PI. 86, figs. 1-6; PI. 87, figs. 2-4; text-fig. 9).

Case C (text-fig. 10c, 1-4). The number of nucleation centres (Stage Cl) is sparser than in Case B.
Post-mortem recrystallization produces a pattern of packed, perforated rosettes (Stage C4) (e.g. PI. 85,
figs. 9-11).

Case D (text-fig. IOd, 1-4). Here, centres of nucleation are very sparse and irregularly distributed
(Stage Dl). As in Cases A-C, spherulites grow outwards in all directions by peripheral accretion
(stages D2, D3). The calcification process terminates before all the platelets could coalesce (Stage D3),
because of the relatively large distance between nucleation centres. Post-mortem, acicular recrystal-
lization then produces isolated rosettes (e.g. PI. 85, figs. 3 and 8; text-fig. 8) or clusters of rosettes (e.g.
PI. 85, figs. 1, 2, 4, 5; PI. 87, fig. 7).

Variation in the distribution and density of nucleation centres could explain the occurrence of
various sized, closely packed platelets on the same valve (e.g. PI. 87, figs. 3-5).

The model proposed (text-fig. 10) can be used to interpret cases where possible conspecific valves,
often from the same horizon, locality, and (even) slab, show a variety of microstructures. For example,
some specimens of B. cf anomala are ‘smooth’ (PI. 87, fig. 8), others show extensive patches of
polygonal platelets (PI. 86, fig. 5; text-fig. 7b), and yet others have a complete polygonal pattern
developed (PI. 86, figs. 2 and 4; text-fig. 7a). Most important and revealing as far as the interpretation
of their various microstructures is concerned, are those conspecific valves occurring adjacent to each
other on a single slab (PI. 86, fig. 6). In this case the ‘smooth’ valves might possibly represent newly
moulted individuals (text-fig. 10, stages Al, Bl, Cl, Dl) with very small amounts of cuticle

EXPLANATION OF PLATE 88

Specimens of Recent myodocopid ostracodes showing various degrees of valve calcification.

Figs. 1-3. Amholeberis antyx Kornicker, 1981. Cylindroleberidid, right valve, USNM 157625. Recent, Grand
Recif, Madagascar, West Indian Ocean; 21 m depth. 1, lateral view showing numerous circular platelets within
the shell, x 45. 2, detail of platelets, some coalescing x 125. 3, closely packed platelets in anterior part of the
valve, below rostrum, x 245.

Fig. 4. Halocypris inflata (Dana, 1849). Halocypridid, broken shell fragment. Recent, North Atlantic Ocean;
0-10 m depth. Shell surface under plane polarized light showing crystalline nature of circular platelets;
growth of platelets is thought (Bate and Sheppard 1982) to take place by coalescence and by accretion around
platelet perimeter, x 300.

Fig. 5. Tetraleheris maddocksae Kornicker, 1981. Cylindroleberidid, left valve, USNM 157626. Recent, Grand
Vasque Lagoon, Madagascar, Western Indian Ocean; 18 m depth. Lateral view, x 24. Juvenile specimen in
process of moulting; shows a few large platelets sparsely distributed within the shell.

Figs. 6 and 7. Leuroleberis sharpei Kornicker, 1981. Cylindroleberidid. Recent, Monterey Bay, California; 36 m
depth. 6, right valve USNM 139286, showing radiating concretionary structure on dorsal part of the valve
from which surface layers have peeled (specimen boiled for 15 min. in dilute potassium hydroxide), x 18.
7, anterior part of left valve, USNM 156930, showing abundant oval fossae on external surface and closely
packed platelets of various sizes within the shell, x 46.

Fig. 8. L. mackenziei Kornicker, 1981. Cylindroleberidid, left valve, USNM 156967. Recent, New South Wales
Coast, Australia. Lateral view showing platelets coalescing to form large calcified areas within the shell, x 17.

Figs. 1-3, 5-8 from Kornicker 1981; 4 from Bate and Sheppard 1982. USNM = National Museum of Natural
History, Smithsonian Institution, Washington.

PLATE 88

SIVETER, VANNIER and PALMER, Recent myodocopid ostracodes

810

PALAEONTOLOGY, VOLUME 30

Patterns of Initial calcification Further accretion

nucleation centres producing platelets producing packed

polygons

I 2 3

Post-mortem changes
including acicular
crystalisation

4

B

l2f

0 • V

-

*\ • * V • 1

TEXT-FIG. 10. Model to explain the possible mode of calcification and formation of
patterns of shell microstructures in Silurian myodocopid ostracodes. a, regular,
densely distributed nucleation centres, b, c, less dense, d, irregular, sparsely
distributed nucleation centres.

calcification (for detailed observations on the development of newly formed cuticle in ostracodes, see
Rosenfeld 1979; Okada 1982). In the same way, patches of platelets on the valve surface might indicate
an intermediate stage of calcification (text-fig. 10, stages A2, B2, C2, D2) or poor calcification, and
specimens with relatively closely packed polygonal patterns have possibly completed their cuticle
calcification prior to moulting anew (text-fig. 10, stages A3, B3, C3, D3). Such an interpretation is
supported by observations (Bate and Sheppard 1982) on sympatric specimens of the Recent
myodocopid H. inflata: some were fully calcified (having coalesced platelets forming a completely
rigid shell), and others were either poorly calcified (patches of platelets) or lacked calcification (no
platelets). Although not explicitly stated by Bate and Sheppard, these variations could reflect diflerent
states of calcification between moults. Another example, and particularly convincing in this context is
the poorly calcified, ‘in-moult’ specimen of the Recent myodocopid Tetraleberis maddocksae figured
by Kornicker (1981, pi. 34, fig. a; herein PI. 88, fig. 5).

SIVETER ET AL.\ SILURIAN MYODOCOPID OSTRACODES

811

Acknowledgements. We thank Professor Charles Holland and Drs Helga Groos-Uffenorde and Louis Kornicker
for critically reading the manuscript. Natural Environment Research Council award GR3/4399, for work on
Silurian myodocopids, is gratefully acknowledged by D. J. S., as is a British Council scholarship (J. M. C. V.) and
a Royal Irish Academy/Royal Society award (D. P.) for collaborative study at Leicester University. D. J. S.
thanks Drs M. Kruta and J. Kfiz for help with studies and fieldwork in Czechoslovakia, and Pierre Paillard is
thanked for his assistance with fieldwork (J. V.) in the Armorican Massif.

REFERENCES

BARRANDE, J. 1872. Systime silurien du centre de la Boheme. Vol. I supp., xxx + 647 pp., 35 pis. Prague-Paris.
BARROis, c. 1886. Apergu de la constitution geologique de la rade de Brest. Bull. Soc. geol. Fr. 3, 14, 678-707.
BATE, R. H. and SHEPPARD, L. M. 1982. The shell structure of Flalocypris inflata (Dana, 1849). In bate, r. h.,
ROBINSON, E. and SHEPPARD, L. M. (eds.). Fossil and Recent ostracods, 25-50. Ellis Horwood and the British
Micropalaeontological Society.

BENSON, R. H. 1971. A new Cenozoic deep-sea genus Abyssocythere (Ostracoda: Trachyleberididae). Smithson.
Contr. Paleobiol. 7, 20 pp.

1973. On Australicythere polylyca (Muller). Stereo- Atlas Ostracod Shells, 1, 3 (32), 169-172.

BLESS, M. J. M. 1973. Radiicypridina a new cypridinid ostracod genus from the Upper Carboniferous. Revta esp.
Micropaleont. 5, no. 2, 247-252.

BOUCEK, B. 1936. Die Ostracoden des bohmischen Eudlows (Stufe e6). Neues Jb. Miner. Geol. Paldont. Beil Bd.
Abt. B, 76, 31-98.

CANAVARi, M. 1900. Fauna dei calcari nerastri con Cardiola ed. Orthoceras di Xea Sant’Antonio in Sardegna.
Palaeontogr. ital. (for 1899), 5, 187-210.

CHAUBET, M. c. 1937. Contribution d Fetude geologique du Gothlandien du versant meridional de la Montagne-
Noire, 213 pp. Thesis (unpublished). University of Montpellier.

CLEMENTS, R. G. 1974. On Timiriasevia mackerrowi Bate. Stereo-Atlas Ostracod Shells, 2, 2 (20), 117-124.
DABARD, M. p. 1983. Etude petrographique et geochimique des roches, sedimentaires paleozoiques du Massif
Armoricain (presquile de Crozon, flanc nord du Bassin de Laval), 207 pp. Thesis (unpublished). University of
Rennes.

and PARIS, F. 1986. Palaeontological and geochemical characteristics of Silurian black shale formations

from the central Brittany domaine of the Armorican Massif (Northwestern France). Chem. Geol. 55, 17-29.
DAS GUPTA, T. 1932. The Salopian graptolite shales of the Fong Mountain and similar rocks of Wenlock Edge.
Proc. Geol. Ass. 43, 325-363.

DORUK, N. 1974. On Urocythereis labyrinthica Uliczny. Stereo-Atlas Ostracod Shells, 2, 1 (8), 49-52.

GNOLi, M., PAREA, G. c., RUSSO, F. and SERPAGLi, E. 1979. Paleoecological remarks on the 'Orthoceras limestone' of
south-western Sardinia (Middle-Upper Silurian). Memorie Soc. geol. Ital. 20, 405-423.

and SERPAGLI, E. 1984. Evidence of phyllocarid remains from Silurian-Devonian boundary beds in

Southwestern Sardinia. Neues Jb. Geol. Paldont. Mh. 5, 257-268.

GOODAY, A. J. 1983. Entomozoacean ostracodes from the Eower Carboniferous of south-western England.
Palaeontology, 26, 755-788.

GUILLIER, A. 1886. Geologie du departement, de la Sarthe, xii-l-430 pp. Le Mans and Paris.

and TROMELiN, G. DE 1874. Note sur le terrain silurien de la Sarthe. Bidl. Soc. Agric. Sci. Arts Sarthe,

22, 585-594.

GURiCH, G. 1896. Das Palaeozoicum im Polonischen Mittelgebirge. Zap. ross. miner. Obshch. 2, 32, 159 pp.
HASAN, M. 1983. On Polycope choane Hasan sp. nov. Stereo-Atlas Ostracod Shells, 10, 1 (11), 67-70.

JAEGER, H. 1976. Silur und Unterdevon vom Thiiringischen Typ in Sardinien und seine regionalgeologische
Bedeitung. Nova Acta Leopoldina, nf 45, 224, 263-299.

JONES, T. R. 1861. In HOWELL, H. H. and GEiKiE, A. The Geology of the neighbourhood of Edinburgh; appendix
of fossils by J. W. Salter. Mem. geol. Surv. U.K. (Map 32), p. 137, pi. 2, fig. 5.

1873. Notes on the Palaeozoic bivalved Entomostraca. X. Entomis and Entomidella. Ann. Mag. nat. Hist.

4, 11, 413-416.

1884. Notes on Palaeozoic bivalved Entomostraca. XVIII. Some species of the Entomididae. Ibid. 5, 14,

391-403.

KERFORNE, F. 1901. Etudc de la Region Silurique occidentale de la Presqu’ile de Crozen (Finistere). Bull. Soc.
Sclent, med. Ouest, 10, 234 pp.

KIRK, N. H. 1947. Geology of the anticlinal disturbance of Breconshire and Radnorshire: Pont Faen to Presteign.
Ph.D. Thesis (unpublished). University of Cambridge.

812

PALAEONTOLOGY, VOLUME 30

KONiNCK, L. G. DE 1876. Rechetches sur les fossiles paleozoiques de la Nouvelle-Galles du Sud (Australia). Mem.
Soc. r. Sci. Liege, 2, 6, 140 pp.

KORNiCKER, L. s. 1975. Antarctic Ostracoda (Myodocopina). Smithson. Contr. Zool. 163, 720 pp.

1981. Revision, Distribution, Ecology and Ontogeny of the Ostracode Subfamily Cyclasteropinae

(Myodocopina: Cylindroleberididae). Ibid. 319, 548 pp.

and CARAiON, F. E. 1978. West African Myodocopid Ostracoda (Sarsiellidae, Rutidermatidae). Ibid. 250,

110 pp.

and SOHN, i. g. 1976. Phylogeny, Ontogeny, and Morphology of living and fossil Thaumatocypridacea

(Myodocopa; Ostracoda). Ibid. 219, 124 pp.

KRiz, J. 1969. Genus Biitovicella Kfiz, 1965 in the Silurian of Bohemia (Bivalvia). J. geol. Sci., Palaeont. 10,
105-139.

1984. Autecology and ecogeny of Silurian Bivalvia. In bassett, m. g. and lawson, j. d. (eds.). Autecology

of Silurian organisms. Spec. Pap. Palaeont. 32, 183-195.

MARTiNSSON, A. 1962. Ostracodes of the Family Beyrichiidae from the Silurian of Gotland. Bull. geol. Instn
Univ. Uppsala, 41, 369 pp.

NEWELL, N. D. 1969. In MOORE, R. c. (ed.). Treatise on Invertebrate Paleontology, Part N, Mollusca 6, N243-
N248. University of Kansas Press, Lawrence.

NYE, o. B. JR., BROWER, J. c. and WILSON, s. E. 1975. Hitchhiking clams in the Marcellus Sea. Bull. Am. Paleont. 67,
287-298.

OKADA, Y. 1981. Development of cell arrangement in ostracod carapaces. Paleobiology, 7, 276-280.

1982. Structure and cuticle formation of the reticulated carapace of the ostracode Bicornucythere bisanensis.

Lethaia, 15, 85-101.

PALMER, D. 1970. A stratigraphical synopsis of the Long Mountain. Montgomery and Shropshire. Proc. geol.
Soc. Lond. 1660, 341-346.

and GNOLi, M. 1985. A preliminary report on new micropaleontological discoveries in the Silurian of

Southwest Sardinia. Boll. Soc. paleont. Ital. 23, 2 (1984), 205-209.

PARIS, F. 1977. Les Formations siluriennes du synclinorium du Menez-Belair; comparaison avec d’autres
formations siluriennes du Massif Armoricain. Bull. Bur. Rech. geol. min. Paris, 1, 2, 75-87.

1981. Les Chitinozoaires dans le Paleozoique du Sud-Ouest de I’Europe (Cadre geologique-Etude syste-

matique — Biostratigraphie). Mem. Soc. geol. miner. Bretagne, 26, 412 pp.

PENEAU, J. 1936. Presence d’Estheria dans les Schistes atribues au Gothlandien du synclinal de laval. Bull.
Soc. Sci. nat. Quest. Fr. 5, 6, 219-226.

PERKINS, R. D. and HALSEY, s. D. 1971. Geologic significance of microboring fungi and algae in Carolina shelf
sediments. J. sedim. Petrol. 41, 3, 843-853.

PETRANEK, and KOMARKOVA, E. 1953. Orientace schranek hlavonozcu ve vapencich Barrandienu a jeji paleo-
geograficky vyznam. Sb. Ustred. Ust. Geol. 20, 129-147.

ROSENFELD, A. 1979. Structure and secretion of the carapace in some living ostracodes. Lethaia, 12, 353-360.
SCHALLREUTER, R. E. L. 1978. On Karinutatia crux Schallreuter gen. et sp. nov. Stereo-Atlas Ostracod Shells, 5,
1 (6), 45-48.

1980. On Klimphores planus Schallreuter. Ibid. 7, I (2), 9-16.

1982fl. On Bilobatia serralobata Schallreuter. Ibid. 9, 1 (2), 9-16.

19826. On Concavhithis latosulcatus Schallreuter gen. et sp. nov. Ibid. 9, 2 (18), 101-104.

1982c. On Gellensia nodoreticulata Schallreuter sp. nov. Ibid. 9, 2 (19), 105 108.

1983. On Reginea reginae gen. et sp. nov. Ibid. 10, (5), 21-24.

SIVETER, D. J. 1976. On Rennibeyrichia mulciber gen. et sp. nov. Ibid. 3, 1 (1), 1-8.

1978. The Silurian. In bate, r. h. and robfnson, j. e. (eds.). A stratigraphical index of British Ostracoda.

Geol. J. Spec. Issue, 8, 57-100.

1980. British Silurian Beyrichiacea (Ostracoda). Monogr. Palaeontogr. Soc. 1, 1-76, pis. 1-27.

1982. Casts illustrating hne ornament of a Silurian ostracod. In bate, r. h., robinson, e. and Shephard,

L. M. (eds.). Fossil and Recent ostracods, 105-122. Ellis Horwood and the British Micropalaeontological
Society.

1984. Habitats and modes of life of Silurian ostracodes. Spec. Pap. Palaeont. 32, 71-85.

SOHN, I. G. 1977. Radiate shell structures in Palaeozoic myodocopid and palaeocopid ostracodes are epigenetic.
J. Res. US geol. Surv. 5, 1, 125-133.

and KORNICKER, L. s. 1969. Significance of calcareous nodules in myodocopid ostracod carapaces. In

NEALE, J. w. (ed.). The taxonomy, morphology and ecology of Recent Ostracoda, 99-108. Oliver and Boyd,
Edinburgh.

SIVETER ET AL.: SILURIAN MYODOCOPID OSTRACODES

813

SYLVESTBR-BRADLEY, p. c. 1961. In MOORE, R. c. (ed.). Treatise on Invertebrate Paleontology, Pt. Q, Arthropoda 3:
Crustacea: Ostracoda, Q387-Q406. University of Kansas Press, Lawrence.

and BENSON, r. h. 1971. Terminology for surface features in ornate ostracodes. Lethaia, 4, 249-286.

THORNTON, s. E. 1984. Basin model for hemipelagic sedimentation in a tectonically active continental margin:
Santa Barbara Basin, California Continental Borderland. In stow, d. a. v. and piper, d. j. w. (eds.).
Fine-grained sediments: deep-water processes and facies. Spec. Pubi. geol. Soc. Lond. 15, 377-394.

TUREK, V. 1 983. Hydrodynamic conditions and the benthic community of Upper Wenlockian calcareous shales
in the western part of the Barrandian (Kosov quarry). Cas. Miner. Geol. 28, 3, 245-260, 10 pis.

VAN MORKHOVEN, F. p. c. M. 1962. Post-Paleozoic Ostracoda: their morphology, taxonomy and economic use, 1,
240 pp. Elsevier, Amsterdam.

WARREN, p. T., PRICE, D., NUTT, M. J. c. and SMITH, E. G. 1984. Gcology of the country around Rhyl and Denbigh.

Mem. geol. Surv. UK (Sheets 95, 107, 94, and 106), 217 pp.

WATKINS, R. 1978. Bivalve ecology in a Silurian shelf environment. Lethaia, 11, 41-56.

WEiTSCHAT, w. 1983. Ostracodcn (O. Myodocopida) mit Weichkorper-Erhaltung aus der Unter-Trias von
Spitzbergen. Paldont. Z. 57, 309-323.

WESTERMANN, G. E. G. 1985. Post-mortem descent with septal implosion in Silurian nautiloids. Ibid. 59,

79-97.

DAVID J. SIVETER

Department of Geology
University of Leicester

University Road
Leicester LEI 7RH

England

J. M. C. VANNIER

Institute de Geologic
Universite de Rennes
35042 Rennes Cedex
France

Typescript received 18 August 1986
Revised typescript 1 February 1987

D. PALMER

Department of Geology
Trinity College
Dublin 2
Eire

A RE-EVALUATION OF THE PLANTS TINGIA AND
TINGIOSTACHYA FRQM THE PERMIAN OF
TAIYUAN, CHINA

by GAO ZHIFENG Cmd B. A. THOMAS

Abstract. The two characteristic Cathaysian Carboniferous Permian genera Tingia and Tingiostachya are
reviewed and rediagnosed in the light of new specimens from Taiyuan, northern China. Epidermal features
of Tingia include bands of longitudinally arranged stomata. Tingiostachya is shown to have spirally arranged
sporophylls with each bearing one spheroidal sporangium, rather than the previously described whorls of
four sporophylls and tetralocular sporangia. Tingiostachya tetralocularis and Tingia elegans are rediagnosed.
The morphological and taxonomic relationships between the two genera are discussed, and it is proposed
that a new family Tingiostachyaceae includes Tingiostachya and the satellite taxon Tingia.

Four floras can be recognized in Carboniferous and Permian strata: the Euramerian flora, the
Angaran or Kusnezk flora, the Glossopteris flora, and the Cathaysian flora (Chaloner and Meyen
1973; Li and Yao 1982, 1983). The Cathaysian flora extended over what is' now China, Korea,
Japan, and as far as Sumatra and New Guinea. The flora started to diversify from the Euramerian
flora in the late Carboniferous, reached its greatest diversity in the early Permian, and persisted
into the late Permian. The Cathaysian flora can itself be divided roughly by the present latitude
of 35° N. into two distinct floras, the Northern Cathaysian flora and the Southern Cathaysian
flora (Li and Yao 1983).

In the Northern flora, plant assemblages are found throughout the sequence of strata from the
late Carboniferous (Stephanian) to the late Permian (Lee 1963, 1964; Li and Yao 1983). Here the
late Carboniferous consists mainly of shales, coal-seams, and limestones with the Neuropteris
ovata-Lepidodendron posthumii assemblage. The assemblage is recognized as the oldest division of
the Cathaysian flora because of the presence of Tingia and the abundance of distinctive lycophytes.
The early Permian Shanxi (Shansi) Series is mainly of terrestrial origin although it includes
several thin marine layers with Lingula spp. The Emplectopteris triangiilaris-Taeniopteris spp.-
Emplectopteridium alatwn assemblage characterizes this series. Only thin coal seams are present
in the overlying Lower Shihezi (Shihhotse) Series and the plants constitute the Emplectopteris
triangularis-Taeniopteris s\yp.-Cathaysiopteris whitei assemblage. There are no coal seams in the
Upper Shihezi Series except in the southern marginal areas and the plant assemblage has changed
to what has been interpreted as the main late Cathaysian flora with the major characteristic
elements being Gigantonoclea hallei, Eascipteris spp., and Lohatannidaria eusifolia.

The plant assemblage sequence in the Southern flora, unlike that of the Northern flora, is
discontinuous and the stratigraphic succession is interrupted by volcanic and marine deposits.

The Lower Permian plants described here are from exposures in Simugedong approximately
5 km north-east of East Hill Mine, Taiyuan, Shanxi province, northern China (text-fig. 1). Here
the Permian strata are well exposed and widely used as an index section in China. The assemblages
of the Northern flora mentioned above were mainly established from work in the Taiyuan area as
the local strata contain one of the best-known Lower Permian floras of the northern hemisphere.
The main groups of plants previously described from this flora are pteridophytes, cycadophytes,
and the Noeggerathiales (including Tingia and Tingiostachya) with Tingia being one of the
characteristic genera of this flora.

I Palaeontology, Vol. 30, Part 4, 1987, pp. 815-828, pis. 89-90.|

© The Palaeontological Association

816

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 1. Map showing distri-
bution of Tingia and Tingiostachya
in south-east Asia. 1, Halle 1925
and 1 927, Konno and Asama 1951,
and present specimens. 2, Konno,
1929. 3, Stockmans and Mathieu
1939. 4, Bohlin 1971. 5, Konno
et al. 1971. 6, Feng et at. 1977.
(1-3 and 6 both genera; 4 and 5
Tingia).

MATERIAL AND METHODS

The specimens described here are preserved as compressions and impressions in a grey shale. The irregular
surface produced during the splitting of the shale necessitated some ciegagement of many of the specimens.
This was done with the aid of tungsten needles sharpened by heating and subsequently dipping into molten
sodium nitrite. Maceration of portions of leaves, sporophylls and sporangia unfortunately gave neither
cuticle nor spore coats, so both fragments of the specimens and their impressions were prepared for high-
resolution studies by scanning electron microscope. Both ‘Silflo’ dental rubber and cellulose acetate
(nitrocellulose dissolved in amyl acetate) pulls were used to prepare these impressions. Specimens were coated
with gold using an Emscope sputter coater prior to examination with a Cambridge S600 Scanning Electron
Microscope.

The specimens will be deposited in the Beijing Graduate School, China Institute of Mining, Beijing, China
(numbers prefixed by GP).

HISTORY OF RESEARCH

Tingia. This genus was instituted by Halle (1925) for some Permian leafy shoots from northern China, which
had previously been described by Schenk (1883) under Pterophyiium Brongniart. Halle placed these specimens
in a new genus as they had anisophyllous shoots with four ranks of leaves, unlike the simpler two ranked

GAO AND THOMAS: CHINESE PERMIAN TINGIOST ACH Y ANS

817

TEXT-FIG. 2. Tingia elegans Konno. A, two larger leaves showing lines of crescent-shaped depressions in the
furrows situated between the veins, GP0089, x 5. b, part of GP0089 showing the smaller leaves; the numbers
indicate their acropetal order. L indicates impressions of the larger leaves, x 5. c, part of GP0089 showing
the arrangement of the crescent-shaped depressions, x 20. D, schematic drawing of part of figure c showing
crescent-shaped depressions within a furrow, each depression being down towards the leaf apex. Arrows
indicate the positions where stomata are observed normally, x 5. e g, schematic sections through the crescent-
shaped depressions as indicated in d. The thick line represents the compression; hatched area the underlying
rock; dotted areas the portions of the sediment trapped in the depression during the splitting of the shale.

The arrow in e is directed towards the leaf apex.

arrangement of leaves in Pterophyllum. Halle’s original diagnosis of Tingia was 'Dorsiventral, frond-like
anisophyllous shoots with thick axis. Leaves apparently arranged in four rows, two on the upper and two
on the lower side of the axis. Leaves of the two rows on one (the upper?) side smaller, directed forward at
narrow angle to the axis, those of the other two rows (on the lower side?) larger, spread out in one plane
and forming a more open angle to the axis: each lateral half of the shoot thus provided with two rows of
dissimilar leaves. Leaves of the larger (normal) type varying from broadly cuneate-obovate to oblong or
linear, with entire lateral margins but more or less deeply lobed at the apex. Several veins entering each leaf,
dichotomizing mostly in the lower part of the leaf, all branches continuing to the apex.’ Originally Halle
diagnosed two species— T/ng/a carbonica (Schenk) and T. crassinervis sp. nov. Subsequently Halle (1927)
added T. partita sp. nov. on revising the genus. Previously he had been unsure of the orientation of the
shoots, but in 1927 stated that the two ranks of smaller leaves were on the lower side of the shoot. He also
pointed out for the first time that the smaller leaves were often deeply dissected.

Konno (1929) described two further species, T. hamaguchii and T. elegans together with new specimens of
T. partita and T. (cf.) carbonica from the Lower Permian of North Korea. He revised Halle’s generic diagnosis
by adding several new morphological characters. The foliage shoot was shown to become gradually narrower
towards the base, while ‘contracted at the top into a pointed or bluntly pointed apex’. The axis was ‘straight
or slightly falcate, unbranched, gradually thinned at the upper and abruptly thickened at the lower end, with
several longitudinal striations’. Larger leaves were described as highly variable in form, generally alternate,
although sub-opposite in their lower part, and more crowded in the upper region of the shoot. The smaller
leaves were generally of a much more regular form.

Other specimens have been described subsequently from the Permian of northern China, the far north of
North Korea, and Malaysia. They are T. laciniata Kawasaki (1934), T. kikkawai Kawasaki (1934), T. gerardi
Stockmans and Mathieu (1939), T. trUobata Stockmans and Mathieu (1939), T. minor Konno and Asama
(1951), T. acuminifissa (Krasser) Bohlin (1971), T. subcarbonica Konno, Asama, and Rajah (1971), T. oblonga

818

PALAEONTOLOGY, VOLUME 30

(Sze) in Gu and Zhi (1974), and T. yichiianensis Feng (in Feng et al. 1977). However, it is important to realize
that amongst these only T. subcarbonica, T. yicinumeiisis, and T. laciniala were described as having the
characteristic four ranks of anisophyllously arranged leaves. Konno et al. (1971) did not describe the smaller
leaves on their T. subcarbonica nor mention them in their diagnosis. Unfortunately, their illustrations are
also not convincing enough to prove that the specimens really do possess these smaller leaves. The other
species seem to have been included in Tingia on the basis of the apparent similarities of their overall
appearance to Halle’s specimens; that is on the features of their larger leaves.

Similar specimens found outside Asia were described by Darrah (1938) as belonging to Tingia. However,
these were later shown by Mamay (1968) to be without the two ranks of smaller leaves and he then removed
them to his new genus Russellites. They were subsequently moved to Yiiania by Du and Zhu (1982).

There is clearly a difficulty in identifying specimens as Tingia when the main generic character of possessing
small leaves may be simply concealed by the alignment of the specimens. This problem is returned to later.

Tingiostachya. T. tetralocularis was named by Konno (1929) for some specimens from the Lower Permian
of the Jido Series in the far north of North Korea. These specimens, described as reproductive organs, were
found associated with either Tingia hamaguchii or T. elegans. Konno’s original diagnosis of the genus was
‘Fertile shoot (in the geno-type, T. tetralocularis) consists of two parts; a long(?) slender axis and cylindrical
terminal cones. Axis slender, forked dichotomously at apex, longitudinally ribbed with small uniform leafy
scales. Cone terminal, developed on each of the dichotomized branchlets of the axis, with numerous
sporophylls in four vertical series. Sporophylls uniform, more or less elongated, but only slightly modified
from the foliage leaf in Tingia. A large tetralocular synangium, hemispherical, placed directly on the upper
surface of each bract, apart from the cone-axis, receiving one group of bundles from the axis which runs
along the median zone of the bract.’ Konno described the spore as ‘often well preserved, usually of obovate
contour, with the larger diameter from 150 /un to 130 /<m’.

Another two specimens of Tingiostachya have been described by Stockmans and Mathieu (1939) from
Kaiping, northern China, although both cones were thought to be specifically indeterminable. The sporophylls
of one of these cones (associated with their new foliage species Tingia trilobata) seem lobed like the ordinary
large leaves of Tingia, although they were much smaller and not clearly discernible. Using a pull technique,
Stockmans and Mathieu recovered triradiate spores from both cones. These were very similar, about 100 /;m
in diameter, with smooth surface and said to resemble closely the microspores of Noeggerathiostrobus figured
by Nemejc (1935).

Examination of the new Taiyuan material has yielded new information that permits a re-evaluation and a
rediagnosis of the two genera. Because we have no conclusive evidence that Tingiostachya came from the
same parent plant as Tingia we propose to follow the classification system proposed by Thomas and Brack-
Hanes (1984) in their discussion of the lycophytes. They proposed that families should be based on
reproductive characters alone, with other organs only included as so-called satellite taxa. This allows a
classification system to be constructed to show the most probable relationships between fertile and sterile
organs without broadening the family definition to include isolated sterile organs of doubtful affinity. We
therefore propose that the family Tingiostachyaceae be used to include the one genus Tingiostachya. Tingia
should be taken as a satellite taxon within the Tingiostachyaceae.

SYSTEMATIC PALAEONTOLOGY
Family tingiostachyaceae fam. nov.

Diagnosis. Pedunculate cones with spirally attached sporophylls. Single sporangia attached to
proximal part of the adaxial surface of sporophylls with upwardly extended lateral margins
(alations).

Genus. Tingiostachya Konno.

Satellite taxon. Tingia Halle.

Genus tingiostachya Konno (1929)

Type species. Tingiostachya tetralocularis Konno (1929).

Emended diagnosis. Fertile shoot consisting of a peduncle, with spirally arranged leaves and a
terminal cone once dichotomized at its base. Leaves on peduncle ensiform, spirally arranged. Cone

GAO AND THOMAS: CHINESE PERMIAN TINGIOSTACH Y ANS

819

axis slender. Sporophylls spirally arranged, pedicels arising at right angles, with upwardly extending
lateral margins (alations). Laminae lanceolate, parallel to cone axis. Sporangia spheroidal, attached
to proximal part of adaxial surface of pedicel. Sporangia with small spores.

Tingiostachya telralocularis Konno (1929)

Plate 89, fig. 10; Plate 90, figs. 1 7; text-fig. 3a-e

1929 Tingiostachya tetralocularis Konno, p. 120; pi. XXIII, fig. 5c; pi. XXIV, figs. 4 and 5;
pi. XXVII.

Emended diagnosis. Cone about 13-0 cm long, individual cones 8 0-12 0 mm broad. Sporophyll
pedicel about 0-5 mm high and 3-0 mm long with upward lateral extensions (alations) about 1-5
mm high. Laminae 5-7-0 mm long, 2-3-0 mm broad. Sporangia 10-1-5 mm in diameter. Spores
20 /rm in diameter. Peduncle 2-3 0 mm broad; Leaves ensiform about 9-0 mm long and 1-0 mm
broad.

Neotype. GP0094, from the Lower Permian, 5 km north-east of East Hill Mine, Taiyuan, Shanxi province,
northern China; the whereabouts of the figured specimens of Konno (1929) are unknown (K. Asama, pers.
comm.)

Distribution. North Korea: Anji-ri, Daido-gun, S. Heiando (Jido Series and Kobosan Series, Permian). China:
East Hill, Taiyuan (Lower Shihezi (Shihhotse) Series, Lower Permian).

Descriptions oj new specimens. Those from Taiyuan suggest a rather dififerent morphological interpretation
to that proposed by Konno (1929). They are still interpreted as strobili, but their sporophylls are clearly
arranged spirally instead of in whorls of four as suggested by Konno. His conclusion that they possessed

TEXT-FIG. 3. Tingiostachya tetralocularis Konno. A, outline drawing showing the apical arrange-
ment of the sporophylls. GP0094, x 2. b, single sporophyll from GP0095 showing a sporangium,
x20. c, single sporophyll from GP0095 showing the two alations (u = upper; 1 = lower; la =
lamina; p = pedicel), x 15. d, schematic cross-section (x-y) of lamina in c showing the
compression as embedded in the shale. E, schematic cross section of the same sporophyll, as in

life.

820

PALAEONTOLOGY, VOLUME 30

their leaves in fours was possibly influenced by the fact that Tingia shows four ranks of leaves. It is hard to
imagine how the fossilization of an organ having the whorled construction shown by Konno (1929, text-
fig. 4a) could possibly result in the appearance of the compression shown in his plate XXIV, fig. 4 and plate
XXVII, fig. 2. These photographs show strobili with sporophylls appearing to be in spirals.

Among the new specimens of Tingiostachya from Taiyuan, the three most complete show a slender axis,
with elongated leaves (PI. 90, fig. 1), bearing a terminal cone that dichotomizes once just above its base (PI.
90, figs. 1-3). One shows the cone apex (PI. 89, fig. 10) to be rounded with the terminal sporophylls gently
curved around it. No cone is complete; the longest is 6-8 cm. Their widths are all between 8-0 and 12 0 mm.
The axis of the cone is slender, 2-3 0 mm broad and longitudinally ribbed. The sporophylls are all regularly
arranged in a spiral on the axis (PI. 90, fig. 3; text-fig. 3a). Their pedicels are at right angles to the axis and
about 3 0 mm long. The laminae turn abruptly upwards to be parallel to the axis and there is a very small
heel protruding downwards from the distal end of each pedicel. The laminae are lanceolate-triangular in
outline with acute apices, about 6-0 mm long and up to 1-8 mm broad in their basal region, and have entire
margins.

Those specimens which are split longitudinally and roughly through the middle reveal certain undescribed
features of sporophyll construction. In some laterally compressed sporophylls a layer of light coloured
sediment can be seen between two layers of compression material (PI. 90, fig. 5; text-fig. 3c-e). This suggests
that the sporophyll pedicel was not flat, but that its sides were extended and curved upwards. These extensions
appear to be approximately equal to, or slightly more than the height of the pedicel. The rounded sporangia
(PI. 90, figs. 4, 6, 7; text-fig. 3b) are about 11 L6 mm in diameter, being probably compressed from an
originally spheroidal shape. Each appears to be attached individually to the proximal part of the adaxial
surface of a sporophyll pedicel. They were probably protected by the sporophyll alations in life but revealed
in the fossil by the splitting of the compression. Scanning electron microscope observation of several
impressions of sporangia showed many flattened circular structures which were all approximately 30 /im in
diameter. These are interpreted as representing casts, or possibly moulds, of spores. Unfortunately no
structural details nor features of surface ornamentation could be seen on any of these putative spores. Some
sporangia also possess what appear to represent impressions of the sporangial wall cells. These are about
20 /;m broad and radially elongated around the edges of the sporangia.

There are small circular marks on the cone axis which are very similar in size to the sporangia, but are
flat and rough in appearance unlike the raised and smooth circles, with concave centres, of the sporangia.
They are best interpreted as false pedicel scars formed by the forced removal of the sporophylls during the
splitting of the matrix. Judging from the appearance of the present specimens, Konno could have mistaken
some of these false scars for sporangia, for he described the sporophylls as having four sporangia, or a
tetralocular synangium attached to the upper surfaee.

Comparisons. The main means of identifying reproductive organs as Tingiostachya are the way in
which the terminal cones dichotomize once at their very base, their spirally arranged sporophylls,
and their spheroidal sporangia attached to the proximal part of the adaxial surface of the sporophyll
pedicel. There are no other described genera that can be thought to be closely comparable.

EXPLANATION OF PLATE 89

Figs. 1-9. Tingia elegans Konno. 1, GP0087, leafy shoot showing acropetal overlapping arrangement of the
deeply lobed larger leaves and their reduction in size towards the base of the shoot, x 2. 2, GP0088, leafy
shoot showing basipetal overlapping arrangement of leaves and basal part of the shoot, x 1-5. 3, SEM
of a rubber peel of a leaf impression from GP0096 showing the furrows between the veins, x 40. 4, SEM
of an amyl acetate peel of the lower surface of a larger leaf compression from GP0097 showing the
epidermal cells over a vein, x 200. 5, GP0090,the apical part of a leafy shoot showing size reduction
towards the apex of the shoot, x 2. 6, GP0089, leafy shoot showing smaller leaves, x 5. 7, GP0098, a
larger leaf showing its toothed upper margin and apex and its dichotomous venation, x 8. 8, GP0046,
part of a leafy shoot showing the smaller leaves and the overlapping larger leaves, x 2. 9, SEM of a rubber
peel of impression showing a putative stoma on the edge of furrow on the lower surface of leaf from
GP0096, x700.

Fig. 10. Tingiostachya tetralocularis Konno, GP0091, the apex of a cone with its spirally arranged and
incurved sporophylls, x 10.

PLATE 89

GAO and THOMAS, Tingia, Tmgiostachya

822

PALAEONTOLOGY, VOLUME 30

Noeggerathiostrobus Nemejc (1928) has often been thought to be quite similar but this was
through the mistaken interpretation that Tingiostachya possessed tetralocular synangia. In fact
Noeggerathiostrobus has a single terminal cone with its bract-like units arranged in semicircular
discs around an axis. The sporangia are located on the upper surface of the discs in rows and have
been shown to produce two sizes of spores (Nemejc 1928; Halle 1954).

As mentioned earlier, the only named species of the genus is T. tetralocidaris Konno (1929). As
the Taiyuan specimens appear very closely comparable in size and in sporophyll morphology to
Konno’s figured specimens, we refer them to his species. The specimens of Stockmans and Mathieu
(1939, 1957) are much smaller so we prefer to continue to keep them separate.

Genus tingia Halle (1925)

Type species. Tingia carbonica (Schenk) Halle (1925)

Emended diagnosis. Dorsiventral, anisophyllous, unbranched foliage shoots; two rows of larger
leaves becoming smaller towards apex and base, two rows of smaller leaves. Larger leaves obovate
to linear with toothed apical margins and decurrent bases, alternately arranged on axis, spreading;
veins dichotomizing, ending in marginal teeth. Smaller leaves almost parallel to shoot axis.

Tingia elegans Konno (1929)

Plate 89, figs. 1-9; text-fig. 2a-g

Emended diagnosis. Larger leaves obovate, about 10 mm long and 6 mm broad; apex rounded
with teeth about 1 mm long; lower margin straight with decurrent base; veins dichotomize twice
or more. Epidermal cells longitudinally elongate along the veins and approximately 100 pm long
and 1 5 pm wide. Stomata visible on edges of veins, approximately 25 /<m long and 20 pm broad.
Smaller leaves about 4 mm long, 1 mm wide.

Neotype. GP0087, from the Lower Permian, 5 km north-east of East Hill Mine, Taiyuan, Shanxi province,
northern China. The whereabouts of the figured specimens of Konno (1929) are unknown (K. Asama, pers.
comm. 1986).

Distribution. Riajin-ri, Daido-gun, south Heiando, North Korea (Jido Series and Kobosan Series, Permian).
East Hill, Taiyuan, China (Lower Shihezi Series, Lower Permian) and all the other localities indicated for
the genus on the accompanying map.

Descriptions of new specimens. About forty specimens collected by Gao from the East Hill region of Taiyuan
were studied. Most are compressions with many lacking their counterpart impressions. Unfortunately all are
fragmentary with only one showing an apex (PI. 89, fig. 5) and three their basal regions (PI. 89, figs. 1 and
2). The longest fragment of shoot is 12 cm, although we estimate their original lengths to be in the order of
20 cm or more. The axes are longitudinally striated and vary in width from 1 mm to 3 mm in different
speeimens. The shoots are dorsiventral and anisophyllous with two ranks of larger leaves and two ranks of
mueh smaller leaves (PI. 89, figs. 6 and 8).

It is difficult to see the arrangements of both kinds of leaves on most specimens. There are two ways in
which the larger leaves overlap each other on the shoot. If the overlapping is directed towards the apex

EXPLANATION OF PLATE 90

Pigs. 1-7. Tingiostachya tetralocularis Konno. 1, GP0093, a once dichotomized, longitudinally split cone on
its leafy pedicel, x 1-5. 2, GP0093, a once dichotomized cone on its leafy pedicel, x 2. 3, GP0094, a once
dichotomized eone showing its spirally arranged lanceolate sporophylls, x 2. 4, GP0093, single sporangium,
X 30. 5, GP0095, single isolated sporophyll (drawn as text-fig. 3c), x 10. 6, GP0095, single sporophyll
with a sporangium (drawn as text-fig. 3b), x 10. 7, GP0094, part of sporangium hidden under the

overlapping compression of the sporophyll alation, x 20.

PLATE 90

GAO and THOMAS, Tingiostachya

824

PALAEONTOLOGY, VOLUME 30

(PI. 89, fig. 1), that is with the lower margin of each leaf being hidden by the upper margin of the subsequent
leaf, the smaller leaves are angled down into the matrix where they are hidden unless uncovered. In those
specimens where the overlapping is directed towards the base the smaller leaves are directed upwards.
However, only one specimen has been found that shows both types of leaves in this arrangement (PI. 89, fig.
8). Most have lost their smaller leaves which were ripped off with the counterpart portion of matrix during
the splitting of the shale (PI. 89, fig. 2).

The larger leaves are usually obovate, sometimes slightly elongate in form, with their oblique and decurrent
bases attached to the axis by short petioles. The length of the larger leaves varies from 8 to 15 mm and the
median width from 4 to 7 mm in different specimens. The angle of the lower leaf margin to the axis varies
from 40° to 80°; being constant within a specimen but varying between them. Occasionally the two ranks
are attached at different angles, although we attribute this to compression effects during preservation. The
lower margin of the leaves are straight and the upper margins somewhat semicircular. The apical portion of
the upper margin and the rounded apex are toothed with the longest teeth in the former area (PI. 89, fig. 7).
The veins are usually dichotomizing and terminate either singly, or very occasionally in pairs, in the marginal
teeth. There are about ten to eighteen veins in the broadest part of the leaf and about the same number or
slightly more reach the margin.

The splitting of the shale has resulted in two basic kinds of specimens. One shows the larger leaves as dark
compressions with either discontinuous light furrows and/or more occasionally discontinuous light or dark
ridges. The other kind show the larger leaves as impressions with discontinuous light ridges and/or light or
dark furrows running along their length. These latter ridges and furrows equate to those visible on those
leaves preserved as compressions. These furrows are positioned between the veins and are most obvious in
the middle and apical parts of the leaves. (PI. 89, fig. 3; text-fig. 2a). Distinctive features have been observed
in the furrows of some specimens in which the leaves overlap each other towards the apex. These furrows
consist of many closely packed but individual depressions; each being crescent-shaped with its convex side
directed towards the leaf apex (text-fig. 2a, c, d). These crescent-shaped depressions are also angled down
towards the leaf apex (text-fig. 2 e-g) and very occasionally they are joined together in the middle. Scanning
electron microscope observations of epoxy and rubber pulls from impressions of the leaves show epidermal
cell outlines (PI. 89, fig. 4) that are longitudinally elongated along the veins and about 50 150 /im long and
8-32 |zm wide. Stomata are also visible on the sides of the furrows, that is on the very edges of the well-
preserved veins. We suspect that there were many stomata in the crescent-shaped depressions, but the
retention of quantities of rock matrix in these depressions prevents us from making the necessary observations.
The probable guard cell-like structures (PI. 89, fig. 9) that can be seen are about 25 /mi long and 10 /m
broad. Regular hollows, mostly in the furrows, that have similar sizes to the stomata are most probably
evidence of stomata in which the guard cells were not preserved.

The smaller leaves are ensiform with tapering margins, 7 mm long, 1-2 mm wide, and directed apically
along the axes (PI. 89, figs. 6 and 8; text-fig. 2b).

The basal part of the shoot shows a rather different arrangement to the main leafy part. The larger leaves
become basipetally smaller and are arranged more acutely to the axis. The smallest are 3 0 mm long, 1-5 mm
broad, and dissected into three lobes, thereby appearing very similar to the smaller leaves in both size and
shape.

Comparisons. The generic identification of Tingia is based on the dorsiventral and anisophyllous
arrangement of its shoots with two ranks of larger leaves on one side of the shoot and two ranks
of much smaller leaves on the other. The toothed apex and apical part of the upper margin of the
larger leaves and the numerous dichotomizing veins that terminate in the teeth are also characteristic
features. The outwardly similar genera Noeggerathia Sternberg, Russellites Mamay, and Plagiozam-
ites Zeiller with their dorsiventrally flattened large leaves are not anisophyllous as they possess no
ranks of smaller leaves. As the new Taiyuan specimens are both dorsiventral and anisophyllous
they clearly belong to Tingia.

Fifteen species of Tingia have been described to date, excluding the two removed by Mamay to
Russellites. However, only eight are known to have the anisophyllous arrangement of four ranks
of leaves. The morphological details of all fourteen species are summarized in Table 1 as described
by their authors. The closest comparable species to the new specimens are T. hamaguchii Konno
(1929), T. partita Halle (1927), and T. elegans Konno (1929).

T. hamaguchii has similar shaped leaves to the new specimens although they are about twice the
size. It also has a looser leaf arrangement (about 1 leaf per cm instead of 1-3-3 leaves per cm as

GAO AND THOMAS: CHINESE PERMIAN TINGIOSTACH Y ANS

825

.2

&

aj

&

aj

a

s

o

u

O

O

4=

c/:)

O 4=

C T3

a

<

4= „
£

E

Oh

4:2

OH'S
< 2

^

3 6

"o ^

o 2
H S

ri

tu

ol 34

o S3 u

c 3 c ^ ^

i2 o io o 2 o

<D G (U ^ <U 03 ^

cn r_) t/5 ^ ^ O ^

3 c 2 g 5 c g

- ^ - o-g 2 o

u O -»-• s-H

•X- ITT o ^

0(^)rn —

O 0^ I I I o

O AOi/^OOrOf^
^ /\ r-1 ^ ^ uo — ^

c c

■ S 2 0,

<D

gig'g 230^^3
c

o o

X) X» X) -D X>

oooooiSxo

i:: X

[ O 1*0

O O

^ r^i (N r-

I I \ ^ C'i *

o o A A

rt r-j — < /N m

O

c

X
o

^ ? rT

o s;

c/3

s «

C3 ^

*s; ro

s

&c S «
a 5 ■=
s bo-S

I a

-s:

a

C/!}

(T-- C-- IT- (T- (T- O- C^-

g -3

c g 3 3

2 o

^ ^ c3

C
D

O "^ * *

^ ^ uo 00

« cd
cd ^ 3 ^
^ cd (Di) cd
> C >
cd ^

.& o

D

O O
■*d-

un O O O O
rj- m X 1/^ ^

c^. (N

o
r-^ ^

I I r-- * r-- *

r- o 10 10

C-. ^ m r-^ m

On

QJ P

CX

C/D

c/3 <3

(U

^ -O ^

g-s §
s

00

g §

o -S:
-o S
o K

E^' E**s

*0

c

=5

o

— ^

Oi) =5 =S
G 0£) 04)
O C C
^ Cd cd

O O -b -b

cd

>

O

X

o

o

X

r-4

•S

s

a

Oh

(1) Halle 1925. (2) Halle 1927. (3) Konno 1929. (4) Kawasaki 1934. (5) Konno et al. 1971. (6) Stockmans and Mathieu 1939. (7) Konno and
Asama 1951. (8) Gu and Zhi 1974. (9) Bohlin 1971. (10) Feng et al. 1977.

* Not mentioned in published description, data from photographs.

826

PALAEONTOLOGY. VOLUME 30

in the Taiyuan specimens) and the leaves are more closely aligned to the axis (their lower margins
being at 25°-50° instead of the 40°-80° in the Taiyuan specimens). T. partita differs from the
present specimens in having larger and more triangular-shaped leaves with more divergent veins
and deeply dissected apices that are most closely comparable to T. elegans (which was included as
a synonym of T. hamaguchii by Lee et al. (Gu and Zhi 1974)), and in having similar leaf
morphologies and leaf densities (Konno’s specimens average 2 leaves per cm). The figures of
Konno’s earlier specimens seem to show more elongated leaves (Konno 1929, pi. XXV, figs. 1~8),
whereas Konno and Asama’s later specimens appear virtually identical to ours (Konno and Asama
1951, pi. 9, figs. 5-7). The differences in appearance of the leaf shape of Konno’s original specimens
can be interpreted as due to the different manner of leaf overlap. Here it is the upper margin which
is hidden by the leaf above, thereby producing an elongated appearance to the otherwise obovate
leaf.

DISCUSSION

Tingia and Tingiostachya have been classified in various ways by different authors. This has resulted
from the varied interpretations and preferred emphasis of vegetative or reproductive characters.
Furthermore, it has been generally assumed that the two organs came from the same parent plant
because of their consistent association at different sites, even though they have never been found
in organic attachment to each other or to any other organ. Indeed, as they have never been found
other than as isolated organs we have no knowledge of the overall morphology of the parent plant.

Halle (1925) proposed Tingia to be an aquatic plant with the two ranks of larger leaves floating
on the surface and the two smaller leaves submerged. Konno rejected Halle’s proposal, suggesting
instead that Tingia grew in the same manner as extant Selaginella and Lycopodium although he
did note that it was much larger and never branched. He concluded that Tingia shoots were most
probably foliage branches given off laterally or almost horizontally on the ground from a creeping
stem.

From the morphological characters of the Taiyuan specimens we conclude that they represent
long shoots, rather than pinnate frond-like organs, that were shed from woody plants. They were
most likely to have been orientated with their smaller leaves uppermost as in the anisophyllous
species of Selaginella. This we deduce from the fact that it is the surfaces of the larger leaves
furthest away from the smaller leaves that have the crescent-shaped grooves. If these really are
stomatal grooves then they are virtually certain to have been on the underside of the leaves. The
larger leaves also appear to be quite thick and this together with their possession of ‘stomatiferous’
crescent-shaped depressions suggests that they were quite fleshy. Furthermore, it can be taken to
suggest that the plants were growing in a relatively dry environment.

Halle (1927) suggested that there was a close relationship between the genera Tingia, Noeggeratlna,
and Plagiozamites. He therefore proposed that Tingia be used as the basis for the family Tingiaceae
that included these three genera. Halle of course had found no reproductive organs. Konno
compared Tingiostachya with the Sphenophyllales and Lycopodiales before classifying it with the
extant Psilotaceae on account of the presumed tetralocular synangia even though he noted several
major morphological differences. Nemejc (1931) subsequently rejected Konno’s interpretations of
the plant’s affinity and followed Halle in grouping Tingia, Noeggeratlna, and Plagiozamites together
on the basis of common foliage organization, but placed them in the Noeggerathiales instead of
using the family Tingiaceae.

Browne (1933) suggested the affinities of Tingia and Tingiostachya should lie with the
Sphenophyllales. This was based on the supposed verticillate arrangement of the sporophylls in
Tingiostachya and that the leaves of Tingia were similar to the undivided wedge-shaped leaves of
the Euramerian genus Sphenophyllum. Browne also suggested that the Psilotales should be included
in the Sphenopsida.

Stockmans and Mathieu (1939, 1957) and Lee (1963) did not attempt to put either Tingia or
Tingiostachya into a family but included them both as incertae sedis.

GAO AND THOMAS: CHINESE PERMIAN TINGIOSTACH VANS

827

Most recent accounts accept some relatively close relationship between Tingia, Tingiostachya,
and Noeggerathia, although there is differing opinion on their classifications. Boureau (1964)
includes the Tingiales {Tingia and Tingiostachya) in the Noeggerathiophytina. Bohlin (1971) and
Lee et al. {in Gu and Zhi 1974) both take a wide view of the Noeggerathiales as containing Tingia
and several other leaf genera including Plagiozamites, Yuania, and Concophylluni. Lee et al. (Gu
and Zhi 1974) also included the reproductive genus Discinites and Bohlin (1971 ) the other leaf genus
Ginkgophyllum. Taylor (1981) merely points out the similarity between Tingia and Noeggerathia and
includes them both in the progymnosperms. Beck (1981) suggested a relationship between Tingia,
Noeggerathia, and Archaeopteris including the former two in the Noeggerathiopsida.

Both Tingia and Tingiostachya appear to us to suggest a quite unique plant; assuming of course
that they really did belong to the same parent plant. We do not accept that there can be any
meaningful classification system established on the basis of rather broad similarities of vegetative
organs. Neither do we believe that there is any close relationship between Noeggerathiostrobus and
Tingiostachya as we have reinterpreted it here. Indeed, we cannot even say if the parent plant was
a pteridophyte, a gymnosperm, or even a progymnosperm.

Acknowledgements. We thank Professor Tian Baolin (Beijing Graduate School, China Institute of Mining,
Beijing) for kindly donating some of the specimens and Dr Dianne Edwards (Department of Plant Science,
University College, Cardiff ) for her assistance and helpful comments. Gao also gratefully acknowledges the
receipt of a Chinese Government student grant which made this work possible.

REFERENCES

BECK, c. B. 1981. Archaeopteris and its role in vascular plant evolution. In niklas, k. j. (ed.). Paleobotany,
paleoecology and evolution, 1, 193-220. Praeger Publishers, New York.

BOHLIN, B. 1971. Late Palaeozoic plants from Yuerhhung, Kansu, China. The Sino-Swedish Expedition,
Publication 51, 150 pp. Stockholm.

BOUREAU, E. 1964. Trade de paleobotanique. III, Sphenophyta, Noeggerathiophyta, 544 pp. Masson et Cie,
Paris.

BROWNE, I. 1933. The Noeggerathiae and Tingiae. The effects of their recognition upon the classification of
the Pteridophyta: An essay and a review. New Phytol. 32, 344-358.

CHALONER, w. G. and MEYEN, s. V. 1973. Carboiiiferous and Permian Floras of the northern continents. In
HALLAM, A. (ed.). Atlas of palaeobiology, 169-186. Elsevier Publishing Co., Amsterdam.

DARRAH, w. D. 1938. The occurrence of the genus Tingia in Texas. Bot. Mus. Leafl. Harv. Univ. 5, 173-188.

DU, X. and ZHU, j. 1982. The emendation of a cycad genus Yuania and the discovery of Y. chinensis sp. nov.
Mem. Beijing Nat. Hist. Mus. 17, 1-6.

FENG, s., CHEN, G., XI, Y. and ZHANG, c. 1977. Plants. In hupei institute of geological science and other
INSTITUTES (eds.). FossU Atlas of Central-South China, II, 622-674. Geological Publishing House, Peking.
[In Chinese.]

GU and ZHI. 1974. Chinese Plant fossils, I. Palaeozoic Plants from China, 277 pp. Sci. Press, Peking. [In
Chinese.]

HALLE, T. G. 1925. Tingia, a new genus of fossil plants from the Permian of China. Bull. geol. Surv. Peking,
China, 173-188.

1927. Palaeozoic plants from central Shansi. Palaeont. Sin. Ser. A, 2, 1-317.

1954. Notes on the Noeggerathianeae. Svensk hot. Tidskr. 4, 368-380.

KAWASAKI, s. 1934. The flora of the Heian system. Bull. geol. Surv. Chosen, 6, 47-31 1; Atlas (1931), pis. XV-
XCIV.

konno, e. 1929. On genera Tingia and Tingiostachya from the Lower Permian and Triassic Beds in Northern
Korea. Jap. J. Geol. Geogr. 6 (5/6), (3-4), 113 147.

and ASAMA, K. 1951 . On the genus Tingia Halle 1925 from Permian beds in south Manchuria and Shansi,

China. Short Pap. Inst. Geol. Palaeont. Tdhoku Univ. 3, 59-69.

and RAJAH, s. s. 1971. The later Permian Linggin Flora from the Gunong Blumut Area, Johore,

Malaysia. In kobayashi, t. and toriyama, o. (eds.). Geology and palaeontology of southeast Asia, 9, 1 86.
University of Tokyo Press, Tokyo.

828

PALAEONTOLOGY, VOLUME 30

KRASSER, F. 1900. Die von W. A. Dorutsehew in China und central Asia 1893-1894 Gesammelta fossilen
Pflanzen. Denkschr. Akad. Wiss. Wien. 70, 139-154.

LEE, H. H. 1963. Fossil plants of the Yuemenkou Series, North China. Palaeont. Sin. 148, New Ser. A(6),
1-185.

1964. The succession of Upper Palaeozoic plant assemblages of Northern China. C. R. Congr. Av. Et.

Strut. Geol. Carbon^ 5me, 2, 531-537.

LI, X. (LEE, H. H.) and YAO, z. 1982. A review of recent research on the Cathaysia flora in Asia. Am. J. Bot.
69, 479-486.

1983. Carboniferous and Permian flora provinces in East Asia. In lu, y. (ed.). Palaeobiological

provinces of China, 74-82. Science Press, Beijing. [In Chinese.]

MAMAY, s. H. 1968. Russellites, new genus, a problematical plant from the Lower Permian of Texas. Prof.
Pap. U.S. Geol. Surv. 591-1, 115.

NEMEJC, F. 1928. The Noeggerathiae and Archaeopterides in the Coal Basins of central Bohemia-Palaeontogra-
phica Bohemiae. Ceska Akad a Umeni, 12, 47-82.

1931. The morphology and systematic relations of the Carboniferous Noeggerathiae with regard to the

‘genera’ Tingia and Plagiozamites of Eastern Asia. Preslia, 10, 111114.

1935. Notes on the spores and leaf cuticles of Noeggerathia foliosa Stbg. Bull. int. Acad. Sci. Boheme,

61 63.

SCHENK, A. 1883. Pflanzen aus der Steinkohlen Eormalion. In Richthofen, f. von. China, 4, 221-244. Berlin.
STOCKMANS, F. and MATHIEU, F. F. 1939. La for e paleozoicjiie du bassin Houiller de Kaiping (China), 49-120.
Musee royal d'Histoire naturelle de Belgique, Bruxelles.

1957. La flore paleozoi'que du Bassin Houiller de Kaiping (Chine) (Deuxieme partie). Association

pour r etude de la paleontologie et de la Stratigraphie Houilleres. Pub. 32, 89 pp.

TAYLOR, T. N. 1981. Pcileobotany: an introduction to fossil plant biology, 589 pp. McGraw-Hill, New York.
THOMAS, B. A. and BRACK-HANES, s. D. 1984. A new approach to family groupings in the lycophytes. Taxon,
33, 247-255.

GAO ZHIFENG

Department of Plant Science
University College
Cardiff CFl IXL

B. A. THOMAS

Typescript received 14 August 1986
Revised typescript received 2 Eebruary 1987

Department of Botany
National Museum of Wales
Cathays Park
Cardiff CFl 3NP

NEW CRETACEOUS FISH FOSSILS
FROM SEYMOUR ISLAND,
ANTARCTIC PENINSULA

by LANCE GRANDE and SANKAR CHATTERJEE

Abstract. Based on Late Cretaceous fossils from Seymour Island, hexanchiform sharks and a beryciform
teleost are reported from the Antarctic region for the first time. The hexanchiform is a species of Noticlanockm,
and the beryciform is a new genus and species of Trachichthyidae. Fragmentary orthacodontid teeth
(Sphenodiisl sp.) are also reported. An abundance and variety of indeterminate teleost bone fragments (e.g.
isolated teeth and centra) were also found. During the Late Cretaceous and Early Tertiary of the Antarctic
region, the diversity of fossil teleosts is much less apparent than that of the fossil sharks; but this is probably
only an artefact of preservation, because skeletal fragments (isolated teeth in particular) normally are much
more diagnostic (i.e. identifiable) for chondrichthyans than they are for teleosts.

Seymour Island, located on the north-east side of the Antarctic Peninsula, is potentially of
great significance to studies on the history and biogeography of the region’s fish fauna. Cretaceous
and Tertiary deposits there include the only known fossil teleosts, neoselachians (‘modern’
chondrichthyans), and holocephalans from the Antarctic region. Nearly all of these fossils belong
to extant families or higher taxa that do not live in the Antarctic region today (i.e. Odontaspididae,
Lamnidae, Squalidae, Pristiophoridae, Squatinidae, Myliobatidae, Chimaeriformes, Siluriformes)
(Grande and Eastman 1986). Two of the Cretaceous fossils described here also belong to higher
taxa not found in the region today (Hexanchiformes and Beryciformes).

There is relatively little higher taxonomic diversity in the Recent fish fauna of the Antarctic
region. The fauna is dominated by the percomorph suborder Notothenioidei, endemic to the region
and comprising at least 67% of the species and 90% of the individuals in the Antarctic region
(Grande and Eastman 1986; De Witt 1971). The incidence of endemism for notothenioids is 86%
at the species level with only a few species found in New Zealand, southern South America, the
Ealkland Islands, and Australia. The only chondrichthyans reported from the Recent fauna are
in the skate family, Rajidae. In the fossil faunas, to date, no rajids have been discovered, and the
only report of a notothenioid (Woodward 1908) was found to be probably in error (Grande and
Eastman 1986).

Woodward (1908) was the first to describe fossil fishes from Seymour Island, and he described
both Cretaceous and Tertiary material, consisting of isolated centra and scales. He referred some
of the material to Ptychodus sp., but this was later (Welton and Zinsmeister 1980; Grande and
Eastman 1986) referred only to Elasmobranchii indeterminate and Teleostei iucertae sedis. The
Tertiary elasmobranchs were later reported by Elliot et al. (1975), De Valle el al. (1976), Clone et
al. (1977), Welton and Zinsmeister (1980), and Grande and Eastman (1986). Grande and Eastman
also reported a Tertiary holocephalan and siluriform, and reviewed the work of all previous
workers; therefore, previous work will not be reviewed in depth here.

The only Cretaceous fish to have previously been described from Seymour Island, other than
the Woodward material described here, is the lamnid shark hums sp. (Grande and Eastman 1986:
p. 122). The specimens described below are from the same formation and locality as that
specimen.

{Palaeontology, Vol. 30, Part 4, 1987, pp. 829-837.}

© The Palaeontological Association

830

PALAEONTOLOGY. VOLUME 30

TEXT-FIG. 1. Map showing Seymour Island and Antarctic Peninsula. Circles indicate the Cretaceous fossil
fish localities. Ages of rock units: Lopez de Bertodano, Late Cretaceous; Sobral, Late Cretaceous-Palaeocene;
Cross Valley, Palaeocene; La Meseta, Eocene; Basaltic dikes, ?Late Tertiary (after Zinsmeister 1982).

GEOLOGICAL SETTING

The thick sequence of fossiliferous marine and non-marine elastics on Seymour Island represent
the most complete and well-exposed section of the Upper Cretaceous to Lower Tertiary rocks
known in the southern hemisphere. The succession has been divided into two groups: the lower
Marambio Group comprised of Lopez de Bertodano and Sobral Formations, and the upper
Seymour Island Group consisting of Cross Valley and La Meseta Formations (Bibby 1966; Rinaldi
et al. 1978; Zinsmeister 1982).

The physical setting of Seymour Island makes it in some ways an ideal place for fossil collection
in Antarctica. The island is small— about 20 km long and 9 km wide— and is virtually snow free
the year round. The oldest sequence, the Lopez de Bertodano Formation, crops out in the southern
two-thirds of Seymour Island (text-fig. 1). It consists of 1200 metres of loosely consolidated greyish
sandstones and sandy siltstones, dipping gently eastward. Concretionary horizons occur throughout
the sequence. The invertebrate fauna is diverse, consisting of ammonites, echinoids, bivalves,
gastropods, arthropods, serpulid worms, and foraminiferans, and suggests Middle Campanian to
Maestrichtian age (Howarth 1966; Spath 1953; Macellari and Huber 1982).

Recently a large collection of vertebrate fossils was made from the Lopez de Bertodano
Formation. The new material includes the remains of bony fish, sharks, plesiosaurs, mosasaurs,
pterosaurs, and possibly birds (Chatterjee and Zinsmeister 1982; Chatterjee et al. 1984).

PREPARATION METHODS

The beryciform skull described here was preserved in a hard, calcite-cemented, coase-grained sandstone. The
bone is badly fractured and much softer than the matrix, making detailed preparation extremely difficult.
Minor preparation was done with needles under a dissecting microscope after immersion in 8% formic acid
to soften the matrix.

GRANDE AND CHATTERJEE; ANTARCTIC CRETACEOUS FISH

831

NEWLY REPORTED CRETACEOUS MATERIAL

Class CHONDRICHTHYES
Subclass ELASMOBRANCHII
Order hexanchiformes
Family hexanchidae (sensii Ward, 1979)

Genus notidanodon Cappetta, 1975
Notidanodon sp.

Referred material. EMNH PF10724 (text-fig. 2a, b), FMNH PFI0725 (text-fig. 2c, d), and TTU P9178, 9180,
9182-9185 (all partial isolated tooth fragments). TTU P9182 is in a block of matrix that also has some
prismatic cartilage preserved (text-fig. 2e).

Description. The three most complete specimens are FMNH PF10724, 10725, and TTU P9182 (text-fig.
2a-e), and the following description is based on these specimens. Terminology and systematics used for
Hexanchidae follow Ward (1979). The teeth are labio-lingually compressed with apically or apico-distally
directed principal cusp and distal cusplets. The mesial cusplets are almost as large as the distal cusplets and
are mesio-basal to, and distinct and separate from, the principal cusp. Thus, following Ward (1979, p. 122)
the tooth fits the diagnosis for the genus Notidanon Cappetta 1975. One specimen (TTU P9182: text-fig. 2e)
is in a block of matrix that also has some prismatic cartilage preserved, apparently from the same animal.
Morphologically, this material most closely resembles N. lanceolatiis, N. pectinatus, and N. dentatus (see
Ward and Thies 1987) and probably belongs in one of these three very similar looking species. The material
differs from N. loozi (e.g. Ward 1979, pi. 3, fig. 8) in that the last mesial cusplet points nearly parallel with
the principal cusp (text-fig. 2a, b). The species of this genus, known only by isolated teeth, are discussed in
Cappetta, 1987.

This is the first report of a hexanchiform from the Antarctic region. The genus is also known from the
Palaeocene of Europe, USSR, and from the Upper Cretaceous of North America and New Zealand (Cappetta
1987). Based on available material, monophyly of this ‘isolated-tooth genus’ seems unsubstantiated.

Order hexanchiformes? (sensti Cappetta, 1987)

Family orthacodontidae? (sensu Cappetta, 1987)

Genus sphenodusI Agassiz 1843
Sphetwdusl sp.

Referred new material. FMNH PF10723 (text-fig. 2f, g) and TTU P9188, 9189, 9191, and 9192 (all partial
teeth missing root and base).

Description. Because of incomplete preservation, precise identification of these teeth is not possible. All are
somewhat labio-lingually compressed with sharp unserrated mesial and distal cutting edges and sharp points.
The teeth were nearly symmetrical in lateral outline, and not curved as in Isnriis. Their outer surface is
smooth, although where the shiny enamel layer is worn off the tooth is striated underneath. They closely
resemble fossil species described as Orthacodus Woodward, 1889 (e.g. see pi. 1, figs. J, n-p in Uyeno et al.
1981 and various figures in a review of the genus by de Beaumont 1960). Orthacodus ( = Sphenodus) is known
only by isolated teeth, apart from some skeletal remains from Soinhofen (Cappetta 1977), and the assignment
of this poorly known genus to Hexanchiformes is debatable (see Cappetta 1977, p. 50).

Order indeterminate
Family indeterminate

Indeterminate selachian vertebral centra (first reported by Woodward 1908, pp. 1-2)

Referred new material. TTU P9193, 9194, FMNH PF1I920, 11921 (two small and two large centra; one of
the large illustrated in text-fig. 2h, i).

Description. These centra are identical to those described by Woodward (1908), which he referred to Ptychodus
[not Ptychodus according to Welton and Zinsmeister (1980) and Grande and Eastman (1986)]. One (text-fig.
2h) is so close in size ( 10 cm diameter) and appearance to a specimen described and illustrated by Woodward
that it could possibly be from the same individual. For detailed description see Woodward 1908).

/

TEXT-FIG. 2. A-E, Notidanodon antarcti n. sp., from Upper Cretaceous deposits of Seymour Island, a, b,
holotype, nearly complete tooth (FMNH PFI0724), x 2-4. a, lateral view (anterior facing right), b, medial
view (anterior facing left), c, d, partial tooth (FMNH PF 10725), x 2. c, lateral view (anterior facing right).
D, medial view (anterior facing left). E, nearly complete tooth in matrix with prismatic cartilage preserved
(arrow) (TTU P9182), x 2, anterior facing up. This is the first report of a hexanchiform from the Antarctic
region, f, g, Orthacodiis sp., partial tooth from Upper Cretaceous deposits of Seymour Island (FMNH
PF10723), x2-3. f, anterior view. G, lateral view (anterior facing right), h, indeterminate selachian vertebral
centrum (FMNH PFl 1920), x 2-5. i, dorsal view of FMNH PFl 1920, x 2-2, from Upper Cretaceous deposits

of Seymour Island.

GRANDE AND CHATTERJEE: ANTARCTIC CRETACEOUS FISH

833

Class OSTEICHTHYES
Subclass ACTINOPTERYGII
Subdivision teleostei
Order beryciformes {sensu Zehren, 1979)

Superfamily trachichthyoidea {sensu Gayet, 1982)

Family trachichthyidae {sensu Gayet, 1982)

Subfamily trachichthyinae {sensu Gayet, 1982)

Genus antarctiberyx gen. nov.

Antarctiheryx seymouri n. sp.

Holotype. TTU P92I0. A poorly preserved anterior skull section with partial dentary attached. This is the
only known specimen.

Diagnosis. This species differs from all other trachichthyids in the extreme development of the
ornamented nasal bridge and in the pattern of radiating spines and serrations on the dorsal surface
of the bridge (text-fig. 3d, e).

Etymology. Antarcti and seymouri refer to type locality; beryx-—a fish (Latin).

Description. The specimen (text-fig. 3a-e) is poorly preserved and incomplete. It is missing the braincase and
gill-cover regions, the postcranial skeleton, and the left side of the skull. Many other bones of the skull are
either incomplete, or hidden by overlying rock or bones. Based on the limited information available (described
below), the specimen still appears to be identifiable as a trachichthyid of the subfamily Trachichthyinae {sensu
Gayet 1982).

Bone terminology below is largely after Patterson (1964). The most prominent features of the skull roof
(text-fig. 3c, D) are large, well-developed mucus cavities (features derived for 'trachichthyoid beryciforms’
according to Rosen 1973, p. 477, and for trachichthyids according to Gayet, 1982, fig. 15).

The frontals (f) are nearly complete, but these arc the posteriormost elements of the skull roof that are
preserved in the specimen. The frontals are complicated by high crests surrounding part of the mucus cavities.
The main (medial) crest of each frontal runs forward beyond the median, posteriormost cavity, and curves
medially to meet its counterpart above the anterior edge of the orbit. The crests are highest anteriorly, and
they bear small spines along their edges (text-fig. 3d).

The nasals (na) are relatively large bones closely sutured to the frontals. The medial edges of the nasals
form a high, spiny crest which is divided anteriorly, forming a V-shaped space which houses the ascending
processes of the premaxillae. The wide lateral cavities on the frontal continue anteriorly to the nasals. A
bridge over the nasal groove (nabr) is formed anteriorly by a process extending from the lateral and medial
edges of the nasal. The lateral edge of the nasal and the dorsal surface of the nasal bridge are highly
ornamented with spines and serrations (text-fig. 3e). Of all ‘beryciforms’ examined here, the form of the
nasals in Antarctiheryx most closely resemble those of the living trachichthyid, Paratrachichthys (e.g. see
figs. 49 and 50 in Zehren 1979). The development of the nasal bridge (text-fig. 3d) is more pronounced in
Antarctiberyx than in any other taxon examined here.

A small, slender unornamented bone (ubl) is present between the nasal and the first antorbital (text-fig.
3b). Although it is possible that it is an antorbital bone ( = accessory nasal of Starks 1904, p. 61 1 ), it is much
deeper in the specimen than the infraorbitals, and positive identification of this bone is not possible.

Of the vomer (vo), only the head is visible on the specimen, but the lateral facets for articulation with the
maxillae are visible. No teeth are visible on the ventral surface of the vomer, but this could be an artefact of
preservation).

The infraorbital bones (io) are not clearly preserved in the specimen, but the first three (or four?) are visible
(text-fig. 3a, b). The flanges overhanging the groove for the sensory canal are ornamented. The ventral
borders of the infraorbitals are incomplete. The subocular shelf is not visible in the specimen, possibly due
to lack of preservation.

Most of the maxilla (mx) is missing, exposing much of the suspensorium. Of the hyomandibula (hy), only
a ventral piece is preserved (text-fig. 3b).

The symplectic (s) is a small rod-shaped bone (text-fig. 3b) inclined forward at about forty-five degrees to
the ventral limb of the hyomandibula, and articulating with the quadrate. (Due to poor preservation, the
details of this articulation are not apparent.)

The quadrate (q) is triangular, with a large condyle for the articular. The ectopterygoid (ecp) joins its

834

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 3. A-E, Antcifctiberyx seymouri n. sp., holotype, from Upper Cretaceous deposits of Seymour Island
(TTU P9210). A, B, right lateral view of skull (photo and drawing), xl. c, d, dorsal view of skull
(photo and drawing), xO-75. E, enlarged photograph of bridge over nasal groove showing ornamentation

pattern, x 2-5.

anterior edge and the metapterygoid (mpt) its dorsal edge. The preopercle (and the rest of the opercular
series) is missing.

The ectopterygoid (ecp) is long and curved bone tapering to a point against the quadrate. The anterior
edge of the bone is not clear because there are several cracks in the bone, but one could well be a suture
with the palatine (as described for other trachichthyids in Patterson 1964).

The metapterygoid (mpt) is a sheet of smooth bone. The anterior and dorsal margins of this bone are
hidden (text-fig. 3b).

There is very little left of the upper jaw in the specimen. Only what appears to be the ascending process
(apm) and the articular process (arpm) of the premaxilla are preserved on the right side (text-fig. 3a, b), and
on the left side there is also part of the body of the premaxilla (text-fig. 3c, d). The ascending process is long
and stout, and well separated from the articular process. The maxillary head is too fragmentary to allow
detailed description.

GRANDE AND CHATTERJEE: ANTARCTIC CRETACEOUS FISH

835

The lower jaw, though not well preserved, is nearly complete (text-fig. 3a, b). The dentary (d) tapers from
a high coronoid process, and appears to have been toothed to its tip. Near the symphysis, the teeth (indicated
primarily by empty sockets in the specimen) extend to the lateral face of the bone (text-fig. 3b). The articular
(art) has a large, concave facet for the quadrate posteriorly. This is considered by Gayet (1982) and Stewart
(1984) to be a character of trachichthyids.

Two bones of the hyoid arch are also visible on the specimen (text-fig. 3a, b), a triangular posterior
ceratohyal (eh) which is closely articulated with the anterior ceratohyal (ch). The dorsal and anterior margins
of the anterior ceratohyal are obscured by the lower jaw and the suspensorium.

Scale (sc) fragments are also present in the matrix on the specimen (text-fig. 3b), and are probably from
the same individual. They are ornamented and bear spines on their posterior margins.

Taxonomic placement. Following recent revisions of ‘beryciforms’ by Gayet (1982) and Stewart (1984),
Antarctiberyx appears to belong in Trachichthyidae based primarily on two characters: the teeth on the labial
side of the mandible and the presence of the large mucus cavities on the top of the head. The appearance of
the nasal bridge and frontals is also very similar to that in extant trachichthyids (e.g. see illustrations of
Paratracliiclithys sp. in Zehren 1979, figs. 49 and 50).

Order Indeterminate
Indeterminate teleost scraps

Referred new material. About one bunded specimens, mostly isolated centra and bone fragments (deposited
at TTU), but including some poorly preserved partial skulls. The presence of such material indicates that
more diagnostic material will eventually be found in the Cretaceous deposits of Seymour Island.

DISCUSSION

It seems clear, based on fossil evidence from Seymour Island (Grande and Eastman 1986; and
above), that the Antarctic region had a much more diverse chondrichthyan fauna during Cretaceous
and Early Tertiary time than it does today. The diversity of fossil teleost fauna is much less
apparent, but this may be only an artefact of preservation. The fossils from Seymour Island are
mostly isolated teeth, scales, and skull fragments. The most complete teleost found so far {A.
seymouri n. sp.) is only a partial anterior region of a head (text-fig. 3a-e). The fact that isolated
teeth are much more diagnostic for chondrichthyans than they are for teleosts may be the main
reason why there appears to be a much larger diversity of fossil chondrichthyans than of fossil
teleosts. A wide variety of teleost centra is present in the Tertiary and Cretaceous deposits of
Seymour Island (many UCR specimens listed in Grande and Eastman 1986; and many TTU
specimens mentioned above), but these appear to be identifiable only as Teleostei indeterminate.
It is hoped that with continued collecting on Seymour Island, the Tertiary and Cretaceous teleost
faunas will become better known.

Acknowledgements. We thank Mort D. Turner, W. J. Zinsmeister, M. O. Woodburne, B. Small, M. W.
Nickell, C. E. Macellari, B. Huber, and many others who assisted in the field work and Colin Patterson,
David Ward, and J. D. Stewart for comments on the manuscripts. This research was supported by National
Science Foundation (DPP81-071 52, DPP82-14686, and DPP84-43847) and Field Museum of Natural History.
W. Simpson helped the authors prepare the Antarctiberyx specimen.

ABBREVIATIONS

In text-figures: ang, angular; apm, ascending process of premaxilla; arpm, articular process of premaxilla;
art, articular; ch, anterior ceratohyal; d, dentary; ecp, ectopterygoid; eh, posterior ceratohyal; f, frontal; hy,
hyomandibula; io, infraorbital; 1, lacrimal (first infraorbital); le, lateral ethmoid; Imt, area on labial side of
mandible showing traces of teeth and tooth sockets; mpt, metapterygoid; mx, maxilla; na, nasal; nabr, bridge
over nasal groove (see text); pm, premaxilla; q, quadrate; s, symplectic; sc, scale; ubl, unornamented bone;
vo, vomer.

Institutional: FMNH PF, Fossil Fish Collection, Department of Geology, Field Museum of Natural
History, Chicago, Illinois; TTU, Museum of Texas Tech University, Fubbock, Texas.

836

PALAEONTOLOGY, VOLUME 30

REFERENCES

BiBBY, J. s. 1966. The stratigraphy of part of north-east Graham Land and the James Ross Island Group.
Br. Antarc. Siirv., Sci. Rep. 53, I 37.

CAPPETTA, H. 1975. Selaciens et Holocephale du Gargasien de la region de Gargas (Vaucluse). Geol. Mediterr.
2, 115-134.

1987. Chondrichthyes II: Mesozoic and Cenozoic Elasmobranchii. In schultz, h. p. (ed.). Handbook of

Paleoicththyology, 1- 193. Gustav Fischer Verlag.

CHATTERJEE, s., SMALL, B. J. and NiCKELL, M. w. 1984. Late Cretaceous marine reptiles from Antarctica.
Antarct. J. US, Ann. Rev. 19 (5), 7-8.

and ziNSMEiSTER, w. J. 1982. Late Cretaceous marine vertebrates from Seymour Island, Antarctic

Peninsula. Ibid. 17 (5), 66.

CIONE, A. L., DE VALLE, R. A., DE VALLE, c. A. RINALDI, c. A. and TONI, E. p. 1977. Nota preliminar sobre
lospinguinos y tiburones del Terciario inferior de la Isla Vicecomodoro Marambio, Antarctica. Inst. Antart.
Argent. Contrib. 213, 3-21.

DE BEAUMONT, G. 1960. Contribution a I’Etude des Genres Orthacodm Woodw. et Notidanm Cuv. (Selachii).
Schweiz-palaeont . Abb. 77, 1 46.

DE VALLE, R. A., FOURCADE, H. H. and RINALDI, c. A. 1976. Nota sobre el hallazgo de una nueva especies
denautiloideo del genero Eutrephoceras Owen, en la Isla Vicecomodoro Marambio, Antarctica. Inst. Antart.
Argent. Contrib. 208, 110.

DE WITT, H. H. 1971. Coastal and deep-water benthic fishes of the Antarctica. Antarctic map folio series, folio
15, 1 10. American Geographical Society, New York.

ELLIOT, D. H., RINALDI, C., ZINSMEISTER, W. J., TRAUTMAN, T. A., BRYANT, W. A. and DE VALLE, R. 1975.

Geological investigations on Seymour Island, Antarctic Peninsula. Antarct. J. US, 10 (4), 182-186.

GAYET, M. 1982. Essai de definition des relations phylogenetiques des Holocentroidea nov. et des Trachich-
thyoidea nov. (Pisces, Acantopterygii, Beryciformes). Bull. Mas. natn. Hist, nat., Paris, (4) 4, C,
21-41.

GRANDE, L. and EASTMAN, J. T. 1986. A review of the Antarctic ichthyofaunas in light of new fossil discoveries.
Palaeontology, 29, 113-137.

HOWARTH, M. K. 1966. Ammonites from the Upper Cretaceous of the James Ross Island Group. Bull. Br.
Antarct. Surv. 10, 55-69.

MACELLARi, c. E. and HUBER, B. 1982. Cretaceous stratigraphy of Seymour Island, East Antarctic Peninsula.
Antarct. J. US, Ann. Rev. 17 (5), 68 70.

PATTERSON, c. 1964. A review of Mesozoic acanthopterygian fishes, with special reference to those of the
English Chalk. Phil. Trans. R. Soc. B 247, 213-482.

RINALDI, c. A., MASSABiE, A., MORELLi, J., ROSEMAN, H. L. and DE VALLE, R. A. 1978. Geology of Vicecomodoro
Marambio Island. Direccion Nacional del Antarctica, Inst. Antarct. Argent. Contrib. 217, 5-17.

ROSEN, D. E. 1973. Interrelationships of higher euteleostean fishes. In greenwood, p. h., miles, r. s. and
PATTERSON, c. (eds.). Interrelationships of Fishes, 397-513. Academic Press, New York.

SPATH, c. F. 1953. The Upper Cretaceous cephalopod fauna of Graham Land. Falkland Islands Dependencies
Survey, Sei. Reps, 3, 1-60.

STARKS, E. c. 1904. The osteology of some berycoid fishes. Proc. US natn. Mus. 27, 601-619.

STEWART, J. D. 1984. Taxonomy, paleoecology and stratigraphy of the halecostome-inoceramid associations
of the North American Upper Cretaceous epicontinental seaway. Ph.D. thesis (unpublished). University
of Kansas.

UYENO, T., MiNAKAWA, T. and MATSUKAWA, M. 1981. Upper Cretaceous elasmobranchs from Matsuyama,
Ehime Prefecture, Japan. Bull. natn. Sci. Mus., Tokyo, 7 (2), 81 86.

WARD, D. J. 1979. Additions to the fish fauna of the English Palaeogene. 3. A review of the Hexanchid sharks
with a description of four new species. Tertiary Res. 2 (3), 111-129.

and THIES, D. 1987. Hexanchid shark teeth (Neoselachii, Vertebrata) from the Lower Cretaceous of

Germany and England. Tertiary Res. (in press).

WELTON, B. J. and zinsmeister, w. j. 1980. Eocene neoselachians from the La Meseta Formation, Seymour
Island, Antarctic Peninsula. Contrib. Sci. nat. Hist. Mus. Los Angeles Co. 329, 1-10.

WOODWARD, A. s. 1889. Catalogue of the fossil fishes in the British Museum, Part I, 474 pp. British Museum
(Natural History), London.

1908. On fossil fish-remains from Snow Hill and Seymour Islands. IViss. Ergebn. schwed. Sikipolar

E.xped., 1901-1903, 3 (4), 1 4.

GRANDE AND CHATTERJEE: ANTARCTIC CRETACEOUS EISH

837

ZEHREN, s. J. 1979. The comparative osteology and phytogeny of the Beryciformes (Pisces: Teleostei).
Evolutionary Monographs, 1, 1-389.

ziNSMEiSTER, w. J. 1982. First US Expedition to the James Ross Island area, Antarctica Peninsula. Antarct.
J. US, Ann. Rev. 17 (5), 63-64.

LANCE GRANDE

Department of Geology
Field Museum of Natural History
Chicago, Illinois 60605

SANKAR CHATTERJEE
The Museum

Typescript received 17 March 1986 je^h University

Revised typescript received 30 September 1986 Lubbock, Texas 79409

AN UNUSUAL OSTEOLEPIFORM FISH FROM THE
LATE DEVONIAN OF VICTORIA, AUSTRALIA

by J. A. LONG

Abstract. A new osteolepiform fish, Beelaronqia patrichae gen. et sp. nov., is described from the Late Devonian
(Frasnian) lacustrine Mount Howitt site, eastern Victoria. Beelaronqia is cosmine-covered, has a broadly haired
parietal shield with large extratemporal bones, cheek plate with an additional small postorbital bone and broad
lateral extrascapulars. The orbits are very small. The scales are rhombic with high dorsal processes. Beelaronqia is
considered closely related to Canowindra grossi Thomson because of the shape of the parietal shield, presence of
an accessory postorbital bone behind the small orbit, and broad extrascapular series.

Until recently almost our entire knowledge of the osteolepiform fishes was based on studies of
material from the Northern Hemisphere which formed part of the Laurasian landmass in
mid-Palaeozoic times. New discoveries from Australia and Antarctica, which formed East
Gondwana, or the antipodes to Laurasia during the Devonian, reveal a diverse osteolepiform fauna,
some elements of which are unexpectedly different from the Northern hemisphere faunas. Aside from
fragments attributed to Strepsodiis by Woodward (1906) the first fossil crossopterygian described
from Australia was Canowindra grossi Thomson (1973) from the Upper Devonian Mandagery
Sandstone of New South Wales. Thomson described the single complete fish from a natural mould,
but was uncertain how to classify it. Since then Long (19856) has redescribed Canowindra from new
preparation of the original specimen, and regards it as an osteolepiform which is representative of a
new higher taxon. Young and Gorter (1981) described IGyroptychius cf. G. australis from the Middle
Devonian Hatchery Creek fauna, New South Wales, and Long (1985a) described a primitive
eusthenopterid, Marsdenichthys longioccipitus, from the Late Devonian Mount Howitt locality,
Victoria. Other osteolepiforms from Australia include Gogonasiis andrewsi from the Upper Devonian
Gogo Formation, Western Australia (Long 1985c) and new species of Megalichthys from the Lower
Carboniferous of Queensland (Long and Turner, 1984). In August 1986 a complete skull of Gogonasus
was discovered at Gogo and this material is currently under study. The new form described here
is significant in that it shares several distinctive characters with Canowindra, and together with
another undescribed form from the Middle Devonian of Antarctica may represent an endemic
Gondwanan group of Osteolepiformes. In this paper the Mount Howitt form is described in detail.
The erection of higher taxonomic ranks will be postponed until all the new material is fully
described.

The material of Beelarongia was collected in the original excavations of the Mount Howitt site over
the 1970/1 field seasons by Professor Jim Warren (Monash University, Zoology Dept.). As most of the
bone was weathered away the pieces of cranium were recognized as natural impressions in the black
shale and assembled together as best possible despite the absence of some pieces. The specimen was
then cast with latex. All of the pieces are thought to belong to the one individual, although as they were
found separately throughout the collection it cannot be determined exactly when they were found and
if they came from the same horizon.

Terminology for dermal bones used herein follows Jarvik ( 1 980) as discussed in Borgen ( 1 983). The
words ‘breadth, length, and height’ are abbreviated as L, B, and H respectively. All measurements
were taken across the surface of bones, thus incorporating some degree of flattening where present.
Indices are expressed as the product of two linear dimensions multiplied by 100.

[Palaeontology, Vol. 30, Part 4, 1987, pp. 839-852, pi. 91. j

© The Palaeontological Association

840

PALAEONTOLOGY, VOLUME 30

SYSTEMATIC PALAEONTOLOGY
Order osteolepiformes (osteolepidida)

Remarks. Despite recent arguments that the Osteolepiformes could be a paraphyletic group (Rosen
et al. 1981; Lauder and Liem 1983; Gardiner 1980), I have presented arguments that the group is
probably monophyletic (Long 1985u), but admit that the evidence for this hypothesis is scarce.
Osteolepiform fishes uniquely possess a pectoral girdle with a large ornamented anocleithrum
separating the cleithrum from contact with the supracleithrum. There are large basal scutes present at
fin bases (Andrews 1973). The cheek has a bar-like preopercular bone which forms the rear border to
the single, large irregularly-shaped squamosal, and is steeply inclined. The squamosal is unique
amongst Osteichthyes in occupying most of the cheek plate area and having six sides of which five
suture to other cheek bones. This character also applies to some early tetrapods which are here
regarded as a subgroup of the Osteolepiformes, probably of close affinity to the Panderichthyidae
(Schultze 1970; Schultze and Arsenault 1985) and not immediately related to the Dipnoi as suggested
by Rosen et al. (1981). The only group of Osteolepiformes in which cheek patterns may be confused
with those of tetrapods is the panderichthyids (Schultze and Arsenault 1985). The cheeks of
panderichthyids are poorly known. They differ slightly from those of other osteolepiforms in contact
margins between the lachrymal and jugal (Elpistostege) and in having a slightly smaller lachrymal
[Panderichthys). In the primitive labyrinthodonts which retain an osteolepiform-like cheek pattern
(Ichthyostega, Acanthostega, Crassigyrinus) the preopercular is much reduced and the jugal has
become much larger than the squamosal (Jarvik 1980; Panchen 1973, 1985).

Beelarongia gen. nov.

Etymology. From the aboriginal word Eeelarong' meaning shiny (Reed 1975), an allusion to the cosmine covered
dermal bones and scales.

Diagnosis. A presumably cosmine-covered osteolepiform with a fronto-ethmoid shield/parietal shield
ratio of approximately 80; index of lengths of frontal/parietal bones about 60. Parietal shield
anteriorly narrow but posteriorly broad with a B/L index of 180, posterior margin strongly pointed
posteriorly. Extratemporal very large, being almost one third as broad as one side of the parietal
shield. Median extrascapular having a short anterior margin, one fifth the extent of that of the broad
lateral extrascapulars. Cheek with long postorbital and one small accessory postorbital bone. Orbits
very small. Opercular large, rounded. Cleithrum with broad unornamented anterior flange. Scales
rhombic with large dorsal processes.

Remarks. Beelarongia is distinguished from all other osteolepiforms, except Canowindra, by its very
small orbits which have a small accessory postorbital bone behind them, the unusually large
extratemporal bones and very broad posterior division on the parietal shield. It is distinguished from
Canowindra, which also has the above features, by the presumed presence of cosmine, narrower shape
of the extratemporal bones, the rhombic scales with high dorsal processes, and the simpler cheek
pattern with only one accessory postorbital bone (Canowindra has two accessory postorbitals. Long
19856). Cosmine is presumed to be present on Beelarongia because of the smooth surface of the bones
and scales from the natural mould. The thick rhombic scales of Beelarongia are almost identical to
those of other cosmine-covered osteolepiforms.

Type species. B. patrichae sp. nov.

Beelarongia patrichae sp. nov.

Plate 91; text-figs. 1-5, 6b

Etymology. After Dr Pat Rich (Monash University, Earth Sciences Dept.) who has given me much assistance and
encouragement.

LONG: DEVONIAN OSTEOLEPIFORM FISH FROM AUSTRALIA

841

Na

TEXT-FIG. 1. Beelarongia pathchae gen. et sp. nov., sketch ofholotype NMV P160872, cranium in dorsal view. The
rest of the specimen, containing most of the left opercular and pectoral girdle is not shown here. Cross hatching
indicates damaged areas filled in by latex. ac.PO, accessory postorbital; Clth, cleithrum; De, dentary; Dsp,
dermosphenotic; ET, extratemporal; Fr, frontal; Ju, jugal; La, lachrymal; L.Ex, lateral extrascapular; 11c, main
lateral line canal pores; Max, maxilla; mpl, middle pit-line groove; Na, nasal; nar, external nostril; nl, n2,
notches on posterior margin of frontal; OP, opercular; Par, parietal shield; pinf, pineal foramen or elevation;
PO, postorbital; POP, preopercular; r, basal ridge of scales; sc, scales; soc, supraorbital sensory-line pores;

Sq, squamosal.

842

PALAEONTOLOGY, VOLUME 30

H olotype. NM V P 1 60872 (PI. 9 1 ), an almost complete head preserved as a natural mould in dorsal view, with part
of the pectoral girdle and fin attached. Housed in the Museum of Victoria, Melbourne (NMV).

Material. NMV P160873, an isolated section of scales with part of a fin attached; NMV P160874, part of the
squamation, and NMV P160875, cleithrum and clavicle.

Occurrence. From the lower mudstone outcropping along the Howqua River at the base of Mount Howitt
(Marsden 1976); Avon River Group, Late Devonian (Frasnian, Long 1983).

Diagnosis. As for genus, only species.

Description. Beelarongia was a moderate-sized osteolepiform having an estimated maximum cranial length of
100 mm from the single known skull. The shape of the head can be easily determined from the holotype, being
broad posteriorly but rapidly narrowing anteriorly, and quite depressed in cross-section. The sections of body
squamation give no indication of the body shape or the disposition of the fins.

Fronto-ethnioid shield. The fronto-ethmoid shield is approximately as broad as long, with the dorsal face being
flat posteriorly, arching convexly at the rostral margin (text-figs. 1 and 2). Few bones can be distinguished due to
the poor preservation. The large frontals are separate from the other bones forming the shield, with a distinct
concave lateral margin on the right frontal for the dermosphenotic. The supraorbital canal is assumed to run off
the frontal to the dermosphenotic, indicated on the specimen by a series of fractures where the canal would be
expected. Approximately halfway along the median suture between the frontals there is a median rough area on
the cast; this is either a pineal thickening or a foramen (pinf).

The frontals (Fr) comprise three quarters of the estimated length of the fronto-ethmoid shield, being about
twice as long as broad. In the right frontal a row of pores representing the supraorbital canal (soc) can be seen
close to the rostral margin. The posterior margin of the fronto-ethmoid shield has an irregular shape with two
notches (nl. n2) present on the margin of the right frontal (text-fig. 1), possibly for ligaments as Jarvik (1972) has
suggested for similar notches in the fronto-ethmoid shield of Glyptolepis. The dermosphenotic (Dsp) is seen on the
left side, displaced from the skull roof table to lie next to the postorbital (text-figs. 1 and 2). It is almost rhombic in
form with rounded margins, comprising three quarters the length of the frontals and is half as broad as long. The
mesial margin is smoothly convex to fit into the frontal, but the lateral margin is straighter with a long posterior
division contacting the postorbital and a short anteromesially directed anterior side for the supraorbital. The
dermosphenotic has an extensive overlap surface for the anterior part of the postorbital (oa.Dsp, text-fig. 2b)
indicating that the cheek was firmly attached to at least the anterior division of the skull.

An indistinct foramen, possibly the nasal opening (nar), is visible on the left side of the fronto-ethmoid shield
(text-figs. 1 and 2a). It is an almost circular opening situated just anterior to the lachrymal as in other
osteolepiforms. The orbital notch is not clearly defined on the fronto-ethmoid shield. The close proximity of the
bones surrounding the orbit, and the reconstruction of these bones in the articulated cheek indicate that the
orbits of Beelarongia were very small, similar to those of porolepiforms in being even smaller than those of
small-eyed osteolepiforms such as Glyptopomus (Jarvik 1950) or Latvius (lessen 1966, 1973).

Parietal shield. The parietal shield is well preserved although only the extratemporal bones (ET) are clearly
differentiated from the rest of the cosmine-covered shield (PL 91; text-fig. 1). In cosmine-covered osteolepidids the
sutures between the parietal, intertemporal, and supratemporal are occasionally not visible (Jarvik 1948, 1985;
Jessen 1966) but are distinct in forms which have no cosmine. The absence of such sutures in C anowindra is
believed to reflect the actual condition of fusion between these bones rather than the sutures being merely masked
by the external cosmine layer (Long 19856). As Beelarongia is closely related to Canowindra, as discussed below,
the condition of the partietal shield in Beelarongia is also thought to be similar to that of Canowindra. The overall
shape of the parietal shield is similar to that of Latvius grewingki (Gross 1956, pi. 1 ), but differs in having an even
broader posterior region with larger extratemporals. The posterior margin has two clearly defined angular
notches, the first being just mesial to the extratemporal, the second being situated about midway along the

EXPLANATION OF PLATE 91

Figs. 1-3. Beelarongia patrichae gen. et sp. nov. 1, holotype cranium NMV PI 60872 in dorsal view. 2, cleithrum
in lateral view with clavicle in ventrolateral view, NMV P160875. 3, part of the squamation showing scale with
high dorsal process, NMV P160874. Latex casts whitened with ammonium chloride.

PLATE 91

LONG, canowindridoid osteolepiforms

844

A

PALAEONTOLOGY, VOLUME 30

pop) Ju

TEXT-FIG. 2. Beelarongia patrichae gen. et sp. nov., cheeks in lateral view, from holotype NMV P160875. a, left
cheek, b, right cheek. Parts of the dorsal edge of the cheek were not well preserved and did not cast very well.
Abbreviations as for text-hg. 1 except for: ifc, infraorbital sensory-line canal pores; oa.a.PO, overlap area for
accessory postorbital bone; oa.Dsp, overlap area for dermosphenotic; orb, orbit; popl, preopercular sensory-line

canal.

margin. These notches are also seen on C. grossi and a few other osteolepiforms (Thursius moythomasi, Jarvik
1948, pi. 19; Gyroptychius dolichotaliis, Jarvik 1985, fig. 12). The anterior margin of the parietal shield is not
completely shown in the specimen, but is assumed to have an irregular form to comply with the notches for
ligamentous attachments seen on the posterior margin of the fronto-ethmoidal shield. The lateral margins of the
parietal shield are smoothly curved with a strong median concavity before splaying out laterally near the contact
with the extratemporal. The parietal suture courses in an irregular zig-zag fashion close to the posterior margin.
The transverse (middle) and posterior parietal pit-lines are situated close to the posterior margin of the shield,
meeting at a point slightly closer to the extratemporal than to the median suture.

The extratemporal (ET) is a large, triangular bone approximately as long as broad. It is a third as long as the
parietal shield and almost a third as broad as one half of the shield. The mesial and posterior margins of
the extratemporal are almost straight, but the lateral margin is strongly convex. Although the shape of the
extratemporal is similar to that of several other osteolepiforms, most of which have a high degree of variation in

LONG: DEVONIAN OSTEOLEPIFORM FISH FROM AUSTRALIA

845

the shape of that bone (e.g. G. dolichotatus, Jarvik 1 985, fig. 1 2) none of the described forms approach Beelarongia
or Canowindra in the proportional size of the extratemporal to the rest of the parietal shield. Porolepis has a
similarly sized extratemporal but differs in the shape of this bone, which is more elongated (Jarvik 1972).

Extrascapulars. The left lateral extrascapular (L.Ex) is preserved on the holotype. The imperfect bone preserved
on the right side which shows an overlap surface is not the right lateral extrascapular but part of the right
opercular which has been displaced over the rear of the skull roof. Overlap relationships are not known for the
extrascapulars, although it may be noted that in Canowindra there is normal osteolepiform overlap of the median
bone by the lateral extrascapulars. The lateral extrascapular is a large rounded bone which is about one and a half
times as broad as long. The anterior margin is strongly convex with two angular prominences. The lateral margin,
considerably shorter than the mesial margin, is gently convex (PI. 91a). The median extrascapular would have
been very narrow at its contact with the parietal shield. There is no indication of the course of the main lateral line
canal or the occipital canal on the material.

Cheek. The cheek of Beelarongia is well preserved on the left and part of the right side of the holotype (text-
figs. 1 and 2). The anterior division is poorly known because of post-mortem collapse of the bones around the
orbit. It is clear that the normal pattern of cheek bones was present: squamosal (Sq), jugal (Ju), postorbital (PO),
lachrymal (La), preopercular (POP), and probably the quadratojugal although this is not seen in the material.
The posterior region of the cheek is not at all well preserved, but part of the preopercular is preserved on the left
side of the specimen. The infraorbital series of Beelarongia differs from that of most osteolepiforms in that the
lachrymal is quite smaller and narrower than the jugal and postorbital elements (text-fig. 6).

The lachrymal can be seen on both sides of the holotype immediately ventral and posterior to the orbit. It is
two thirds the length of the jugal, and slightly more than half as long as the fronto-ethmoid shield. The dorsal
margin is not well preserved on either of the two examples, presumably fitting below the orbit, but not necessarily
participating in the orbit. Open pores of the infraorbital canal are visible close to the ventral margin of the
lachrymal.

The jugal is a large bone contacting the lachrymal anteroventrally, the postorbital dorsally, the squamosal
posteriorly and the maxilla ventrally. The anterior contact is obscure, as if the bones are place in approximate
juxtaposition there is a space immediately posterior to the orbit for a small accessory postorbital bone. This is
further supported by the anterior margin of the postorbital which shows a broad overlap flange for another
dermal element (text-fig. 2a). There is a small area of bone preserved anterior to the postorbital which is identified
as the accessory postorbital (ac.PO, text-fig. 2a). This bone is a small almost square plate which corresponds in
size to the anterior accessory postorbital of Canowindra, but differs markedly from that element in Canowindra by
its shape. However, it is alternatively possible that this small bone in front of the postorbital is part of the
supraorbital which has fallen down from the skull roof. In view of the anterior overlap area on the large
postorbital and the general fit of the bones around the orbit it is most likely that an additional postorbital element
is indeed present, as restored in text-fig. 5. The jugal has an elongate form, being slightly more than twice as long
as high. The dorsal margin is gently curved, turning steeply at the anterior margin to form an anterior point. The
posterior margin is strongly convex, curving downwards to meet the straight ventral margin. There are numerous
pores present indicating the presence of the preopercular sensory line, although its exact course and junction with
the infraorbital canal are not clear.

The postorbital (text-figs. 1 and 2) is a large bone dorsal to the jugal, lateral to the dermosphenotic and anterior
to the squamosal. It is unusual in having a broad anterior end with two overlap flanges, a dorsal one for the
dermosphenotic (oa.Dsp) and a ventral flange for an additional small accessory postorbital bone (oa.a.PO), as
also occurs in Canowindra in which two accessory postorbitals are found (Long 19856). The anterior part of the
dorsal margin is relatively straight, but moderately inclined anteroventrally; the anterior margin being pointed
with a concave facet bearing broad overlap surfaces on either side of the point. The posterior margin is indented
for overlap of the squamosal. Overall the postorbital is slightly longer than high, and marginally shorter than the
jugal.

The anterior region of the squamosal (Sq) is preserved on both sides of the specimen (text-fig. 2). The anterior
margin has a dorsal pointed division for contacting the postorbital and a smoothly concave ventral division for
the jugal. From the extent of the whole cheek, restored by the relative position of the opercular and
fronto-ethmoidal shield, the squamosal was particularly large relative to the other cheek bones, a character
typifying the osteolepiforms.

The dorsal region of the preopercular can be seen on the left side of the specimen. All that can be said about this
bone is that it was situated right at the rear of the cheek complex with a steeply inclined long axis as in other
osteolepiforms.

846

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 3. Beelarongia patrichae gen. et sp. nov., left pectoral girdle and fin, from the holotype NMV P160875. a,
external view of fin, counterpart to b, internal view of cleithrum showing scapulocoracoid and humerus, c, sketch
interpretation of scapulocoracoid and humerus as preserved, d, f, attempted restoration of the humerus in (d)
ventral and (f) anterior views, e, left humerus of Sterropterygion in ventral view (after Rackoff, 1980). G, left
humerus of Emthenopteron in ventral view (after Jarvik, 1980). but.a, but.d, but.p, anterior, dorsal and posterior
buttress of scapulocoracoid; ca.Hu, caput humeri; Clth, cleithrum; corp, corpus of humerus; del, crest for
deltoideus muscle; ent, entepicondyle; f.ent, entepicondylar foramen; f.sc, supracoracoid fossa; fsgl, supraglenoid
foramen; ?f.sub, area of subscapular fossa; gr.com, commissural groove; gr.ect, ectepicondylar groove; scha, shelf
for attachment of scapulohumeroideus muscle; sup., ridge for supinator muscle; See, medial division of the

scapulocoracoid.

LONG: DEVONIAN OSTEOLEPIFORM FISH FROM AUSTRALIA

847

Operculogiilar series. The opercular bone (OP) is the only element of this series preserved on the specimen, being
almost completely preserved on the left side. Text-fig. 1, which shows part of the opercular, does not show that
part of the specimen which joins onto the opercular as this contains the pectoral girdle and cannot be properly
studied if glued back to the block containing the cranium. The opercular is subrectangular in shape with an
estimated B/L index around 112 (restoring the dorsal margin). The anterior margin is straight and comprises
54% of the opercular length in its extent. The dorsal margin is not preserved, but can be restored to fit the
posterolateral margin of the skull table. This gives a deeply concave dorsal margin. The posterior margin is
strongly convex, curving smoothly round to the ventral margin without a marked posteroventral corner
developed.

Pectoral girdle and fin. The holotype shows the proximal region of the left pectoral fin and shoulder girdle
preserved (text-fig. 3), and NMV P160875 shows the external surfaces of the cleithrum and clavicle well preserved
(PI. 91; text-fig. 4). It should be noted that the left pectoral girdle and fin attach to the side of the skull in text-fig. 1,
completing the posterior region of the opercular and showing the humerus and scapulocoracoid resting inside the
flattened cleithrum.

The cleithrum (Clth; text-figs. 1, 3, 4) has a large dorsal rectangular division which is slightly more than twice as
long as the ventral triangular division (dividing line taken as the line passing through the broadest section of the
bone). The externally exposed ornament is restricted to a narrow triangle on the dorsal rectangular division,
occupying less than half of the surface area. The dorsal margin is straight, and the anterodorsal corner has a
depressed curved area for the overlap of the opercular (oa.OP). The posterior margin has a slightly concave
dorsal division with a well rounded bulge for the scapulocoracoid where the pectoral fin inserted (f.att). The
inwardly bent smooth area anterior to the ornament has a broad outer region (ia), probably for the membranous
cover of the posterior wall of the branchial cavity (Jarvik 1948, p. 98) and a slightly narrower inner strip which was
presumably covered by soft tissues at the rear of the gill cover (gc). The visceral surface of the cleithrum is partially

TEXT-FIG. 4. Beelarongia patrichae gen. et sp. nov., cleithrum in lateral view
with associated clavicle in ventrolateral view, NMV P160875. asc.pr, ascend-
ing process of clavicle; Clav, clavicle; Clth, cleithrum; f.att, area where
pectoral fin attaches to scapulocoracoid; gc, inner smooth lamina of
cleithrum forming part of the gill chamber; ia, inner smooth area of cleithrum;
oa.Cl, overlap area for cleithrum; oa.OP, overlap area for opercular; oa.s,
overlap area for scales; vl, ventral lamina of clavicle.

848

PALAEONTOLOGY, VOLUME 30

preserved on the holotype, being smooth and featureless aside from the endoskeletal attachment area, hidden on
the specimen by the scapulocoracoid (See). In having a relatively short ventral triangular division the cleithrum of
Beehirongia is typical of osteolepiforms, although the development of a broad, smooth, inwardly bent area on the
external surface is a feature of porolepiform cleithra (Jarvik 1972, p. 125).

The clavicle (Clav) is partially preserved on NMV PI 60875 (text-fig. 4). It has an ascending process (asc.pr) with
a distinct overlap flange for the cleithrum (oa.Cl), and features a broad ventral lamina (vl). In overall length it
approaches that of the cleithrum (noting the absence of the tip of the ascending process). The dorsolateral vertical
portion of the clavicle is a narrow, smooth region which meets the ventral, ornamented wall along a pronounced
lateral ridge.

The scapulocoracoid (See) and humerus (Hu) are poorly preserved inside of the left cleithrum. In general form
the scapulocoracoid (text-fig. 3c) is elongate, being just under half as broad as long, and slightly broader than
deep (height relative to mesial wall of cleithrum). The mesial or visceral surface is quite flat, merging smoothly
forward to meet the mesial wall of the cleithrum, although as it is studied from a latex cast the exact nature of the
subscapular fossa (?f.sub) cannot be determined. There appears to be anterior (but. a) and dorsal buttresses (but.d)
supporting the central body of the scapulocoracoid, as in other osteolepiforms (Andrews and Westoll 1970«,
19706; Janvier 1980), although the space between the posterior buttress (but.p) and the wall of the cleithrum is
not differentiated on the latex cast. At the point where the humerus meets the scapulocoracoid, at the glenoid
fossa, the mesial surface of the scapulocoracoid is most furthest from the inner surface of the cleithrum.
Both the supracoracoid foramen (f.ssc) and supraglenoid fossa (f.sgl) are seen. Overall the scapulocoracoid of
Beehirongia, as far as can be described, conforms to the usual osteolepiform type, but may differ in the
absence of a well-developed subscapular fossa if this absence is not an artifact of latex casting, as may be
the case here.

The humerus (Hu, text-fig. 3c, d, F)can be seen to be a broad bone, more than twice as broad as the long axis of
the shaft, as preserved, but probably much broader as the entepicondylar process is incomplete. To facilitate
comparisons I have figured the humeri of two other osteolepiforms (text-fig. 3e, g) in which it is well known: the
osteolepid, Sterropterygion {Ra.cko[( 1980), and theeusthenopterid, Eiisthenopteron (Andrews and Westoll 1970a;
Jarvik 1980). The distal surface is not well preserved on the specimen as it is covered by scales of the fin web which
reach high up on the fin-lobe. In overall form the humerus is similar to that of Eusthenopteron (text-fig. 3g) in the
long entepicondyle (ent.) and well defined commissural groove (gr.com). Processes for attachment of the
scapulohumeroideus (scha), supinator (sup), and deltoideus (delt.) muscles are present, but the latter two are not
as well developed as on the osteolepid Sterropterygion (text-fig. 3e). The deltoideus crest is seen as a small bump
separated from the dorsal side of the humerus which bears the supinator muscle ridge by a well formed groove
(text-fig. 3f). The entepicondylar foramen (Lent) is a deeply excavated pit seen on the ventral surface of the
entepicondyle. This differs from the condition in both Sterropterygion and Eusthenopteron. In Eusthenopteron
the entepicondylar foramen or canal runs almost parallel with the corpus of the humerus to open posteriorly near
the articulatory facet for the ulna, and in Sterropterygion the entepicondylar foramen is either not noticeable or
absent (Rackoff 1980, figs. 5 and 6). This major difference in the morphology of the humerus indicates a possible
early dilTerentiation of the pectoral fin of Beelarongia from that of other osteolepiforms.

The proximal region of the pectoral fin can be seen on the holotype (text-fig. 3a, b). The scale covered lobe is
short, relative to cranial length, as in osteolepiforms (although it should be noted that the length of the pectoral
fin is unknown for other primitive ‘crossopterygians’ such as Porolepis, Powichthys, or Youngolepis'. Jarvik 1972;
Jessen 1980; Chang 1982). Lepidotrichia (f.r) are segmented and bifurcation occurs only once in the preserved
area. Each lepidotrichial segment has a rectangular cosmine face separated from the cylindrical basal bone by a
restricted neck.

Scales and squamation. There are a few specimens showing the scale morphology and parts of the intact
squamation (holotype, NMV P160873, P160874; PI. 91, fig. 3; text-fig. 1) but none of these specimens have the
bone preserved well enough to show histological structure.

The typical flank scales are rectangular, being about three-quarters as long as deep. The thick cosmine covered
external surface has a narrow groove running parallel to the dorsal and anterior margins for an interlocking fit
with the overlapping scales. The dorsal end of each scale features a large process which is approximately one third
as high as the whole scale, similar to the condition in palaeoniscoid scales. Anterior to the cosmine surface is a
smooth overlap area equal to about one fifth of the scale breadth. The basal surface of each scale has a vertical
thickening running parallel to the anterior margin and a large depressed area in the ventral half of the scale for
receiving the dorsal process of the scale below. Scales covering the pectoral fin are rhombic, decreasing in size
distally from the fin base. Along the leading edge of the fin there is an elongated narrow basal fin scute. Minute fin
scales close to the junction with lepidotrichial rows are rectangular in form.

LONG: DEVONIAN OSTEOLEPIFORM FISH FROM AUSTRALIA

849

TEXT-FIG. 5. Beelarongia patrichae gen. et sp. nov. Attempted restoration of the head in a, dorsal view, and b,

lateral view, after the holotype, NMV P160875.

850

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 6. Comparison of the heads in lateral view (a-d) of two closely related East Gondwana osteolepiforms
(A, B) and two typical osteolepiforms (c, d). Accessory postorbital bones stippled, a, Canowimira grossi (after
Long 1985b). b, Beelcirougia patrichae gen. et sp. nov. c, the eusthenopterid Eiisthenopteron foordi (after Jarvik
1980). D, the osteolepidid Osteolepis macrolepidotiis (after Jarvik 1948). Abbreviations as for text-fig. 1.

The squamation is known from isolated blocks of scale impressions of uncertain position on the body. Scale
rows are near-vertically disposed on the body, as seen close to the cranium and indicated by the angles of contact
between externally exposed scale surfaces.

DISCUSSION

Within the Osteolepiformes Beelarongia shares the following features with C anowindra grossi:
small orbits, parietal shield with a very broad posterior division, large extratemporal bones, possible
fusion between the parietal-intertemporal-supratemporal bones, extrascapulars which are unusually
broad and short, and an additional cheek element anterior to the postorbital bone (text-figs. 5 and
6a, b). The taxonomic distribution of these characters within the Osteichthyes has been discussed for
C anowindra (Long 19856) and this will not be reiterated here, except to say that on the basis of that
work Beelarongia is here regarded as a sister taxon to Canowindra. Both Beelarongia and C anowindra
are regarded as true osteolepiforms because of the nature of the cheek plate (having a single large
squamosal, lacking a prespiracular bone behind the postorbital, no preoperculo-submandibular
present, and bar-like preopercular which is steeply inclined to form most of the posterior margin of the
cheek plate), the overlap relationships of the extrascapulars (median extrascapular overlapped by
lateral extrascapulars), the presence of only one pair of external nares (known with certainty only in
Canowindra), and in the structure of the shoulder girdle (large externally exposed anocleithrum,
cleithrum with short ventral division). In addition Beelarongia has basal scutes, another characteristic
feature of Osteolepiformes (Andrews 1973). The suggestion that Beelarongia and Canowindra are

LONG: DEVONIAN OSTEOLEPIFORM FISH FROM AUSTRALIA

851

closely related implies that the transition from rhombic to cycloid scales and the loss of cosmine in
Canowindra occurred independently of other osteolepiform lineages.

Unfortunately we do not have information on the braincase or palate of these two interesting taxa,
and lack of such data inhibits discussion of their relationships within the Osteolepiformes. Current
views on osteolepiform interrelationships (e.g. Vorobyeva 1977) require revision in the light of recent
discoveries (Long 1985c). Further remains of osteolepiforms which share derived characters with
Beelaronyia and Canowindra are now known from the Middle Devonian of Antarctica (Mount Crean
fauna) and the Late Devonian of Victoria (South Blue Range fauna), supporting the suggestion (Long
1985h, 1985t/) that these species probably belong to a new higher taxon of Osteolepiformes endemic to
the East Gondwana Province (scnsu Young 1981). The new material includes some preservation of the
braincase and palate and, when described will enable further discussion of this problem.

Acknowledgements. Sincere thanks to Professor Ken Campbell (Geology Department, The Australian National
University), Dr Peter Forey (British Museum of Natural History), and an anonymous referee for reading and
commenting on the manuscript. Professor Jim Warren (Zoology Department, Monash University) allowed me
to study the Mount Howitt material. Dr Gavin Young (Bureau of Mineral Resources, Canberra) is thanked for
helpful discussion of the work. Dr Pat Rich (Earth Sciences Department, Monash University) supervised the first
part of the work which formed part of a Monash University doctorate dissertation. The work was carried out at
the Australian National University under the award of a Rothmans Fellowship, and at the University of Western
Australia under a National Research Fellowship— Queen Elizabeth II Award.

REFERENCES

ANDREWS, s. M. 1973. Interrelationships of crossopterygians. In greenwood, p. h., miles, r. s. and patterson, c.
(eds.). Interrelationships of Fishes. 137-177. Academic Press, London, New York,

and WESTOLL, x. s. 1970a. The postcranial skeleton of Eiisthenopteron foordi Whiteaves. Trans. R. Soc.

Edinb. 68, 207-329.

— 1970h. The postcranial skelton of rhipidistian fishes excluding Eiisthenopteron. Ibid. 68, 391-489.

BORGEN, u. J. 1983. Homologizations of exocranial roof bones between tetrapods and osteolepiform fishes.
Palaeontology, 26, 735-754.

CHANG MEE-MAN. 1982. The braincase of Youngolepis, a Lower Devonian crossopterygian from Yunnan, south-
western China. Paps Dept. Geol. Univ. Stockholm, 113 pp.

GARDINER, B, G. 1980. Tetrapod ancestry: a reappraisal. In panchen, a. l. (ed.). The Terrestrial Environment
and the Origin of Land Vertebrates, 177-185. Academic Press, London and New York.

GROSS, w. 1956. Uber Crossopterygier und Dipnoer aus dem baltischen Oberdevon im Zusammenhang einer
vergleichenden Untersuchung des Porenkanalsystems palaozoischer Agnathen und Fische. K. svenska
Vetensk-Akad. Handl. 5 (6), 1-140.

JANVIER, p. 1980. Osteolepid remains from the Devonian of the Middle East, with particular reference to the
endoskeletal shoulder girdle. In panchen, a. l. (ed.). The Terrestrial Environment and the Origin of Land
Vertebrates, 223-254. Academic Press, London and New York.

JARVIK., E. 1948. On the morphology and taxonomy of the Middle Devonian osteolepid fishes of Scotland.
K. svenska Vetensk-Akad. Handl. 3 (25), 1, 1-301.

1950. On some osteolepiform crossopterygians from the Upper Old Red Sandstone of Scotland. Ibid. 2(4),

1-35.

1972. Middle and Upper Devonian Porolepiformes from East Greenland with special reference to Glypto-

lepis groenlandica n.sp. Meddr. Gronland, 187, 1-295.

1980. Basic Structure and Evolution of Vertebrates. Vol. 1. Academic Press, London and New York.

1985. Devonian osteolepiform fishes from East Greenland. Meddr. Gronland., Geosci. 13, 52 pp.

JESSEN, H. 1966. Die Crossopterygier des Oberen Plattenkalkes (Devon) der Bergisch-Gladbach PafTrather
Mulde (Rheinisches Scheifergebirge) unter Berucksichttigung von amerkianischem und europaischem
Onychodus-Matenal. Ark. Zool. (2) 18, 305-389.

1973. Weitere Fischreste aus dem Oberen Plattenkalk der Bergisch-Gladbach-PalTrather Mulde

(Oberdevon, Rheinisches Schiefergebirge). Palaeontographica, 143A, 159-187.

1980. Lower Devonian Porolepiformes .from the Canadian Arctic with special reference to Powichthys

thorsteinsonni lessen. Palaeontographica, 167 A, 180-214.

852

PALAEONTOLOGY, VOLUME 30

LAUDER, G. V. and LiEM, K. 1983. The evolution and interrelationships of the actinopterygian fishes. Bull. Mus.
comp. Zool. Harv. 150, 95-197.

LONG, J. A. 1983. New bothriolepid fishes from the Late Devonian of Victoria, Australia. Palaeontology, 26,
295-320.

1985u. The structure and relationships of a new osteolepiform fish from the Late Devonian of Victoria,

Australia. Alcheringa, 9, 1-22.

1985h. New information on the head and shoulder girdle of C anowUulra grossi Thomson, from the Upper

Devonian Mandagery Sandstone, New South Wales. Rees. Aust. Mus. 37, 91-99.

1985c. A new osteolepidid fish from the Upper Devonian Gogo Formation, Western Australia. Rees. West.

Au.st. Mus. 12, 361-377.

1985d. The evolution of the Crossopterygii: a study based on new material from the Devonian of

Australia and Antarctica. (Abstract.) Geol. Surv. New Zealand, Record, 9, 70-72.

and TURNER, s. 1984. A checklist and bibliography of Australian fossil fishes. In archer, m. and clayton, g.

(eds.). Vertebrate Zoogeography and Evolution in Australasia, 325-354. Hesperian Press, Perth.

MARSDEN, M. A. H. 1976. Upper Devonian-Carboniferous. In Douglas, j. and ferguson, j. (eds.). Geology of
Victoria. Geol. Soc. Aust. Spec. Pub. 5, 77-124.

PANCHEN, A. L. 1973. On Crassigyrinus scoticus Watson, a primitive amphibian from the Lower Carboniferous
of Scotland. Palaeontology, 16, 179-193.

— - 1985. On the amphibian Crassigyrinus scoticus Watson from the Carboniferous of Scotland. Phil. Trans.
R. Soc. B 309, 505-568.

RACKOFF, J. s. 1980. The origin of the tetrapod limb and the ancestry of tetrapods. In panchen, a. l. (ed.). The
Terrestrial Environment and the Origin of Land Vertebrates, 255-92. Academic Press, London and New York.

REED, A. 1975. Aboriginal words and place names. Rigby, Sydney.

ROSEN, D. E., FOREY, p. L., GARDINER, B. G. and PATTERSON, c. 1981. Lungfish, tetrapods, palaeontology and
plesiomorphy. Bull. Am. Mus. nat. Hist. 167, 159-276.

SCHULTZE, H.-p. 1970. Folded teeth and the monophyletic origin of the tetrapods. Am. Mus. Novit. 2408, 1-10.

and ARSENAULT, M. 1985. The panderichthyid fish Elpistostege: a close relative of tetrapods? Palaeontology,

28, 293-309.

THOMSON, K. s. 1973. Observations on a new rhipidistian fish from the Upper Devonian of Australia. Palaeonto-
graphica, 143A, 209-220.

VOROBYEVA, E. 1977. Morfologiya i osobennosti evolyutsii kisteperykh ryb. (Morphology and the nature of
evolution of crossopterygian fish). Trudy paleont. inst., Nauka, Moscow, 163, 1-239.

WOODWARD, A. s. 1906. On a Carboniferous fish fauna from the Mansfield District. Mem. natn Mus. Viet. 1,
1-32.

YOUNG, G. c. 1981. Biogeography of Devonian vertebrates. Alcheringa, 5, 223-245.

and GORTER, ). D. 1981. A new fish fauna of Middle Devonian age from the Taemas/Wee Jasper region of

New South Wales. Bull. Bur. Miner. Resour. Geol. Geophys. Aust. 209, 83-147.

J. A. LONG

Department of Geology
University of Western Australia

Typescript received 4 September 1985 Nedlands

Revised typescript received 12 March 1987 Western Australia 6009

WHY THE RHYNCHONELLID BRACHIOPODS
SURVIVED AND THE SPIRIFERIDS
DID NOT: A SUGGESTION

by DEREK V. AGER

ABSTRACT. It is Suggested that the rhynchonellid brachiopods survived to the present day whilst all other
forms with spirolophous lophophores became extinct because the latter were restrained by calcareous spiralia
whereas the former were able to extend their lophophores outside their shells.

Alan Hoverd has shown recently (1985) that the living rhynchonellid brachiopod Notosaria
nigricans (Sowerby) can uncoil and extend its lophophore or feeding organ beyond its shell. This
had only previously been recorded, in Hemithiris psittacea (Gmelin), by Morse (1869) and by
implication in other rhynchonellid species by Davidson (1887), who noted that the lophophore
might extend more than four times the length of the shell. However, this observation appears to
have been subsequently overlooked by other brachiopod workers. Such a capability is obviously
advantageous for a suspension feeder, since it greatly extends the area for trapping food particles
in the water. Hoverd has shown that it is also advantageous in that the lophophore acts as a brood
chamber for the larvae and then liberates them into the water when extended. It is possible that
brachiopods adopted a different feeding strategy in the juvenile stage, before the development of
calcareous brachidia, but we cannot know about this in the extinct spiriferids.

Rhynchonellid brachiopods have a longer history than any other articulate group, since they
range from the Ordovician to Recent, whereas their present-day contemporaries, the terebratuloids,
only range back to the Devonian. The rhynchonellids are particularly interesting in that they are
the only living group with a spirally coiled fleshy lophophore. However, a similar spiral lophophore
was clearly present in the past in the hugely diverse group of brachiopods included within the
Order Spiriferida. All these other forms, i.e. the Suborders Atrypidina, Retziidina, Athyridina, and
Spiriferidina (including the punctate forms), had calcareous supports in the form of spiralia for
their lophophores (text-fig. 1). With such rigid supports it would clearly have been impossible for
them to extend their lophophores beyond the shell. Only in the rhynchonellids is the lophophore
support limited to simple rods or crura. Generally speaking the crura can only have supported the
proximal ends of the lophophore and would still have permitted the main spiral part to extend
like a spring. Certain groups of rhynchonellids, for example the stenocismatids, which abound in
Permian rocks, have more complex structures, but this group soon became extinct as did those
such as Rhynchonellina, of the Early Mesozoic, with very long crura.

Among the spire-bearers with calcareous spiralia the most interesting in this connection are the
atrypids, which are very common in Silurian and Devonian rocks, but suddenly became extinct,
all over the world, at the top of the Frasnian stage (Copper 1966). The distinguishing feature of
these atrypids was that their spiralia were directed in a dorsal direction, as are the lophophores in
modern rhynchonellids, and not laterally as in all other spire-bearers. I have earlier suggested
(Ager 1968) that there was a ‘take-over’ by the rhynchonellids from the atrypids at the end of
Frasnian times. I later demonstrated in the Silurian and Devonian rocks of Morocco (Ager et al.
1976) how the rhynchonellids of the Famennian Stage, at the end of Devonian times, may have
taken over an exactly similar ecological niche to that previously occupied by atrypids, in this case
a colonial development on a substrate of algal turf on a local rise within a shallow muddy sea. It

I Palaeontology, Vol. 30, Part 4, 1987, pp. 853-857.)

© The Palaeontologieal As.sociation

854

PALAEONTOLOGY, VOLUME 30

TEXT-FIG. 1. Form of the spiralia in (a) the Sub-Order Spiriferidina, (b) the Sub-Order Retziidina, (e) the
Sub-Order Athyridina, (d) the Sub-Order Atrypidina, (e) the Family Thecospiridae, and (/’) the known form

of the lophophore in the Order Rhynchonellida.

AGER: RHYNCHONELLID BRACHIOPODS

855

is noteworthy that after playing a comparatively minor role in earlier shelf faunas (as far back as
the Ordovician), the rhynchonellids suddenly proliferated in the Famennian following the extinction
of the atrypids, as illustrated in many papers by Sartenaer (e.g. 1968, 1969) on Famennian
rhynchonellids around the world. The only significant difference between the atrypids and the
rhynchonellids is the presence in the former of calcareous lophophore supports or spiralia.

The rhynchonellids would have had an obvious advantage over the atrypids if they could then,
as now, protrude their lophophores beyond their valves for feeding purposes and as part of the
reproductive process. Hoverd also pointed out (1985) that modern rhynchonellids have the further
unusual ability of regenerating their lophophores if the ends are snapped off (by the sudden closing
of the shell in an emergency when the lophophore is extended).

Of course, we know nothing about the lophophores of other important groups such as the
orthids and pentamerids, which are long extinct and which have no indication of the lophophore
shape in their brachidia. Flowever, it may be significant that just as the atrypids may have been
replaced ecologically by the ‘ordinary’ rhynchonellids, so the pentamerids may have been replaced
by the stenocismatid rhynchonellids. Both have a spondylium-type structure in one or both valves
and both seem to have prospered in the vicinity of reefs. The stenocismatids began just when the
pentamerids declined to extinction in the Mid to Late Devonian and they had a final burst of
abundance before their own extinction in the Late Permian. They are particularly common, for
example in the vicinity of the Magnesian Limestone reefs of Durham (N. Hollingworth, pers.
comm. 1985).

After the extinction of the atrypids, the other forms with laterally directed calcareous spiralia
continued to flourish through the Carboniferous, Permian, and even Triassic Periods. Athyrids
such as Tetractinella were locally abundant, for example in the Italian Middle Triassic, and many
other forms lasted as late as the Rhaetian (Pearson 1977). There must have been some advantage
in laterally directed lophophores, presumably in the separation of inhalant and exhalant feeding
currents. Ager and Wallace (1966) suggested that they permitted a more efficient lateral flow of
water rather than that conventionally assumed. Lfowever, the laterally directed spiral lophophores
still had the limitation of a rigid framework supporting the fleshy lophophores and this presumably
could not compete with the more flexible lophophores of the increasingly abundant rhynchonellids.
In other words most of the spire-bearers opted for strength rather than flexibility.

In Palaeozoic times there was also the hugely successful order of brachiopods the Strophomenida,
which ranged from the early Ordovician to the early Jurassic and which included the chonetids
and productids that dominated the brachiopod world of the late Palaeozoic. There is clear evidence
(from impressions or ridges inside the valves of genera such as Leptaenisca and Davidsonia) of
spiral lophophores directed dorsally but without calcareous supports. Only in one small group the
Thecospiriidae are calcareous spiralia preserved, directed ventrolaterally (text-fig. \e). It may be
significant that this family is restricted to rocks of Triassic age, the last of all the Strophomenida
apart from the very problematical and isolated superfamily, the Cadomellacea of the Toarcian.

With the coming of Jurassic times the rigid spire-bearers were clearly in a decline and after the
widespread extinctions at the end of the Triassic, only one genus— 5’/?/>7/cn>7o— survived, in
progressively decreasing diversity, in the Early Jurassic. By the end of the Pliensbachian they had
all but disappeared and they only lingered on, very locally, into the earliest Toarcian.

The progressive extinction of the last of the spire-bearers was detailed by Thomas (1978). It is
noteworthy that the spire-bearers suffered a major eclipse in the widespread sulphurous black mud
conditions at the end of the Triassic in the areas where they were still endemic. Then their final
extinction came in the similar conditions of the Toarcian when they were even more localized. Of
particular interest is one of the last species; S. adscendens (Eudes-Deslongchamps) of the late
Pliensbaehian, in which Thomas found ventrally directed spiralia (text-fig. 2). No other species is
known to have this type of internal structure after the extinction of atrypids in the Late Devonian.
One can surmise that this was evolutionary convergence with the contemporaneous rhynchonellids
and was perhaps an advantage in the special environment of turbid black sulphurous bottom
waters. Miguel Mancenido in fact suggested (in Thomas, p. 291) that this species may have opened

TEXT-FIG. 2. Ventrally directed spiralia in the probably unique species Spiriferina adscendem (Eudes-
Deslongchamps). By kind permission of Dr Alun Thomas.

ACER: RHYNCHONELLID BRACHIOPODS

857

its valves as wide as possible withdrawing the spiralia and presenting the unprotected lophophore
into the surrounding waters for feeding. However, it is not clear how this could be done and it
would appear that this rare species also could not escape the calcareous straight-jacket of its own
spiralia so that it, and the whole group of spiralia-bearing brachiopods, soon after became extinct.

Acknowledgements. I should like to record my sincere thanks to Alan Hoverd for useful discussions when I
was in New Zealand and to Alun Thomas for permission to use unpublished information from his thesis.
John Edwards kindly drew the diagrams in text-fig. 1 and Mrs V. Jenkins typed the manuscript.

REFERENCES

ACER, D. V. 1968. The Famennian takeover. Circ. Palaeont. Ass. 54a, 1.

cossEY, s. p. J., MULLiN, p. R. and WALLEY, c. D. 1976. Biachiopod ecology in mid-Palaeozoic sediments

near Khenifra, Morocco. Palaeogeog. Palaeodimat. Palaeoecol. 20, 171 185.

and WALLACE, p. 1966. Flume experiments to test the hydrodynamic properties of certain spiriferid

brachiopods with reference to their supposed life orientation and mode of feeding. Proc. geol. Soc. Land.
1635, 160-163.

COPPER, p. 1966. The Atrypa zonata brachiopod group in the Eifel, Germany. Senckenherg. letli. 47, 1 55.
DAVIDSON, T. 1887. A monograph of Recent Brachiopoda. Part H, 75 182, pis. 14 25. Trans. Linn. Soc.
Land., Ser. 2, Vol. IV (Zoology).

HOVERD, w. A. 1985. Histological and ultrastructural observations of the lophophore and larvae of the
brachiopod, Notosaria nigricans (Sowerby 1846). Jl nat. Hist. 19, 831 850.

MORSE, E. s. 1869. Note on the extension of the coiled arms in Rhynchonclla. Amer. Jl Sci. Arts, 17, 257.
PEARSON, D. A. B. 1977. Rhactiaii Brachiopods of Europe. Nene Denkschr. natiirhist. Museums Wien, 1, 1 85.
SARTENAER, p. 1968. Dc I’importance stratigraphique des rhynchonelles famenniennes situees sous la zone a
Ptycliomaletoechia omaliusi (Gosselet, J., 1877). Sixieme note . . . Bull. Inst. roy. Sci. nat. Belg. 44, 1-32.

1969. Fate Upper Devonian (Famennian) rhynchonellid brachiopods from western Canada. Bull. geol.

Surv. Can. No. 169, 269 pp.

THOMAS, A. R. 1978. The ecology, evolution and extinction of Spirijerina in the Lower Jurassic, 352 pp. Ph.D.
thesis (unpublished). University College of Swansea.

DEREK V. ACER

Department of Geology

Typescript received 7 October 1986 University College

Revised typescript received 10 November 1986 Swansea SA2 8PP

V '

' , ■ .. - ' , i

”w

\ ' ■■ "

- .•»■..'"<• !■* ^>- ■ '-r

•' ' ■ -I' ' ' ;«]

)

1

?'

1

THE PALAEONTOLOGICAL ASSOCIATION

ANNUAL REPORT OF COUNCIL FOR 1986

Membership and Subscriptions. Membership totalled 1,368 on 31 December 1986, a decrease ot 10 from
the previous year. There were 869 Ordinary Members, a decrease of 6; 70 Retired Members, an increase of 2;
123 Student Members, an increase of 6; and 306 Institutional Members, a decrease of 12. Total individual
and institutional subscriptions to Palaeontology through Marston’s agency numbered 442, an increase of 15.

Subscriptions to Special Papers in Palaeontology numbered 116 individuals, an increase ot 3; and 122
Institutions, a decrease of 6. Orders through Marston’s agency for Special Papers numbered 322 volumes.

Sales of back parts of Palaeontology via the Membership Treasurer realized £475- 15. Sales of back parts
of Palaeontology under the special offer for clearance of surplus backstock, yielded £7, 168-51 gross (less at
least £687-23 postage). Sales of back numbers of Special Papers in Palaeontology to individuals yielded
£1,039-58. Institutions purchased 3 copies of Palaeontology and 47 Special Papers. Sales of 4 copies ol the
Atlas of the Burgess Shale through the Marketing Manager yielded £60; and sales of Fossil Plants of the
London Clay through him numbered 57 copies, yielding £341 and $24. Sales of the Atlas of Invertebrate
Macrofossils under the special offer to members of the Association yielded £442 net. Royalties from Longman,
from sales of the latter, amounted to £1,512. A special sale of Special Paper 22 to members of the Geological
Curators’ Group yielded £360-74 net.

Finance. During 1986 the Association published Volume 29 of Palaeontology at an estimated cost of £57,873
(including postage and distribution). Special Papers 35 and 36 were published at an estimated cost of £9,177
and £7,114 respectively (including postage and distribution). The Association is grateful to all those who
made donations to offset the cost of publishing Palaeontology.

Publications. Volume 29 of Palaeontology for 1986, published m lour parts, contained 877 pages and 75
plates. Special Papers 35 {Studies in palaeohotany and palynology in honour of N. F. Hughes, edited by D. .1.
Batten and D. E. G. Briggs), and 36 {Campanian and Maastrichtian ammonites from northern Aquitaine,
France, by W. J. Kennedy) were published in December.

Meetings. Nine meetings were held m 1986. The Association is indebted to the organizers, hosts, and held
leaders of these.

a. Review Seminar on 'Palynological approaches to palaeoenvironmental analysis’, held on 26 February,
at the British Geological Survey, Keyworth, Nottingham. About 70 people attended the meeting, which
was convened by Dr Rex Harland.

b. Joint Seminar with the Malacological Society on The species concept in living and fossil molluscs’, held
on 12 March at the British Museum (Natural History). Dr G. Oliver and Dr J. D. Taylor convened
the meeting, which was atttended by 51 people.

c. Twenty-Ninth Annual General Meeting, held in the Lecture Theatre of the Geological Society of London
on 19 March. Professor D. M. Raup delivered the Annual Address, on ‘Extinction’. The Sylvester-
Bradley Award was made to Mr D. K. Loydell.

d. Workshop on ‘Microcomputers in palaeontology’ held at the University of Dundee on 3 April. The
meeting was attended by 26 people, and was convened by Dr A. W. Owen. A software package for the
treatment of palaeontological data, ‘Palstat’, written by Dr D. A. T. Harper and Dr P. D. Ryan was
introduced at the meeting, and is published by Lochee Publications, on behalf of the Association.

e. Field Meeting, organized by the Carboniferous Group, to the Llandudno and Anglesey area. North
Wales, on 18 21 April. The meeting was led by Dr I. D. Somerville, and attended by over 60 members.

/. First Lyell Meeting, held jointly with the Geological Society of London, on ‘Gondwana and Tethys’ at
the Society’s rooms in Piccadilly, on 7-8 May. Over 150 attended the symposium, which was convened
by Professor A. Hallam and Professor M. G. Audley Charles. Papers are to be published as a volume
by the Geological Society.

860

THE PALAEONTOLOGICAL ASSOCIATION

g. "Progressive Palaeontology' Meeting, held at the Open University, Milton Keynes, on 29 May. The
meeting was convened by Rachel Wood, and was attended by 42 people.

h. Review Seminar on 'Palaeobiology of benthic molluscs', held on 12 November at the University of
Cambridge. More than 60 people attended the meeting, which was convened by Dr J. A. Crame and
Dr S. Conway Morris.

/. The Annual Conference, held at the University of Leicester, on 18 21 December, comprised an open
meeting followed by a field trip to the Jurassic of the district, led by Dr J. D. Hudson and Dr R. G.
Clements. The conference was attended by over 180 people, and the Local Secretary was Dr David J.
Siveter. The President’s Award was made to Mr P. A. Allison

Council. The following members served on Council following the Annual General Meeting on 19 March
1986. President'. Dr L. R. M. Cocks; Vice-Presidents'. Dr M. G. Bassett, Dr D. E. G. Briggs; Treasurer: Dr
M. Romano; Membership Treasurer: Dr A. T. Thomas; Secretary: Dr P. W. Skelton; Marketing Manager:
Dr V. P. Wright; Institutional Membership Treasurer: Dr A. W. Owen; Editors: Dr D. Edwards, Dr P. A.
Selden, Dr P. R. Crowther, Dr L. B. Halstead, Dr T. J. Palmer, and Dr C. R. C. Paul; Circular Reporter:
Dr Derek J. Siveter; Public Relations Officer: Dr M. J. Benton; other members: Dr M. E. Collinson,
Dr H. A. Armstrong, and (by co-option to vacant positions) Professor B. M. Eunnell and Dr P. D. Taylor.

Circulars. Pour Circulars, numbers 123 6, were distributed to Ordinary, Student, and Retired Members,
and, on request, to over 90 Institutional Members.

Council Activities. There has been much fruitful liaison with other bodies in 1986. Early in the year, the
Geological Society of London, together with the Association, the British Micropalaeontological Society, and
the Palaeontographical Society formally set up the ‘Joint Committee for Palaeontology’, with the purpose
of strengthening relationships between the sponsoring societies and promoting the subject of palaeontology
in general. Like the other constituent bodies, the Association is represented on the J.C.P. by two members
of Council. The J.C.P. has already been active, for example, in preparing a general submission concerning
palaeontology to the University Grants Committee review on university provision for the Earth Sciences,
alongside those made separately by its constituent societies.

A special offer on surplus backparts of Palaeontology was made to members of the Association, and
subsequently to members of a number of other palaeontological societies, and was a great success. Council
is indebted to Dr M. G. Bassett and his staff at the National Museum of Wales for putting the offer into
effect.

Council granted financial contributions towards the projected re-launching of the Cephalopod Newsletter
and the 1987 issue of the Reef Newsletter, and also made an award from its Conservation Fund to the work
of the West London Wildlife Group in the conservation of some 3,000 tons of tip material of Westphalian
age, at Lower Writhlington, Avon, in which much well-preserved arthropod, plant, and some vertebrate
material is present.

A new series of annual joint meetings with the Geological Society, the ‘Lyell Meetings’ was launched in
May, with a two-day symposium on Gondwana and Tethys. The next, planned for February 1987, will be
on ‘Catastrophes and the history of life’. Thereafter the Lyell meetings will be co-ordinated by the J.C.P.

A successful joint symposium with the Malacological Society was held in March, and in 1987 there are to
be joint meetings with the Society for Experimental Biology (‘Biomechanics in evolution’), the Systematics
Association and the Linnean Society (‘The phytogeny and classification of the tetrapods’), and the
Conservation Committee of the Geological Society and the Geological Curators’ Group (‘The use and
conservation of palaeontological sites’), while a further meeting with the Linnean Society, on the co-evolution
of organisms and the atmosphere, is planned for December 1987.

Other plans for 1987 include the publication of the Association’s first software package, ‘Palstat’, which
was introduced for experimental testing in April 1986 and displayed in its final form at the Annual Conference
in December 1986. The next volume in the ‘Fossils of the X Formation’ series, that on Fossils of the Chalk
will be published in mid to late 1987, and further titles are in the pipeline. The Association’s 30th anniversary
year is also to be marked by a celebratory dinner following the Annual Address in March, and by a ‘long
weekend’ field trip to West Germany, in September, with an exciting programme of visits to several sites
with exceptionally preserved biotas.

BALANCE SHEET AND ACCOUNTS FOR THE
YEAR ENDING 31 DECEMBER 1986

Balance Sheet as at 31 December 1986

1985

£

£

£

£

48,521

Investments at Cost (see schedule)

Current Assets

51,917

2,000

Sundry debtors .........

2,829

45,363

Cash at bank .........

76,330

2,029

Sylvester-Bradley Fund .......

1,974

Loans ......

450

49,392

81,583

Current Liabilities

5,811

Subscriptions received in advance .....

Provision for cost of:

2,854

13,500

Palaeontology .........

16,000

Special Papers .........

16,291

3,025

Sundry creditors .........

2,973

1,000

Loan from Royal Society .......

23,336

38,118

26,056

43,465

£74,577

£95,382

Represented by:

Publications Reserve Account

45,365

Balance brought forward .......

Excess of income over expenditure for the year transferred

69,917

69,917

24,552

from Income and Expenditure Account ....

20,860

90,777

Sylvester-Bradley Eund

1,816

Balance brought forward .......

2,029

200

Donation ..........

213

Interest ..........

145

2,029

(200)

Grant awarded .........

(200)

1,974

2,631

Meeting Reserve .........

2,631

£74,577

£95,382

Income and Expenditure Account

1985

£

44,753

32,041

8,853

483

696

1,800

2,554

62

11,602

235

INCOME

Subscriptions

1986

1985

Palaeontology

Sales .......

Donations ......

Special Papers

Sales

Donations

Burgess Shale Portfolio
Sales

Fossil Plants of the London Cla y

Sales

Postage/Stationery ....

Atlas of Invertebrate Macrofossils
Sales .......

Postage/Stationery ....

Offprints

Profit of Sales of Investments
Investment Income (see schedule)
Sundry Income

£ £

41,474

368

41,840

34,486

694

35,180

8,319

582

8,901

169

169

335

(29)

306

2,074

(120)

1,954

303

4,209

12,180

30

£103,079

£105,072

FOR THE Year Ended 31 December 1986

1985

£

53,698

14,594

2,093

3,719

4,423

£78,527

£24,552

EXPENDITURE

Cost of Publication of Palaeontology
Volume 29— Part 1 ....

Part 2 ... .

Part 3 . . . .

Part 4 (provisional)

Other provision for 28/4

Cost of Publication of Special Papers

No. 35

No. 36

Warehousing OF Publications .

Grants

Cost of circulars

Preparation .....

Postage ......

Credit ......

Administrative Costs
Postage and stationery
Editorial expenses ....
Meeting expenses ....

Audit fee

Membership of Societies

13.668

14,292

13,913

16,000

(129)

57,744

9,177
7,1 14

16,291

2,280

200

2,917

1,044

(480)

3,481

610

386

2,910

250

60

£84,212

Excess of Income over Expenditure for the Year Transferred to
Publications Reserve Account £20,860

Schedule of Investments and Investment Income as at 31 December 1986

Gross Income

£2,000

1 1% Exchequer Stock 1991

Cost

£

1,991

for Year
£

150

£12,000

135% Exchequer Stock 1987

11,520

1,590

£1,000

9% Treasury Stock 1992/1996

992

90

£1,000

9% Treasury Stock 1994

955

90

£4,000

8% Treasury Stock 2002/2006

2,192

320

£5,357

135% Treasury Stock 1997

5,000

710

£3,280

13^% Exchequer Stock 1996

3,000

435

5,270

M. & G. Charifund Units

4,073

1,265

10,000

New Throgmorton Trust (1983) p.l.c. 25p Income Shares

1,706

474

700

Clarke, Nicholls & Coombs p.l.c. 25p Shares

668

62

6,180

M.E.P.C. p.l.c. 65% Convertible Unsecured Loan Stock 1995/2000

4,943

402

374

M.E.P.C. p.l.c. 25p Shares

703

58

1,140

National Westminster Bank p.l.c. £1 Ordinary Shares ....

3,929

260

10,150

Agricultural Mortgage Corporation Ltd. 7|% Debenture Stock 1991/
1993

8,251

787

1,865

Imperial Group p.l.c. 25p Ordinary Shares— sold March 1986

—

176

1,460

Saatchi & Saatchi 6-3% Convertible Cumulative Redeemable Preference
£1 Shares

1,994

129

Bank Interest ............

51,917

6,998

5,182

12,180

Market Value at 31 December 1986 (1985— £74,498) .... £83,255

Report of the Auditor to the Members of
The Palaeontological Association

In my opinion, the Accounts as set out on pages 861-864 give a true and fair view of the state of the affairs of the
Association at 31 December 1986 and of its income and expenditure for the year ended on that date.

March 1987

Market Harborough, Leicestershire

G. R. Powell
Chartered Accountant

INDEX

Pages 1-206 are contained in Part 1; pages 207-428 in Part 2; pages 429-645 in Part 3; pages 646 868 in Part 4. Figures in

Bold Type indicate plate numbers.

A

Acanlhocliacroiliuml Uisselii, 614, 71
Acritarchs; Ordovician, China, 613

Acroleulhis {Boreioleulhis) ruwsoni sp. nov., 640; (B.) stolleyi
sp. nov., 641

Adams-Tresman, S. M. The Callovian (Middle Jurassic)
marine crocodile Meuiorhynchiis from central England,
179; the Callovian (Middle Jurassic) teleosaurid marine
crocodiles from central England, 195
Adorfia c{. finna, 61 5, 70

Ager, D. V. Why the rhynchonellid brachiopods survived
and the spiriferids did not: a suggestion, 853
Agnoslid gen. et. sp. indet., 216, 25
Allocriocenis dentonense, 62, 10; larvatnm, 63, 10
Amholeheris antyx, 88

Ammonites: Turonian, west Texas, 27; type Santonian,
France, 765

Ammonoid: oldest ‘colour’ patterns, 299
Amphibia: Anthracosauria, Northumberland Coal Meas-
ures, 15

Annual Address: Raup, D. M. Mass extinction: a commen-
tary, 1

Antarctica: Cretaceous Dimitobelidae, 147; Cretaceous
wood, 233; Cretaceous fish, 829
Aphelaspis cantori sp. nov., 218, 25
Arbusadidium fdamentosum, 615, 70
Archaeocanihus howerbunki, 595, 65
Argentina: staurikosaurid dinosaur, 493
Australia: Devonian conodonts, 251; Tertiary echinoids,
319; Devonian fish, 839

B

Baltisphaeridium longispinosun} longi.spmo.sun], 615, 72
Bathy.sc/udia wetherelli, 609, 67
Beelarongia patrichae gen,, et sp. nov., 840, 91
Belemnitida: Cretaceous, Antarctica. 147; Cretaceous,
Mozambique, 311; Acroleuihi.s, north-west Europe, 635
Belodcdla devonica, 41; re.sima, 41; iriangulciri.s, 41
Beltanella gilesi, 653, 73
BeltanelUfonnis bnmsae, 664, 74, 75
Benton. M. J. See Wright, A. D. and Benton, M. J.
Berg-Madsen, V. A new cyclocystoid from the Lower Ordo-
vician of Oland, Sweden, 105
Berganiia johan.s.soni sp. nov,, 94, 13

Bivalvia; island hopping rudist, Oman, 505; Ordovician,
Wales, 677

Blows, W. T. The armoured dinosaur Polaainthu.s fo.xi from
the Lower Cretaceous of the Isle of Wight, 557
Boehmocera.s loescheri. 111 ,Kl

'Bolbozoe' cf. cmomala. 85, 86, 87; cf bohemica, 84, 87; sp.
nov. A, 83, 84

Bolbozoid gen. et sp. nov. A, 83, 85

Botrioides broeggeri, 86, 12; hronnii, 81, 11; efflore.scen.s, 91,
\3; foveolatu.s, 90, 12, 13; impo.ster sp. nov., 85, 11; marge
sp. nov., 91, 13, 14; .simplex sp. nov., 87, 12; sp. A, 88, 12;
sp. B, 90, 12

Brachiopoda: rhynchonellids, 853

Brinkman. D. B. and Sues. JF-D. A staurikosaurid dinosaur
from the Upper Triassic Ischigualasto Formation of
Argentina and the relationships of the Staurikosauridae,
493

Bromley, R. C. See D’Alessandro, A. and Bromley, R. C.
Bryozoa: Ordovician, trepostome, 691
Bull, E. E. Upper Llandovery dendroid graptolites from the
Pentland Hills, Scotland, 1 17

C

Cambrian: Irilobites, Tasmania, 207; cornute, Wales, 429
Canada: Ediacaran biota, 647

Carboniferous: Anlhraco.saiiru.s from Northumberland Coal
Measures, 15

Chahfa, Y. See Raab. M. and Chalifa, Y.

Chatterjee, S. See Grande, L. and Chatterjee, S.

China: Ordovician acritarchs, 613; Permian plants. 815
Clack, J. A. Two new specimens of Anthracosaurus (Am-
phibia: Anthracosauria) from the Northumberland Coal
Measures, 15

Conodont: Devonian, Australia, 251
Coremagraptus imperfeetus. 132, \9;kallu.si, 131. 20
Cornute: Cambrian, Wales, 429
Coryphidiwn bohemicum, 616, 72

Cretaceous: basal Turonian ammonites, west Texas, 27;
Dimitobelidae, Antarctic Peninsula, 147; preservation of
conifer wood, Antarctica, 233; belemnites, Mozambique,
311; island hopping rudist bivalve, 505; armoured dino-
saur Polacanllni.s, 557; ammonites. 765; Antarctic fish, 829
Cri.staUinium deniatum, 618, 68
Crocodile: Jurassic central England, 179, 195
Crustacea: Eocene, England, 581
Cyclocystoid: Ordovician, Sweden. 105
Cyclomedii.sa plana, 656, 73; sp., 658, 73
Cymatiogalea cf. eri.siata, 618. 72; cuviUierl 618, 72
‘Cypridinid’ sp., 85; gen. et sp. nov. A, 84, 87; B, 87

D

D'Alessandro, A. and Bromley, R. C. Meniscate trace fossils
and the Muen.steria Taenidium problem, 743
Denagnostus corbetti sp. nov., 213, 24, 25
Devonian: conodonts, Australia, 251; osteolepiform fish.
Australia, 839

Dictyonema pentlandiea sp. nov., 122, 17
Dimitobelu.s dayi sp. nov., 169, 23; cf. dayi sp. nov., 170, 23;
dipiychus, 161, 22; sp. nov. aff. diptychu.s, 162, 22;

866

INDEX

Dimitohelus (cont.):

praelindsayi sp. nov., 167, 23; stimulus, 163, 22; cf. stimulus
var. extremis, 166, 22, 23; sp. nov. A, 172, 23; sp. nov. B,
172,23; sp. nov. C, 173,23

Dinosaur: theropod. South Africa, 141; staurikosaurid,
Argentina, 493; armoured. Lower Cretaceous, 557
Dokinocephalid gen. et sp. indet., 219, 25
Donovan, S. K. See Jefferies, R. P. S., Lewis, M. and Dono-
van, S. K.

Doyle, P. The Cretaceous Dimitobelidae (Belemnitida) of
the Antarctic Peninsula region, 147; Early Cretaceous
belemnites from southern Mozambique, 31 1
Drepanodus sp. 41

Drygant, D. M. See Selden, P. A. and Drygant, D. M.

E

Echinoids: Tertiary, Australia, 319
England: Jurassic crocodile, 179, 195
Eocene: Crustacea, 58 1
Eiigonocare sp., 219, 26
Eulophoceras austriacum, 776, 82

Extinction: mass extinction: a commentary. Twenty-ninth
annual address by D. M. Raup, 1

F

Fagesia catinus, 5 1 , 7, 8
Favositids: review of their affinities, 485
Fish: Cenomanian, echodontid, 717; Cretaceous, Antarctic,
829; Devonian osteolepiform, Australia, 839
France: Ammonites of the type Santonian, 765

G

Gao Zhifeng and Thomas, B. A. A re-evaluation of the
plants Tingia and Tingiostachya from the Permian of
Taiyuan, China, 815
Glyphea scahra, 594, 64
Goniophora sp., 679, 76

Grande, L. and Chatterjee, S. New Cretaceous fish fossils
from Seymour Island, Antarctic Peninsula, 829
Graptolites: Upper Llandovery, Scotland, 1 17
Graptoloid: classification of the Diplograptacea, 353

H

Halocypris inflata, 88

Flancock, J. M. See Kennedy, W. J., Wright, C, W. and
Hancock, J. M.

Hemiaster {Bolhaster) callidus sp. nov., 346, 47, 48; dolosus
sp. nov., 338, 45; planedeclivis, 342, 46, 47; .subidus sp.
nov., 366, 44, 48; verecundus sp. nov., 340, 45, 46
Hickey, D, R. Skeletal structure, development, and elemen-
tal composition of the Ordovician trepostome bryozoan
Peronopora, 691

Hofmann, H. J. See Narbonne, G. M. and Hofmann, H. J.
Homarus gammarus, 64; morrisi sp. nov., 585, 64
Hoploparia gammaroides, 588, 68; victoriae sp. nov., 592, 64;
wardi sp. nov., 591, 64

1

Ichthyosaur: swimming and buoyancy, 531; case study, 543
Israel: Cenomanian fish, 717

J

Jago. J. B. Idamean (Late Cambrian) trilobites from the
Denison Range, south-west Tasmania, 207
JelTeries, R. P. S., Lewis, M. and Donovan, S. K. Prutocy-
stites menevensis~& stem group chordate (Cornuta) from
the middle Cambrian of South Wales, 429
Jefferson, T. H. The preservation of conifer wood: examples
from the Lower Cretaceous of Antarctica, 233
Jurassic: marine crocodile Metriorhynchus, England, 179;
telesaurid marine crocodile, England, 195

K

Kamenmoceras calvertense, 33, 3

Kat, P. W. Biogeography and evolution of African fresh-
water molluscs: implications of a Miocene assemblage
from Rusinga Island, Kenya, 733
Kennedy, W. J. Ammonites from the type Santonian
and adjacent parts of northern Aquitaine, western France,
765

Kennedy, W. J., Wright, C. W. and Hancock, J. M. Basal
Turonian ammonites from west Texas, 27
Kenya: freshwater molluscs, 733
Kullingidl sp,, 659, 73

L

Leiosphueridia tenuissima, 618, 68
Leuroleheris mackenziei, 88; sharpei, 88
Lewis, M. See JelTeries, R. P. S., Lewis, M. and Donovan,
S. K.

Li Jun. Ordovician acritarchs from the Meitan Formation
of Guizhou Province south-west China, 613
Limiparus eocenicus, 600, 66; scyllariformis, 603, 66
Long, J. A. An unusual osteolepiform fish from the Late
Devonian of Victoria, Australia, 839

M

Mammites powelli sp. nov., 42, 3, 4
Mammitinae gen. et sp. indet., 44, 2

Mapes, R. H. and Sneck, D. A. The oldest ammonoid
'colour' patterns: description, comparison with Nautilus,
and implications, 299

Martin, D. M. A taphonomic and diagenetic case study of a
partially articulated ichthyosaur, 543
Mateer, N. J. A new report of a theropod dinosaur from
South Africa, 141

Mawson, R. Early Devonian conodont faunas from Buchan
and Bindi, Victoria, Australia, 251
McNamara, K. J. Taxonomy, evolution, and functional
morphology of southern Australian Tertiary hemiasterid
echinoids, 319

Medusinites asteroides, 660, 73
Meniscate trace fossils, 743
Micragnostus sp. 1, 209. 24; sp. 2, 210, 24
Miocene: freshwater Mollusca, 733

Mitchell, C. E. Evolution and phylogenetic classification of
the Diplograptacea, 353
Mollusca: freshwater, Kenya, 733
Monocveloides oelandicus sp. nov., 109, 15, 16
Mozambique: Cretaceous, belemnites, 31 1
Multiplicisphaeridium cf. irregulare, 619, 70

INDEX

867

Mutterlose, J., Pinckney, G. and Rawson, P. F, The belem-
nite Acroteuthis in the HiboUtes beds (Hauterivian-
Barremian) of north-west Europe, 635
Myodakryotus deigryn gen. et sp. nov., 685, 77

N

Nadalina yukonensis sp. nov., 661 , 73

Narbonne. G. M. and Hofmann, H. J. Ediacaran biota of
Wernecke Mountains, Yukon, Canada, 647
NeohihoUtes ewcddi, 3 12, 43
Neoptychites sp., 56, 9

O

Oman: island hopping rudist, 505
Ophthalmosaunis sp., 63

Ordovician: trinucleid trilobites, Scandinavia, 75; a new
cyclocystoid from Sweden, 105; acritarchs, China, 613;
bivalves, Wales, 677; trepostome bryozoan, 691
Orthodesma sp,, 688, 76
Ostracodes: Silurian myodocopids, 783
Oidodus nmn indidensis, 269. 31

Owen. A. W. The Scandinavian middle Ordovician tri-
nucleid trilobites, 75

Ozarkodina huchanensis, 282, 37; excavatu excavala, 286, 31;
linearis, 288, 38; prolala sp. nov., 290, 39

P

Palaeodictyota pergraciiis, 1 36, 20

Palmer, D. See Sivetcr. D. J., Vannier, J. M. C. and
Palmer, D.

Panderodns recurvatus. 41; unicostatus, 41; valgus, 41
Pandormellina exigua exigua, 292, 40
Pasternakevia podoiica gen. et sp. nov., 539
Peratobelus foersteri sp. nov., 313. 43; cf. a\n'5, 154. 21; ?
sp., 155, 21

Permian: Tingiostachya, China, 815
Peronopora decipiens, 78, 79; vera, 78, 79
Petaioferidium fiorigerum, 619, 72
Peteinosphaeridium trifurcatum intermedium, 619, 71
Pinckney, C. See Mutterlose, J., Pinckney, C. and Rawson,
P. F.

Pirea sinensis sp. nov., 620, 70
Placenticeras parapianutn, 769, 80
Plants: Permian, China, 815

Polygnathus inver,sus, 274, 33, 36; labiosus sp. nov., 274, 35,
36; nothoperbonus sp. nov., 276, 32, 33, 36; perbonus, 276,
34, 36; pseudoserotinus sp. nov., 277, 35, 36; dehlscens
abyssus, 272, 34, 36; dehiscens, 270, 32, 36; serotinus ‘delta
morphotype', 280, 33, 36
Poiygortium gracile, 620, 72

Proceratopvge gordonensis sp. nov., 222, 26, 27; sp., 224,

27

Protocystites menevensis, 436, 54, 55, 56, 57, 58, 59, 60
Psephoaster apokryphos sp. nov., 348, 47, 48; klydonos sp.

nov., 350, 48; lissos sp. nov., 347, 47
Pseudarca celtica sp. nov., 688, 76
Pseudaspidoceras flexosum, 34, 2

Pseudoagnostus (Pseudoagnostus) idalis denisonensis subsp.

nov., 210, 24; cf. idaiis sagittus, 212, 25; sp., 212, 25
Pseudoyuepingia vanensis sp. nov., 227, 27

Q

Quayle, W. J. English Eocene Crustacea (lobsters and
stomatopod), 581
Quitmaniceras reaseri, 30, 1

R

Raab, M. and Chalifa, Y. A new enchodontid fish genus
from the upper Cenomanian of Jerusalem, Israel, 717
Raup, D. M. Mass extinction: a commentary (Twenty-ninth
annual address, delivered 19 March 1986), 1
Rawson, P. F. See Mutterlose, J., Pinckney, G. and Rawson,
P. F.

Reedollthus carniatus, 96, 14; cpiebecensis, 14; sp., 14
Rhaetic: trace fossils, 407

Rhopaliophere membrana sp. nov., 621, 71; palmata, 621, 71
Rugoconltesl sp., 73

S

Scandinavia: Ordovician trinucleid trilobites, 75
Schizodiacrodium'i muitiramiferum sp. nov., 622, 69
Sciponoceras sp., 64, 2

Scotland: Upper Elandovery dendroid graptolites. 1 17
Scrutton, C. T. A review of favositid affinities, 485
Scyllaiides tubercuiatus, 606, 67

Selden, P. A. and Drygant, D. M. A new Silurian xiphosuran
from Podolia, Ukraine, USSR, 537
Silurian: dendroid graptolites, Scotland, 117; xiphosuran,
USSR, 537; myodocopid ostracodes, 783
Siveter, D. J., Vannier, J. M. C. and Palmer. D. Silurian
myodocopid ostracodes: their depositional environments
and the origin of their shell microstruclures, 783
Skelton. P. W. and Wright, V. P. A Caribbean rudist bivalve
in Oman: island-hopping across the Pacific in the Late
Cretaceous, 505

Sneck. D. A. See Mapes, R. H. and Sneck, D. A.

South Africa: theropod dinosaur, 141
Spriggia annulata, 661, 73; wadeae, 662, 73
Striatotheca principalis parva, 622, 70; quieta, 624, 70; cf.
transformata, 624. 71

Sues, II.-D. See Brinkman. D. B. and Sues. H.-D.

Sweden: Ordovician cyclocystoid, 105

T

Tasmania: Cambrian trilobites, 207

Taylor. M. A. A reinterpretation of ichthyosaur swimming
and buoyancy, 53 1
Tectitheca cf. additionalls, 626, 72
Tertiary: echinoids, Australia, 319

Tetrabelus seclusus, 156, 21; whitehousei sp. nov.. 159. 21;

willeyi sp. nov., 158, 21
Tetraleberis maddocksae, 88
Texanites ( Texanites) gallicus, 770, 80, 81
ThaUograptus arborescens, 127, 18; inequaiis, 130. 18
Thomas, B. A. See Gao Zhifeng and Thomas, B. A,
Thomasites adkinsi, 61,9
Tingiostachya tetralocularis, 819, 89, 90
Tongzia meitana gen. et sp, nov., 626, 69
Torreites sanchezi, 508, 61, 62
Trace Fossils: Rhaetic, 407, 743
Triasiana sp., 662, 74
Triassic: staurikosaurid dinosaur, 493

868

INDEX

Trilobites: Ordovician, Scandinavia, 75; Cambrian, Tas-
mania, 207

Trinucleid gen. et sp. indet., 100, 4

Tunnicliff, S. P. Caradocian bivalve molluscs t'rom Wales,
677

U

USSR: Silurian xiphosuran, 537
V

Vannier, J, M. C. See Siveter, D. J., Vannier, J. M. C. and
Palmer, D.

Vanuxemia bulla. 680, 76, 77
Vascoceras propriuni. 46, 4, 5, 6
Veryhachium trispinosum, 628, 68

W

Wales: Cambrian cornute, 429; Ordovician bivalves, 677
Wood: Cretaceous, Antarctica, 233
Worlhoceras cf. vermiculus. 64, 8

Wright, A. D. and Benton, M. J. Trace fossils from Rhaetic
shore-face deposits of Staffordshire, 407
Wright, C. W. See Kennedy, W. J., Wright, C. W. and
Hancock. J. M.

Wright, V. P. See Skelton, P. W, and Wright, V. P.
Wrightocerus muneri, 58. 10

X

Xiphosuran: Silurian, USSR, 537

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 CFl IXL, 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 1987 is £33 (U.S. S58) for Institutional
Members, and £16-50 (U.S. S30) 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 ILB, England.

RECENT PALAEONTOLOGICAL ASSOCIATION

PUBLICATIONS

Special Papers in Palaeontology

Numbers 1 19 are still in print and are available (post free) together with those listed below:

20. (for 1977): Fossil Priapulid Worms, by s. CONWAY morris. 155 pp., 99 text-ftgs., 30 plates. Price £16 (U.S. S24).

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., I plate. Price £30 (U.S. $45).

24. (for 1980): Dinoflagellate Cysts and Acritarchs from the Eocene of Southern England, by j. p. bujak, c. downie, g. l.
EATON anclG. L. williams. 100 pp., 24 text-figs., 22 plates. Price £15 (U.S. $23).

25. (for 1980): Stereom Microstructure of the Echmoid Test, b\ a. b. smith. 81 pp., 20 text-figs., 23 plates. Price £15
(U.S. $23).

26. (for 1981): The Fine Structure of Graptolite Periderm, bv 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 £1 5 (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., \Q 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 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., \6 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).

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).

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-F241 pp. Price £13-95. Available in the USA from Halsted Press at U.S. $24.95.

© The Palaeontological Association, 1987

Palaeontology

VOLUME 30 ■ PART 4

CONTENTS

Ediacaran biota of the Wernecke Mountains, Yukon,

Canada

G. M. NARBONNE and H. J. HOFMANN 647

Caradocian bivalve molluscs from Wales

S. P. TUNNICLIFF 677

Skeletal structure, development, and elemental composition

of the Ordovician trepostome bryozoan Peronopora

D. R. HICKEY 691

A new enchodontid fish genus from the upper Cenomanian
of Jerusalem, Israel

M. RAAB and Y. CHALIFA 717

Biogeography and evolution of African freshwater molluscs:
implications of a Miocene assemblage from Rusinga Island,

Kenya

P. W. KAT 733

Meniscate trace fossils and the Muensteria-Taenidium
problem

A. d’alessandro and r. g. bromley 743

Ammonites from the type Santonian and adjacent parts of
northern Aquitaine, western France

W. J. KENNEDY 765

Silurian myodocopid ostracodes: their depositional

environments and the origin of their shell microstructures
D. J. SIVETER, J. M. C. VANNIER and D. PALMER 783

A re-evaluation of the plants Tingia and Tingiostachya from
the Permian of Taiyuan, China

GAO ZHIFENG and B. A. THOMAS 815

New Cretaceous fish fossils from Seymour Island, Antarctic
Peninsula

L. GRANDE and S. CHATTERJEE 829

An unusual osteolepiform fish from the late Devonian of
Victoria, Australia

J. A. LONG 839

Why the rhynchonellid brachiopods survived and the
spiriferids did not: a suggestion

D. V. AGER 853

Printed in Great Britain at the University Printing House, Oxford 0031-0239

by David Stanford, Printer to the University

o

o

* z — (A :z. U) -z. U) _

illSNI NVINOSHilWS S3iyVMan libraries SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHill

— , , CO 2 5 ~

I I I v-''fd'sVAr\ / f\ /^^SO\7^S. UJ , _ Xf^vTSOA/^S. Ll_l .XTuCVaT^ 2!

CR I ES^SMITHS0NIAN"'|NSTITUTI0N NOIirUllSNl“‘NVmOSHil('JS S3iaVdan’^LJBRARI ES^SMITHSONI,

§ i i ” xV..v i I -

^ m

_ CO —

iliSNI NVINOSHillMS S3IHVaan L

2: CO z

S I I

V/ § y^y / ' '' t

m ' fT,

CO ± (/, _

BRARIES SMITHSONIAN INSTITUTION NOlifUliSNI NVINOSHill

Z « CO z

s

> , 5 > • s

\R I ES'SMITHSONIAN INSTITUTION NOlifUliSNI _ NVINOSHilWS^^ S 3 I d V H a IT L I B R A R I E s'^SMlTHSONI;

^ :2 \ ^ ^ CO

CO

iliSNI NVlNOSHillAIS S3IHVdan LIBRARIES SMITHSONIAN INSTITUTION NOliOiliSNI NVINOSHill

f“ z r- z r- Z •“

.■Si^ O X\v<svJ>:v _ O O .

CO

IRIES SMITHSONIAN INSTITUTION NOliOiliSNl NVINOSHilWS SaidVdail LIBRARIES SMITHSONI

iliSNI NVINOSHillMS S3ldVaan LIBRARIES SMITHSONIAN INSTITUTION NOliOiliSNI NVINOSHill

CO CO

o

CD

za

> lp(

.. \'/'l

a za \%}i

j:.y —

m -V^'’ ^ \jvA5iiV^ rn

CO ' ± CO

iliSNI NviNOSHilws S3iavaan libraries Smithsonian institution NoiiniusNi nvinoshti

z c, z V CO z

CO

O
2

2 ■ CO 2 in ' ^ (O

Hvyan libraries Smithsonian institution NoiiniiiSNi nvinoshiiws saiuvaan librai

— CO “ ~

Q ^ Q '•'^ ZI XlOiujJX O

riTUTION N0liniliSNl“'NVIN0SHillNS^S3 1 dVaa Ill'Ll B RAR I ES^SMITHS0NIAN“‘|NSTITUTI0N NOliOi!

z I-, 2 I- 2 [;^ 2

o xtivfiro^K ~ ,VV m 2 m 2

m

in — cn — to

BRARIES SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHiUMS SBiavaaiT LIBRA!
to 21 y cn 2 to 2

to

3, >• 5 > 5

2 CO •■■" 2 (O 2 (O

riTUTiON NOliniliSNI NviNOSHiiiNS saiavaaiT libraries Smithsonian institution Noiini

CO — CO ~ to

avaan libraries Smithsonian institution NoiiniiiSNi NviNOSHims S3iavaan libra

PO
>

m V<CL£:^ ^ m

to CO f: CO _

TITUTION NOliniliSNI NVINOSHillMS S3iavaaiT LIBRARIES SMITHSONIAN INSTITUTION NOlinJ
W 2 to 2 . to 2 CO

.< 2 .< ,vV^\.; 2 <

t E 2

> 2 3, ^

2 CO 2 CO - 2 C/3 ‘2

avaan libraries Smithsonian institution NoiiniiiSNi nvinoshtiws saiavaan libra

~ to — to — CO 2

CO

>\ —

to 4^ -■ m 5 ' X* “ 5

2 "• _l 2

riTUTION NOliniliSNI NVINOSHilWS S3iavaail libraries SMITHSONIAN INSTITUTION NOliOi
2 t^,, 2 r- 2 2

2 m ^ n O n-i x^'to'vPx O

■I'iN — 03 — A 03 ^

I— /-r^mlrMyi-X h- Aripi(""A^\ /VX«tVr»i. A 1— /A

__ m XlJ(os»ii>P' 2 ^ ^

ava a n _Li B RAR I Es Smithsonian institution NoiiniiiSNi~NviNOSHims S3 i ava a n ~li b rai

CO 2 to 2 ,

riTUTION

CO

ooT^oX tu
oX —
DC

2 /

> ' 2 X^osv^X^ > 5

, CO •'■ 2 CO 2 CO

,iiSNI_NVlNOSHilWS S3iaVaail libraries SMITHSONIAN INSTITUTION NOlini
to — CO r; CO